Organ-on-a-chip: Engineered Microenvironments for Safety and Efficacy Testing [1 ed.] 0128172029, 9780128172025

Organ-on-a-Chip: Engineered Microenvironments for Safety and Efficacy Testing contains chapters from world-leading resea

257 69 27MB

English Pages 546 [494] Year 2019

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Chap1.pdf
1 Need for alternative testing methods and opportunities for organ-on-a-chip systems
Introduction
Why we need alternative and improved methods
Requirements for in vitro alternatives to animal testing
Meeting functional requirements for improved in vitro methods
Correct and physiologically relevant human biology
Robust and repeatable methods
Ability to scale
Low capital and consumable costs
Remaining technical challenges
Scaling the biology down to the microlevel
Coculturing cells
Physiological relevance
Regulatory view and validation
Choice of materials for organ-on-a-chip systems
Conclusions
Outlook: stimulating the adoption of new technologies to replace animal testing
References
Chap2
2 Cell sources and methods for producing organotypic in vitro human tissue models
Introduction
Human tissue/cell sources and isolation methods
Continuous (immortal) cell lines
Primary cell cultures and early-passage cell lines with finite lifespan
Human tissue sources
Induced pluripotent stem cells
Methods for producing three-dimensional “organotypic” tissue cultures
Use of controlled substrate rigidity and nanopatterned culture substrates to enhance morphological and functional different...
Use of microporous membrane substrates to prepare cultures with polarized barrier, transport, and differentiated properties
Techniques for preparing three-dimensional spheroids/organoids with differentiated organotypic functions
Three-dimensional bioprinting
Summary/Outlook
References
Further reading
Chap3
3 Organs-on-a-chip engineering
Introduction
Definition
Engineering challenges
Design challenges
Manufacturing challenges
Microengineering
Introduction
Design
Materials
Fabrication process
Engineering fluid control for organ-on-chips
Liquid actuation
Pipetting robots
Gravity-driven flow
Peristaltic pumps
Syringe pumps
Pressure controller
Membrane or diaphragm pumps
Viscous drag pumps
On-chip pumps
Membrane pumps
Surface-driven flow or capillary pumps
Valves and bubble traps
Flow sensing
Flow-rate sensors
Thermal sensors
Coriolis mass flow meters
Differential pressure-based flow sensors
Imaging-based sensors
Pressure sensors
Other sensors
Industrial implementation
Stimulation and sensing
Optical
Mechanical
Stimulation
Shear stress
Stretch: tensile strain
Compression
Sensing
Electrical
Electrochemical sensors
Electrochemical enzyme biosensors
Electrochemical immunosensors
pH- and ion-sensitive sensors
Electrochemical oxygen sensors
Electrical biosensors
Electrical impedance spectroscopy
Transepithelial electrical resistance sensors
Commercial sensors
References
Chap4
4 Lung-on-a-chip platforms for modeling disease pathogenesis
Introduction
Anatomy and physiology of the respiratory system
The burden of respiratory diseases: classification, impact on global health, and statistics
In vitro and in vivo models of respiratory disease pathogenesis
The in vivo approach: animal models of lung diseases
In vivo models of the main lung diseases
Limitations of animal models
Ex vivo approach
Use of in vitro models
Current lung-on-a-chip systems
Mechanically active alveolar–capillary interface
Mimicking the pulmonary parenchymal environment
Simulating surface tension stresses
Complex organotypic cocultures
Organ-on-a-chip systems for modeling pathological conditions
Microfluidic precursors as lung pathology models: liquid plugs in small airways
Modeling lung inflammation, asthma, and chronic obstructive pulmonary disease in small-airway chips
Pulmonary edema and intravascular thrombosis modeled in alveolus-on-a-chip devices
Modeling cystic fibrosis in a microfluidic device
Lung tumor-on-a-chip devices
Improvements needed in lung-on-a-chip platforms for disease modeling and lung regeneration
Limitations of current organ-on-a-chip systems
Improvements to organ-on-a-chip technology
Innovative strategies and fabrication methods for improving standard microfluidic devices
Organoids-on-a-chip and of the need for multiorgan platforms
Three-dimensional printing for microfabrication and bioprinting
Conclusions
Acknowledgments
References
Chap5
5 Requirements for designing organ-on-a-chip platforms to model the pathogenesis of liver disease
Introduction
Liver function and structure
Liver function
Hepatic lobules and sinusoids
Cellular components of the sinusoids
Zonation
Drug-induced liver injury
Limitations of conventional nonclinical studies and clinical trials
In vitro cell-based assay
Hepatocytes
Microphysiological systems for drug-induced liver injury
Long-term repeated-dose toxicity tests
Assessment of toxicity by cholestasis
Site-specific hepatic toxicity in the liver (liver zonation)
Idiosyncratic drug-induced liver injury
Liver fibrosis
Hepatic stellate cells
Microphysiological systems for evaluating fibrosis
Organs-on-a-chip
Gut–liver axis
Enterohepatic circulation
Conclusion
References
Chap6
6 Brain-on-a-chip systems for modeling disease pathogenesis
Introduction
Need for microphysiological brains-on-chips
Prevalent diseases and disorders
Design and biofabrication approaches for systems-on-chips
State-of-the-art brains-on-chips
Microfluidic models
Compartmentalized neuronal models
Hydrogel-based models
Spheroid models
Higher-order system-on-a-chip functionality
Spatiotemporal control of chemoattractants, drugs, and other biologics in the microenvironment via integrated drug release ...
Integration of bioelectronic interfaces with microphysiological neural systems for in situ sensing and stimulation capabilities
Molecular-level interface between electronic materials and neural tissue toward constructing next-generation neural interfa...
Manufacturing approaches
Mold lithography
Contact printing
Hydrogel casting
Three-dimensional printing
Future outlook
References
Chap7
7 Kidney-on-a-chip
Introduction
Structure and function of the kidney
Kidney pathologies
Challenges associated with developing in vitro kidney models
Microphysiological kidney models
Two-dimensional versus three-dimensional systems
Cellular models
Kidney-on-a-chip models
Glomerulus-on-a-chip
Proximal tubule-on-a-chip
Distal tubule-/Collecting duct-on-a-chip
Kidney-on-a-chip: future perspectives
High-throughput technologies
Applications of kidney-on-a-chip platforms
Opportunities and challenges
References
Chap8
8 Heart-on-a-chip
Introduction
Anatomy and physiology of the heart
Dimension, location, and envelope
Heart wall
Heart cavities
The cardiac valves
Microscopic anatomy of the heart muscle
Mechanism of cardiac contraction
Physiology of the heart muscle
Spontaneous production of action potential
Excitation and action potential propagation
Heart–nerve connections
Electrocardiography
In vitro models of the heart
Traditional two-dimensional cell culture heart models
Two-dimensional cell culture heart models for ischemia
Two-dimensional cell culture heart models for drug discovery
3D cell culture models of the heart
The future of 3D cell culture heart models: bioprinting
Limitations of traditional two-dimensional model and 3D heart models
Organ-on-a-chip models of the heart: a new opportunity to mimic cardiac physiology, pathology, and toxicity
Effects of heart-on-a-chip models on cell differentiation
Final remarks and future directions
Acknowledgments
References
Chap9
11 Caenorhabditis elegans-on-a-chip: microfluidic platforms for high-resolution imaging and phenotyping
Introduction
Limitations of current toxicity screening models
Small animal models: Caenorhabditis elegans
Conception of Caenorhabditis elegans-on-a-chip
Caenorhabditis elegans as a whole-animal model in scientific research
Advantages of Caenorhabditis elegans  as a model for neuroscience and neurotoxicity
Challenges in using Caenorhabditis elegans for large-scale studies
Recent advances in the development of Caenorhabditis elegans-on-a-chip
The use of microfluidic platforms for Caenorhabditis elegans research
Serial versus parallel platforms
First-generation serial microfluidic platforms
Second-generation serial microfluidic platforms
Parallel microfluidic platforms
Disadvantages of current platforms
Large-scale microfluidics for phenotyping multiple populations of Caenorhabditis elegans nematodes
Development of the vivoChip
Practical applications of the vivoChip
Future directions
Acknowledgments
References
Chap10
9 Gut-on-a-chip microphysiological systems for the recapitulation of the gut microenvironment
Introduction
Gut models with 3D structures mimicking intestinal epithelial layer topology
Microfluidics-based gut models for mimicking the dynamic environment
First-pass metabolism models
Gut microbe coculture models
Conclusion
Acknowledgments
References
Chap11
10 Computational pharmacokinetic modeling of organ-on-chip devices and microphysiological systems
Introduction
Drug discovery and development protocols
Organ-on-chip technology and prospects for human-on-a-chip devices
Computational models for in vitro and in vivo pharmacology
Computational models for designing organ-on-a-chip devices
Modeling transport in conventional in vitro cell cultures
Modeling metabolic support in in vitro cell cultures
Modeling perfusion in organ-on-a-chip devices
Modeling peripheral components of microfluidic organ-on-a-chip devices
Models of drug pharmacokinetics in organ-on-a-chip devices
Compartmental models of transport in organ-on-a-chip devices
Spatial models of transport in organ-on-a-chip devices
Multiscale models of organ-on-a-chip devices
Models of drug pharmacokinetics in microphysiological systems
Models based on in vitro-to-in vivo translation
Conclusions and future perspectives
Acknowledgments
References
Further reading
Chap12
12 Design and engineering of multiorgan systems
Motivation for in vitro multiorgan systems
Scope of the chapter
Concepts of multiorgan systems
Interconnection of chip perfusion chamber
Monolithic design
Microfluidic system with plug-and-play organs
Building blocks for multiorgan systems
Organ models
Suspension models
Plug-in models
Integrated models
Interconnections
Flow actuation and circulation
Monitoring and sensors
Discussion
Expanding organ-on-chips toward multiorgan systems
Physiological scaling, order, and flow distribution
Interconnection versus monolithic systems
Well-orchestrated fabrication and maturation of individual organ models
Quality control prior to interconnection (reproducibility and system yield)
Timing and setup (logistics, supply chain, dependency on third-party delivery)
Modularity
Readout methods
Optical access
Access to medium
Access to cells
Operations complexity and robustness
Scalability and parallelization
Industry standards
Conclusion
References
Chap13
13 Human body-on-a-chip systems
Introduction
Why we need organismal models on a chip
Design principles
Size matters
From monolayer fluidics to multi-compartmental platforms
Enhancing the circulatory system
Opportunities
Challenges
Conclusion
References
Chap14
14 Automation and opportunities for industry scale-up of microphysiological systems
High-throughput versus high-content systems
History of laboratory automation
The purpose of microphysiological system and their suitability for automation
Automation of device production
Automation of tissue preculture and loading
Automation of system operation
Medium circuit flow
Oxygen supply
Incubation and pH-stabilization
Minimizing contamination and evaporation
Executing handling steps
General observation of constant cell culture conditions
Automation of monitoring and sensing
Summary and outlook
References
Further reading
Chap15
15 How to build your multiorgan-on-a-chip system: a case study
Introduction
Selecting the appropriate models and coculture medium
Choosing the appropriate biological model
Identifying and evaluating a coculture medium
Multi-organ-on-a-chip development project
Analysis of project needs and rationalization of solutions
Statement of need
Functional analysis of the system in its environment
Functional analysis related to system state change
Rational choice of solutions identified during functional analyses
Functional block diagram of the multi-organ-on-a-chip system
Function analysis system technique method
Project realization using agile development methods
Adaptation of agile method to product design
Redistribution of functional blocks into sprints
Sprints realized for the multiorgan-on-a-chip system
Group 1: pump unit development
First sprint: proofs of concept of a single pump unit in its environment
Second sprint: proofs of concept of the pump unit in its environment
Fourth sprint: multiplication of pump units per incubator
Group 2: pump controller development
Third sprint: flow rate control for a single pump unit
Fourth sprint: flow direction control for multiple pump units
Group 3: chip plate development
First sprint: proofs of concept of a chip plate design and biocompatibility testing
Third sprint: biocompatibility of the chip plate
Fifth sprint: testing a new well design adapted for culturing liver spheroids
Sixth sprint: catalog of chip plates
Final version of the multiorgan-on-a-chip system
Testing the multiorgan-on-a-chip system
Evaluating tissue stability
Fit-for-purpose testing of the multiorgan-on-a-chip system
Conclusion
Acknowledgments
References
Recommend Papers

Organ-on-a-chip: Engineered Microenvironments for Safety and Efficacy Testing [1 ed.]
 0128172029, 9780128172025

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

CHAPTER

Need for alternative testing methods and opportunities for organ-on-a-chip systems

1

J. Malcolm Wilkinson Technology For Industry Ltd., Chesterfield, United Kingdom

Introduction Two-dimensional (2D) in vitro cell culture systems are a poor representation of human or animal physiology (Kirkpatrick et al., 2007), because they fail to replicate the complexity of the physiological environment in Petri dishes or microplates (Zhang, 2004). Cells are sensitive to their microenvironments, which are rich in molecular signals from the extracellular matrix, other cells, and mechanical stimuli induced by flow, concentration gradients, and movement. These mechanical and biochemical signals are almost completely absent from static cultures in well plates. One method for recreating the three-dimensional (3D) environment is to seed cells at a higher density on scaffolds. However, at this higher cell density, the supply of nutrient and oxygen becomes critical, particularly for culture experiments that last several days. Media flow can be introduced to overcome this limitation but renders the design of the cell culture chamber far more complex to predict and control flow-induced stress. With flow systems, practical issues, such as avoiding leakage and blockages, must also be overcome. Once the flow is introduced, multiple chambers can be coupled to enable the construction of more sophisticated coculture models and studies of crosstalk between various tissues (Mazzei et al., 2010). The interest in flow and coculture has developed parallelly with the concept of organ-on-a-chip (OOC) devices that incorporate microfluidics. Because of the widespread industrial use of 96 and 384 well plates or microtiter plates, it was considered that a worthwhile goal would be to scale the cell culture chambers to similar small dimensions. Although there are intense research-and-development efforts in this direction, it has proved difficult to translate experimental methods from the millimeter to the micrometer scale because of practical problems such as blockages, air bubbles, and loading cells into microscopic chambers. Since OOC devices do not actually aim to recapitulate a complete organ, an alternative description, “microphysiological systems,” is coming into use. For biologists and laboratory technicians to embrace these new, physiologically more relevant culture methods, the transition from current wells and dishes Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00001-2 © 2020 Elsevier Inc. All rights reserved.

1

2

CHAPTER 1 Need for alternative testing methods and opportunities

to other tools must be simple and inexpensive. Ideally, the use of existing protocols and equipment should be maximized to allow third-party laboratories or academic laboratories to adopt microscale devices. Some organ-on-a-plate approaches, scaled slightly larger than OOC systems, are being developed by TissUse GmbH in Germany (Dehne et al., 2017) and Kirkstall Ltd. in the United Kingdom (Ahluwalia et al., 2011). Multiple cell types have been successfully cultured in these devices, including hepatocytes (Vinci et al., 2011), Caco-2 gut cells (Ucciferri et al., 2013), adipocytes, and endothelial cells (Vinci et al., 2012). Current work is extending the range of applications and cell types to skin, kidney, respiratory epithelium, and the blood brain barrier. The companies and laboratories developing smaller scale OOC devices are also making rapid progress in widening the range of cell models used in-house, but with less success in transferring these developments to the third parties.

Why we need alternative and improved methods The justification for change stems from economic, ethical, and scientific arguments. There is a clear market need for improvements in the drug discovery and development process in the pharmaceutical industry. Although the development of a drug takes, on average, 13.5 years and costs $2.5 billion, 92% of drugs fail in human-clinical trials and never reach the market (Maschmeyer, 2019). Systemic, human cell-based models that better reflect human physiology are therefore urgently needed, and organ-on-a-plate and OOC devices may save hundreds of millions of dollars. The ethical arguments relate to the use of animals, a large number being sacrificed in experiments. This involves not only discomfort and suffering for the animals but also stress for the human researchers carrying out these experiments. The animal experiments have an enormous economic cost invested (Bottini and Hartung, 2009) in breeding, housing, and disposal burdens. The scientific arguments arise from the recognition that there are differences between human and animal biology, even for primates. Hence, the findings obtained from animal tests do not translate into the clinic (Pistollato et al., 2014). The years of wasted research also exert an economic cost. Many animals are bred in sterile conditions and neither do they develop an immune response that is necessary to model disease (Landhuis, 2016) nor do they possess a balanced gut microbiome, affecting drug metabolism (Simon et al., 2019). In vitro testing of the activity (toxicity or efficacy) of chemical compounds needs to accurately predict what will happen in the clinic. Problems arise when a test gives a false positive (toxic effect where the compound is actually safe) or false negative (no adverse reaction detected where the compound is toxic). Since many compounds are safe at low dose but toxic at high dose, the sensitivity of the test is critical. These issues have been reviewed by Proctor et al. (2017) for

Requirements for in vitro alternatives to animal testing

the particular case of liver toxicity. Few of the in vitro models contain the full complement and functionality of metabolic enzymes and transporters present in human hepatocytes in vivo. 2D cultures of plated primary human hepatocytes rapidly lose liver phenotype and CYP450 activity in traditional monolayer cultures. These factors significantly limit the ability of these platforms to detect metabolite-induced cytotoxicity as well as the effects of the parent drug and its metabolites on bile-acid homeostasis/intrahepatic cholestasis and mitochondrial impairment. Several improvements in in vitro methods have been identified to yield more physiologically relevant results. These include the transform to 3D cultures, the use of human primary cells, the introduction of flow and mechanical stimulation, and coculturing multiple cell types. 3D in vitro methods are now more widely adopted (Gaskell et al., 2016) and have been shown to be more effective as toxicity screens than simple 2D cultures. Better methods for testing drugs, nutraceuticals, and cosmetics are still needed, however, and the shift to patient-specific medicines and individually tailored therapies will demand new methodologies as well.

Requirements for in vitro alternatives to animal testing An alternative method must meet the following requirements:

• • • • •

Fulfilling the required function Exhibiting correct and physiologically relevant human biology Robust and repeatable Ability to scale to the required throughput (e.g., the number of compounds that can be tested at a given time and at a given cost) Low startup and recurring consumable costs, to justify the change to a new methodology

Meeting functional requirements for improved in vitro methods A growing body of evidence shows that the use of animal cells in vitro contributes to the poor performance of the current methods. Even the use of whole animal models does not replicate the in vivo human situation, so it is not surprising that animal cells in an in vitro environment yield misleading results (Zeeshan et al., 2018). The choice to use animal cells is often driven by convenience rather than scientific reasons. Human cells are difficult to obtain, are often derived from a single-diseased patient, and are not representative of a larger pool of donors. Cell lines derived from human cells are more readily available, but the cell lineage may be problematic. Tumor-derived cell lines proliferate readily, but their functionality may differ from that of healthy tissue, and their robustness may undermine a sensitivity test for toxicity of a chemical or drug. Even when a

3

4

CHAPTER 1 Need for alternative testing methods and opportunities

representative supply of cells has been secured, the models may be inadequate. Current research indicates that 2D static cell cultures with no medium flow are not as good at predicting toxicity as 3D cell cultures. Perfusion (flow) of media over or through the cells has been shown to produce a better prediction of the half-maximal inhibitory concentration of a drug than static immersion in medium (Davidge and Bishop, 2017). Building on this research, we can set out a list of requirements for any advanced in vitro method, including OOC devices.

Correct and physiologically relevant human biology Animal cells may be easier to obtain and maintain than human primary cells, but in no way can they advance our understanding of human disease and toxicity mechanisms. Human-tumor-derived cell lines are easy to culture but are not representative of healthy tissue. Human-induced pluripotent stem cells appear promising but are currently expensive to culture and require long, complex protocols to derive the differentiated cells needed for organ models. Human-donor tissue could be considered the gold standard, but cryopreservation is required to store such tissues prior to experiments and can compromise the cellular function. A review of the cell types used in OOC models is available elsewhere (Esch et al., 2015). Once the appropriate cells have been selected, they must be cultured under conditions that produce physiologically relevant organoids. Cells under static conditions (no flow) grown on plastic rapidly change their shape and function and no longer represent human tissue (Maltman et al., 2010). 3D cultures on scaffolds or with an extracellular matrix exhibit better performance. In the body, cells do not exist in isolation but exchange molecular signals with cells from other organs. An ideal model of human biology would then need to include connected organoids, so that the system models the whole organism. There is a clear trade-off in OOC platforms between complexity and accuracy. High-throughput screening (HTS) typically relies on short-term culturing and exposure to the compounds of interest. Long-term culturing and homeostasis are important in repeat-dose testing or testing low-clearance compounds. Models for studying cancer and lung and neurological diseases must support longer testing periods and ideally be fully immunocompetent.

Robust and repeatable methods For any new technology to achieve regulatory acceptance, it must demonstrate robustness and repeatability. Many OOC methods are a long way from this goal, being complex to set up and operate and therefore precluding widespread adoption. Kirkstall Ltd. has designed its Quasi Vivo organ-on-a-plate platform to be easy to use and fast to set up in the laboratory. This system is targeted at the academic market and has already shown that results are repeatable across multiple laboratories, with a current academic user base of more than 70 universities.

Requirements for in vitro alternatives to animal testing

Mundane practical issues can occasionally derail sophisticated equipment. In microfluidic systems connected by capillaries or channels smaller than 100 µm in diameter, air bubbles can disrupt the flow, and cellular material can cause blockages.

Ability to scale Fig. 1.1 shows the stages in the drug discovery and development process where OOC devices could be leveraged. There is a clear divergence between the requirements for HTS and the animal replacement. The former application needs to screen thousands of compounds and improves accuracy (fewer false positives and false negatives). The latter needs to test tens of compounds in depth and replace hundreds of animals used in preclinical screening. Most of the current OOC developments indicate that screening large numbers of compounds is their commercial goal. In contrast, TissUse and Kirkstall have opted for 24-well plate-size chambers that should be more suited to in-depth studies and a focus on animal replacement.

Low capital and consumable costs Since so few of the OOC projects have reached the market, it is difficult to assess the costs likely involved. Many could be scaled to volume production and so, in

FIGURE 1.1 Drug discovery and development process. The different stages in the process that are being targeted by OOC developers. Developed from an original by Clerk, S., Villien, M., 2017. Organs-On-Chips 2017, Market & Technology. Report available from Yole Developpement, Le Quartz, 75 Cours Emile Zola, 69100 Lyon-Villeurbanne, France. Yole Developpement.

5

6

CHAPTER 1 Need for alternative testing methods and opportunities

theory, should meet customer expectations on cost. The economics of animal replacement have been thoroughly researched and summarized (Bottini and Hartung, 2009). The acceptable capital and consumable costs for a product to replace animal testing are likely higher than costs to replace HTS. However, the throughput (number of compounds that can be tested at a given time) is considerably higher, suggesting that the market for advanced cell cultures and OOC devices could become segmented into high-throughput approaches and highcontent (lower throughput) methods.

Remaining technical challenges Scaling the biology down to the microlevel Individual cells can be seeded into the micrometer-sized space in an OOC device, but there is not much scope for growing functional organoids or tissue slices. Once the cells form a 3D structure resembling tissue, insufficient oxygen and nutrient levels will influence cellular function and can even lead to a necrotic core at the center of the cell mass (Berger et al., 2018). Perfusing with a medium can largely overcome this but raises other questions for which data are not necessarily available yet: what flow stresses, velocities, and pressures are actually experienced by cells and tissue in vivo? Gut and kidney models are being used to investigate these questions on the ranges of flow characteristics (Lennon et al., 2014). Microscale devices do limit the amount of biological material available for measuring simple endpoints; this may be acceptable in HTS, which has long used 96-well plates, but other techniques, such as Western blot, require more cellular material.

Coculturing cells Coculturing cells or tissues is essential to constructing in vitro models that are representative of the in vivo situation, as cells communicate to induce and modulate functions and metabolism in other parts of the body. A simple way to culture multiple cell types together is to place them in a single chamber, but this is less representative of the in vivo situation than placing cells in separate chambers connected by perfusion (flowing media). Perfusion also introduces other aspects, such as nutrient supply and laminar flow stimulation that better mimic and recapitulate the in vivo environment in vitro.

Physiological relevance A workshop held in the United Kingdom addressed the progress of OOC devices toward more physiologically relevant models (National Centre for the Replacement, Refinement & Reduction of Animals in Research and Medical

Remaining technical challenges

Research Council Centre for Drug Safety Science, 2018) and included an overview of OOC technology and its utility presented by Gianni Dal Negro. Producing a representative, validated, and qualified 3D cell model is challenging, but this holds the promise of a positive impact across the drug-discovery pipeline from target identification and validation through efficacy and safety assessment. The model must be relevant to the specific biological question under study, and one model cannot answer all questions. Currently available models typically lack integrated physiology and longitudinal (time course) measurement capacity. The challenges to overcome include physically relevant cell interactions, scaling ratios between organs, incorporation of immune or endocrine systems, and the requirement for a common signal-carrying medium flowed appropriately between the components. Malcolm Haddrick reviewed the progress toward connected systems. Although 4-, 7-, and 10-organ cultures have been connected in a microfluidic setup involving subcircuits and tunable flow rates, prolonged viability has only been demonstrated thus far for individual organs (Esch et al., 2015). Cell models must be specialized and incorporate mixed populations of cells capable of some functions of the organ they represent. Primary tissue is accessible, but reproducibility and scalability are problematic. Induced pluripotent stem-cell-derived cells often exhibit an immature phenotype, limiting their potential as alternatives to primary tissue. To monitor the health of the cell models and interpret their biological responses will require on-chip label-free, real-time biosensors.

Regulatory view and validation At the same workshop, a regulatory view of OOC technologies was outlined by David Jones of the UK Medicines and Healthcare products Regulatory Agency. In the future, human-based OOC devices may be applied in human-clinical trials with improved safety and efficacy profiles. OOC technology is promising but requires validation and improved translational understanding; the systems do not yet fully mimic human-organ physiology because they lack endocrine and immune responses. Moreover, in vivo toxicity and human-disease processes are not fully understood. David Hughes, director of the OOC company CN Bio Innovations, commented that validation of new methods will involve testing appropriate numbers of relevant annotated compounds. In the United States, the National Institutes of Health established tissue-chip testing centers to independently generate data from various platforms to corroborate manufacturer claims and stimulate wider acceptance and use of these devices. Regulatory agencies such as the US Food and Drug Administration have also begun to evaluate these technologies.

Choice of materials for organ-on-a-chip systems Historically, most OOC systems have used the silicon-based polydimethylsiloxane, as it is easy to mold. However, this material can interact with both the

7

8

CHAPTER 1 Need for alternative testing methods and opportunities

biological sample and quite a number of drugs. It is also gas-permeable, which precludes experiments where oxygen tension should be controlled, as the medium in the chamber will equilibrate with ambient gas in the incubator within seconds. Varying the oxygen tension to create aerobic or anaerobic conditions is a key requirement for a zonal liver model, for example (Tomlinson et al., 2019). The choice of materials for OOC is extensively covered in the Chapter 3 of this book.

Conclusions It is apparent that the OOC technology is developing rapidly fueled by the significant government grants and commercial venture capital funding. There are real but potentially diverging requirements between the need for incremental improvements for in vitro methods and the potentially disruptive shift away from the use of animals. In common with many earlier technology-driven developments, there is a clear need for standards to emerge. These will facilitate the adoption of the technology and ensure that users have a choice between competing methods to meet the specific application needs. Much of the current interest and excitement about OOC technology has been fueled by marketing hype and may soon be replaced by disillusionment unless practical working systems are developed.

Outlook: stimulating the adoption of new technologies to replace animal testing A possible divergence in the market for OOC systems was mentioned earlier in the chapter. Here, we focus on the animal-replacement opportunity in the academic community. Although there is a clear need in the industrial sector for improvements in the field of drug discovery and development process, the situation and motivation in universities is different. There is a surprising level of inertia in the academic world; many careers have been built on the use of animal models, and it is possible that only a new generation of researchers will adopt different methods. The career paths of academics are driven by their ability to attract grants and to publish their work. Hence, researchers depend on the availability of grants and willingness of reviewers to consider alternatives to animal testing. The peer-review system for awarding grants and controlling approval for publication will take time to change. It is only over the past decade or so that academic centers of excellence have emerged to support animal-replacement technologies. The Center for Alternatives to Animal Testing at Johns Hopkins University in the United States was one of the first, and now has a satellite location at the University of Konstanz in Germany. The United Kingdom has the Animal Replacement Centre of Excellence at Queen Mary University, London, and

Outlook: stimulating the adoption of new technologies

Canada launched the Canadian Centre for Alternatives to Animal Methods at the University of Windsor. These centers act as nuclei for creating further awareness, funding research, evaluating technology, and supporting industry. Their most effective contribution may be to train a new generation of researchers in this transformative technology as agents for change. Activities in academia and industry can synergize to support the shift to nonanimal methods. Fig. 1.2 describes an innovative approach to synthesizing a transformative change using a number of small, incremental steps. The foundation begins with good science from a few opinion leaders in academia and continues with the creation of centers of excellence that eventually drive widespread adoption of the new paradigm. Of note, academic research accounts for 49% of the

FIGURE 1.2 A roadmap and strategy for accelerating the adoption of alternative methods, emphasizing the important role of the academic community and the incremental steps in the paradigm shift.

9

10

CHAPTER 1 Need for alternative testing methods and opportunities

total number of animals used in the United Kingdom (UK Government, 2017). Industry adoption follows, but it is slower at first because of the need for extensive evidence to support claims of superiority for the new technology. The early evidence comes from academia and is followed by the development of robust protocols by contract research organizations. The pharmaceutical industry is increasingly using these organizations to perform validation and development work that may previously have been done in-house.

References Ahluwalia, A., Vozzi, F., Montemurro, F., Wilkinson, M., 2011. Hepatotoxicity of diclofenac in a Quasi-Vivot multicompartment bioreactor. Toxicol. Lett. 205 Supplement, S270. Berger, E., Magliaro, C., Paczia, N., Monzel, A.S., Antony, P., Linster, C.L., et al., 2018. Millifluidic culture improves human midbrain organoid vitality and differentiation. Lab Chip 18, 3172 3183. Bottini, A.A., Hartung, T., 2009. Food for thought . . . on the economics of animal testing. Altern. Anim. Exp. 26 (1), 3 16. Available from: https://doi.org/10.14573/ altex.2009.1.3. Davidge, K.S., Bishop, J., 2017. Improved IC50 prediction using Quasi Vivo® SOT 2017 Conference Proceedings, Poster 3230. Baltimore, USA. Dehne, E.M., Hasenberg, T., Marx, U., 2017. The ascendance of microphysiological systems to solve the drug testing dilemma. Future Sci. 3 (2), FSO0185. Esch, E.W., Bahinski, A., Huh, D., 2015. Organs-on-chips at the frontiers of drug discovery. Nat. Rev. Drug Discovery 14 (4), 248. Gaskell, H., Sharma, P., Colley, H.E., Murdoch, C., Williams, D.P., Webb, S.D., 2016. Characterization of a functional C3A liver spheroid model. Toxicol. Res. 5 (4), 1053 1065. Kirkpatrick, C.J., Fuchs, S., Hermanns, M.I., Peters, K., Unger, R.E., 2007. Cell culture models of higher complexity in tissue engineering and regenerative medicine. Biomaterials 28, 5193 5198. Landhuis, A., 2016. Lab mice are poor models of the human immune system. Sci. Am. 315, 12 13. Lennon, R., Randles, M.J., Humphries, M.J., 2014. The importance of podocyte adhesion for a healthy glomerulus. Front. Endocrinol. 5, 160. Maltman, D.J., Przyborski, S.A., 2010. Developments in three-dimensional cell culture technology aimed at improving the accuracy of in vitro analyses. Biochem. Soc. Trans. 38 (4), 1072 1075. Available from: https://doi.org/10.1042/BST0381072. Maschmeyer, I., 2019. Is there an end in sight for animal testing? Altern. Anim. Exp. 36 (1), 142. Mazzei, D., Guzzardi, M.A., Giusti, S., et al., 2010. A low shear stress modular bioreactor for connected cell culture under high flow rates. Biotechnol. Bioeng. 106, 127 137. National Centre for the Replacement, Refinement & Reduction of Animals in Research and Medical Research Council Centre for Drug Safety Science, 2018. Organ-on-a-chip

References

technologies (OOAC): current status and translatability of data. Retrieved from ,[email protected].. Pistollato, F., et al., 2014. Alzheimer disease research in the 21st century: Past and current failures, new perspectives and funding priorities. Oncotarget 7 (26). Available from: https://doi.org/10.18632/oncotarget.9175. Proctor, W.R., Foster, A.J., Vogt, J., Summers, C., Middleton, B., Pilling, M.A., et al., 2017. Utility of spherical human liver microtissues for prediction of clinical druginduced liver injury. Arch. Toxicol. 91, 2849. Available from: https://doi.org/10.1007/ s00204-017-2002-1. Simon, J.-C., Marchesi, J.R., Mougel, C., et al., 2019. Host-microbiota interactions: from holobiont theory to analysis. Microbiome 7, ISSN:2049-2618. Tomlinson, L., Hyndman, L., et al., 2019. In vitro Liver Zonation of Primary Rat Hepatocytes. Front. Bioeng. Biotechnol 18. Available from: https://doi.org/10.3389/ fbioe.2019.0001. Ucciferri, N., Collnot, E.M., Gaiser, B.K., Tirella, A., Stone, V., Domenici, C., et al., 2013. In vitro toxicological screening of nanoparticles on primary human endothelial cells and the role of flow in modulating cell response. Nanotoxicology 8, 697 708. UK Government, 2017. Annual Statistics of Scientific Procedures on Living Animals, Great Britain 2017 Ref: ISBN 978-1-78655-687-5, HC1369 2018-19. Vinci, B., Duret, C., Klieber, S., Gerbal-Chaloin, S., Sa-Cunha, A., Laporte, S., et al., 2011. Modular bioreactor for primary human hepatocyte culture: medium flow stimulates expression and activity of detoxification genes. Biotechnol. J. 6, 554 564. Vinci, B., Murphy, E., Iori, E., Meduri, F., Fattori, S., Marescotti, M.C., et al., 2012. An in vitro model of glucose and lipid metabolism in a multicompartmental bioreactor. Biotechnol. J. 7, 117 126. Zeeshan, A., Chandrasekera, C., Pippin, J., 2018. Animal research for type 2 diabetes mellitus, its limited translation for clinical benefit, and the way forward. Altern. Lab. Anim. 46 (1), 13 22. Zhang, S., 2004. Beyond the petri dish. Nat. Biotechnol. 22, 151 152.

11

CHAPTER

Cell sources and methods for producing organotypic in vitro human tissue models

2 Patrick J. Hayden

MatTek Corporation, Ashland, MA, United States

Introduction The goal of organ-on-a-chip (OoC) technology is to reproduce key human organ systems using miniaturized in vitro cultures that are equivalent to at least the smallest functional unit of each organ (Dehne et al., 2017; Ronaldson-Bouchard and and Vunjak-Novakovic, 2018; Huh et al., 2011). The organ equivalents incorporated into the chips do not necessarily resemble their in vivo counterparts in a visual sense, but will reproduce the essential functions of the organ, and will ideally also incorporate any relevant physical/mechanical features [e.g., threedimensional (3D) extracellular environment/architecture, stretching, contraction, fluid flow, and shear forces] that contribute to organotypic differentiation and organ functions such as breathing, cardiac beating, and blood flow. OoC platforms should also incorporate sensors or features that allow measurement of relevant functional parameters (e.g., compatibility with imaging devices, sensors for measuring real-time conditions, and ports for removal of media/tissues for downstream analysis). A wide variety of OoC platforms have been developed, ranging from systems that incorporate multiple repetitions of a single organ for highthroughput screening applications to systems that incorporate several interacting organs (Fig. 2.1). The ultimate vision for OoC platforms is to create a human-ona-chip device that will replicate all key organ systems and physiologically relevant interactions of the human body (Fig. 2.2). OoC technologies are expected to facilitate the transition from animal-based models to human-based models, provide faster and more predictive human toxicity assessments, and lead to faster development of effective human therapeutics. A requisite for development and widespread use of human OoC technologies is reliable access to human tissues and cells. This chapter provides a detailed survey of human tissue/cell sources and isolation methods, including methods that use induced pluripotent stem cells (iPSCs) currently available to researchers. State-of-the-art culturing devices and techniques that can be applied to OoC Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00002-4 © 2020 Elsevier Inc. All rights reserved.

13

FIGURE 2.1 (A) OrganoPlate 3-lane microfluidic tissue culture device containing 40 independent microfluidic culture chips. (B) TissUse Multi-Organ-Chip platform containing 4 interacting organ models. (A) (Courtesy MIMETAS); (B) Reproduced with permission from Maschmeyer I., Lorenz A.K., Schimek K., Hasenberg T., Ramme A.P., Hu¨bner J., et al., A four-organ-chip for interconnected long-term co-culture of human intestine, liver, skin and kidney equivalents, Lab Chip 15(12), 2015a, 2688 2699. https://doi. org/10.1039/c5lc00392j.

Human tissue/cell sources and isolation methods

FIGURE 2.2 Human-on-a-chip concept: in vitro platform reproducing all key organ tissues and physiological interactions. (Courtesy TissUse, Gmbh)

systems—such as the use of nanopatterned substrates, tunable elastic substrates, air liquid interface cultures, spheroid/organoid culture techniques, and 3D bioprinting—are also described.

Human tissue/cell sources and isolation methods Continuous (immortal) cell lines Human somatic cells are in general capable of only a limited number (40 60) of cell divisions in culture before they become senescent and lose their ability to divide (Hayflick and Moorhead, 1961). However, immortal somatic cell lines with a capacity for unlimited division potential can be obtained by a number of processes. Genetically mutated cells derived from cancerous tissues are a common source for establishing immortal cell lines. In rare cases, cultured normal (noncancer-derived) cells may spontaneously acquire genetic mutations that provide the ability for unlimited growth. Normal cells may also be transformed into immortalized cells by the introduction of viral oncogenes such as EBV, SV40LT, HPV16 E6/E7, and Ad5 E1A (Honegger, 2001; Freshney, 2016). Induction of telomerase activity by transduction of human Telomerase reverse transcriptase (hTERT) into cells can induce immortal transformation while retaining more normal cell phenotypes than viral-induced transformations (Freshney, 2016). The development of methods for establishing and maintaining continuous cell lines in culture marked a revolutionary advancement in biology. Since the establishment of HeLa cells as the first immortal human cell line in 1952

15

16

CHAPTER 2 Cell sources and methods

(Gey et al., 1952), continuous cell lines have become widely used as indispensable and inexpensive tools for basic biological research, chemical metabolism and toxicity tests, and production of biological compounds such as vaccines, antibodies, and therapeutic proteins. Numerous immortal cell lines derived from a wide variety of tissues are now readily available. Key advantages of immortal cell lines are that they are affordable, well characterized, and easy to culture. However, immortal cell lines generally exhibit significant genotypic and phenotypic abnormalities that may limit their ability to reproduce normal cell behavior and may undergo additional genotypic or phenotypic drift with continued long-term passaging. Furthermore, many continuous cell lines have been misidentified or have become contaminated with mycoplasma or other cell lines over time (Geraghty et al., 2014; Lorsch et al., 2014). Authentication of cell lines is now recommended or required for publication or submission of research results to regulatory authorities (Geraghty et al., 2014). Despite their shortcomings, immortal cell lines remain in widespread use and will continue to be important biological tools that may be suitable or advantageous for specific OoC applications. Table 2.1 presents a list of commonly used immortal human cell lines derived from a variety of organs. These cell lines, as well as others, are also available as authenticated and quality-controlled resources from a number of nonprofit repositories (Table 2.2).

Primary cell cultures and early-passage cell lines with finite lifespan Cells obtained directly from fresh tissue are commonly termed as primary cells. Advancements in the development of defined culture media, culture conditions, and matrix requirements have led to an increasing ability to culture many types of normal (nonimmortal) primary cells. With the exception of hematopoiesis-derived cells, which may be cultured as cell suspensions, most primary cells require attachment to a substrate to survive and proliferate. Adherence-dependent primary cell cultures may be initiated by explanting small pieces of tissue into a culture plate with the appropriate medium and allowing cells to migrate and proliferate as monolayer cultures. Alternate methods of initiating primary adherent cell cultures involve mechanical and/or enzymatic disaggregation of tissue to form a cell suspension, following by plating the suspension at a low density onto cell culture plates or flasks. Coating the culture plate with various forms of extracellular matrix material or the presence of an established feeder cell layer may be required to support culture establishment when using either the explant or disaggregation methods (Honegger, 2001; Freshney, 2016). Primary cultures obtained by seeding cells or explanted tissue fragments directly after isolation from fresh tissue will consist of those cells that are capable of attachment and survival under the culturing conditions (e.g., culture medium and extracellular matrix coating). The primary culture will consist of a mixed

Table 2.1 Commonly used continuous (immortal) human cell lines. Cell type

Origin

Name

Reference

Endothelial Hepatic

Liver Liver

SK HEP-1 HepaRG

Epithelial Epithelial Epithelial Keratinocyte Type I pneumocyte Type II pneumocyte Type II pneumocyte Epithelial Epithelial Epithelial Epithelial Epithelial Epithelial Epithelial Glial

Liver Breast Breast Epidermis Lung

HepG2 MCF-7 ZR-75-1 HaCaT hAELVi

Heffelfinger et al. (1992) Parent et al. (2004), Guillouzo et al. (2007), and Takahashi et al. (2015) Diamond et al. (1980) Brooks et al. (1973) Engel et al. (1978) Boukamp et al. (1988) Kuehn et al. (2016)

Lung

A549

Giard et al. (1973)

Lung

NCI-H441

Brower et al. (1986)

Lung Lung Kidney Ovary Colon Colon Cervix Glioma

Reddel et al. (1988) Fogh and Trempe (1975) Graham et al. (1977) Tsuruo et al. (1986) Fogh et al. (1977) Fogh and Trempe (1975) Gey et al. (1952) Balmforth et al. (1986)

Glial Lymphocytic Myeloid Myeloid Myeloid Myeloid

Brain Blood Blood Blood Blood Blood

BEAS-2B Calu-3 HEK-293 A2780 Caco-2 HT-29 HeLa MOG-GCCM U-251 MG EB-3 K562 HL-60 THP-1 U937

Ponten and Macintyre (1968) Epstein and Barr (1964) Andersson et al. (1979) Olsson et al. (1981) Tsuchiya et al. (1980) Sundström and Nilsson (1976)

Adapted from Freshney, R.I. (2016). Culture of Animal Cells: A Manual of Basic Technique and Specialized Applications (7th ed.). NJ: Wiley-Blackwell, with additions.

Table 2.2 Nonprofit immortal cell line repositories. Institution

Headquarters

Website

ATCC

Manassas, VA, United States Australia, Sydney, Australia, Salisbury, Wiltshire, United Kingdom

www.atcc.org

CellBank ECACC

The Leibniz Institute DSMZ: German Collection of Microorganisms and Cell Cultures GmbH (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH) JCRB Cell Bank

Braunschweig, Germany

Ibaraki City, Osaka, Japan

www.cellbankaustralia. com www.pheculturecollections.org.uk/ collections/ecacc.aspx www.dsmz.de

www.cellbank.nibiohn. go.jp

ATCC, American Type Culture Collection; ECACC, The European Collection of Authenticated Cell Cultures; JCRB, Japanese Collection of Research Bioresources.

18

CHAPTER 2 Cell sources and methods

population of cells at various stages of differentiation and, for certain tissue types, tissue-specific stem cells with proliferative potential. Cells that are already committed to terminal differentiation may attach and remain viable but will not proliferate further in culture. The proliferation of tissue-specific stem cells will continue until space in the culture vessel becomes limited, at which point the cells may be harvested and passaged into fresh culture vessels. Cells are typically passaged while still in log-phase growth and before reaching confluence, to avoid loss of proliferative capacity owing to contact inhibition that may occur when the cultures become too densely packed. Adherent cells are generally harvested from culture using enzymatic reagents (trypsin or Accutase) or by manual scraping. Primary cells that have been passaged are thereafter termed primary cell lines. The passaged lines will be enriched in cells that have adapted well to the specific culture conditions and those that retain proliferative capacity, while differentiated/nonproliferative cells will die off. With continued culture and successive passaging, the cell lines will at first continue to adapt and become further enriched in cells with proliferative capacity. However, the cells will eventually exhaust their ability to proliferate and become senescent. If the culture conditions are not specifically tuned to promote only the growth of the desired cell type, the cultures may become overgrown with unwanted cell types such as fibroblasts. Conditions that allow in vitro proliferation of many types of primary human cells [e.g., epithelial cells, stromal cells (fibroblasts, stellate cells, pericytes, astrocytes), and endothelial cells] have been successfully developed. However, certain types of cells from key organ systems (e.g., cardiomyocytes, hepatocytes, neurons, islet cells, and monocytes) can only be maintained for a limited time as primary cultures and do not have significant proliferative capacity in vitro. A number of comprehensive texts on the subject of animal and human cell isolation and cell culture techniques are available (Honegger, 2001; Randell and Fulcher, 2013; Picot, 2005; Mitry and Hughes., 2012). Table 2.3 lists protocols in the literature for the isolation and culture of organ-specific human cell types.

Human tissue sources While protocols for isolation of many organ-specific human cell types have been developed, availability and access to fresh human tissue samples may be a significant limitation for researchers outside of clinical research university or hospital settings. Access to fresh human tissues for research requires informed consent of the tissue donor and institutional review board approval (Pirnay et al., 2015). Even for researchers with access to fresh human tissues, cell isolation protocols require specialized techniques that may be difficult to master and are a time-consuming endeavor that may not be feasible or desirable for many laboratories. As an alternative, a number of vendors offer primary human cells (Table 2.4).

Human tissue/cell sources and isolation methods

Table 2.3 Human organ-specific primary cell isolation protocols. Cell type

Reference

Keratinocytes Melanocytes Breast epithelial cells Oral epithelial cells Olfactory neuroepithelial cells Female reproductive tract epithelium Prostate epithelium Coroid plexus Osteoblasts Chondrocytes Myoblasts Fibroblasts Adipocytes

Randell and Fulcher (2013) and Picot (2005) Picot (2005) and Mitry and Hughes (2012) Randell and Fulcher (2013) Randell and Fulcher (2013) Randell and Fulcher (2013)

Mononuclear phagocytes Peripheral blood natural killer cells Neuronal cells Schwann cells Tracheal/bronchial epithelial cells Alveolar epithelial cells Colon epithelial cells Hepatocytes Kupffer cells Glomerular epithelial cells Renal cortical epithelial cells Parathyroid cells Islets of Langerhans cells Corneal and conjunctiva cells Retinal pigment epithelial cells Fetal gastric epithelial cells

Randell and Fulcher (2013) Randell and Fulcher (2013) Randell and Fulcher (2013) Picot (2005) and Mitry and Hughes (2012) Picot (2005) and Mitry and Hughes (2012) Picot (2005) Picot (2005) and Mitry and Hughes (2012) Picot (2005), Mitry and Hughes (2012), Carswell et al. (2012), Brooks et al. (2017), and Ruiz-Ojeda et al. (2016) Picot (2005) Picot (2005) Picot (2005), Gordon et al. (2013), Peng et al. (2013), and Mains and Patterson (1973) Picot (2005) Randell and Fulcher (2013) and Picot (2005) Picot (2005) and Mitry and Hughes (2012) Picot (2005) Randell and Fulcher (2013), Picot (2005), and Mitry and Hughes (2012) Dixon et al. (2013) and Alabraba et al. (2007) Picot (2005) and Mitry and Hughes (2012) Randell and Fulcher (2013), Picot (2005), Mitry and Hughes (2012), Ichimura et al. (2008), and Valente et al. (2011) Picot (2005) and Mitry and Hughes (2012) Picot (2005) and Mitry and Hughes (2012) Randell and Fulcher (2013) and Picot (2005) Randell and Fulcher, 2013 Picot (2005) and Mitry and Hughes (2012) (Continued)

19

20

CHAPTER 2 Cell sources and methods

Table 2.3 Human organ-specific primary cell isolation protocols. Continued Cell type

Reference

Intestinal crypt and villus cells Ovarian cells Vascular smooth muscle cells Endothelial cells Mesenchymal stem cells Peripheral blood mononuclear cells Dendritic cells Regulatory T-cells

Mitry and Hughes (2012), Chopra et al. (2010), Gjorevski et al. (2016), Tong et al. (2018), and Holmberg et al. (2017) Mitry and Hughes (2012) Mitry and Hughes (2012) Mitry and Hughes (2012) Mitry and Hughes (2012) Mitry and Hughes (2012) Fahrbach et al. (2007) Mitry and Hughes (2012)

Table 2.4 Commercial sources of continuous cell lines and primary human cells. Vendor

Headquarters

Website

AllCells ATCC BioIVT Biopredic International Cell Systems, iXCells Biotechnologies

Emeryville, CA, United States Manassas, VA, United States Westbury, NY, United States Saint Grégoire, France Kirkland, WA, United States San Diego, CA, United States Frederick, MD, United States

www.allcells.com www.atcc.org www.bioivt.com www.biopredic.com www.cell-systems.com www.ixcellsbiotech. com www.lifelinecelltech. com www.lonza.com www.mattek.com www.emdmillipore.com

Lifeline Cell Technology Lonza MatTek Corporation MilliporeSigma Novabiosis PromoCell ScienCell Research Laboratories Sekisui XenoTech, LLC Stemcell Technologies Thermo Fisher Scientific Zen-Bio

Basel, Switzerland Ashland, MA, United States Burlington, MA, United States Morrisville, NC, United States Heidelberg, Germany Carlsbad, CA, United States Kansas City, KS, United States Vancouver, BC, Canada Waltham, MA, United States Durham, NC, United States

ATCC, American Type Culture Collection.

www.novabiosis.com www.promocell.com www.sciencellonline. com www.xenotech.com www.stemcell.com www.thermofisher.com www.zen-bio.com

Methods for producing three-dimensional “organotypic” tissue cultures

Induced pluripotent stem cells A seminal advancement in cell culture and regenerative medicine occurred in 2006 with the development of methods for generating iPSCs from differentiated somatic cells via induced expression of four transcription factors (Takahashi et al., 2007a,b; Yu et al., 2007; Sayed et al., 2016). Because adult cells are used, iPSCs avoid the restrictions and controversy surrounding human embryonic stem cells. Human somatic tissues, fluids, and cell types such as fibroblasts, blood cells, and urine have been used to generate iPSCs. The initial iPSC protocols used retroviral and lentiviral systems to integrate transcription factors into the host genome. Recently developed protocols allow the use of nonintegrating systems, including Sendai virus, episomal reprogramming factors, and microRNAs, to generate iPSCs without integrating the reprograming factors into the genome (Fusaki et al., 2009; Warren et al., 2010). Once generated, iPSCs are theoretically capable of differentiation into any cell type; iPSCs are therefore a valuable source for generation of large numbers of cells that normally do not proliferate in vitro (e.g., cardiomyocytes, hepatocytes, and neuronal cells) (McKernan and Watt, 2013). In addition, iPSCs allow researchers to recreate in vitro models of inherited genetic human diseases and enable the derivation of multiple types of organ models from the same donor (“body-on-a-chip” or “you-on-a-chip” devices). Development of protocols to drive differentiation of iPSCs into various tissue-specific lineages is an area of current intense effort. The growing list of organspecific cell types that have been generated to date from iPSCs includes hepatic, cardiac, neuronal, endothelial (including blood brain barrier), pancreatic, lung, renal, and intestinal cells (Table 2.5). iPSC technology currently requires significant expertise and is a time-consuming process. The science is still developing, and current protocols do not yet reproduce fully differentiated organ-specific cell phenotypes. A growing number of iPSC sources currently exist (McKernan and Watt, 2013; De Sousa et al., 2017; Kim et al., 2017; Ntai et al., 2017). The California Institute for Regenerative Medicine (CIRM) hPSC Repository is the world’s largest, containing iPSCs from over 3000 individuals. The CIRM iPSC lines are produced by nonintegrating episomal reprogramming. Demographic and clinical data are available for a variety of diseases or conditions affecting the brain, heart, lung, liver, and eyes. The Coriell Institute for Medical Research (Camden, NJ, United States) offers dozens of iPSC lines related to Parkinson’s disease, amyotrophic lateral sclerosis, and Huntington’s disease, and supplies iPSCs to other repositories. European-based cell banks include the European Bank for Induced Pluripotent Stem Cells (De Sousa et al., 2017), the Human Induced Pluripotent Stem Cell Initiative, and StemBANCC (Table 2.6).

Methods for producing three-dimensional “organotypic” tissue cultures Traditional cell isolation techniques and culture methods for adherence-dependent cells typically involve submersion cultures of cell monolayers on two-dimensional

21

22

CHAPTER 2 Cell sources and methods

Table 2.5 Cell types and disease models developed from induced pluripotent stem cells to date. Tissue type

References

Hepatic

Meier et al. (2017), Yamashita et al. (2018), Matoba et al. (2018), Wang et al. (2017), and Takayama and Mizuguchi (2017) Sallam et al. (2014), Jiang et al. (2016), Kamdar et al. (2015), Tanaka et al. (2015), Talkhabi et al. (2016), Ronaldson-Bouchard et al. (2018), Matsa et al. (2016), and Sharma et al. (2017) Schwartz et al. (2015), Lancaster et al. (2013), Centeno et al. (2018), Lee et al. (2017), Gabriel and Gopalakrishnan (2017), Klaus et al. (2019), Pamies et al. (2017), and Hofrichter et al. (2017) Bao et al. (2015), Wong et al. (2012), Sone and Nakao (2013), Qian et al. (2017), and Stebbins et al. (2016) Kondo et al. (2018a), Loo et al. (2018), Bose and Sudheer (2016), Yabe et al. (2017), and Kim et al. (2016) Wang et al. (2016) and Wilkinson et al. (2018) Sinagoga et al. (2018), Kondo et al. (2018b), Miura and Suzuki (2018), Takahashi et al. (2018), and Rahmani et al. (2019) Takasato and Little (2017) and Schutgens et al. (2019)

Cardiac

Neuronal

Endothelial (including blood brain barrier) Pancreatic Lung Intestinal

Renal

Table 2.6 Repositories and other sources of induced pluripotent stem cells. Institution

Website

American Type Culture Collection Boston University Center for Regenerative Medicine California Institute for Regenerative Medicine Coriell Institute for Medical Research European Bank for Induced Pluripotent Stem Cells Harvard Stem Cell Institute New York Stem Cell Foundation StemBANCC Tempo Bioscience Human Induced Pluripotent Stem Cell Initiative UK Stem Cell Bank US National Institute of Mental Health US National Institute of Neurological Disorders and Stroke

www.atcc.org www.bu.edu/dbin/stemcells/ iPSC_bank.php www.cirm.ca.gov www.coriell.org www.ebisc.org www.hsci.harvard.edu www.nyscf.org www.stembancc.org www.tempobioscience.com www.hipsci.org www.nibsc.org/ukstemcellbank www.nimhgenetics.org www.nindsgenetics.org

Methods for producing three-dimensional “organotypic” tissue cultures

(2D) plastic substrates. These methods were developed to promote proliferation of cells and generally lead to loss of differentiated cellular functions. Moreover, traditional 2D culture environments lack the important cell cell, cell matrix, 3D architecture, and mechanical cues (e.g., stretch, strain, shear forces, nanotopography, and substrate compliance) that are found in the in vivo environment of the cells and that are essential for functional differentiation (Huh et al., 2011; Schmeichel and Bissell, 2003; Alhaque et al., 2018). Recognition of the inherent limitations of 2D culture environments has motivated efforts to develop 3D cell culture conditions that better replicate in vivo tissue architectures and provide more physiologically relevant “organotypic” in vitro tissue models of normal function and disease. A variety of 3D organotypic tissue culture techniques and models have been developed over the past decades. Several of these techniques are used as key building blocks for producing tissues that may be incorporated into OoC platforms described in this book, which add organ connectivity and mechanical features such as medium flow and shear forces. The remainder of this chapter covers current approaches to producing 3D organotypic tissue models. Table 2.7 provides a list of commercial sources of cultureware, devices, and supplies. Table 2.8 lists the current commercial providers of ready-to-use 3D organotypic tissue models.

Use of controlled substrate rigidity and nanopatterned culture substrates to enhance morphological and functional differentiation Functional organotypic differentiation depends on the presence of an appropriately structured microenvironment. The mechanical properties of the matrix, such as elasticity and nanotopography, provide fundamental cues that regulate and guide cell migration, morphology, assembly and alignment, signal transduction pathways, and gene transcription (Kim et al., 2013; Park et al., 2012; Kshitiz et al., 2012). Nanopatterned culture surfaces have been designed to provide a cellular microenvironment that mimics the architecture of the native extracellular matrix. Cells can align, elongate, grow, and migrate along the nanopatterned surfaces, leading to more physiologically representative structural and functional phenotypes. For example, iPSC-derived cardiomyocytes cultured on nanopatterned culture surfaces exhibit polarized expression of gap junction proteins such as Cx43 and develop anisotropic cell shape, striated sarcomeres, and tissue-level alignment, as well as achieving enhanced baseline electrophysiology such as faster longitudinal conduction velocity and lower resting membrane potential (Fig. 2.3; Kim et al., 2010). Nanopatterned culture substrates have also been used to enhance the morphological and functional differentiation of vascular smooth muscle cells (Chaterji et al., 2014), endothelial cells (Jeon et al., 2015), osteocytes (You et al., 2010), and skeletal muscle cells (Smith et al., 2016) in vitro. Traditional tissue culture plasticware is much stiffer than the natural extracellular microenvironments found in tissues and organs. The elastic modulus of

23

24

CHAPTER 2 Cell sources and methods

Table 2.7 Commercial suppliers of culture devices and products for organotypic culture models. Company

Headquarters

Products

Website

Accela, Strašnice Advanced BioMatrix

Czech Republic Carlsbad, CA, United States

www.accela.eu www. advancedbiomatrix. com

Corning

Corning, NY, United States

ExCellness Biotech SA

Lausanne, Switzerland

Greiner Bio-One GmbH

Kremsmünster, Austria

Matrigen Life Technologies MatTek Corporation

Brea, CA, United States Ashland, MA, United States

MilliporeSigma

Burlington, MA, United States

NanoSurface Biomedical

Seattle, WA, United States

Rotary bioreactors Extracellular matrix material, scaffolds, CytoSoft elastic modulus plates, bioinks Cell culture plasticware and reagents, microporous membrane inserts, spheroid-forming plates, attachment plates Elastic substrate cultureware and products Cell culture plasticware and reagents, microporous membrane inserts, spheroid-forming plates Softwell elastic substrate plates Glass-bottomed culture dishes, nanopatterned dishes Cell culture plasticware and reagents, microporous membrane inserts, spheroid-forming plates Nanopatterned dishes

Synthecon

Houston, TX, United States Waltham, MA, United States

ThermoFisherScientific

Rotary bioreactors Cell culture plasticware and reagents, microporous membrane inserts, spheroid-forming plates

www.corning.com/ worldwide/en/ products/lifesciences.html

www.excellness. com www.gbo.com/ en_US.html

www.matrigen. com www.mattek.com

www.emdmillipore. com

www. nanosurfacebio. com www.synthecon. com www.thermofisher. com

Methods for producing three-dimensional “organotypic” tissue cultures

Table 2.8 Commercial suppliers of three-dimensional organotypic culture models. Company

Headquarters

Products

Epithelix

Geneva, Switzerland Schlieren, Switzerland Ashland, MA, United States

Lung epithelial models

InSphero MatTek Corporation Organovo

Phenion Episkin StemoniX

San Diego, CA, United States Düsseldorf, Germany Lyon, France Maple Grove, MN, United States

Liver, pancreas, and cancer spheroid models Skin, lung, corneal, intestinal, vaginal, and oral epithelial models; primary human cells Bioprinted liver and kidney models

Skin models Skin, corneal, and other epithelial models Cardiac and brain spheroid models; two-dimensional brain models

www.epithelix. com www.insphero. com www.mattek. com www.organovo. com www.phenion. com www.episkin. com www.stemonix. com

FIGURE 2.3 Alignment of cardiomyocytes on nanopatterned culture well. (Courtesy NanoSurface Biomedical).

typical cell culture plastic is on the order of 1 3 107 kPa, whereas tissues and organs have much lower values, typically 0.2 64 kPa (Wells, 2008; Charrier et al., 2018). Tissue stiffening has been shown to be a major component of aging and fibrotic diseases (Lampi and Reinhart-King, 2018; Asano et al., 2017;

25

26

CHAPTER 2 Cell sources and methods

Liu et al., 2010). Substrate stiffness has also been found to regulate proliferation and differentiation of stem cells (Engler et al., 2006; Gilbert et al., 2010; Smith et al., 2017) and the growth and migration of cancer cells (Tilghman et al., 2010). Polyacrylamide, polydimethylsiloxane, and silicon substrates can be engineered to match the compliance of the in vivo environments of various organs and disease states (Mih et al., 2011; Smith et al., 2017). Several companies offer cultureware that features physiologically compliant substrates, including Matrigen (www. matrigen.com), Advanced BioMatrix (www.advancedbiomatrix.com), and ExCellness Biotech (www.excellness.com).

Use of microporous membrane substrates to prepare cultures with polarized barrier, transport, and differentiated properties Many epithelial and endothelial tissues exist in vivo as polarized sheets or tubes that are bounded on the basolateral side by a basement membrane and have an external (e.g., skin) or luminal (e.g., airways, intestine, kidney tubules, and blood vessels) side that may be exposed to air or liquid. These types of polarized tissues are commonly modeled in vitro by means of a “raft” culture on metal mesh grids (Anacker and Moody, 2012) or on specialized microporous membrane support wells (Fig. 2.4; Adler et al., 1987; Be´rube´ et al., 2010) to create air liquid interface or liquid liquid interface environments. Tissues cultured under these conditions typically form tight junctions between cells at the apical layers and provide selective barrier function. Morphologically, these cultures may range from flattened or simple cuboidal monolayers to multilayered 3D stratified or pseudostratified epithelial structures. Differentiated organotypic functional

FIGURE 2.4 Schematic illustration of air-liquid interface tissue culture using microporous membrane culture insert. (Courtesy MatTek Corp.)

Methods for producing three-dimensional “organotypic” tissue cultures

FIGURE 2.5 Examples of organotypic air-liquid interface human skin, intestine and airway epithelial tissues. (Courtesy MatTek Corp)

attributes of these models include a stratum corneum for epidermal models (Bell et al., 1981; Asselineau et al., 1985; Cannon et al., 1994), mucus-secreting goblet cells for airway (Bolmarcich et al., 2018) or intestine (Ayehunie et al., 2018), brush border for intestine (Ayehunie et al., 2018), and beating cilia for airway (Be´rube´ et al., 2010). Multiple cell types, such as epithelial, stromal, endothelial, and immune cells, may be incorporated together to produce more complex cocultures with additional in vivo like functionality (Fizes¸an et al., 2019; Marescotti et al., 2019). Fig. 2.5 displays representative examples of epidermal, bronchial, and intestinal epithelial tissue models produced on microporous membrane inserts. These organotypic systems provide isolated apical and basolateral compartments that allow realistic in vivo like drug/environmental exposures and are very useful for studies of drug transport and efficacy (Kaluzhny et al., 2018; Ayehunie et al., 2018), toxicology and safety (Hayden et al., 2015; Gordon et al., 2015; Maschmeyer et al., 2015a,b; Peters et al., 2019; Hoppensack et al., 2014), and viral or bacterial infections (Bai et al., 2015; Maldonado-Contreras et al., 2017). Human diseases such as asthma (Bai et al., 2015; Hackett et al., 2011) and psoriasis (Chamcheu et al., 2015) have also been successfully modeled with these organotypic systems. Recently, membranes fabricated from polydimethylsiloxane were used to add in vivo like mechanical stretching features to air liquid interface systems (Fig. 2.6; Novak et al., 2018; Felder et al., 2019).

Techniques for preparing three-dimensional spheroids/organoids with differentiated organotypic functions A variety of methods have been developed for inducing scaffold-free selfassembly of cells into 3D spheroids or organoids that mimic key structural and functional attributes of the original organ (Alhaque et al., 2018; Sato and

27

28

CHAPTER 2 Cell sources and methods

FIGURE 2.6 (A) Air-liquid interface PDMS Lung-on-Chip epithelial tissue allows dynamic medium and air flow, shear stress and stretching. (B) Air-liquid interface Lung-on-Chip platform containing 12 alveolar epithelial tissues featuring membrane with in vivo-like expansion and contraction. (A) Reproduced with permission from Huh D., Matthews B.D., Mammoto A., Montoya-Zavala M., Hsin H.Y. and Ingber D.E., Reconstituting organ-level lung functions on a chip, Science 328 (5986), 2010, 1662 1668. https://doi.org/10.1126/science.1188302; (B) Courtesy Alveolix.

Clevers, 2015). These methods make use of the hanging drop technique, rotary reactor devices, embedding in hydrogel matrices, or nonadherent plates or molds. Organoid culture techniques have been applied to produce models of the liver (Meier et al., 2017; Lee et al., 2013; Bell et al., 2016; Proctor et al., 2017;

Methods for producing three-dimensional “organotypic” tissue cultures

FIGURE 2.6 (Continued).

Messner et al., 2018), pancreas (Bose and Sudheer, 2016; Kim et al., 2016; Zuellig et al., 2014), brain (Schwartz et al., 2015; Lancaster et al., 2013; Lee et al., 2017; Gabriel and Gopalakrishnan, 2017; Klaus et al., 2019; Pamies et al., 2017; Hofrichter et al., 2017), kidney (Schutgens et al., 2019), heart (Beauchamp

29

30

CHAPTER 2 Cell sources and methods

FIGURE 2.7 A process for producing cerebral organoids from hIPSCs using a spinning bioreactor. Reproduced with permission from Lancaster M.A., Renner M., Martin C.-A., Wenzel D., Bicknell L.S., Hurles M.E., et al., Cerebral organoids model human brain development and microcephaly, Nature 501, 2013, 373 379

et al., 2015; Rismani Yazdi et al., 2015), intestine (Miura and Suzuki, 2018; Takahashi et al., 2018; Rahmani et al., 2019; Sachs et al., 2017; Mu´nera and Wells, 2017), and cancer (Schmeichel and Bissell, 2003; Plummer et al., 2019; Herter et al., 2017). Coculture models of the liver have included hepatocytes, stellate cells, and Kupffer cells and are reported to maintain drug metabolism capabilities long-term and to demonstrate in vivo like inflammatory responses (Lee et al., 2013; Messner et al., 2018). Brain organoids have been constructed with glial cells, pericytes, and astrocytes, and demonstrate extensive neurite outgrowth, electrical connectivity, and myelination (Schwartz et al., 2015; Lancaster et al., 2013; Pamies et al., 2017). The size of organoid structures should be limited to diameters of ,500 nm, the limit of oxygen diffusion, to maintain cell viability and avoid cell necrosis from lack of oxygen. An example of cerebral organoid production from human iPSCs is depicted in Fig. 2.7 (Lancaster et al., 2013).

Three-dimensional bioprinting The application of 3D printing techniques to biological systems has provided unprecedented opportunities to reproduce heterogeneous tissues with a complex 3D architecture (Xia et al., 2018; Kolesky et al., 2018; Vanderburgh et al., 2017). Cells, growth factors, and matrix materials combined together in so-called “bioinks” are deposited layer by layer to produce the 3D structures. Bioprinting has leveraged inkjet, laser-assisted, and extrusion printing techniques; computeraided design files are used to direct the deposition of the cells and biomaterials

Summary/Outlook

with microscale resolution. Bioinks may be synthetic or biologically derived polymers. Common synthetic polymers used in bioink include modified poly(ethylene glycol), poly(lactic acid), poly(lactic-co-glycolic acid), and poly(e-caprolactone). Bioinks prepared from thermosensitive polymers (e.g., Pluronic F127) that can later be liquefied and removed have been used to create vasculature or tubule-like networks within bioprinted tissues (Herter et al., 2017). Biologically derived polymers commonly used in preparing bioinks include collagens, gelatin, hyaluronic acid, fibrin, alginate, and decellularized extracellular matrices (Herter et al., 2017). These polymers are often chemically modified or combined with ceramic, glass, or hydroxyapatite to provide improved mechanical strength and rheological properties. Bioprinting has been achieved and described for both hard tissues such as bone and cartilage (Kolesky et al., 2018; Mu¨ller et al., 2017; Nguyen et al., 2017) and soft tissues such as skin (Xia et al., 2018), liver (Xia et al., 2018; Ma et al., 2016; Norona et al., 2019), kidney (Xia et al., 2018; King et al., 2017; Homan et al., 2016; Homan et al., 2019; Lin et al., 2019), brain (Xia et al., 2018; Espinosa-Hoyos et al., 2018), lung (Xia et al., 2018), heart (Lind et al., 2017), intestine (Madden et al., 2018), and vasculature (Xia et al., 2018; Kolesky et al., 2016; Zhu et al., 2017). Bioprinting processes enable precise placement of multiple cell types and heterologous material components, thereby providing tremendous promise for replicating functional vasculature, tubules, and glands. Remarkable progress has been made in bioprinting, and techniques continue to develop, along with new bioink materials that are both printable and biocompatible. A variety of companies now offer bioprinting devices, bioinks, and related supplies (Table 2.9).

Summary/Outlook The fields of in vitro cell culture and organotypic model development are experiencing an exciting new era brought about by key advancements. Methods for creating immortal cell lines are well established, and methods for isolation and maintenance of primary human cells have been developed for most tissues and organs. A wide variety of immortal and primary human cells are now readily available to the research community. Use of iPSC technology enables largescale generation of differentiated cell types that were not previously available for research because of proliferation limitations. As the methods for culturing of normal human cells and differentiation of organ-specific cells from iPSCs continue to advance, reliance on immortal cell lines will likely decrease. Key advancements in novel culture substrates and scaffold materials, culturing techniques, spheroid/organoid models, and bioprinting are facilitating rapid progress in the development of in vitro models that replicate complex in vivo like structures and multicellular interactions. Major ongoing needs include

31

32

CHAPTER 2 Cell sources and methods

Table 2.9 Commercial suppliers of bioprinter devices, bioinks, and related services and supplies. Company

Headquarters

Website

3D Bioprinting Solutions Advanced Solutions Life Sciences Allevi

Moscow, Russia Louisville, KY, United States Philadelphia, PA, United States Vancouver, BC, Canada

www.bioprinting.ru www.lifesciences. solutions www.allevi3d.com

Aspect Biosystems Cellbricks CELLINK Cyfuse Biomedical K.K. EnvisionTEC nScrypt Poietis regenHU ROKIT Healthcare

Berlin, Germany Gothenburg, Sweden Tokyo, Japan Dearborn, MI, United States Orlando, FL, United States Pessac, France Villaz-Saint-Pierre, Switzerland Seoul, Korea

www.aspectbiosystems. com www.cellbricks.com www.cellink.com www.cyfusebio.com/en www.envisiontec.com www.nscrypt.com www.poietis.com www.regenhu.com www.rokithealthcare.com

improved protocols for differentiating iPSCs to more faithfully recapitulate organ-specific phenotype and functions, and universal culture media (i.e., blood surrogates) that can support multiple cell and organ types. Continued progress in these areas will lead to better organotypic differentiation, more faithful reproduction of in vivo organ function, and physiologically relevant modeling of human disease. OoC platforms and technologies will leverage advancements in cell and organotypic tissue culture to model organ organ interactions by providing organ connectivity, medium flow (artificial blood), and mechanical forces. These combined technologies are expected to greatly advance our understanding of the biology and interactions of human organs and their functions under normal and disease conditions. Organotypic human models in OoC platforms will also enable rapid in vitro assessment of absorption, distribution, metabolism, excretion, and toxicity of chemicals and drugs, thereby facilitating faster and more efficient development of human therapeutics.

References Adler, K.B., Schwartz, J.E., Whitcutt, M.J., Wu, R., 1987. A new chamber system for maintaining differentiated guinea pig respiratory epithelial cells between air and liquid phases. BioTechniques 5, 462 466.

References

Alabraba, E.B., Curbishley, S.M., Lai, W.K., Wigmore, S.J., Adams, D.H., Afford, S.C., 2007. A new approach to isolation and culture of human Kupffer cells. J. Immunol. Methods 326, 139 144. Alhaque, S., Themis, M., Rashidi, H., 2018. Three-dimensional cell culture: from evolution to revolution. Philos. Trans. R. Soc. London, Ser. B: Biol. Sci. 373 (1750), pii: 20170216. Review. Available from: https://doi.org/10.1098/rstb.2017.0216. Anacker, D., Moody, C., 2012. Generation of organotypic raft cultures from primary human keratinocytes. J. Visualized Exp. 22 (60), pii: 3668. Available from: https://doi. org/10.3791/3668. Andersson, L.C., Nilsson, K., Gahmberg, C.G., 1979. K562—a human erythroleukemic cell line. Int. J. Cancer 23 (2), 143 147. Asano, S., Ito, S., Takahashi, K., Furuya, K., Kondo, M., Sokabe, M., et al., 2017. Matrix stiffness regulates migration of human lung fibroblasts. Physiol. Rep. 5 (9), pii: e13281. Available from: https://doi.org/10.14814/phy2.13281. Asselineau, D., Bernhard, B., Bailly, C., Darmon, M., 1985. Epidermal morphogenesis and induction of the 67 kD keratin polypeptide by culture of human keratinocytes at the liquid-air interface. Exp. Cell Res. 159 (2), 536 539. Ayehunie, S., Landry, T., Stevens, Z., Armento, A., Hayden, P., Klausner, M., 2018. Human primary cell-based organotypic microtissues for modeling small intestinal drug absorption. Pharm. Res. 35 (4), 72. Available from: https://doi.org/10.1007/s11095018-2362-0. Bai, J., Smock, S.L., Jackson Jr, G.R., MacIsaac, K.D., Huang, Y., et al., 2015. Phenotypic responses of differentiated asthmatic human airway epithelial cultures to rhinovirus. PLoS One 10 (2), e0118286. Available from: https://doi.org/10.1371/journal. pone.0118286. eCollection 2015. Balmforth, A.J., Ball, S.G., Freshney, R.I., Graham, D.I., McNamee, H.B., Vaughan, P.F., 1986. D-1 dopaminergic and beta-adrenergic stimulation of adenylate cyclase in a clone derived from the human astrocytoma cell line G-CCM. J. Neurochem. 47 (3), 715 719. Bao, X., Lian, X., Dunn, K.K., Shi, M., Han, T., Qian, T., et al., 2015. Chemically-defined albumin-free differentiation of human pluripotent stem cells to endothelial progenitor cells. Stem Cell Res. 15 (1), 122 129. Available from: https://doi.org/10.1016/j. scr.2015.05.004. Beauchamp, P., Moritz, W., Kelm, J.M., Ullrich, N.D., Agarkova, I., Anson, B.D., et al., 2015. Development and characterization of a scaffold-free 3D spheroid model of induced pluripotent stem cell derived human cardiomyocytes. Tissue Eng., C: Methods 21 (8), 852 861. Available from: https://doi.org/10.1089/ten.TEC.2014.0376. Bell, E., Ehrlich, H.P., Buttle, D.J., Nakatsuji, T., 1981. Living tissue formed in vitro and accepted as skin-equivalent tissue of full thickness. Science 211 (4486), 1052 1054. Bell, C.C., Hendriks, D.F., Moro, S.M., Ellis, E., Walsh, J., Renblom, A., et al., 2016. Characterization of primary human hepatocyte spheroids as a model system for druginduced liver injury, liver function and disease. Sci. Rep. 6, 25187. Available from: https://doi.org/10.1038/srep25187. Be´rube´, K., Pitt, A., Hayden, P., Prytherch, Z., Job, C., 2010. Filter-well technology for advanced three-dimensional cell culture: perspectives for respiratory research. Altern. Lab. Anim. 38 (Suppl 1), 49 65. Bolmarcich, J., Wilbert, S., Jackson, G.R., Oldach, J., Bachelor, M., Kenney, T., et al., 2018. In vitro human airway models for study of goblet cell hyperplasia and mucus

33

34

CHAPTER 2 Cell sources and methods

production: effects of Th2 cytokines, double-stranded RNA, and tobacco smoke. Appl. In Vitro Toxicol. 4 (4), 332. Available from: https://doi.org/10.1089/aivt.2017.0001. Bose, B., Sudheer, P.S., 2016. In vitro differentiation of pluripotent stem cells into functional β islets under 2D and 3D culture conditions and in vivo preclinical validation of 3D islets. Methods Mol. Biol. 1341, 257 284. Available from: https://doi.org/10.1007/ 7651_2015_230. Boukamp, P., Petrussevska, R.T., Breitkreutz, D., Hornung, J., Markham, A., Fusenig, N. E., 1988. Normal keratinization in a spontaneously immortalized aneuploid human keratinocyte cell line. J. Cell Biol. 106 (3), 761 771. Brooks, J.C., Judd, R.L., Easley, C.J., 2017. Culture and sampling of primary adipose tissue in practical microfluidic systems. Methods Mol. Biol. 1566, 185 201. Available from: https://doi.org/10.1007/978-1-4939-6820-6_18. Brooks, S.C., Locke, E.R., Soule, H.D., 1973. Estrogen receptor in a human cell line (MCF-7) from breast carcinoma. J. Biol. Chem. 248 (17), 6251 6253. Brower, M., Carney, D.N., Oie, H.K., Gazdar, A.F., Minna, J.D., 1986. Growth of cell lines and clinical specimens of human non-small cell lung cancer in a serum-free defined medium. Cancer Res. 46, 798 806. Cannon, C.L., Neal, P.J., Southee, J.A., Kubilus, J., Klausner, M., 1994. New epidermal model for dermal irritancy testing. Toxicol. In Vitro 8 (4), 889 891. Carswell, K.A., Lee, M.J., Fried, S.K., 2012. Culture of isolated human adipocytes and isolated adipose tissue. Methods Mol. Biol. 806, 203 214. Centeno, E.G.Z., Cimarosti, H., Bithell, A., 2018. 2D versus 3D human induced pluripotent stem cell-derived cultures for neurodegenerative disease modelling. Mol. Neurodegener. 13 (1), 27. Available from: https://doi.org/10.1186/s13024-018-0258-4. Chamcheu, J.C., Pal, H.C., Siddiqui, I.A., Adhami, V.M., Ayehunie, S., Boylan, B.T., et al., 2015. Prodifferentiation, anti-inflammatory and antiproliferative effects of delphinidin, a dietary anthocyanidin, in a full-thickness three-dimensional reconstituted human skin model of psoriasis. Skin Pharmacol. Physiol. 28 (4), 177 188. Available from: https://doi.org/10.1159/000368445. Charrier, E.E., Pogoda, K., Wells, R.G., Janmey, P.A., 2018. Control of cell morphology and differentiation by substrates with independently tunable elasticity and viscous dissipation. Nat. Commun. 9 (1), 449. Available from: https://doi.org/10.1038/s41467-01802906-9. Chaterji, S., Kim, P., Choe, S.H., Tsui, J.H., Lam, C.H., Ho, D.S., et al., 2014. Synergistic effects of matrix nanotopography and stiffness on vascular smooth muscle cell function. Tissue Eng., A 20 (15 16), 2115 2126. Available from: https://doi.org/10.1089/ ten.tea.2013.0455. Chopra, D.P., Dombkowski, A.A., Stemmer, P.M., Parker, G.C., 2010. Intestinal epithelial cells in vitro. Stem Cells Dev. 19 (1), 131 142. Available from: https://doi.org/ 10.1089/scd.2009.0109. Dehne, E.M., Hasenberg, T., Marx, U., 2017. The ascendance of microphysiological systems to solve the drug testing dilemma. Future Sci. OA 3 (2), FSO185. Available from: https://doi.org/10.4155/fsoa-2017-0002. De Sousa, P.A., Steeg, R., Wachter, E., Bruce, K., King, J., Hoeve, M., et al., 2017. Rapid establishment of the European Bank for induced Pluripotent Stem Cells (EBiSC)—the Hot Start experience. Stem Cell Res. 20, 105 114. Available from: https://doi.org/ 10.1016/j.scr.2017.03.002.

References

Diamond, L., Kruszewski, F., Aden, D.P., Knowles, B.B., Baird, W.M., 1980. Metabolic activation of benzo[a]pyrene by a human hepatoma cell line. Carcinogenesis 1 (10), 871 875. Dixon, L.J., Barnes, M., Tang, H., Pritchard, M.T., Nagy, L.E., 2013. Kupffer cells in the liver. Compr. Physiol. 3 (2), 785 797. Available from: https://doi.org/10.1002/cphy. c120026. Engel, L.W., Young, N.A., Tralka, T.S., Lippman, M.E., O’Brien, S.J., Joyce, M.J., 1978. Establishment and characterization of three new continuous cell lines derived from human breast carcinomas. Cancer Res. 38 (10), 3352 3364. Engler, A.J., Sen, S., Sweeney, H.L., Discher, D.E., 2006. Matrix elasticity directs stem cell lineage specification. Cell 126 (4), 677 689. Epstein, M.A., Barr, Y.M., 1964. Cultivation in vitro of human lymphoblasts from Burkitt’s malignant lymphoma. Lancet 1, 252 253. Espinosa-Hoyos, D., Jagielska, A., Homan, K.A., Du, H., Busbee, T., Anderson, D.G., et al., 2018. Engineered 3D-printed artificial axons. Sci. Rep. 8 (1), 478. Available from: https://doi.org/10.1038/s41598-017-18744-6. Fahrbach, K.M., Barry, S.M., Ayehunie, S., Lamore, S., Klausner, M., Hope, T.J., 2007. Activated CD34-derived Langerhans cells mediate transinfection with human immunodeficiency virus. J. Virol. 81 (13), 6858 6868. Felder, M., Trueeb, B., Stucki, A.O., Borcard, S., Stucki, J.D., Schnyder, B., et al., 2019. Impaired wound healing of alveolar lung epithelial cells in a breathing lung-on-a-chip. Front. Bioeng. Biotechnol. 7, 3. Available from: https://doi.org/10.3389/fbioe.2019.00003. Fizes¸an, I., Cambier, S., Moschini, E., Chary, A., Nelissen, I., Ziebel, J., et al., 2019. In vitro exposure of a 3D-tetraculture representative for the alveolar barrier at the airliquid interface to silver particles and nanowires. Part. Fibre Toxicol. 16 (1), 14. Available from: https://doi.org/10.1186/s12989-019-0297-1. Fogh, J., Fogh, J.M., Orfeo, T., 1977. One hundred and twenty-seven cultured human tumor cell lines producing tumors in nude mice. J. Natl. Cancer Inst. 59 (1), 221 226. Fogh, J., Trempe, G., 1975. New human tumor cell lines. In: Fogh, J. (Ed.), Human Tumor Cells In Vitro, 115 141. Plenum Press, New York. Freshney, R.I., 2016. Culture of Animal Cells: A Manual of Basic Technique and Specialized Applications, Seventh ed. Wiley-Blackwell, Hoboken, NJ. Fusaki, N., Ban, H., Nishiyama, A., Saeki, K., Hasegawa, M., 2009. Efficient induction of transgene-free human pluripotent stem cells using a vector based on Sendai virus, an RNA virus that does not integrate into the host genome. Proc. Jpn. Acad., Ser. B: Phys. Biol. Sci. 85, 348 362. Gabriel, E., Gopalakrishnan, J., 2017. Generation of iPSC-derived human brain organoids to model early neurodevelopmental disorders. J. Visualized Exp. 122, e55372. Available from: https://doi.org/10.3791/55372. Geraghty, R.J., Capes-Davis, A., Davis, J.M., Downward, J., Freshney, R.I., Knezevic, I., et al., 2014. Guidelines for the use of cell lines in biomedical research. Br. J. Cancer 111 (6), 1021 1046. Available from: https://doi.org/10.1038/bjc.2014.166. Gey, G.O., Coffman, W.D., Kubicek, M.T., 1952. Tissue culture studies of the proliferative capacity of cervical carcinoma and normal epithelium. Cancer Res. 12, 264 265. Giard, D.J., Aaronson, S.A., Parks, W.P., 1973. In vitro cultivation of human tumors: establishment of cell lines derived from a series of solid tumors. J. Natl. Cancer Inst. 51, 1417 1423.

35

36

CHAPTER 2 Cell sources and methods

Gilbert, P.M., Havenstrite, K.L., Magnusson, K.E., Sacco, A., Leonardi, N.A., Kraft, P., et al., 2010. Substrate elasticity regulates skeletal muscle stem cell self-renewal in culture. Science 329 (5995), 1078 1081. Available from: https://doi.org/10.1126/ science.1191035. Gjorevski, N., Sachs, N., Manfrin, A., Giger, S., Bragina, M.E., Ordo´n˜ez-Mora´n, P., et al., 2016. Designer matrices for intestinal stem cell and organoid culture. Nature 539 (7630), 560 564. Available from: https://doi.org/10.1038/nature20168. Gordon, J., Amini, S., White, M.K., 2013. General overview of neuronal cell culture. Methods Mol. Biol. 1078, 1 8. Gordon, S., Daneshian, M., Bouwstra, J., Caloni, F., Constant, S., Davies, D.E., et al., 2015. Non-animal models of epithelial barriers (skin, intestine and lung) in research, industrial applications and regulatory toxicology. Altern. Anim. Exp. 32 (4), 327 378. Available from: https://doi.org/10.14573/altex.1510051. Graham, F.L., Smiley, J., Russell, W.C., Nairn, R., 1977. Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J. Gen. Virol. 36 (1), 59 74. Guillouzo, A., Corlu, A., Aninat, C., Glaise, D., Morel, F., Guguen-Guillouzo, C., 2007. The human hepatoma HepaRG cells: a highly differentiated model for studies of liver metabolism and toxicity of xenobiotics. Chem. Biol. Interact. 168 (1), 66 73. Hackett, T.L., Singhera, G.K., Shaheen, F., Hayden, P., Jackson, G.R., Hegele, R.G., et al., 2011. Intrinsic phenotypic differences of asthmatic epithelium and its inflammatory responses to respiratory syncytial virus and air pollution. Am. J. Respir. Cell Mol. Biol. 45 (5), 1090 1100. Available from: https://doi.org/10.1165/rcmb.2011-0031OC. Hayden, P.J., Bachelor, M., Ayehunie, S., Letasiova, S., Kaluzhny, Y., Kandarova, H., 2015. Application of MatTek in vitro reconstructed human skin models for safety, efficacy screening, and basic preclinical research. Appl. In Vitro Toxicol. 1 (3), 226 233. Hayflick, L., Moorhead, P.S., 1961. The serial cultivation of human diploid cell strains. Exp. Cell Res. 25, 585 621. Heffelfinger, S.C., Hawkins, H.H., Barrish, J., Taylor, L., Darlington, G.J., 1992. SK HEP-1: a human cell line of endothelial origin. In Vitro Cell. Dev. Biol. 28A (2), 136 142. Herter, S., Morra, L., Schlenker, R., Sulcova, J., Fahrni, L., Waldhauer, I., et al., 2017. A novel three-dimensional heterotypic spheroid model for the assessment of the activity of cancer immunotherapy agents. Cancer Immunol. Immunother. 66 (1), 129 140. Available from: https://doi.org/10.1007/s00262-016-1927-1. Hofrichter, M., Nimtz, L., Tigges, J., Kabiri, Y., Schro¨ter, F., Royer-Pokora, B., et al., 2017. Comparative performance analysis of human iPSC-derived and primary neural progenitor cells (NPC) grown as neurospheres in vitro. Stem Cell Res. 25, 72 82. Available from: https://doi.org/10.1016/j.scr.2017.10.013. Holmberg, F.E., Seidelin, J.B., Yin, X., Mead, B.E., Tong, Z., Li, Y., et al., 2017. Culturing human intestinal stem cells for regenerative applications in the treatment of inflammatory bowel disease. EMBO Mol. Med. 9 (5), 558 570. Available from: https://doi.org/10.15252/emmm.201607260. Homan, K.A., Kolesky, D.B., Skylar-Scott, M.A., Herrmann, J., Obuobi, H., Moisan, A., et al., 2016. Bioprinting of 3D convoluted renal proximal tubules on perfusable chips. Sci. Rep. 6, 34845. Available from: https://doi.org/10.1038/srep34845. Homan, K.A., Gupta, N., Kroll, K.T., Kolesky, D.B., Skylar-Scott, M., Miyoshi, T., et al., 2019. Flow-enhanced vascularization and maturation of kidney organoids in vitro.

References

Nat. Methods 16 (3), 255 262Epub 2019 Feb 11. Available from: https://doi.org/ 10.1038/s41592-019-0325-y. Honegger, P., 2001. Overview of cell and tissue culture techniques. Curr. Protoc. Pharmacol. 4, 12.1.1 12.1.12. Available from: https://doi.org/10.1002/0471141755. ph1201s04. Hoppensack, A., Kazanecki, C.C., Colter, D., Gosiewska, A., Schanz, J., Walles, H., et al., 2014. A human in vitro model that mimics the renal proximal tubule. Tissue Eng., C: Methods 20 (7), 599 609Epub 2014 Jan 3. Available from: https://doi.org/10.1089/ten. TEC.2013.0446. Huh, D., Hamilton, G.A., Ingber, D.E., 2011. From 3D cell culture to organs-on-chips. Trends Cell Biol. 21 (12), 745 754Epub 2011 Oct 25. Review. Available from: https://doi.org/10.1016/j.tcb.2011.09.005. Ichimura, T., Asseldonk, E.J., Humphreys, B.D., Gunaratnam, L., Duffield, J.S., Bonventre, J., 2008. V. Kidney injury molecule-1 is a phosphatidylserine receptor that confers a phagocytic phenotype on epithelial cells. J. Clin. Invest. 118 (5), 1657 1668. Available from: https://doi.org/10.1172/JCI34487. Jeon, H., Tsui, J.H., Jang, S.I., Lee, J.H., Park, S., Mun, K., et al., 2015. Combined effects of substrate topography and stiffness on endothelial cytokine and chemokine secretion. ACS Appl. Mater. Interfaces 7 (8), 4525 4532. Available from: https://doi.org/ 10.1021/acsami.5b00554. Jiang, W., Lan, F., Zhang, H., 2016. Human induced pluripotent stem cells for inherited cardiovascular diseases modeling. Curr. Stem Cell Res. Ther. 11 (7), 533 541. Kaluzhny, Y., Kinuthia, M.W., Truong, T., Lapointe, A.M., Hayden, P., Klausner, M., 2018. New human organotypic corneal tissue model for ophthalmic drug delivery studies. Invest. Ophthalmol. Visual Sci. 59 (7), 2880 2898. Available from: https://doi.org/ 10.1167/iovs.18-23944. Kamdar, F., Klaassen Kamdar, A., Koyano-Nakagawa, N., Garry, M.G., Garry, D.J., 2015. Cardiomyopathy in a dish: using human inducible pluripotent stem cells to model inherited cardiomyopathies. J. Card. Failure 21 (9), 761 770. Available from: https:// doi.org/10.1016/j.cardfail.2015.04.010. Kim, D.H., Lipke, E.A., Kim, P., Cheong, R., Thompson, S., Delannoy, M., et al., 2010. Nanoscale cues regulate the structure and function of macroscopic cardiac tissue constructs. Proc. Natl. Acad. Sci. U.S.A. 107 (2), 565 570Published online 2009 Dec 16. Available from: https://doi.org/10.1073/pnas.0906504107. Kim, H.N., Jiao, A., Hwang, N.S., Kim, M.S., Kang, D.H., Kim, D.H., et al., 2013. Nanotopography-guided tissue engineering and regenerative medicine. Adv. Drug Delivery Rev. 65 (4), 536 558. Available from: https://doi.org/10.1016/j.addr.2012.07.014. Kim, J.H., Kurtz, A., Yuan, B.Z., Zeng, F., Lomax, G., Loring, J.F., et al., 2017. Report of the International Stem Cell Banking Initiative Workshop Activity: current hurdles and progress in seed-stock banking of human pluripotent stem cells. Stem Cells Transl. Med. 6 (11), 1956 1962. Available from: https://doi.org/10.1002/sctm.17-0144. Kim, Y., Kim, H., Ko, U.H., Oh, Y., Lim, A., Sohn, J.W., et al., 2016. Islet-like organoids derived from human pluripotent stem cells efficiently function in the glucose responsiveness in vitro and in vivo. Sci. Rep. 6, 35145. Available from: https://doi.org/ 10.1038/srep35145. King, S.M., Higgins, J.W., Nino, C.R., Smith, T.R., Paffenroth, E.H., Fairbairn, C.E., et al., 2017. 3D Proximal tubule tissues recapitulate key aspects of renal physiology to enable

37

38

CHAPTER 2 Cell sources and methods

nephrotoxicity testing. Front. Physiol. 8, 123. Available from: https://doi.org/10.3389/ fphys.2017.00123. Klaus, J., Kanton, S., Kyrousi, C., Ayo-Martin, A.C., Di Giaimo, R., Riesenberg, S., et al., 2019. Altered neuronal migratory trajectories in human cerebral organoids derived from individuals with neuronal heterotopia. Nat. Med. 25, 561 568. Available from: https://doi.org/10.1038/s41591-019-0371-0. Kolesky, D.B., Homan, K.A., Skylar-Scott, M., Lewis, J.A., 2018. In vitro human tissues via multi-material 3-D bioprinting. Altern. Lab. Anim. 46 (4), 209 215. Kolesky, D.B., Homan, K.A., Skylar-Scott, M.A., Lewis, J.A., 2016. Three-dimensional bioprinting of thick vascularized tissues. Proc. Natl. Acad. Sci. U.S.A. 113 (12), 3179 3184. Available from: https://doi.org/10.1073/pnas.1521342113. Kondo, Y., Toyoda, T., Inagaki, N., Osafune, K., 2018a. iPSC technology-based regenerative therapy for diabetes. J. Diabetes Invest. 9 (2), 234 243. Available from: https:// doi.org/10.1111/jdi.12702. Kondo, S., Mizuno, S., Hashita, T., Iwao, T., Matsunaga, T., 2018b. Using human iPS cellderived enterocytes as novel in vitro model for the evaluation of human intestinal mucosal damage. Inflammation Res. 67 (11 12), 975 984. Available from: https:// doi.org/10.1007/s00011-018-1193-0. Kshitiz, Park, J., Kim, P., Helen, W., Engler, A.J., Levchenko, A., et al., 2012. Control of stem cell fate and function by engineering physical microenvironments. Integr. Biol. 4 (9), 1008 1018. Kuehn, A., Kletting, S., de Souza Carvalho-Wodarz, C., Repnik, U., Griffiths, G., Fischer, U., et al., 2016. Human alveolar epithelial cells expressing tight junctions to model the air-blood barrier. Altern. Anim. Exp. 33 (3), 251 260. Available from: https://doi.org/ 10.14573/altex.1511131. Lampi, M.C., Reinhart-King, C.A., 2018. Targeting extracellular matrix stiffness to attenuate disease: from molecular mechanisms to clinical trials. Sci. Transl. Med. 10 (422), eaao0475. Available from: https://doi.org/10.1126/scitranslmed.aao0475. Lancaster, M.A., Renner, M., Martin, C.-A., Wenzel, D., Bicknell, L.S., Hurles, M.E., et al., 2013. Cerebral organoids model human brain development and microcephaly. Nature 501, 373 379. Lee, C.T., Bendriem, R.M., Wu, W.W., Shen, R.F., 2017. 3D brain Organoids derived from pluripotent stem cells: promising experimental models for brain development and neurodegenerative disorders. J. Biomed. Sci. 24 (1), 59. Available from: https://doi.org/ 10.1186/s12929-017-0362-8. Lee, S.A., No da, Y., Kang, E., Ju, J., Kim, D.S., Lee, S.H., 2013. Spheroid-based threedimensional liver-on-a-chip to investigate hepatocyte-hepatic stellate cell interactions and flow effects. Lab Chip 13 (18), 3529 3537. Available from: https://doi.org/ 10.1039/c3lc50197c. Lin, N.Y.C., Homan, K.A., Robinson, S.S., Kolesky, D.B., Duarte, N., Moisan, A., et al., 2019. Renal reabsorption in 3D vascularized proximal tubule models. Proc. Natl. Acad. Sci. U.S.A. 116 (12), 5399 5404. Available from: https://doi.org/10.1073/ pnas.1815208116. Lind, J.U., Busbee, T.A., Valentine, A.D., Pasqualini, F.S., Yuan, H., Yadid, M., et al., 2017. Instrumented cardiac microphysiological devices via multimaterial threedimensional printing. Nat. Mater. 16 (3), 303 308. Available from: https://doi.org/ 10.1038/nmat4782.

References

Liu, F., Mih, J.D., Shea, B.S., Kho, A.T., Sharif, A.S., Tager, A.M., et al., 2010. Feedback amplification of fibrosis through matrix stiffening and COX-2 suppression. J. Cell Biol. 190 (4), 693 706. Available from: https://doi.org/10.1083/jcb.201004082. Loo, L.S.W., Lau, H.H., Jasmen, J.B., Lim, C.S., Teo, A.K.K., 2018. An arduous journey from human pluripotent stem cells to functional pancreatic β cells. Diabetes Obes. Metab. 20 (1), 3 13. Available from: https://doi.org/10.1111/dom.12996. Lorsch, J.R., Collins, F.S., Lippincott-Schwartz, J., 2014. Cell biology. Fixing problems with cell lines. Science 346 (6216), 1452 1453. Available from: https://doi.org/ 10.1126/science.1259110. Ma, X., Qu, X., Zhu, W., Li, Y.S., Yuan, S., Zhang, H., et al., 2016. Deterministically patterned biomimetic human iPSC-derived hepatic model via rapid 3D bioprinting. Proc. Natl. Acad. Sci. U.S.A. 113 (8), 2206 2211. Available from: https://doi.org/10.1073/ pnas.1524510113. Madden, L.R., Nguyen, T.V., Garcia-Mojica, S., Shah, V., Le, A.V., Peier, A., et al., 2018. Bioprinted 3D primary human intestinal tissues model aspects of native physiology and ADME/Tox functions. iScience 2, 156 167. Available from: https://doi.org/10.1016/j. isci.2018.03.015. Mains, R.E., Patterson, P.H., 1973. Primary cultures of dissociated sympathetic neurons. I. Establishment of long-term growth in culture and studies of differentiated properties. J. Cell Biol. 59 (2 Pt 1), 329 345. Maldonado-Contreras, A., Birtley, J.R., Boll, E., Zhao, Y., Mumy, K.L., Toscano, J., et al., 2017. Shigella depends on SepA to destabilize the intestinal epithelial integrity via cofilin activation. Gut Microbes 8 (6), 544 560. Available from: https://doi.org/ 10.1080/19490976.2017.1339006. Marescotti, D., Serchi, T., Luettich, K., Xiang, Y., Moschini, E., Talikka, M., et al., 2019. How complex should an in vitro model be? Evaluation of complex 3D alveolar model with transcriptomic data and computational biological network models. Altern. Anim. Exp. 2019, [Epub ahead of print]. Available from: https://doi.org/10.14573/ altex.1811221. Maschmeyer, I., Lorenz, A.K., Schimek, K., Hasenberg, T., Ramme, A.P., Hu¨bner, J., et al., 2015a. A four-organ-chip for interconnected long-term co-culture of human intestine, liver, skin and kidney equivalents. Lab Chip 15 (12), 2688 2699. Available from: https://doi.org/10.1039/c5lc00392j. Maschmeyer, I., Hasenberg, T., Jaenicke, A., Lindner, M., Lorenz, A.K., Zech, J., et al., 2015b. Chip-based human liver-intestine and liver-skin co-cultures—a first step toward systemic repeated dose substance testing in vitro. Eur. J. Pharm. Biopharm. 95 (Pt A), 77 87. Available from: https://doi.org/10.1016/j.ejpb.2015.03.002. Matoba, N., Yamashita, T., Takayama, K., Sakurai, F., Mizuguchi, H., 2018. Optimal human iPS cell culture method for efficient hepatic differentiation. Differentiation 104, 13 21. Available from: https://doi.org/10.1016/j.diff.2018.09.005. Matsa, E., Burridge, P.W., Yu, K.H., Ahrens, J.H., Termglinchan, V., Wu, H., et al., 2016. Transcriptome profiling of patient-specific human iPSC-cardiomyocytes predicts individual drug safety and efficacy responses in vitro. Cell Stem Cell 19 (3), 311 325. Available from: https://doi.org/10.1016/j.stem.2016.07.006. McKernan, R., Watt, F.M., 2013. What is the point of large-scale collections of human induced pluripotent stem cells? Nat. Biotechnol. 31, 875 877. Erratum. Nature Biotechnology 31, 1148. Available from: https://doi.org/10.1038/nbt.2710.

39

40

CHAPTER 2 Cell sources and methods

Meier, F., Freyer, N., Brzeszczynska, J., Kno¨spel, F., Armstrong, L., Lako, M., et al., 2017. Hepatic differentiation of human iPSCs in different 3D models: a comparative study. Int. J. Mol. Med. 40 (6), 1759 1771. Available from: https://doi.org/10.3892/ijmm.2017.3190. Messner, S., Fredriksson, L., Lauschke, V.M., Roessger, K., Escher, C., Bober, M., et al., 2018. Transcriptomic, proteomic, and functional long-term characterization of multicellular three-dimensional human liver microtissues. Appl. In Vitro Toxicol. 4. Available from: https://doi.org/10.1089/aivt.2017.0022. Mih, J.D., Sharif, A.S., Liu, F., Marinkovic, A., Symer, M.M., Tschumperlin, D.J., 2011. A multiwell platform for studying stiffness-dependent cell biology. PLoS One 6 (5), e19929. Available from: https://doi.org/10.1371/journal.pone.0019929. Mitry, R., Hughes, R. (Eds.), 2012. Human Cell Culture Protocols. Humana Press, NY. Miura, S., Suzuki, A., 2018. Brief summary of the current protocols for generating intestinal organoids. Dev. Growth Differ. 60 (6), 387 392. Available from: https://doi.org/ 10.1111/dgd.12559. ¨ ztu¨rk, E., Arlov, Ø., Gatenholm, P., Zenobi-Wong, M., 2017. Alginate Mu¨ller, M., O sulfate-nanocellulose bioinks for cartilage bioprinting applications. Ann. Biomed. Eng. 45 (1), 210 223. Available from: https://doi.org/10.1007/s10439-016-1704-5. Mu´nera, J.O., Wells, J.M., 2017. Generation of gastrointestinal organoids from human pluripotent stem cells. Methods Mol. Biol. 1597, 167 177. Available from: https://doi. org/10.1007/978-1-4939-6949-4_12. Nguyen, D., Ha¨gg, D.A., Forsman, A., Ekholm, J., Nimkingratana, P., Brantsing, C., et al., 2017. Cartilage tissue engineering by the 3D bioprinting of iPS cells in a nanocellulose/alginate bioink. Sci. Rep. 7 (1), 658. Available from: https://doi.org/10.1038/ s41598-017-00690-y. Norona, L.M., Nguyen, D.G., Gerber, D.A., Presnell, S.C., Mosedale, M., Watkins, P.B., 2019. Bioprinted liver provides early insight into the role of Kupffer cells in TGF-β1 and methotrexate-induced fibrogenesis. PLoS One 14 (1), e0208958. Available from: https://doi.org/10.1371/journal.pone.0208958. Novak, R., Didier, M., Calamari, E., Ng, C.F., Choe, Y., Clauson, S.L., et al., 2018. Scalable fabrication of stretchable, dual channel, microfluidic organ chips. J. Visualized Exp. (140), . Available from: https://doi.org/10.3791/58151. Ntai, A., Baronchelli, S., La Spada, A., Moles, A., Guffanti, A., De Blasio, P., et al., 2017. A Review of research-grade human induced pluripotent stem cells qualification and biobanking processes. Biopreserv. Biobanking 15 (4), 384 392. Available from: https://doi.org/10.1089/bio.2016.0097. Olsson, I., Olofsson, T., Mauritzon, N., 1981. Characterization of mononuclear blood cellderived differentiation inducing factors for the human promyelocytic leukemia cell line HL-60. J. Natl. Cancer Inst. 67 (6), 1225 1230. Pamies, D., Barreras, P., Block, K., Makri, G., Kumar, A., Wiersma, D., et al., 2017. A human brain microphysiological system derived from induced pluripotent stem cells to study neurological diseases and toxicity. Altern. Anim. Exp. 34 (3), 362 376. Available from: https://doi.org/10.14573/altex.1609122. Parent, R., Marion, M.J., Furio, L., Tre´po, C., Petit, M.A., 2004. Origin and characterization of a human bipotent liver progenitor cell line. Gastroenterology 126 (4), 1147 1156. Park, J., Kim, H.N., Kim, D.H., Levchenko, A., Suh, K.Y., 2012. Quantitative analysis of the combined effect of substrate rigidity and topographic guidance on cell morphology.

References

IEEE Trans. Nanobiosci. 11 (1), 28 36. Available from: https://doi.org/10.1109/ TNB.2011.2165728. Peng, Y., Xiong, W.C., Mei, L., 2013. Culture of dissociated hippocampal neurons. In: Zhou, R., Mei, L. (Eds.), Neural Development: Methods and Protocols, Vol. 1018. Humana Press, Totowa, NJ, Methods in Molecular Medicine. Peters, M.F., Landry, T., Pin, C., Maratea, K., Dick, C., Wagoner, M.P., et al., 2019. Human 3D gastrointestinal microtissue barrier function as a predictor of drug-induced diarrhea. Toxicol. Sci. 168 (1), 3 17. Available from: https://doi.org/10.1093/toxsci/ kfy268. Picot, J., 2005. Second ed. Human Cell Culture Protocols, vol. 107. Humana Press, NJ, Methods in Molecular Medicine. Pirnay, J.P., Baudoux, E., Cornu, O., Delforge, A., Delloye, C., Guns, J., et al., 2015. Access to human tissues for research and product development: from EU regulation to alarming legal developments in Belgium. EMBO Rep. 16 (5), 557 562. Available from: https://doi.org/10.15252/embr.201540070. Plummer, S., Wallace, S., Ball, G., Lloyd, R., Schiapparelli, P., Quin˜ones-Hinojosa, A., et al., 2019. A Human iPSC-derived 3D platform using primary brain cancer cells to study drug development and personalized medicine. Sci. Rep. 9, 1407. Available from: https://doi.org/10.1038/s41598-018-38130-0. Ponten, J., Macintyre, E.H., 1968. Long term culture of normal and neoplastic human glia. Acta Pathol. Microbiol. Scand. 74, 465 486. Proctor, W.R., Foster, A.J., Vogt, J., Summers, C., Middleton, B., Pillings, M.A., et al., 2017. Utility of spherical human liver microtissues for prediction of clinical druginduced liver injury. Arch. Toxicol. 91 (8), 2849 2863. Available from: https://doi. org/10.1007/s00204-017-2002-1. Qian, T., Maguire, S.E., Canfield, S.G., Bao, X., Olson, W.R., Shusta, E.V., et al., 2017. Directed differentiation of human pluripotent stem cells to blood-brain barrier endothelial cells. Sci. Adv. 3 (11), e1701679. Available from: https://doi.org/10.1126/ sciadv.1701679. Rahmani, S., Breyner, N.M., Su, H.M., Verdu, E.F., Didar, T.F., 2019. Intestinal organoids: a new paradigm for engineering intestinal epithelium in vitro. Biomaterials 194, 195 214. Available from: https://doi.org/10.1016/j.biomaterials.2018.12.006. Second ed. Randell, S.H., Fulcher, M. (Eds.), 2013. Epithelial Cell Culture Protocols, vol. 945. Humana Press, NJ, Methods in Molecular Medicine. Reddel, R.R., Ke, Y., Gerwin, B.I., McMenamin, M.G., Lechner, J.F., Su, R.T., et al., 1988. Transformation of human bronchial epithelial cells by infection with SV40 or adenovirus-12 SV40 hybrid virus, or transfection via strontium phosphate coprecipitation with a plasmid containing SV40 early region genes. Cancer Res. 48 (7), 1904 1909. Rismani Yazdi, S., Shadmani, A., Bu¨rgel, S.C., Misun, P.M., Hierlemann, A., Frey, O., 2015. Adding the “heart” to hanging drop networks for microphysiological multi-tissue experiments. Lab Chip 15 (21), 4138 4147. Available from: https://doi.org/10.1039/ c5lc01000d. Ronaldson-Bouchard, K., Ma, S.P., Yeager, K., Chen, T., Song, L., Sirabella, D., et al., 2018. Advanced maturation of human cardiac tissue grown from pluripotent stem cells. Nature 556 (7700), 239 243. Available from: https://doi.org/10.1038/s41586-0180016-3.

41

42

CHAPTER 2 Cell sources and methods

Ronaldson-Bouchard, K., Vunjak-Novakovic, G., 2018. Organs-on-a-chip: A fast track for engineered human tissues in drug development. Cell Stem Cell 22 (3), 310 324. Available from: https://doi.org/10.1016/j.stem.2018.02.011. Ruiz-Ojeda, F.J., Rupe´rez, A.I., Gomez-Llorente, C., Gil, A., Aguilera, C.M., 2016. Cell models and their application for studying adipogenic differentiation in relation to obesity: a review. Int. J. Mol. Sci. 17 (7), 1040. Available from: https://doi.org/10.3390/ ijms17071040. Sachs, N., Tsukamoto, Y., Kujala, P., Peters, P.J., Clevers, H., 2017. Intestinal epithelial organoids fuse to form self-organizing tubes in floating collagen gels. Development 144 (6), 1107 1112. Available from: https://doi.org/10.1242/dev.143933. Sallam, K., Kodo, K., Wu, J.C., 2014. Modeling inherited cardiac disorders. Circ. J. 78 (4), 784 794. Sato, T., Clevers, H., 2015. SnapShot: growing organoids from stem cells. Cell 161 (7), 1700 1700.e1. Available from: https://doi.org/10.1016/j.cell.2015.06.028. Sayed, N., Liu, C., Wu, J.C., 2016. Translation of human-induced pluripotent stem cells: From clinical trial in a dish to precision medicine. J. Am. Coll. Cardiol. 67 (18), 2161 2176. Available from: https://doi.org/10.1016/j.jacc.2016.01.083. Schmeichel, K.L., Bissell, M.J., 2003. Modeling tissue-specific signaling and organ function in three dimensions. J. Cell Sci. 116 (Pt 12), 2377 2388. Schutgens, F., Rookmaaker, M.B., Margaritis, T., Rios, A., Ammerlaan, C., Jansen, J., et al., 2019. Tubuloids derived from human adult kidney and urine for personalized disease modeling. Nat. Biotechnol. 37 (3), 303 313. Available from: https://doi.org/ 10.1038/s41587-019-0048-8. Schwartz, M.P., Hou, Z., Propson, N.E., Zhang, J., Engstrom, C.J., Santos Costa, V., et al., 2015. Human pluripotent stem cell-derived neural constructs for predicting neural toxicity. Proc. Natl. Acad. Sci. U.S.A. 112, 12516 12521. Sharma, A., Burridge, P.W., McKeithan, W.L., Serrano, R., Shukla, P., Sayed, N., et al., 2017. High-throughput screening of tyrosine kinase inhibitor cardiotoxicity with human induced pluripotent stem cells. Sci. Transl. Med. 9 (377), eaaf2584. Available from: https://doi.org/10.1126/scitranslmed.aaf2584. Sinagoga, K.L., McCauley, H.A., Mu´nera, J.O., Reynolds, N.A., Enriquez, J.R., Watson, C., et al., 2018. Deriving functional human enteroendocrine cells from pluripotent stem cells. Development 145 (19), dev165795. Available from: https://doi.org/10.1242/ dev.165795. Smith, A.S.T., Davis, J., Lee, G., Mack, D.L., Kim, D.H., 2016. Muscular dystrophy in a dish: engineered human skeletal muscle mimetics for disease modeling and drug discovery. Drug Discov. Today 21 (9), 1387 1398. Available from: https://doi.org/ 10.1016/j.drudis.2016.04.013. Smith, Q., Chan, X.Y., Carmo, A.M., Trempel, M., Saunders, M., Gerecht, S., 2017. Compliant substratum guides endothelial commitment from human pluripotent stem cells. Sci. Adv. 3 (5), e1602883. Available from: https://doi.org/10.1126/sciadv.1602883. Sone, M., Nakao, K., 2013. Vascular research using human pluripotent stem cells and humoral factors. Endocr. J. 60 (4), 397 402. Stebbins, M.J., Wilson, H.K., Canfield, S.G., Qian, T., Palecek, S.P., Shusta, E.V., 2016. Differentiation and characterization of human pluripotent stem cell-derived brain microvascular endothelial cells. Methods 101, 93 102. Available from: https://doi.org/ 10.1016/j.ymeth.2015.10.016.

References

Sundstro¨m, C., Nilsson, K., 1976. Establishment and characterization of a human histiocytic lymphoma cell line (U-937). Int. J. Cancer 17 (5), 565 577. Takahashi, K., Okita, K., Nakagawa, M., Yamanaka, S., 2007a. Induction of pluripotent stem cells from fibroblast cultures. Nat. Protoc. 2, 3081 3089. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., et al., 2007b. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131, 861 872. Takahashi, Y., Hori, Y., Yamamoto, T., Urashima, T., Ohara, Y., Tanaka, H., 2015. 3D spheroid cultures improve the metabolic gene expression profiles of HepaRG cells. Biosci. Rep. 35 (3), e00208. Available from: https://doi.org/10.1042/ BSR20150034. Takahashi, Y., Sato, S., Kurashima, Y., Yamamoto, T., Kurokawa, S., Yuki, Y., et al., 2018. A refined culture system for human induced pluripotent stem cell-derived intestinal epithelial organoids. Stem Cell Rep. 10 (1), 314 328. Available from: https://doi. org/10.1016/j.stemcr.2017.11.004. Takasato, M., Little, M.H., 2017. Making a kidney organoid using the directed differentiation of human pluripotent stem cells. Methods Mol. Biol. 1597, 195 206. Available from: https://doi.org/10.1007/978-1-4939-6949-4_14. Takayama, K., Mizuguchi, H., 2017. Generation of human pluripotent stem cell-derived hepatocyte-like cells for drug toxicity screening. Drug Metab. Pharmacokinet. 32 (1), 12 20. Available from: https://doi.org/10.1016/j.dmpk.2016.10.408. Talkhabi, M., Aghdami, N., Baharvand, H., 2016. Human cardiomyocyte generation from pluripotent stem cells: a state-of-art. Life Sci. 145, 98 113. Available from: https://doi. org/10.1016/j.lfs.2015.12.023. Tanaka, A., Yuasa, S., Node, K., Fukuda, K., 2015. Cardiovascular disease modeling using patient-specific induced pluripotent stem cells. Int. J. Mol. Sci. 16 (8), 18894 18922. Available from: https://doi.org/10.3390/ijms160818894. Tilghman, R.W., Cowan, C.R., Mih, J.D., Koryakina, Y., Gioeli, D., Slack-Davis, J.K., et al., 2010. Matrix rigidity regulates cancer cell growth and cellular phenotype. PLoS One 5 (9), e12905. Available from: https://doi.org/10.1371/journal. pone.0012905. Tong, Z., Martyn, K., Yang, A., Yin, X., Mead, B.E., Joshi, N., et al., 2018. Towards a defined ECM and small molecule based monolayer culture system for the expansion of mouse and human intestinal stem cells. Biomaterials 154, 60 73. Available from: https://doi.org/10.1016/j.biomaterials.2017.10.038. Tsuchiya, S., Yamabe, M., Yamaguchi, Y., Kobayashi, Y., Konno, T., Tada, K., 1980. Establishment and characterization of a human acute monocytic leukemia cell line (THP-1). Int. J. Cancer 26 (2), 171 176. Tsuruo, T., Hamilton, T.C., Louie, K.G., Behrens, B.C., Young, R.C., Ozols, R.F., 1986. Collateral susceptibility of adriamycin-, melphalan- and cisplatin-resistant human ovarian tumor cells to bleomycin. Jpn. J. Cancer Res. 77, 941 945. Valente, M.J., Henrique, R., Costa, V.L., Jero´nimo, C., Carvalho, F., Bastos, M.L., et al., 2011. A rapid and simple procedure for the establishment of human normal and cancer renal primary cell cultures from surgical specimens. PLoS One 6 (5), e19337. Available from: https://doi.org/10.1371/journal.pone.0019337. Vanderburgh, J., Sterling, J.A., Guelcher, S.A., 2017. 3D printing of tissue engineered constructs for in vitro modeling of disease progression and drug screening. Ann.

43

44

CHAPTER 2 Cell sources and methods

Biomed. Eng. 45 (1), 164 179. Available from: https://doi.org/10.1007/s10439-0161640-4. Wang, C., Hei, F., Ju, Z., Yu, J., Yang, S., Chen, M., 2016. Differentiation of urinederived human induced pluripotent stem cells to alveolar type II epithelial cells. Cell. Reprogram. 18 (1), 30 36. Available from: https://doi.org/10.1089/ cell.2015.0015. Wang, Y., Alhaque, S., Cameron, K., Meseguer-Ripolles, J., Lucendo-Villarin, B., Rashidi, H., et al., 2017. Defined and scalable generation of hepatocyte-like cells from human pluripotent stem cells. J. Visualized Exp. (121), . Available from: https://doi.org/10.3791/55355. Warren, L., Manos, P.D., Ahfeldt, T., Loh, Y.H., Li, H., Lau, F., et al., 2010. Highly efficient reprogramming to pluripotency and directed differentiation of human cells with synthetic modified mRNA. Cell Stem Cell 7, 618 630. Wells, R.G., 2008. The role of matrix stiffness in regulating cell behavior. Hepatology 47 (4), 1394 1400. Available from: https://doi.org/10.1002/hep.22193. Wilkinson, D.C., Mellody, M., Meneses, L.K., Hope, A.C., Dunn, B., Gomperts, B.N., 2018. Development of a three-dimensional bioengineering technology to generate lung tissue for personalized disease modeling. Curr. Protoc. Stem Cell Biol. 46 (1), e56. Available from: https://doi.org/10.1002/cpsc.56. Wong, W.T., Huang, N.F., Botham, C.M., Sayed, N., Cooke, J.P., 2012. Endothelial cells derived from nuclear reprogramming. Circ. Res. 111 (10), 1363 1375. Available from: https://doi.org/10.1161/CIRCRESAHA.111.247213. Xia, Z., Jin, S., Ye, K., 2018. Tissue and organ 3D bioprinting. SLAS Technol. 23 (4), 301 314. Available from: https://doi.org/10.1177/2472630318760515. Yamashita, T., Takayama, K., Sakurai, F., Mizuguchi, H., 2018. Billion-scale production of hepatocyte-like cells from human induced pluripotent stem cells. Biochem. Biophys. Res. Commun. 496 (4), 1269 1275. Available from: https://doi.org/10.1016/j. bbrc.2018.01.186. Yabe, S.G., Fukuda, S., Takeda, F., Nashiro, K., Shimoda, M., Okochi, H., 2017. Efficient generation of functional pancreatic β-cells from human induced pluripotent stem cells. J. Diabetes 9 (2), 168 179. Available from: https://doi.org/10.1111/ 1753-0407.12400. You, M.-H., Kwak, M.K., Kim, D.-H., Kim, K., Levchenko, A., Kim, D.-Y., et al., 2010. Synergistically enhanced osteogenic differentiation of human mesenchymal stem cells by culture on nanostructured surfaces with induction media. Biomacromolecules 11 (7), 1856 1862. Available from: https://doi.org/10.1021/bm100374n. Yu, J., Vodyanik, M.A., Smuga-Otto, K., Antosiewicz-Bourget, J., Frane, J.L., Tian, S., et al., 2007. Induced pluripotent stem cell lines derived from human somatic cells. Science 318, 1917 1920. Zhu, W., Qu, X., Zhu, J., Ma, X., Patel, S., Liu, J., et al., 2017. Direct 3D bioprinting of prevascularized tissue constructs with complex microarchitecture. Biomaterials 124, 106 115. Available from: https://doi.org/10.1016/j.biomaterials.2017.01.042. Zuellig, R.A., Cavallari, G., Gerber, P., Tschopp, O., Spinas, G.A., Moritz, W., et al., 2014. Improved physiological properties of gravity-enforced reassembled rat and human pancreatic pseudo-islets. J. Tissue Eng. Regener. Med. 11, 109 120. Available from: https://doi.org/10.1002/term.1891.

Further reading

Further reading Huh, D., Matthews, B.D., Mammoto, A., Montoya-Zavala, M., Hsin, H.Y., Ingber, D.E., 2010. Reconstituting organ-level lung functions on a chip. Science 328 (5986), 1662 1668. Available from: https://doi.org/10.1126/science.1188302. Shi, Y., Inoue, H., Wu, J.C., Yamanaka, S., 2017. Induced pluripotent stem cell technology: a decade of progress. Nat. Rev. Drug Discov. 16 (2), 115 130. Available from: https://doi.org/10.1038/nrd.2016.245.

45

CHAPTER

Organs-on-a-chip engineering

3

Felix Kurth1, Erika Gyo¨rvary1, Sarah Heub1, Diane Ledroit1, Samantha Paoletti1, Kasper Renggli2, Vincent Revol1, Marine Verhulsel3, Gilles Weder1 and Fre´de´ric Loizeau1 1

Centre Suisse d’Electronique et de Microtechnique SA, Neuchaˆtel, Switzerland ETH Zu¨rich, Department of Biosystems Science and Engineering, Basel, Switzerland 3 Fluigent SAS, Le Kremlin-Biceˆtre, France

2

Introduction One of the greatest challenges in health care is the lack of physiologically relevant preclinical models (Moraes et al., 2012). Animal cell lines and animal tests are still used to develop drugs for humans, often leading to incorrect correlations between animal and human physiology or between in vitro and in vivo data. Drug manufacturers often fail in drug development as correctly predicting human responses from these methods is not possible (Pound et al., 2004). This can result in high drug costs, fewer new compounds in the pipeline, compounds with only limited efficacy, and good compound candidates lost in (failed) translation. A new way to investigate organ physiology mechanisms and drug-testing platforms for personalized medicine and disease modeling is required.

Definition Traditional two-dimensional (2D) cell cultures can provide indications of compound efficacy and toxicology but cannot model cell functions and physiology, because they lack the three-dimensional (3D) structures found in intact organs. Recent developments and the convergence of biology, materials sciences, engineering, and microtechnologies are bringing us closer to physiologically relevant preclinical models. These technologies make it possible to establish a new concept for preclinical models, that of an organ-on-a-chip (OOC). OOCs have been described as “a fit-for-purpose microfluidic device, containing living engineered organ substructures in a controlled microenvironment, that recapitulates one or more aspects of the organ’s dynamics, functionality and (patho)physiological response in vivo under real-time monitoring” (Thomassen, 2018). This hard-won definition is the result of several expert interviews carried out in the framework of the European Union’s ORCHID project. OOCs can be classified as single-organ systems, Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00003-6 © 2020 Elsevier Inc. All rights reserved.

47

48

CHAPTER 3 Organs-on-a-chip engineering

focusing on the key parameters of a single organ, or as multiorgan platforms, where several organ models’ interactions and responses can be monitored in a single system simultaneously. Since the invention of OOCs nearly 20 years ago, several manufacturing techniques and materials have been developed and matured to make this technology more accessible to researchers worldwide. Nevertheless, challenges remain and many design rules or trade-off need to be made during the design and fabrication of OOCs.

Engineering challenges OOCs exist at the intersection of two areas, the cell and microtissue culture field and the microfabrication technologies field, which was originally developed for the semiconductor industry. While the former is an organic, 3D, and dynamic world, the semiconductor world is inert and flat and favors immutability. Challenges therefore arise when cells cohabit and must survive alongside materials or processes that were not designed for interaction with biological components. As engineers develop OOCs, seeking physiologically relevant solutions, several points in the design and manufacturing phases should receive particular attention.

Design challenges In the design phase the main objective is to create models that are relevant to answering biological questions. By combining structured soft materials with microactuators and micropumps, it is possible to mimic the basic functions of organs or tissues. For example, a lung-on-a-chip platform exposes lung epithelial cells to an airflow as a microcavity expands following a breathing pattern (Huh et al., 2010). In a kidney-on-a-chip platform, kidney epithelial cells lie on a thin, porous membrane separating two flows to filtrate toxins (Jang et al., 2013). Such designs take advantage of the planar limitation of microfabrication technologies to develop simple yet sufficient models. Besides overall organ or tissue function, OOCs must control the migration, location, and morphology of the cultured cells; otherwise the cell phenotype might not be maintained. Mature vascular smooth muscle cells, for instance, typically display an elongated morphology in their native blood vessel (Rhodin, 2014). Cultured in vitro, they can, however, exhibit less elongation and lose their contractility (Alford et al., 2011). Therefore designers need to find ways to grow and mature these cells in their elongated state on an OOC (Sarkar et al., 2005). Not only should the model mimic the chosen organ functions, it should also measure necessary parameters such as cell activity or secretions. Each parameter can generally be monitored in several ways. Depending on the required sensitivity or speed, not all monitoring solutions are appropriate. pH, for example, can be monitored through electrochemical or optical means, each with its pros and cons. It is the designers’ responsibility to select the most relevant solution for a given experiment.

Introduction

How to mimic organ function, maintain cell phenotype, and measure critical parameters are the key questions that must be answered prior to the manufacturing step.

Manufacturing challenges Vascularization, filtration, and separation are key tissue processes that occur at a scale up to a few hundred micrometers. For OOCs to mimic such processes, they must be fabricated with similar resolutions. There are not many manufacturing methods that can produce features with reliable dimensions on such small scales; the only standard manufacturing method currently available comes from the field of microelectronics. For decades, integrated circuits and sensors have been produced on silicon wafers by the deposition and subsequent local etching of various materials with submicron precision. Hence, researchers have naturally been using the same microfabrication processes to create microfluidic chips and OOCs, although they have certain limitations. Microfabrication processes are carried out in a clean-room environment—an infrastructure that involves high maintenance costs and strict policies for avoiding contamination of the equipment and consumables. Researchers have circumvented this limitation mainly by fabricating a master in the clean room using silicon or glass and then molding and replicating their devices in more bio-friendly and cost-efficient materials outside the clean room. Silicone-based polymers such as polydimethylsiloxane (PDMS) are widely used in this context since they provide the right range of elasticity and a welcome optical transparency. Their tendency to absorb many chemicals and, therefore, trap or release drugs unintentionally is problematic and should be assessed carefully (Shirure and George, 2017). Besides the lack of suitable materials, microfabrication environments also employ quality control systems that may not be relevant for cell culture applications. Adapting those systems might become necessary for the production of OOCs, especially during the later phases of industrialization and commercialization. Because of the limitations of microfabrication, engineers have been investigating new fabrication processes and materials to replace or complement the existing ones. Additive manufacturing methods such as inkjet printing bring with them the materials and flexibility that microfabrication lacks; printers have been depositing microdroplets precisely for decades. With some modifications, printers can now print functional biomaterials such as collagen, proteins, and even cells. While still in their infancy, companies such as RegenHU (Fribourg, Switzerland) or Cellink (Gothenburg, Sweden) are developing and manufacturing bioprinters and bio-inks specifically for OOCs or tissue culture applications, making these emerging technologies more and more available to engineers and researchers. While whole devices can now be 3D-printed, additive manufacturing is too lengthy a process to match the production capabilities of microfabrication. A realistic and promising solution would be to combine the production scale of microfabrication with the materials and flexibility of 3D printing to create hybrid

49

50

CHAPTER 3 Organs-on-a-chip engineering

devices. Mechanical structures with electrical functions can be mass-produced in the clean room while 3D printers add bio-functionalization only where it is needed. Whereas originally only a few infrastructure-intensive and restrictive fabrication methods were available, engineers now have access to more adapted and flexible solutions for manufacturing their OOCs. From the materials to the surface chemistry, engineers have choices when developing OOCs to specifications.

Microengineering Introduction Design, materials selection, and fabrication techniques for manufacturing an OOC platform are not trivial tasks. Several key parameters, such as fluid control, absence of air bubbles, optimal oxygen concentration, and sterility maintenance, must typically be considered. Regardless of these numerous technological challenges, the first thing to do is to clearly define the specific organ-level function(s) to be mimicked. Although several parameters are consciously oversimplified, the engineered microenvironment must remain as close as possible to in vivo conditions. In addition, an OOC platform has to integrate a precisely controlled environment and some monitoring systems. Most physiological tissues are subject to chemical, mechanical, and/or electrical stimulations and exhibit anisotropic behavior, meaning that their properties often depend on the direction of applied stimuli. This complex in vivo condition is tentatively reproduced with the combined use of microfluidic devices with biosurfaces and actuators. It can be integrated onto single chips or parallelized standardized formats such as multiwell plates. Fig. 3.1 shows a general process for the elaboration of an OOC platform. The physicochemical properties of materials—chemistry, topography, and functionalization—exert a significant influence on the cells. In this context, the materials must also fulfill other major requirements, such as compatibility with sterilization methods, visualization tools, and especially fabrication techniques. Built using a fully automated procedure, 3D printing is typically a tool of choice for complex shapes. This section describes the main and critical steps to consider in the development of an OOC platform.

Design The design phase of OOC platforms is often a complex process, as these devices are made of many components that are specific to the different applications intended. For example, with regard to microfluidic architecture, parameters such as the number of channels, 2D/3D conformation, dimensions, geometry, and pattern must be considered. These parameters can play a critical role in establishing an adequate biochemical environment for the cells as well as tissue functionality.

Microengineering

FIGURE 3.1 Process flow diagram for developing an organ-on-a-chip platform.

51

52

CHAPTER 3 Organs-on-a-chip engineering

The incorporation of bio-surfaces such as membranes that can act as barriers or interfaces and thus better mimic the modeled organ/tissue is therefore common. Moreover, actuators and sensors are often desirable in reproducing in vivo conditions and monitoring important physicochemical parameters. Finally, the choice of chip-to-world connections, such as fluidic and electrical connections, must be made according to the applied conditions in the chip and the ease of their implementation. An overview of the elements that might be integrated into an OOC device is given in Fig. 3.2. A microfluidic device consists mainly of channels. The channel network enables the construction of a scaffold and control of the spatial location of the organ/tissue growth in a so-called mini-bioreactor. The features of this bioreactor depend on the biological tissue: monolayer, multilayer, or 3D cell culture. The microfluidic channels associated with fluid control elements enable the provision of a suitable biochemical environment for organ/tissue growth, stimulation, and analysis in the bioreactor. Spatial control of the biological tissue means several interfaces with multiple cell types, culture media, and surfaces. Classical 2D chip devices are often too simple so a 3D chip approach is required. A more complex microarchitecture, closer to the in vivo situation, can be achieved by the superimposition of channels, biointerfaces, and control elements, resulting in an integrated hybrid, multilayer device (Caplin et al., 2015; Zheng et al., 2016; Zhang et al., 2018). Another aspect of the design of the microfluidic architecture is the channel patterning, that is, the fabrication of defined patterns inside the channels (e.g., microgrooves). The functionalization of channels via geometry typically enables immobilization of the organ/tissue in the bioreactor but also serves as a scaffold

FIGURE 3.2 Schematic overview of the components of an organ-on-a-chip microfluidic device that should be considered during the design phase.

Microengineering

for bio-functionalization of the surface (Sun et al., 2011). In addition, the microchannels support the establishment of an organ-like vascularization system (Haase and Kamm, 2017; Bertassoni et al., 2014; King et al., 2004). Finally, differing channel geometry can be produced, the choice of which depends on the desired flow conditions and the mimicked environment. Fig. 3.3 presents a summary of possible options in the design of microfluidic channels. The design of the microfluidic device and its channels can be assisted by computer simulation to calculate the desired dynamic conditions (flow profile, shear, physical actuation). Detailed information on such methods can be obtained from case studies in the literature (Hagmeyer et al., 2013; Oleaga et al., 2016; Zengerle et al., 1995; Sajay et al., 2017; Borenstein et al., 2002; Kaihara et al., 2000). To model the different barriers and tissue interfaces, plastic membranes are often incorporated into the microfluidic channels of an OOC device (Zhang et al., 2018). In some cases the selected membrane can be porous, to better mimic the properties of a particular tissue. Another approach involves the insertion of a membrane that supports the growth of two cell types, one on each side of the membrane. This multilayer channel construction enables the exposure of each cell type to a different environment while mimicking a tissue junction. This format is widely used for modeling multiple organs, such as brain, liver, gut, heart, skin, and lung (Ahadian et al., 2018; Lee and Sung, 2018; Skardal et al., 2016; Ronaldson-Bouchard and Vunjak-Novakovic, 2018). The incorporation of such membranes into OOC devices is discussed in the “Fabrication process” section.

FIGURE 3.3 Schematic description of the most common options applied to the design of channels in an organ-on-a-chip microfluidic device. I, Inlet; O, outlet; P, pressure.

53

54

CHAPTER 3 Organs-on-a-chip engineering

An OOC platform also interacts with its external environment. This includes the surrounding physical parameters such as gas, temperature, humidity, and light, and connected external equipment such as readers, off-chip sensors, a fluidcontrol unit, and analytical instruments. The user may also interact directly with the platform for manual sampling. The environmental parameters are typically maintained by placing the OOC platform in an incubator with a controlled temperature and gas supply and protected from light. The footprint of the device must therefore be considered early in the design phase, as the whole OOC platform should be compatible with incubation chambers. In many OOC devices, dynamic conditions are applied, and fluid control can be part of the microdevice body (Huh et al., 2013; Bhatia and Ingber, 2014). Elements such as micropumps (Chen et al., 2017; Sonntag et al., 2015), valves (Chen et al., 2009; Huh et al., 2011), mixers (Polinkovsky et al., 2009; Brennan et al., 2014), or flow constrictors (Rodrı`guez-Villarreal et al., 2010) can be incorporated by specific channel design. Gradient control can also be assisted by channel geometry. The architecture of the microfluidic device must also comprise the elements required for the control and observation of the physiological aspects that are relevant to the application. In other words, it must integrate specific geometries or building blocks to generate desired stimuli (and sensing). When reproducing a function of the body, it is possible that mechanical, biochemical, or biophysical stimuli are required simultaneously (Park et al., 2015). The design strategies for applying these stimuli to the cells/tissues inside the chip are manifold. A mechanical stimulus, for example, can be achieved by the insertion of a stretchable membrane in skin or lung models (Guenat and Berthiaume, 2018), whereas a biochemical stimulus can be achieved by applying a gradient injection of a growth factor (Yum et al., 2014). Concerning observation, if the material used for the fabrication of the microfluidic device is not optically transparent, the design must include optical access, which can be achieved by the insertion of a transparent material. This also implies that the design must anticipate access to this area using a microscope, camera, light source, or other reader. Additional electrical and/or mechanical biochemical actuators and sensors are often needed for complex functions (see the “Stimulation and sensing” section). Embedded control units and sensors have advantages over external instrumentation. The former often enable frequent time-point or even continuous monitoring of a physicochemical parameter and limit the risk of culture contamination that is associated with sampling. Such units can be inserted into the chip construct either directly on the surface of a channel or as a building block in the device body. These on-chip solutions are principally the focus of research, given that they add complexity to the process of manufacturing the chip. As an example, the use of flexible membranes requires mastery of the bonding of at least three individual layers. Since OOC chips are by definition disposable, any additional on-chip function leads to higher costs of disposables. Beyond functionality, designers of OOC platforms should keep in mind that the cost of a device has to be balanced

Microengineering

by the additional information it provides. In general, a solution that makes use of off-chip components will be more cost-efficient than the same chip with on-chip functionalities. Electrical and fluidic connections also play a crucial role in the chip design (Temiz et al., 2015). An external fluid control unit is typically required to provide a dynamic liquid environment, administer biochemicals, supply gases, manage waste, and support in-line and off-line sample analysis. The various methods for building an external fluid control unit are discussed in the “Engineering fluid control for organ-on-a-chips” section. From a design standpoint, the OOC platform usually requires multiple sources of fluids, and their connections to the device must not leak or clog, so physicochemical parameters of fluid control, such as pressure, flow rate, salinity, or pH, must also be considered. The design of the fluidic interface must also preserve sterility inside the chip. This is particularly critical for applications that require time-point sampling for off-line analysis. Here, the sample needs to be delivered or accessed without endangering the remaining culture inside the OOC device. The fluidic ports are frequently placed on top of the microfluidic device and around an observation window, so as not to disturb optical access. Side connections are rare, as they are not suited to multilayer structures, but technologies that allow for 3D bulk chip fabrication (3D printing, micromachining) give more flexibility regarding the positioning of the fluidic ports, which are then more conveniently placed on the sides. These design approaches offer more opportunities for interconnecting the various elements of the microfluidic device. Finally, the format of the device should be considered. Although microfluidic chips are often used as single chips mimicking one or several organs, it is desirable that OOC platforms be compatible with standard formats used in drug development—such as multiwell plates and glass slides—to achieve high throughputs (Ronaldson-Bouchard and Vunjak-Novakovic, 2018).

Materials Material selection is inextricably linked to the design of the OOC device. In the body, cells are surrounded by other cells, fluids, and diverse materials such as protein fibers or minerals. When building an in vitro model, the practical need for a solid support to grow cells and study the biological model thereby generates an interface that is artificial in terms of its geometry and composition. Besides device architecture, then, material selection is therefore crucial to controlling the effect of the device on the cells and mimicking the in vivo microphysiological environment of the tissue and in particular the biointerfaces. In this section, we guide the reader concerning the selection of appropriate materials from which to build an OOC device, providing insight into the requisite criteria and available options. In practice a device is composed of a bulk material that forms its structure and other materials that can be added to customize the surface bioproperties or add functions (materials used for electrical, chemical, and optical sensors are

55

56

CHAPTER 3 Organs-on-a-chip engineering

discussed in the “Stimulation and sensing” section). It is rare that one single material can be used to fabricate an OOC device, because of the complexity of such systems, and several materials must be used. Multiple criteria must be considered when selecting a material to integrate into an OOC device:

• • • • • •

Biocompatibility Sterilization Physicochemical properties Material function in the device Design and fabrication possibilities Cost

One goal of the OOC approach is to provide a more accurate in vitro model by reconstructing an organ’s environment. Thus even if the experiments are conducted outside the human body, the material selected for the device must be biocompatible. When referring to implantable devices, a biocompatible material would not lead to any undesired or harmful biological effects. Concerning OOC devices and tissue engineering, a biomaterial would be able to perform as a substrate that supports the appropriate cellular activity (Williams, 2008). Common biocompatible materials include silicon substrates, polymers, resins, and hydrogels (Zhang et al., 2018). Sterilization of materials that are in direct or close contact with cells is necessary to avoid microorganismal contamination. Several sterilization techniques are readily available, and their utility is governed by the device-development context. Wet heat (autoclaving), dry heat (flaming, baking), solvents (ethanol), and radiation (ultraviolet light) are common sterilization agents available in most laboratories, while the use of X-rays, gamma rays, or ethylene oxide is generally limited to specialized laboratories or commercialized processes. Sterilization is characterized by its efficiency, measured in microbial inactivation, total sterilization time, operability, and scale-up. When selecting a material, it is important to consider what sterilization methods it would withstand, accounting for material properties such as permeability and thermal or chemical resistance, and verify that its properties or structure will not be altered during the process (dos Santos et al., 2017). As there is no current standard method for either fabricating or sterilizing OOC devices, efforts should be made with regard to the development of appropriate procedures to improve these processes. Various physicochemical properties can be required of the material, depending on the chip design and application. The following properties are essential when developing an OOC device:

• Optical transparency, mostly for observing the culture inside the chip. In practice, the microfluidic chip is either fabricated using an optically transparent material or integrates optical windows into defined observation areas. A material is qualified as optically transparent if it can transmit incident

Microengineering







• •

light with relatively little absorption or reflection. PDMS is typically used for OOC devices, as are glass, hydrogels, and transparent polymers such as poly (methyl methacrylate) (PMMA). Optical microscopy is generally applied as a simple, fast, and noninvasive measurement technique (Yi et al., 2017). Gas permeability, as most human cells require a supply of oxygen. A highly permeable material such as PDMS guarantees that enough oxygen is provided to the cells in the microchannels, while a nonpermeable material such as glass or plastic requires separate oxygenators (Huh et al., 2011) or perfusion. Adsorption of molecules to the surface can potentially alter the physiological response of the organ and the results of drug exposure tests. This issue is recurrent with polymers (Shirure and George, 2017) and is one of the main limitations of PDMS, making its current widespread use questionable (Shirure and George, 2017; Capulli et al., 2014; Berthier et al., 2012). Resistance to chemicals, as a release of compounds caused by degradation could interfere with the mimicked organ functions. Chemical degradation of the material could also cause leakage or the introduction of air plugs, which would spoil the experiment. Thermal resistance, as the material should not undergo dilation or contraction that could alter the properties of the system. Stiffness can also be an important parameter. A material with a stiffness similar to that of the reproduced organ will be more accurate, but such materials are not yet available. For device fabrication, high flexibility of the chip could present both advantages (e.g., the possibility of microfabricating integrated microfluidic channels, valves, or pumps) and disadvantages, such as deformation under experimental conditions such as pressure, causing variations in flow or shear stress (Berthier et al., 2012).

Finally, the costs associated with the production of the microfluidic device with the selected materials are an important factor with regard to the standardization of OOC platforms. Since these platforms mainly target the drug development market, the number of tests required could be considerable. One factor affecting the cost per data point is the availability of the raw material; more important still is the possibility of cost-effective, large-scale production (up-scalable fabrication process). Since the first steps in development of microfluidic devices in the 1950s, many inorganic (silicon, glass, ceramics) and organic (polymers) materials have been used to produce devices in a range of application areas. Extensive reviews of these materials and their advantages and limitations are available in the literature (Berthier et al., 2012; Tsao, 2016; Becker and Locascio, 2002), and details of their properties can be obtained from their manufacturers. The same materials are employed for OOC platforms, and an overview of their relevant properties is provided in Table 3.1. Today, PDMS is mostly used for research purposes and small-scale fabrication. Glass has outstanding physicochemical properties; it remains costly, but

57

Table 3.1 Properties of common materials used in organ-on-a-chip microfluidic devices. Material

Stiffness

Sterilization method

Biocompatibility

Optical transparency

Functionalization

Drug absorption

Cost

Polydimethylsiloxane Polycaprolactone Polylactic acida Glass Silicon Polycarbonatea Poly(methyl methacrylate) Polystyrenea COC/COP/CBC

Elastic Elastic Rigid Rigid Rigid Rigid Rigid

A, D, E E B, C A,B, D, E A,D,E A,B,C B, C

1 1 1 1 1 1 1

11 2 2 11 22 11 11

11 1 1 11 11 0 1

2 1 1 11 11 11 11

0 0 0 2 2 11 11

Rigid Rigid

B, C B, C

1 1

11 11

0 1

11 11

11 1

The symbols are defined as the following: 11 “Excellent,” 1 “Good,” 0 “Neutral,” 2 “Poor,” and 22 “Unrealizable”. References (Moraes et al., 2012; Pound et al., 2004; Thomassen, 2018; Huh et al., 2010, 2011, 2013; Jang et al., 2013; Rhodin, 2014; Alford et al., 2011; Sarkar et al., 2005; Shirure and George, 2017; Caplin et al., 2015; Zheng et al., 2016; Zhang et al., 2018; Sun et al., 2011; Haase and Kamm, 2017; Bertassoni et al., 2014; King et al., 2004; Hagmeyer et al., 2013; Oleaga et al., 2016; Zengerle et al., 1995; Sajay et al., 2017; Borenstein et al., 2002; Kaihara et al., 2000; Ahadian et al., 2018; Lee and Sung, 2018; Skardal et al., 2016; Ronaldson-Bouchard and Vunjak-Novakovic, 2018; Bhatia and Ingber, 2014; Chen et al., 2009, 2017; Sonntag et al., 2010, 2015; Polinkovsky et al., 2009; Brennan et al., 2014; Rodrìguez-Villarreal et al., 2010; Park et al., 2015; Guenat and Berthiaume, 2018; Yum et al., 2014; Temiz et al., 2015; Williams, 2008; dos Santos et al., 2017; Yi et al., 2017; Capulli et al., 2014; Berthier et al., 2012; Tsao, 2016; Becker and Locascio, 2002; Zhou, 2017; Pocock et al., 2016; Hamad et al., 2014; Lee and Cho, 2016; Mills et al., 2006; Su et al., 2011; Yavuz et al., 2016). A, Autoclave; B, ethylene oxide; C, gamma rays; CBC, cyclic block copolymer; COC, cyclic olefin copolymer; COP, cyclic olefin polymer; D, ultraviolet light; E, ethanol. a Properties such as transparency, stiffness, and temperature resistance are tunable according to the specific composition.

Microengineering

industrial players in the biomedical market (Micronit, Dolomite, Illumina, IMT) are focusing their efforts on the development of cost-effective fabrication processes. Thermoplastics unlock upscalable fabrication processes, which makes them attractive for production and standardization (Becker and Locascio, 2002). Recent developments in 3D printing techniques in the field have introduced new materials, mostly resins based on polyester/polyether oligomers that have an acrylate or methacrylate group (e.g., polycaprolactone) and biodegradable compounds derived from methacrylate-functionalized polyesters (Zhou, 2017). Although 3D printing would bring automation at a relatively low cost, the formulation of an ink that would possess all the necessary properties for an OOC device is still challenging. In an effort to recreate the physiological and pathological responses of tissues and organs in OOCs, research is being conducted into the incorporation of a new category of materials—known as smart materials—into the chip. Smart materials can be defined as materials with one or more key properties that can be altered in response to a defined stimulus. These compounds can change, for example, their color, shape, rigidity, opacity, or porosity in response to a stimulus such as an alteration in physiological properties (temperature, pH, enzyme concentrations) or an external stimulation (light, electrical current, magnetic field). Materials with smart properties include polymers, proteins, and nanoparticles, and have been reviewed by Verma et al. (2016). In OOC devices, they can act as actuators, sensors, or on-demand self-assembly agents. By way of illustration, it is possible to reproduce mechanical stimulations applying to different biological tissues by including polypyrrole actuators in devices. When an electric field is applied, polypyrrole can expand, stimulating cells mechanically. The stimulation of epithelial cells using this type of actuator is illustrated in Fig. 3.4. Although the association of smart materials and OOC applications appears promising, especially in modeling cells such as muscles or neurons, the incorporation of such smart materials into the design is challenging. Moreover, there is still room for the development of more efficient smart materials that would be sensitive to specific changes in stimulus but stable with regard to storage and to other fluctuations in the environment.

FIGURE 3.4 Schematic illustration of the principle of the in vitro mechanical stimulation of MadinDarby canine kidney epithelial cells using polypyrrole microactuators. Adapted with permission from Svennersten, K., Berggren, M., Richter-Dahlfors, A., Jager, E.W.H., 2011. Mechanical stimulation of epithelial cells using polypyrrole microactuators. Lab Chip 11, 32873293.

59

60

CHAPTER 3 Organs-on-a-chip engineering

Raw synthetic materials are often not sufficient to recreate the environment within the human body. It is, however, possible to functionalize surfaces to better mimic the 3D microenvironment and to model the interfaces of specific tissues (Ahadian et al., 2018). Hydrogels, which consist of networks of hydrophilic polymer chains that can contain a large amount of water, can be used to attain this goal. They often act as the biomaterial scaffold on which the cells grow, spread, and proliferate. They must therefore possess adequate stiffness to support the cells, permeability to allow oxygen supply, and biocompatibility. Hydrogels can be natural or synthetic. While natural compounds are typically nontoxic and closer to those found in the body, they present several disadvantages compared with synthetic hydrogels, which have physical and chemical properties that can be easily tuned depending on the desired function. For example, the surface of PDMS can be altered by coating with hydrogels (Ergir et al., 2018), attaining a different flexibility to better mimic a given organ. Other modifications can include peptide attachment, photo-crosslinkable moieties, and combination with other materials (Ahadian et al., 2018). It is, however, worth mentioning that synthetic hydrogels are often supplemented with adhesion molecules such as laminin or fibronectin, as they lack natural cell adhesion ligands (Verhulsel et al., 2014). Examples of hydrogels that have been used in the production of OOC platforms and tissue engineering include collagen, Matrigel, poly(ethylene oxide), poly (vinyl alcohol), poly(acrylic acid), gelatin, gelatin methacryloyl, and fibrin (Zhang et al., 2018). One of the most important engineering challenges is to shape such a hydrogel into a stable structure. The choice of fabrication process is therefore critical, with microfabrication and 3D printing presenting very promising prospects (Zhang et al., 2018).

Fabrication process The fabrication of an OOC platform benefits from extensive experience in microfabrication in other fields. It mainly differs from other applications in its greater complexity and in requiring hybrid platforms. Different processes are needed for each of the key elements of an OOC platform, and these elements can be grouped into four categories: the microfluidic chip, the microtissues, the components necessary to generate the desired stimulus, and components for sensors (Ahadian et al., 2018; Yang et al., 2017). We briefly review the classical techniques for microfluidic devices, then focus more closely on two aspects that are inherently linked to OOC platforms, the insertion of a membrane into the device and fabrication using 3D bioprinting, a promising approach that makes it possible to build complete bio-platforms in a continuous manner. Methods for the fabrication of stimulation and sensing features are discussed separately in the “Stimulation and sensing” section.

Microengineering

The fabrication process of the microfluidic device is inextricably linked to its design (multilayer, pattern resolution), its connections with external equipment, and the material selection, where material properties give access to a limited set of techniques. Most often, an OOC device is used for research purposes, which allows for manual production and single use. In such a case, there are numerous possibilities in terms of material selection and available fabrication techniques, and costs are less of a concern. More important constraints are encountered when a device is intended for large-scale production, and costs and upscaling potential are also taken into account (Tsao, 2016; Iliescu et al., 2012). In drug-screening applications the goal is typically to reduce the cost per data point. The manufacturing of the chip body and microstructures is commonly achieved using one of the microfabrication techniques described in Table 3.2. The literature contains helpful reviews that discuss the advantages and limitations of these methods (Ahadian et al., 2018; Faustino et al., 2016; Kuo et al., 2019). In Table 3.3, we provide references to individual elements of the literature that, when combined, describe all the methods used with the most common materials for OOC microfluidic devices. Once the various elements and layers have been produced, the challenge is to ensure both sealing and the connection of the device with the necessary external equipment. These features must be taken into account early in the design phase. In their review, Temiz et al. (2015) describe the different methods for achieving sealing and fluid and electrical connection. The choice of method also depends on the material’s properties and the context in which the device is produced (small vs large scale, single vs multiple use, research vs commercial product). Sealing typically consists of bonding two layers of the same or different materials using an adhesive, or of inserting a gasket to form channel walls. Electrical connections consist of electrodes or contact pads, which should be accessible on one side of the device. The fluidic connections mainly take one of four forms: a gasket, a soft-bonded port, tubing inserted directly into the access hole, or commercially available standard fittings. Apart from conventional 2D chip assembly techniques, OOC platforms often require the addition of a membrane (see the “Stimulation and sensing” section). The selection process of the membrane is critical with regard to the integration of the desired biofunctions into the in vitro model. Depending on the organ model selected, and as part of the general design and fabrication process, the integration of a membrane can be complex. To support readers who would like to develop an OOC device that includes a membrane, we propose a step-by-step approach for membrane selection and integration (Fig. 3.5). Different techniques can be used to combine microfluidic devices and membranes (De Jong et al., 2006). The two main approaches are membrane fabrication during chip fabrication and separate fabrication and later assembly with other chip elements. In this approach the membrane is first produced using a range of processes, depending on the desired material and characteristics. Fig. 3.6 shows the fabrication of a porous membrane to be integrated into a lung-on-a-chip

61

62

CHAPTER 3 Organs-on-a-chip engineering

Table 3.2 Microfabrication techniques for microfluidic devices. Method

Short description

Principal advantages

Principal limitations

Wet etching

Chemical removal of material layers using a mask, in liquid phase

Short run time Cost-efficient

Not suitable for highaspect-ratio features

Dry etching

Physical removal of material layers using a mask, using plasma or gas

High precision High resolution Upscalable process

Clean-room process Safety

Photolithography

Pattern transfer from a mask to a light-sensitive material via exposure to ultraviolet light, then etching

High precision Cost-effective

Clean-room process Material availability Long processing time Requires planar surface

Soft lithography

Pattern transfer from an elastomeric stamp, mold, or master to the substrate

Applicable to rough and flexible substrates

Clean-room process Long processing time Costs Precision

Microinjection molding

Injection of melted plastic into a mold

Large-scale process

Residual stress in the product Not suitable for highaspect-ratio features

Precision Safety

Short run time Low cost Can circumvent sealing step Hot embossing

Microreplication of structures on a heated plastic substrate

Low residual stress in the product Suitable for high-aspectratio features

Long processing time

Micromilling

Mechanical ablation of the material on the micrometer scale

Large-scale process Suitable for high-aspectratio features

Long processing time Surface roughness Limited resolution Requires further sealing steps

3D printing

Layer-by-layer addition of material ink

Accessible to biomaterials Upscalable process Short run time No need for sealing steps Cost-efficient Versatile

Available materials

Requires further sealing steps

Resolution

Microengineering

Table 3.3 Review of microfabrication and sealing methods associated with materials for microfluidic devices. Class

Material

Method

References

Inorganic

Glass and silicon

Review

Polydimethylsiloxane

Review Rapid prototyping Scaffold removal Review Hot embossing Review

Iliescu et al. (2012) and Wang et al. (2018a) Mcdonald et al. (2000) Anderson et al. (2000)

Polymer

Poly(methyl methacrylate) (from the thermoplastics family) Thermoplastics in general

Micromilling Hot embossing Micrometerinjection molding

Saggiomo and Velders (2015) Chen et al. (2008a) Mathur et al. (2009) Becker and Locascio (2002) Guckenberger et al. (2015) Becker and Heim (2000) Attia et al. (2009)

device (Huh et al., 2013); the membrane can then be inserted between two layers of the device body so that all components are sealed together. The sealing step represents a major challenge for this technique, as it needs to guarantee a tight and robust junction at each interface while preserving membrane structure. The assembly methods to bind the membrane and the microfluidic parts are similar to those employed for chip sealing, and include clamping, gluing, and thermal fusion. The choice of the method greatly depends on the material used to fabricate the membrane. For example, PDMS membranes are widely used as PDMS is really easy to seal (please refer to the “Compression” section for detailed information). Materials other than PDMS—for instance, hard plastics—are best assembled using thermal diffusion bonding (Sonntag et al., 2016). Binding can also be improved by specific surface treatments. Pocock et al. (2016) (Murphy and Atala, 2014) were able to assemble a microdevice made up of a polycarbonate membrane situated between two glass substrates by functionalizing the polycarbonate membrane with ammonia, performing a plasma treatment on the substrates, and heating all the layers together at a low temperature (less than the glass-transition temperature of polycarbonate). Membranes can also be produced and integrated during the device’s fabrication. Depending on the desired membrane properties, it can be fabricated as part

63

FIGURE 3.5 General approach for selecting a membrane for inclusion in an organ-on-a-chip device.

FIGURE 3.6 Example of the fabrication of a porous PDMS membrane by molding using a silicon master and cured PDMS counterpart. In this work, the mold pillars are 10 μm in diameter and 50 μm high, and the final PDMS membrane is 10 μm thick. PDMS, Polydimethylsiloxane. Adapted with permission from Huh, D., et al., 2013. Microfabrication of human organs-on-chips. Nat. Protoc. 8, 11, 21352157.

Microengineering

of the chip body by typical microengineering methods or created as a postprocessing step (De Jong et al., 2006). So far we have discussed the fabrication of the microfluidic device and its features. What distinguishes an OOC device from other microfluidic chips is the combination of the device itself with the model tissue. The insertion of this live biological material into a complex system constitutes one of the main challenges in the production of an OOC platform as the cells deposition method has to preserve their viability. In most cases, the microfluidic chip is first fabricated according to the techniques described previously (Table 3.2), and the cells are then manually deposited into the microfluidic channels, for example, using a syringe. The channels are often precoated with the extracellular matrix to promote cell adhesion and provide an appropriate physiological environment. Afterward, the whole chip can be incubated, and an appropriate culture medium is typically flowed through the microchannels to promote cell/tissue growth. Heart-on-a-chip is one example of an initial fabrication of the microfluidic device, here by 3D printing of the chip, sensors, and actuators, followed by seeding the cells (Lind et al., 2017). Here, everything but the cells is printed in a single, continuous step using multiple printer heads, and the self-assembly of the tissue is guided and controlled by printed microstructures. Such postprocessing cell-loading methods are generally suitable for research purposes but involve multiple individual steps that make overall OOC device fabrication more complex. The lack of reproducibility and upscaling limitations constitute a major obstacle when it comes to standardizing OOC platforms, especially when these steps are performed manually. Recent advances in the fabrication process were made possible by the development of 3D bioprinting, a biofabrication strategy of printing viable cells and constituting 3D tissue structures in a single continuous procedure with great accuracy (Avci et al., 2018; Lee and Cho, 2016; Murphy and Atala, 2014; Vijayavenkataraman et al., 2018). The principle is similar to that of conventional 3D printing—the layer-by-layer deposition of bio-inks onto, in this case, a biocompatible scaffold—and can be performed using various printing techniques, including stereo lithography, inkjet, extrusion, and laser-assisted bioprinting. Depending on the process, different cell viabilities, resolutions, and printing speeds can be achieved and inks with different viscosities can be processed. An overview of some of the aforementioned techniques is given in Fig. 3.7. It is worth mentioning that the selected process should be mild enough not to alter the viability of the cells; this is why cells are often encapsulated in biocompatible materials such as hydrogels, to prevent mechanical damage (Yang et al., 2017). The combination of the 3D bioprinting of biological cells and microfluidic devices to build an OOC platform can follow two different approaches (Yi et al., 2017). Two-step fabrication is possible, involving first the production of a microfluidic chip using any conventional microfabrication method and then the 3D bioprinting of cells or organs on the prefabricated chip. This method makes it possible to design OOCs with multiple cellular arrangements and structures. For example, Chang et al. (Malkoc, 2018) were able to fabricate a liver-on-a-chip by

65

66

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.7 Selected biofabrication approaches involving the use of hydrogels as bioinks: (A) inkjet bioprinters, (B) microextrusion bioprinters, and (C) laser-assisted bioprinting. Adapted with permission from Malda, J., et al., 2013. 25th Anniversary article: engineering hydrogels for biofabrication. Adv. Mater. 25, 36, 50115028.

3D extrusion-printing an alginate bioink containing specific liver cells onto a prefabricated PDMS chamber and then assembling it with glass microfluidic channels. They were thus able to show that the direct printing of cells onto the microfluidic device led to better structural adaptability of the cells with respect to the design specifications. Nevertheless, this method has several disadvantages, as it still cannot be completely automated because of its sequence of steps, performed on different machines. Manual intervention is therefore needed to assemble the device, which can lead to a lack of reproducibility and to possible contamination in the culture. The single-step fabrication approach is believed to solve these issues: it consists of printing the entire chip device, including the cells, the chip, the microfluidic channels, and eventual sensors and actuators, in only one continuous and automated process. This can be accomplished using various nozzles and biomaterials in the printer. In the last few years, several groups have achieved the one-step production of OOCs such as a liver-on-a-chip (Yi et al., 2017), a nervous system (Johnson et al., 2016), and a kidney (Homan et al., 2016). It is possible to 3D-print actuators using smart materials and embedded sensors directly onto the chip, leading to a well-equipped and monitored device fabricated in only one step (Yang et al., 2017). While the process and the device may be more complex to design, this method would solve the automation and manual handling problems, making it more time-efficient and reproducible. Despite the potential offered by this technology, it presents diverse challenges that must be overcome. First, a higher printing resolution would be required to produce more complex structures, such as capillary networks. A higher printing speed would also be desirable to improve cell viability and decrease overall processing time. Further development is also required before this technology can be scaled up and standardized (Ma et al., 2018). One of the key limitations of 3D

Table 3.4 Overview of the different bioinks used for the bioprinting of scaffolds (Merceron and Murphy, 2015). Hydrogel type

Bioink

Advantages

Disadvantages

3D printing technique

Natural

Collagen

Highly biocompatible, easy to use, low cost Highly biocompatible

Poor structural and mechanical properties, slow cross-linking Melting temperature ,37 C

Inkjet, extrusion Extrusion

High biocompatibility, quick gelation time Highly biocompatible, highly customizable (when mixed to other hydrogels)

Lack structural and mechanical stability, quick degradation time Highly soluble at room temperature

Inkjet

Alginate

Low cost, abundantly available, shape fidelity

Not biomimetic for mammalian cells, high calcium content

Agarose

Good thermal gelation properties

Matrigel

Good thermal gelation properties, good biomimicry High biocompatibility, easily functionalized, versatile mechanical properties (depending on polymerization degree), low cost Good shape fidelity after printing

Not enzymatically biodegradable and not biomimetic in mammals, difficult to print Unfavorable gelation kinetics

Gelatin

Fibrin Hyaluronic acid

Synthetic

Poly (ethylene glycol) Poloxamers

PEG, Polyethylene glycol.

Low cell adhesion, not enzymatically biodegradable

Low cell adhesion, not enzymatically biodegradable, poor long-term structural properties, potential cytotoxicity

Extrusion (when mixed to gelatin) Inkjet, laser, extrusion Inkjet, extrusion Inkjet, laser Inkjet, extrusion

Extrusion

Example of applications Scaffold Sacrificial material (if pure), scaffold (if mixed), additive Adjuvant Scaffold (if mixed)

Scaffold

Sacrificial material, scaffold Single-layered structures Cell encapsulation, cross-linking agent, scaffold Sacrificial material

68

CHAPTER 3 Organs-on-a-chip engineering

bioprinting is the restricted choice of biomaterials that can be used. Material properties must address requirements such as printability (suitable viscous and rheological properties), biocompatibility, degradation kinetics and byproducts (adequate kinetics, nontoxic byproducts), structural and mechanical properties, and material biomimicry (tissue specificity) (Murphy and Atala, 2014). Table 3.4 gathers a list of bioinks that are currently used for 3D bioprinting. Finally, there are still fundamental issues with building a suitable vasculature, which is required to supply the cells with nutrients and oxygen. These issues include the lack of resolution for 3D bioprinters, as the size of vascular networks can reach down to 3 μm while the one for high-resolution bioprinters is around 20 μm. Moreover, the printing time of such small structures must not compromise cells viability, which is still challenging. Various approaches are currently being developed but are still limited by nozzle dimensions and the complexity of the structure (Malkoc, 2018). Overall, there is a clear need to broaden the spectrum of biomaterials that can be used for 3D bioprinting so that this technology can be applied to the production of OOCs and tissues. For this technique to fulfill its potential, advancements in cell sources, printing technologies, and the combination of techniques are necessary.

Engineering fluid control for organ-on-chips Microfluidic chips present a higher surface-to-volume ratio than traditional Petri dishes and well plates, resulting in a lower volume of medium available for each individual cell. Such devices allow the detection of cell-secreted factors at early time points in the experiment, since the limit of detection is reached earlier. In contrast, toxins and waste products saturate faster and nutrients are depleted earlier; the cell culture medium must consequently be renewed more frequently. To overcome this challenge, various perfusion systems can be plugged into the chip or directly integrated with it. In addition to refreshing the medium, cell perfusion can be implemented to reproduce physiological functions of the body, such as laminar flow in blood vessels, or to recreate living cell environments, such as biochemical gradients or cell signaling (Rothbauer et al., 2018). Each type of application requires the integration of equipment with specific performance, versatility, volumetric capacity, and automation properties. Adapting microfluidics to cell culture has introduced some constraints on microfluidic instruments that were not seen with the traditional bioanalytical and physicochemistry applications that initiated this field. These constraints include the following:

• Robustness for long-term experiments: Typical physicochemical experiments •

last from several minutes to 1 day, compared with several weeks for cell culture experiments. Incubator proof: Cells are typically grown in incubators with controlled temperature and CO2 levels. If chips are impermeable to gas, the instrument

Engineering fluid control for organ-on-chips







itself should guarantee that the medium perfused is buffered with appropriate CO2 and O2 concentrations. The instrument should also withstand exposure to high air humidity. Contamination: Cell culturing must be performed in an aseptic environment. Instruments should be equipped with adapted HEPA filters and all parts in contact with the cell culture medium (such as tubing and syringe) should be disposable or autoclavable for reuse. High-throughput: Given the duration and variability of cell biology experiments, samples are typically prepared with at least three replicates at once rather than one sample at a time for physicochemistry experiments. Instruments should therefore be able to perfuse numerous chips in parallel. Footprint: The instruments should have a reduced footprint to fit into the incubator or on a microscope stage for live-cell imaging.

Thus in comparison with traditional in vitro cell cultures in Petri dishes or flasks, fluid control (actuating, directing, monitoring, and controlling the flow of all liquids) is key in the design of OOC environments. Before reviewing fluidcontrol options, it is important to understand the functions of a fluid control system (Fig. 3.8):

• Medium exchange refers to the process of changing the culture medium at regular intervals.

• Perfusion relates to the passage of fluid through a microfluidic system for

• •

• •

medium exchange and mimicking physiological flow conditions. Perfusion rates are typically low and can be pulsatile, constant, or with a defined flow pattern. Controlling the flow rate is particularly important to stimulate the cells mechanically (see the “Mechanical” section on mechanical stimulation). Flow recirculation involves perfusion in a closed circulatory system. Injection or chemical stimulation denotes the delivery of one or more chemical compounds to simulate drug delivery. Notably, implementing concentration gradients both spatially and temporally is particularly challenging. Sampling relates to the action of removing part of the fluid circulating within the chip, typically to perform measurements. The combination of the abovementioned functions.

The main engineering challenge for fluid control in OOCs is the small volume of fluids (Wikswo et al., 2013). In particular, in applications studying signaling and organ-to-organ interactions (Faley et al., 2008), nonphysiological dilutions of metabolites, hormones, and paracrine signals should be avoided. Since the size of the tissues formed in OOCs is up to 1 million times smaller than organs in the human body, a circulation volume as small as 5 μL must be mastered. This, in turn, leads to stringent requirements for the fluidic components—including, for example, the minimization of dead volumes. Moreover, cost comes into play since the parallelization of experiments over days to weeks must be ensured.

69

70

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.8 Example of a complex organ-on-a-chip, where recirculation is performed among the different organs using a pump and on-chip microvalves. Reproduced with permission from Zhang, Y.S., et al., 2017. Multisensor-integrated organs-on-chips platform for automated and continual in situ monitoring of organoid behaviors. Proc. Natl. Acad. Sci. 114, 12, E2293E2302.

In the following sections, we offer insights into the options available for controlling fluids in OOC systems and assist the reader in evaluating these options for future applications. These sections are organized into four parts. The first is dedicated to liquid actuation options, that is, pumps. The second and third parts

Engineering fluid control for organ-on-chips

review microvalves and flow-control options, respectively. The fourth part discusses practical implementations of flow control in OOCs.

Liquid actuation We list below the microfluidic perfusion systems adapted for cell cultures, their strengths, and their limitations, and indicate some applications for which they are relevant. Table 3.5 gives an overview of the main liquid-actuation methods.

Pipetting robots The simplest fluid actuation solution consists of manual pipetting or pipetting robots to remove and add liquids within the OOC system. Because of the high degree of automation, liquid-handling robots are a competitive solution to exchanging cell culture medium at regular intervals and to administering drugs or removing samples. Airtight ports permit the use of pipetting robots as disposable pumps (CSEM, 2010). Pipetting robots can sequentially serve different microfluidic chips while ensuring sterility (tip exchange, integration into laminar-air-flowcontrolled environments). The flow rate can be easily controlled and altered during the experiment. This approach alone, however, is not adapted for continuous perfusion, since the pipetting head would have to be immobilized during the whole perfusion period and would thus be limited to addressing one chip only over a long time period. Furthermore, pipetting robots generally have a large footprint and are more expensive that alternative solutions.

Gravity-driven flow The flow in the chip results from the difference in liquid height between the fluid inlet and outlet (Ong et al., 2017; Chen et al., 2011a; Ayyapan et al., 2016; Esch et al., 2016). This is the simplest and least-expensive method and can be easily scaled to multiple chips, as the asymmetry in liquid levels is directly set on the chip (Fig. 3.9). Furthermore, chip wells can be positioned so that the spacing is compatible with pipetting robots for industrial use. The main limitation of this technique is that the flow rate is not constant over time, as the liquid height and the associated hydrostatic pressure continuously change. To overcome this obstacle, chips can be placed on a rocking plate and tilted constantly. The frequency and amplitude of rocking can be adjusted to achieve the desired flow rate. One example is the commercial OOC platform OrganoPlate (Mimetas, Leiden, The Netherlands), a 96-well plate format with a related perfusion rocker to control the fluid flow rate and the number of desired rocking cycles. Until recently, the geometry of the flow obtained was limited to bidirectional laminar flow, which may not be physiological for certain tissues and may prevent the formation of the appropriate phenotype for specific cell types, such as endothelial cells, which normally polarize against the flow. Wang et al. have demonstrated long-term recirculating unidirectional perfusion with gravity-driven flow (Wang and Shuler, 2018), circumventing one of the most severe limitations of

71

72

CHAPTER 3 Organs-on-a-chip engineering

Table 3.5 Overview of the most relevant liquid-actuation mechanisms. Actuation principle Peripheral

Manual or robotized pipetting Gravitydriven flow

Peristaltic pump

Syringe pump

Pressure controller

Membrane pump Viscous drag pump On-chip

Membrane pump

Surfacedriven flow

Features

References

Standardized solution for medium exchange. Limited applicability for perfusion. Large footprint and initial costs Simple and cost-efficient. Easily scalable to multiple chips. Bidirectional, tunable flow rates achievable using tilt tables Simplest solution for recirculating medium. Not applicable for circulating cells due to tube pinching. Not applicable for very low fluid volumes Unidirectional flow with direct control of the flow rates independently of chip geometry or liquid properties. Relatively large footprint and costs High flow stability and response time. Can reproduce complex flow patterns. Relatively small footprint, in particular when multiplexing Small footprint, pulsatile flow, relatively low costs Bidirectional, fast response time, pulsation-free, small footprint

CSEM (2010), Tecan (2019), and Hamilton (2019)

Very low dead volumes and flow rates. Well adapted for recirculating low fluid volumes. Use of polydimethylsiloxane membrane leads to adsorption of small molecules. Increases the complexity and cost of the disposable Unidirectional, passive liquid actuation. Low-cost and compact solution. Pumping rate changes with time (no external control)

Ong et al. (2017), Chen et al. (2011a), Ayyapan et al. (2016), and Esch et al. (2016) Caplin et al. (2015), Maschmeyer et al. (2015a), Kieninger et al. (2018), Alexander et al. (2018a), and Moya et al. (2018) Agarwal et al. (2013), Loskill et al. (2015, 2017), Li and Tian (2018), Misun et al. (2016), and Rennert et al. (2015) Fluigent (2019a), Feaver et al. (2013), and Benam et al. (2015)

Jenke et al. (2017) and Wang and Fu (2018) Dubeau-Laramée et al. (2014), Chauvet et al. (2014), and CSEM (2010) Huang et al. (2008), Jeong and Konishi (2008), Lee et al. (2007), Kim et al. (2006), Wang and Lee (2006), and Materne et al. (2015)

Juncker et al. (2002) and Walker and Beebe (2002)

Engineering fluid control for organ-on-chips

FIGURE 3.9 Example of a gravity-flow-driven cell-based assay. (A) Photo of the bottom of the 32-unit perfusion array. (B) Magnified picture of a flow unit. (C) Scheme of a cross-section of the flow unit and the associated gravity-driven flow. Reproduced with permission from Chen, S.Y.C., Hung, P.J., Lee, P.J., 2011a. Microfluidic array for threedimensional perfusion culture of human mammary epithelial cells. Biomed. Microdevices, 13, 4, 753758.

simple gravity-driven flow. As in conventional well-plate cultures, these chips can be placed directly into an incubator, and the open wells ensure CO2 exchange to buffer the medium independently of the chip material. Combined with pipetting robots to fill the inlet reservoirs and dispense test compounds, gravity-driven flow is a powerful method to parallelize and standardize fluid flow for OOC.

Peristaltic pumps Peristaltic pumps are widely used in cell culture as they are compact and easy to use and connect (Caplin et al., 2015; Maschmeyer et al., 2015a; Kieninger et al., 2018; Alexander et al., 2018a; Moya et al., 2018) (Fig. 3.10). They can either be used to pump medium from source to waste or to recirculate medium in a chip. The number of samples that can be perfused in parallel varies from 1 to 24, depending on the number of tubes that can be coiled around the pump rotor. Peristaltic pumps deliver pulsatile flow. The frequency and amplitude of oscillations required to obtain a given flow rate vary between pumps by the number of rollers, their rotation frequency, and the diameter of the tubes. Most peristaltic pumps are software-driven to automate protocols such as periodic injection or simple flow variations, the latter including median flow rate increase or decrease (Mazzocchi et al., 2018). Most complex flow patterns, including sinusoidal flow, are difficult to achieve because of flow pulsation. In most cases, all tubes are coiled around one rotor and therefore deliver the same flow rate in every chip, which can be limiting for certain studies. Tube-pinching by the rollers damages the tubes, which must be replaced periodically, and this type of pump should not be used in experiments involving circulating cells. Most peristaltic pumps are designed to fit within an incubator, and the latest generation of micropumps are highly compact (e.g., MP2 micropumps from

73

74

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.10 Scheme of a six-chamber cell-based assay with a low-volume closed-loop fluidic circuit perfusing each sample. Reproduced under a Creative Commons Attributions 4.0 International License from Mazzocchi, A.R., Rajan, S.A.P., Votanopoulos, K.I., Hall, A.R., Skardal, A., 2018. In vitro patient-derived 3D mesothelioma tumor organoids facilitate patient-centric therapeutic screening. Sci. Rep. 8, 1, 112.

Elemental Scientific, Omaha, NE, United States, and RP-TX microperistaltic pumps from Takasago Fluidic Systems, Kobe, Japan). Recirculating fluid saves culture medium, which can be highly expensive for primary cells, and enriches the medium with cell-secreted factors. Enclosed circulation, however, requires that medium buffering with CO2 is achieved inside the chip, using either a gaspermeable material or a CO2-independent medium for short time periods (typically up to 2 hours).

Syringe pumps Syringe pumps were traditionally the instrument most frequently used in microfluidics (Agarwal et al., 2013; Loskill et al., 2015, 2017; Li and Tian, 2018; Misun et al., 2016; Rennert et al., 2015). They are also used in the biomedical field and in hospitals, for drug dosing and calibrated injections. A wide range of syringe pumps is available on the market, delivering flow rates of 0.012300 mL/min. Most syringe pumps are standardized instruments since they are designed to be compatible with a variety of syringes. Their flow stability and intuitive user experience make them the preferred choice of biologists, but their volume capacity is limited by the volume of the syringe. Their footprint and compatibility with incubators are other restrictive factors, as most syringe pumps are

Engineering fluid control for organ-on-chips

not designed to work in a humid environment. To overcome this limitation, syringe pumps can be placed outside the incubator and aspirate the medium from the chip outlet instead of pushing it into its inlet. The source can be a reservoir, sealed with a permeable membrane or a non-gas-tight lid, which is placed in the incubator, connected to the chip, or even designed as part of the chip. Using such a reservoir at the chip fluid inlet also overcomes the problem of medium buffering, which is complex in perfusion mode because syringes and many types of tubing are gas-impermeable. Throughput is also limited since in most cases, a syringe pump is designed for one to two syringes. Some manufacturers have developed accessories to push up to 10 syringes with one plunger simultaneously at identical flow rates (PHD ULTRA, Harvard Apparatus) and individual but connectable syringe pump units. Another drawback of syringe pumps is related to the constant delivery of cell suspensions; since the syringes are fixed on the syringe pumps, cells in suspension begin to sediment rapidly, compromising the experiment as a homogeneous cell suspension can no longer be supplied. The newest commercially available solutions rotate the syringe itself or stirring bottles connected directly to the pump (Cetoni, 2019).

Pressure controller In microfluidic devices, flows are laminar and follow the equation ΔP 5 R 3 Q, where R is the hydrodynamic resistance of the chip, ΔP is the pressure drop between the chip inlet and outlet, and Q is the volumetric flow inside the chip. This equation demonstrates that the flow can be set directly using flow controllers or by adjusting the pressure with pressure controllers. Pressure controllers such as Flow EZ (Fluigent, Paris, France) apply a controlled pressure at the liquid’s surface, which pushes out the fluid at a controlled flow rate. Fig. 3.11 illustrates a typical implementation, where the microfluidic chip is connected to a reservoir. Pressure controllers have shorter response times and better flow stability than standard syringe pumps. When a flow sensor is added to the system, a feedback loop can allow the pressure controller to constantly adjust the pressure to maintain the desired flow rate in the chip. Such a setup provides additional information to the user, as the system records both the pressure and the flow rate over the course of the experiment. This information is valuable for cell culturing as cells can easily clog the chip, resulting in a pressure increase to maintain the flow rate (Bondot et al., 2012). The responsiveness of pressure controllers is particularly useful when reproducing complex flow patterns such as aortic coronary flow. For sensitive cells exposed to a dynamic environment the controlling of all experimental parameters is of major importance. Endothelial cells, for example, have been shown to be sensitive to complex shear stress frequency harmonics (Feaver et al., 2013). Complex flow patterns can be directly coded and implemented in the software that drives the flow controllers. Working with pressure controllers is also particularly convenient when reproducing airliquid interfaces for OOC models such as skin or lung (Benam et al., 2016).

75

76

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.11 Scheme of a pressure-controlled open fluid circuit. The reservoir is pressurized, resulting in a fluid flow through the microfluidic chip. Reproduced with permission from Fluigent (www.fluigent.com).

Regarding compatibility with the incubator, pressure controllers have a competitive advantage over syringe pumps and peristaltic pumps since the user can adjust the gas composition to pressurize the medium. Furthermore, agitation of the medium can be easily implemented because pressure systems do not restrain tubes in a fixed position. Finally, multiplexing is another major benefit of pressure controllers, as a pressure source or stock solution can be split into multiple pressure or liquid channels, increasing throughput and reducing footprint (Fig. 3.12).

Membrane or diaphragm pumps Membrane and diaphragm pumps are formed from a cavity and two one-way valves. As the membrane or diaphragm is actuated, the volume of the cavity varies, creating repeating cycles or aspiration and expulsion of the liquid within the cavity (Fig. 3.13). Pneumatic, piezoelectric, electrostatic or electroactive, thermal, and electromagnetic forces are used to actuate the flexible membrane. Piezoelectric pumps are the most common, and various products are available commercially for microfluidics. Flow rates ranging from 0.1 μL/min to tens of milliliters per minute can be achieved depending on the size of the cavity, the actuation voltage, and the

Engineering fluid control for organ-on-chips

FIGURE 3.12 Scheme of the use of pressure to multiplex experiments. On the left, a fluidic manifold is connected to the liquid output to perfuse multiple chips from one reservoir. On the right, a pneumatic manifold divides the pressure between multiple sources to perfuse each chip with a dedicated medium. Reproduced with permission from Fluigent (www.fluigent.com, Laura Lelli).

FIGURE 3.13 Example of a piezoelectric membrane pump and its cross-section: (1) piezo; (2) adhesive; (3) actuation diaphragm; (4) inlet valve; (5) outlet valve. Reproduced under a Creative Commons Attributions 4.0 International License from Jenke, C., et al., 2017. The combination of micro diaphragm pumps and flow sensors for single stroke based liquid flow control. Sensors 17, 4, 755.

frequency (Wang and Fu, 2018). Membrane pumps provide a pulsatile flow, and the flow rate can be tuned by changing the actuation frequency.

Viscous drag pumps Viscous drag pumps use the friction between the fluid and a moving surface to convey the fluid from the inlet to the outlet. Fig. 3.14 illustrates a viscous drag pump, with its rotating, disk-shaped surfaces that drag the fluid. A scraping unit conveys fluid to and away from the moving surface. Viscous drag pumps have the advantage of being bidirectional and entirely pulsation-free. Their main drawbacks are the limited pressure generated for aqueous solutions (0.51 bar) and

77

78

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.14 Schematic representation of the working principle of the viscous drag pump. Reproduced with permission from the Swiss Center for Electronics and Microtechnology (CSEM).

FIGURE 3.15 The TURBISC pump designed by the CSEM. CSEM, Swiss Center for Electronics and Microtechnology. Reproduced with permission from Swiss Center for Electronics and Microtechnology (CSEM).

that the flow rate achieved at a given rotation speed depends on the back pressure. Thus a flow sensor is often used in combination with viscous drag pump to enable closed loop control of the flow-rate. Fig. 3.15 shows a pump designed by the Swiss Center for Electronics and Microtechnology (CSEM, 2010). It has been used in different microfluidic flow cells (Dubeau-Larame´e et al., 2014; Chauvet et al., 2014) for its small footprint (width ,8 mm), its pulsation-free flow, and small inner volume (about 100 μL).

Engineering fluid control for organ-on-chips

Because the housing and seal are made out of polyetherimide and polyetheretherketone, respectively, the pump is resistant to most chemicals. In particular, this pump has been used for flow cytometry setups in the International Space Station (Dubeau-Larame´e et al., 2014). Due to the high shear rate induced by this pumping principle, the effect on cell viability needs to be verified when pumping cell solutions. For instance, it has been shown that hepatocytes are damaged during the pumping process while yeast cells remain unaffected.

On-chip pumps Beyond the peripheral (off-chip) pumps mentioned above, different on-chip pumps have been investigated and tested. On-chip pumps offer the option to create self-contained, closed microfluidic systems (Boyd-Moss et al., 2016), have inherently lower dead volumes, and reduce the number of fluidic interconnects on the chip. In general, on-chip pumps achieve very low flow rates, but they tend to increase the complexity of the microfluidic manufacturing processes and add costs to the disposable. On-chip pumps use a wide range of actuation principles, either passive (e.g., surface-driven flow, osmosis-driven flow) or active [e.g., membrane pumps, rotary pumps (Darby et al., 2010), bubble and acoustic pumps (Oskooei and Gu¨nther, 2015), and electrochemical, electrokinetic, and electroosmotic pumps]. Several groups have reviewed micropumps, including Au et al. (2011), Byun et al. (2014), and Wang and Fu (2018). Surface-driven-flow and membrane pumps are the most relevant technologies for OOCs because of their compatibility with cell cultures and their technological maturity.

Membrane pumps On-chip membrane pumps are based on the same concept as peripheral membrane pumps (see the “Membrane or diaphragm pumps” section), except that the whole pump is built into the microfluidic chip (Fig. 3.16). Unidirectionality of flow is achieved using check valves (Kim et al., 2006) or the sequential actuation of consecutive valves, thus pushing the displaced volume in the desired direction (Huang et al., 2008; Jeong and Konishi, 2008; Lee et al., 2007; Kim et al., 2006; Wang and Lee, 2006). Various implementations of pneumatic actuation exist, such as peristaltic pumps based on three consecutively arranged membrane valves (Unger et al., 2000; Studer et al., 2004), peristaltic micropumps based on a serpentine geometry (Huang et al., 2008; Wang and Lee, 2006), and the doormat micropump (Grover et al., 2003) (Fig. 3.17). Single-stroke micropumps are based on a single pneumatic control line (Lai and Folch, 2011). Generally, pneumatic micropumps achieve maximum flow rates in the range of nanoliters per second to hundreds of nanoliters per second for an actuation frequency around 2075 Hz and a driving pressure of some tens to hundreds of kilopascals. The dead volume of these pumps is very low, in the range of hundreds of picoliters.

79

80

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.16 Peristaltic micropump based on the sequential deformation of three pneumatic actuators. Reproduced with permission from Jeong, O.C., Konishi, S., 2008. Fabrication of a peristaltic micro pump with novel cascaded actuators. J. Micromech. Microeng. 18, 025022.

FIGURE 3.17 Scheme of a diaphragm micropump consisting of three doormat microvalves. Reproduced with permission from Grover, W.H., Skelley, A.M., Liu, C.N., Lagally, E.T., Mathies, R.A., 2003. Monolithic membrane valves and diaphragm pumps for practical large-scale integration into glass microfluidic devices. Sens. Actuat. B 89, 315323.

Another common actuation approach is based on piezoelectric materials, as illustrated in Fig. 3.18 (Tracey et al., 2006). Piezoelectric pumps can achieve higher flow rates than pneumatically actuated pumps, since much higher actuation frequencies can be achieved. Despite the increasing use of polymer materials (PDMS or PMMA), however, piezoelectric pumps remain more expensive than other pumping schemes and thus are not well suited for disposable chips. In general, the main drawback of membrane pumps is the choice of flexible materials for the membrane. Many designs have been demonstrated with PDMS, which limits the scalability and applicability of these pumps because of the difficulty of the manufacturing scale-up and the nonspecific binding of small proteins to PDMS. Nonetheless, various membrane micropumps are available commercially, such as those manufactured by TissUse (Berlin, Germany) (Sonntag et al., 2010;

Engineering fluid control for organ-on-chips

FIGURE 3.18 Operation of a single piezoelectric micropump pumping from left to right. Reproduced with permission from Tracey, M.C., Johnston, I.D., Davis, J.B., Tan, C.K.L., 2006. Dual independent displacement-amplified micropumps with a single actuator. J. Micromech. Microeng. 16, 14441452.

FIGURE 3.19 Schemes of capillary pump operation. Reproduced with permission from Juncker, D., et al., 2002. Autonomous microfluidic capillary system. Anal. Chem. 74, 61396144.

Materne et al., 2015; Schimek et al., 2013), or Formulatrix (Bedford, MA, United States) (Formulatrix, 2019b).

Surface-driven flow or capillary pumps Surface-driven flow pumps or capillary pumps use the interplay between surface tension, surface chemistry, and topography to move the fluid in the direction that minimizes the free energies between the vapor, fluid, and surfaces (Fig. 3.19). An IBM group from Switzerland has pioneered the use of such capillary pumps (Juncker et al., 2002). Flow rates in the order of tens to hundreds of nanoliters per second have been demonstrated. Surface-driven flow pumps are compact, simple to integrate, and low-cost (Walker and Beebe, 2002). The main drawbacks are the limited volumes that can be actuated and the limitation to unidirectional flow.

81

82

CHAPTER 3 Organs-on-a-chip engineering

Furthermore, the pumping rate changes with time and cannot be controlled externally.

Valves and bubble traps Valves allow the user to control the fluid flow within a microfluidic chip. Used in combination with liquid actuation, valves permit, for instance, the flow from one channel to another to be redirected or the sequence of fluids flushed through the chip to be alternated. For liquid actuation, it is important to distinguish between off-chip valves and on-chip valves. Many companies offer commercial off-chip products, including solenoid valves. Since the pioneering work of the group led by Andreas Manz in 1993, different technologies have been proposed for on-chip microvalves (Au et al., 2011; Oh and Ahn, 2006). Table 3.6 describes their applicability for OOCs. Table 3.6 Overview of microvalve technologies used in organ-on-a-chip applications. Microvalve category

Applicability for organ-on-achip devices

Electrokinetic

Not applicable because of the use of high voltages and large variability

Pneumatic

Simple integration into microfluidic chips using commercial pressure controllers. However, the use of polydimethylsiloxane may cause issues because of adsorption of hydrophobic molecules

Pinch

Simple integration into microfluidic chips using mechanical actuators

Phase change

Requires integration of heating or cooling elements. Slow actuation time Useful to rectify pressure peaks during cell insertion, thus avoiding damaging the cells Applicable for chip devices using biocompatible electroactive polymers. Easy to control using low-voltage electronics

Burst

Electroactive polymers

References Lee et al. (2008), Kaigala et al. (2008), Jacobson et al. (1999), and Schasfoort et al. (1999) Unger et al. (2000), Studer et al. (2004), Grover et al. (2003), Sundararajan et al. (2005), Hosokawa and Maeda (2000), Yang and Lin (2007), Yoo et al. (2007), Irimia and Toner (2006), and van der Wijngaart et al. (2007) Pemble and Towe (1999), Weibel et al. (2005, 2007), Pilarski et al. (2005), and Formulatrix (2019a) Yang and Lin (2007) and Yoo et al. (2007) Cho et al. (2007) and Chen et al. (2008b) Tanaka et al. (2013), Parker (2015), and Carpi and Smela (2009)

Engineering fluid control for organ-on-chips

FIGURE 3.20 Principle of operation of a bubble trap. The bubbles present in the liquid at the inlet are removed by gas diffusion through a semipermeable membrane.

Bubble traps are often integrated into the microfluidic system to ensure stable operation by removing bubbles circulating in the microfluidics that may block it and potentially disturb the flow or the sensors. Introduced air bubbles are, moreover, lethal for cultured microtissues. Fig. 3.20 illustrates the principle of operation of bubble traps. The bubbles present in the liquid at the inlet are removed by gas diffusion through a semipermeable membrane. The efficiency of bubble removal depends on the contact surface between the liquid and the semipermeable membrane. This effect can be enhanced by using vacuum at the exhaust or by using micropillars in the fluid path to pin the air bubbles (Zhang et al., 2017). Different commercial products are available (Fluigent, 2019b; ElveFlow, 2019a).

Flow sensing Flow-rate and pressure are the two most important parameters to monitor and measure during the operation of OOCs. The flow-rate is indeed directly related to oxygen and nutrients perfusion of the OOC but also the shear stress to which the cells are exposed. Furthermore, exact control of low flow-rates is essential when emulating cell-to-cell interactions or spatiotemporal chemical gradients in order to achieve physiologically relevant concentrations of the different molecules in the medium. Pressure is another important parameter to monitor, both during cell loading into the microphysiological system and during cell perfusion. Cells can be easily damaged during cell loading due to pressure peaks, as reported by Wang et al. (2016). Pressure is also an important aspect of human physiology, in particular when mimicking blood circulation (Chen et al., 2017) or to induce mechanical stimulation (see the “Mechanical” section). Different flow-rate and pressure sensors technologies are introduced and compared in the following paragraphs.

83

84

CHAPTER 3 Organs-on-a-chip engineering

Table 3.7 Overview of flow-rate sensor technologies. Sensor principle

Response time (ms)

References

Optical/Capacitive flank monitoring Differential-pressure based Thermal calorimetry Thermal time-of-flight Coriolis force Gravimetric balance

0.4 ,1 40 625 50200 ,2000

Zengerle et al. (1995) Richter et al. (1999) Sensirion (2019) Ashauer et al. (1999) Bronkhorst (2019) Sartorius (2019)

Flow-rate sensors A large variety of sensors based on different fields of physics are available (Table 3.7). Selecting the appropriate flow meter adapted to the flow regime and fluid is critical for accurate measurements.

Thermal sensors A common technology relies on the calorimetric method. A microheater provides a minimal amount of heat to the medium monitored (around 1 C). Two temperature sensors, located on each side of the heater, detect temperature variations. The flow rate is then calculated based on the spread of heat, which is directly related to the flow rate. This method of monitoring flow is one of the simplest and least intrusive. It is easy to integrate into microelectromechanical devices, as very small heaters and sensors already exist. Nonetheless, it requires knowledge of the fluid density and specific heat capacity. These values should also be constant for the proper functioning of the sensor. In biological experiments the presence of cells or particles in the fluid may affect the fluid’s properties and the measurement. Furthermore, a change in temperature may not be acceptable to the viability of cells in the solution. Several other thermal flow meters are available that function in a similar way (Kuo et al., 2012). The hot wire uses a resistor as a heater and a sensing element. As the resistance is dependent on the temperature, a relationship between applied tension, temperature, and resulting resistance can be established. Other sensors use so-called time-of-flight sensing (Fig. 3.21). This technique uses only one sensor, which is located downstream of the heater. By observing the heat distribution over time, time-of-flight sensing can deduce fluid velocity and thus the flow rate.

Coriolis mass flow meters Use of mass flow meters in microfluidics is growing as the technology is improved for microscale flows. In a mass flow meter operating on the Coriolis principle (Fig. 3.22) the fluid flows on a vibrating channel. The flow rate through its mass will proportionally affect the frequency, phase shift, or amplitude of the initial vibration. The main advantage of this technology is the independence of

Engineering fluid control for organ-on-chips

FIGURE 3.21 Function principle of a time-of-flight thermal sensor. Reproduced under a Creative Commons Attributions 3.0 International License from Kuo, J.T.W., Yu, L., Meng, E., Kuo, J.T.W., Yu, L., Meng, E. Micromachined thermal flow sensors—a review. Micromachines 3, 3, 550573.

FIGURE 3.22 Schematic principle of the Coriolis flow sensor. Reproduced under a Creative Commons Attributions 4.0 International License from Lo¨tters, J.C., Lammerink, T.S., Groenesteijn, J., Haneveld, J., Wiegerink, R.J. Integrated thermal and microcoriolis flow sensing system with a dynamic flow range of more than five decades. Micromachines 2012, 3, 194203.

the measured flow rate from the properties of the liquid. These sensors can monitor gas flow or oils without any specific calibration; however, the technology remains expensive, and the small inner diameter of the fluidic path may not be suitable for biological experiments.

Differential pressure-based flow sensors Differential pressure-based (DPB) sensors calculate the flow rate by measuring the differential pressure across a fluidic restriction of known resistance. The dynamic range and sensitivity of DPB flow sensors can easily be adapted by changing the restriction placed in the channel. The higher the restriction is, the higher is the sensitivity. But the restriction should be selected carefully as a high flow resistance limits the fluidic performance of the system. DPB flow sensors offer a faster response time (,5 ms) than thermal flow sensors. The flow restriction is often realized by using a capillary or a diaphragm placed in the flow path

85

86

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.23 Principle of a pressure sensorbased flow sensor: (1) silicon diaphragm; (2) orifice cone angle; (3) piezo resistors; (4) conducting gasket. Reproduced under a Creative Commons Attributions 4.0 International License from Jenke, C., et al., 2017. The combination of micro diaphragm pumps and flow sensors for single stroke based liquid flow control. Sensors 17, 4, 755.

(Fig. 3.23). The ratio between orifice diameter and diaphragm thickness changes the influence of viscous friction and therefore viscosity and temperature on the differential pressure across the restriction. When the ratio is small, viscous friction dominates (HagenPoiseuille equation), while inertia effects dominate for large ratios (Torricelli’s law). For ratios close to 1, both effects must be considered (Hagenbach correction) (Richter et al., 1999). DPB sensors should typically be calibrated for certain liquid classes to achieve reproducible results.

Imaging-based sensors Various methods based on the imaging of moving or flexible elements in the chip have been demonstrated. For instance, microparticle image velocimetry tracks the displacement of particles within the medium using a digital camera. Since images are acquired at fixed intervals, the velocity of the particles can be calculated. Particle image velocimetry is a well-known method for characterizing flow properties in microfluidic chips (Nguyen and Wereley, 2002; Lindken et al., 2009), with the advantage of being contactless and yielding a velocity distribution. It is not, however, adapted for nonresearch applications, since particles often must be introduced into the medium to visualize the flow. Other imaging-based approaches make use of flexible structures such as pillars or posts to measure fluid flow (Mann et al., 2012).

Pressure sensors Pressure is an important parameter to monitor, both during cell loading into the microphysiological system and during cell perfusion. Cells can be easily damaged during cell loading due to pressure peaks, as reported by Wang et al. (2016). Pressure is also an important aspect of human physiology, in particular when mimicking blood circulation (Chen et al., 2017). For this reason, different pressure sensor solutions have been employed in OOCs. Commercial pressure sensors can be used, but they are not adapted to measure localized pressure and typically have a large dead volume. On-chip sensor solutions have been explored; most solutions are based on the deflection of a flexible membrane such as PDMS (Kartalov et al., 2007; Chung et al., 2009;

Engineering fluid control for organ-on-chips

FIGURE 3.24 Example of a micropressure sensor based on a flexible membrane. Reproduced with permission from Wang, L., et al., 2009. Polydimethylsiloxane-integratable micropressure sensor for microfluidic chips. Biomicrofluidics 3, 3, 034105.

Wang et al., 2009; Liu et al., 2013), which is measured either optically or electrically (Fig. 3.24). An interesting alternative has been proposed by Chen et al. (2017) that is based on the use of a sealed capillary. Pressure is directly related to the position of the liquid meniscus in the sealed capillary and can thus be measured with a video camera. Using this sensor, the authors were able to demonstrate that the pressure profile in their chip was identical to the systolic and diastolic pressure cycles observed in humans.

Other sensors Beyond pressure and flow rate, other parameters may require monitoring to ensure the proper operation of the OOC. For instance, the formation of bubbles within the microfluidic chip because of degassing or cellular activity may have a serious impact. Bubbles change the flow resistance of microfluidic channels and, thus, flow rates, but measuring the change in liquids passing through a microfluidic channel may also be important to monitor the process. Flow-front sensors measure changes of phase in front of the sensor as well as changes in the liquid properties. Flow-front sensors based on optical (Nguyen and Truong, 2005) or electromagnetic (Wrasse et al., 2017) approaches have been developed. One example is bubble sensors that are commercialized by various companies (Elveflow, 2019b).

Industrial implementation The engineering of the right fluid control system depends greatly on the organlevel function that a specific OOC needs to mimic. The complexity of the flow

87

88

CHAPTER 3 Organs-on-a-chip engineering

Table 3.8 Overview of organ-on-a-chip companies and their methods of flow control. Company

Flow control

References

InSphero

Bidirectional, gravity-based flow using a tilt table. Flow rate is calculated from known flow resistance in the microfluidic plate Bidirectional, gravity-based flow using a tilt table. Flow changes path depending on the direction Medium exchange by manual pipetting Membrane-based peristaltic pump

Kim et al. (2015) and InSphero (2019)

Mimetas

AlveoliX TissUse

Emulate

FiberCell Systems CN Bio Innovations Draper Laboratory

Depending on the application, pressure control, peristaltic, or syringe pumps Peristaltic pump Fully automated fluid control

Electrical actuated pumps

Trietsch et al. (2013, 2017) and Mimetas (2019) Stucki et al. (2015, 2018) and Alveolix (2019) Sonntag et al. (2010), Materne et al. (2015), Schimek et al. (2013), and TissUse (2019) Huh et al. (2013), Emulate (2019), and Kasendra et al. (2018) FibreCellSystems (2019) Edington et al. (2018) and Tsamandouras et al. (2017) Coppeta et al. (2017) and Xiao et al. (2017)

control system can be orders of magnitude apart between performing a simple medium exchange or a complex recirculation to simulate metabolism. Flow-control strategies used in commercial OOCs vary. Table 3.8 shows that OOC companies are using conservative approaches for flow control; practically all OOC companies use peripheral (off-chip) liquid actuation strategies. The intent is to achieve high reproducibility and reliability but also to decrease the cost of the disposable chip. Adoption of OOC devices by pharmaceutical companies, however, requires simplicity, easier standardization, and compatibility with standard laboratory automation systems. Only TissUse currently offers an on-chip membrane pump to recirculate the medium, and as their OOC platform aims to stimulate the interplay between urine and blood circuits, liquid actuation with very low dead volumes is beneficial. As a general rule, researchers interested in transferring their solution to the market should consider adopting the platform with the lowest degree of complexity for simulating a particular organ-level function.

Stimulation and sensing

Stimulation and sensing OOC platforms strive for model systems of nearly in vivo quality. To this end and additionally to the already described engineering requirements (the “Microengineering” section), stimulation tools are essential to induce the development of physiological cell functions for each individual organ mimic. Moreover, sensing tools allow measurements with high sensitivity and selectivity for accurate model system interpretation and thus are fundamental for experimental readout, validation, and standardization. Together, stimulation and sensing technologies are key components for the success of OOC and cell analysis platforms. This section discusses optical, mechanical, and electrical methods that enable quantitative analyses specifically for OOC platforms and introduces stimulation strategies used in cell analysis and OOC devices. Since the chemical stimulation of tissue in microfluidic OOCs commonly occurs via liquid perfusion and thus does not require specialized solutions, it is not part of the overview provided here.

Optical Optical assessments are the most commonly used analysis methods for OOC systems. They can take the form either of real-time culture monitoring or of endpoint assays. Whereas real-time culture monitoring is considered noninvasive and measurement is remote, endpoint assays are less time-critical but are often destructive. Optical probes consist of luminescent molecules—that is to saymolecules that emit either a fluorescent or a phosphorescent signal upon light excitation (Lakowicz, 1999). A plethora of luminescent molecules exists, allowing for—for instance—cell viability monitoring, protein expression analyses, and characterizing the impact of engineered devices on biological systems (Johnson and Spence, 2010; Varma et al., 2017). It should be noted that it is best to accurately evaluate molecular probes for optical analysis since they may, under certain conditions, interfere with the biological system studied (Kurth et al., 2018). Most optical assays employed in OOC devices are, in fact, molecular assays developed for conventional 2D cultures; they are already well documented and thus are not discussed here [for an overview, refer to Johnson and Spence (2010) and Varma et al. (2017)]. This section will focus on assays specifically required and partially developed for 3D cell model systems. As it is of the utmost importance for OOC platform control and validation to accurately monitor various parameters in the cell culture environment, there have been increasing efforts to integrate a variety of sensors into the microfluidic cell environment, including dissolved oxygen and CO2 (Grist et al., 2010; Modena et al., 2018). Although optical sensors are available for numerous other parameters and metabolites, such as pH, glucose, and lactate, the optical sensor designs are rather similar: either a hydrophilic fluorophore sensitive to a target

89

90

CHAPTER 3 Organs-on-a-chip engineering

molecule or metabolite is perfused within the culture medium or the sensitive fluorophore is encapsulated within a solid matrix of different formats. Since dissolved oxygen is one of the key parameters for OOC platform engineering, we will exemplarily focus on oxygen-sensitive optical measurements in the following. The features of optical oxygen sensors mainly depend on: 1. Sensing methods and material The molecular oxygen concentration reversibly quenches both the luminescence intensity and the excited-state lifetime of the luminophore so that optical oxygen sensors are sensitive to changes in its intensity as well as its decay time. Higher luminescent intensities correlate to higher oxygen concentrations, whereas higher luminophore lifetimes correlate to lower oxygen concentrations. Despite the fact that intensity-based sensing has been successfully applied for in vivo and cell culture methods, several publications have concluded that lifetime-based recordings are more robust and suitable for OOC monitoring compared to intensity measurements as fluorescent lifetimes are unaffected by bleaching or impaired photon yield due to proliferating tissue increasingly masking the probe (Grist et al., 2010; Oomen et al., 2016). Key parameters for the choice of the luminophore are the efficiency of the quenching process, stability against photobleaching, compatibility with inexpensive and available excitation sources, and the capability of being discriminated from the autofluorescence background. Examples of molecular fluorescent oxygen indicators include ruthenium-based and metallo-porphyrinbased molecules (Grist et al., 2010). 2. Encapsulation Any probe to be incorporated within a cell culture system must be carefully evaluated for its biocompatibility to exclude toxic effects. To prevent direct interaction between the luminophore and the biological material the fluorophore is thus often encapsulated and immobilized in a polymer, solgel, or silica matrix. Important considerations for the choice of the encapsulation matrix are the diffusion constant of oxygen and the solubility of the selected oxygen-sensitive luminophore within it (Oomen et al., 2016). If a sensor is to be reused, it must withstand adequate cleaning and sterilization routines. 3. Format Various formats of optical oxygen sensors exist for microfluidic cell culture devices (Fig. 3.25). The target-sensitive luminescent molecules can be encapsulated within patterned thin films attached to, for instance, the microfluidic channel or chamber bottom surface. The combination of multiple different sensor films next to each other allows for multiparametric readouts in the case that other luminophores with diverse target molecules are incorporated. A more miniaturized but still locally immobilized solution is the

Stimulation and sensing

(A)

(B)

Sensor film

Substrate

Substrate

Sensor film

(C)

(D)

Micro/nanoparticles

Optical fiber Opaque optical isolation Sensor film Aqueous solution (E)

Micro/nanoparticles

(F)

Sensor film

Substrate

Aqueous solution

FIGURE 3.25 The essential advantages of optical sensors for microfluidic bioreactors and OOC systems are the miniature sizes of the optical probes and their remote operation by fluorescence microscopy. To integrate these optical sensors into OOC platforms, molecule-sensitive luminophores can be encapsulated into thin films (A and B), onto optical fibers (C), and into beads of varying sizes (D and E). They are also available as hydrophilic probes dissolved in solution (F). OOC, Organ-on-a-chip. Reproduced with permission from Grist, S.M., Chrostowski, L., Cheung, K.C., 2010. Optical oxygen sensors for applications in microfluidic cell culture. Sensors (Switzerland) 10, 92869316.

packing of the sensor molecule into a matrix at the tip of an optical fiber, which serves as a direct light transmitter with minimal photon losses. Packed mobile probes consist of polymer or silica beads enriched with the luminophore of interest either inside the porous core structure or immobilized onto the outer surface of the beads. These beads can be introduced into the platform on demand in solution and their concentration can be adapted to specific needs. Polymer matrices enveloping the beads result in prolonged operational stability due to a decrease in both luminophore leaching and luminophore bleaching, the latter by interaction with environmental

91

92

CHAPTER 3 Organs-on-a-chip engineering

compounds. Nanoparticle-sized beads become cell permeable and therefore enable, for example, real-time intracellular oxygen monitoring in single cells or across tissues as well as in situ oxygenation studies. Submerging oxygensensitive beads in gel matrices also opens up the possibility of parameter monitoring in often-used gel matrices, such as Matrigel, gelatin methacryloyl, and collagen. Hydrophilic target molecule-sensitive luminophores are, on the contrary, soluble in water and have already served as probes for in vivo imaging. They can easily be dosed in varying concentrations and injected and recovered on demand, when perfusion systems are used. Compared to beads introduced into the device, luminophores do not interfere with bright-field and fluorescence microscopy apart from their emission wavelength. The main disadvantage of this approach is that, since not encapsulated, the probe releases cell-toxic singlet oxygen in its excited state to its environment. The use of such probes must consequently be characterized to prevent metabolic stress. Recent findings have, however, shown that the addition of 10% fetal bovine serum to the solution containing the hydrophilic probe prevents cells from experiencing the toxic side effect (Eyer et al., 2017). 4. System The sensor will finally be part of an optical measurement system consisting of at least a light source to excite the luminescent dye and a detector to detect the luminescent emission. Although detectors can be reduced to the minimum—for instance, a photodiode—most solutions for OOC devices employ imaging setups in order to retrieve the spatial resolution of the organ model. Conventionally used systems include standard fluorescence microscopes, but depending on the OOC device other setups providing sufficient sensitivity and spatial resolution may offer highly affordable alternatives. For example, simple bright field imaging can provide the readout of spherical microtissue sizes as a measure of growth and viability. It is worth mentioning that the use of any system already available in cell culture laboratories will certainly support the establishment of OOC platforms in daily laboratory routines.

Mechanical The principles of mechanical cell stimulation for miniaturized culture models were originally developed for single cells or 2D cultures in the early years of the 21st century (Kim et al., 2009) and have since been further optimized for OOC model systems. While dimensions and durability demanded adaptations, the principles of mechanical force transmission persisted, all with the ultimate goal of inducing mechanotransduction processes unique and relevant for the mimicked in vivo situation. Mechanotransduction processes—that is to say, the translation of an external physical force into an intracellular biochemical signal—are mediated by a plethora of distinct signal cascades (Clapham, 2003; Jaalouk and Lammerding, 2009). Those processes are distinct for unique cell differentiation

Stimulation and sensing

FIGURE 3.26 The principles of mechanical force transmission mimicking the in vivo situation comprise shear forces, strain/stretch, compression, and intracellular contractile forces, as well as gravity. Example models for the microfluidic realization of these scenarios are presented in the right column. The combination of different forces increases complexity but can be beneficial for tissue and multi-tissue applications. Adapted from Kurth, F., Eyer, K., Franco-Obrego´n, A., Dittrich, P.S., 2012. A new mechanobiological era: microfluidic pathways to apply and sense forces at the cellular level. Curr. Opin. Chem. Biol. 16, 400408 with permission from Elsevier.

states as well as terminally differentiated tissues and rely on particular force paradigms (Fig. 3.26): (1) fluid flow-induced shear stress, (2) tensile strain (stretch), (3) compression, (4) contraction, and (5) gravity. The biological relevance of these cases is rather complex, but at the same time, it is fundamentally important to choose the correct force paradigm for the individual 3D cell model system of interest. Good overviews on the mechanobiological principles can be found in Discher et al. (2005), Vogel and Sheetz (2006), Huang et al. (2004), Ingber (2003), Baker and Chen (2012), and Krishnan et al. (2011). Of all the enumerated forces, gravity is perhaps the most underestimated of all, since it is omnipresent and hardly changes. Due to its constant effect on Earth and the technological difficulty of inducing changes, it has however not yet been of interest for OOC devices. The electromechanical stimulation of tissue, for instance by pulsed electromagnetic fields (Parate et al., 2017), is gaining recognition in regenerative medicine and is listed in Table 3.9, together with all mechanically relevant

93

94

CHAPTER 3 Organs-on-a-chip engineering

Table 3.9 Mechanically relevant forces to induce and sense mechanotransduction processes realized in organ-on-a-chip devices and cell analysis platforms. The listed reviews discuss options for mechanical cell stimulation, including by means of magnetic bead incorporation (stretch, compression), acoustic waves, substrate microand nanopatterning, and hyper- and hypogravity. Mechanical force

Objective

Technological solution

Fluid shear force

Induce physiological shear/prevent harmful shear

Accurate flow control

Stretch/strain

Dynamic, cyclic stretching

Soft polymer membranes

Compression

Dynamic, cyclic mechanical loading

Reactor design (metal, hard plastic)

Constant or dynamic, cyclic mechanical loading Sense cell-induced contraction force Induce mechanical stimulation by electromagnetic field

Soft polymer membrane

Contraction Electrical

Cell substrates Reviews

Sense mechanotransduction response Varying substrate stiffness to influence cell development

Soft strain gauge sensors Electronic microactuators, pulse electromagnetic fields Transepithelial electrical resistance Substrates and matrices (e.g., gels)

Platforms for the mechanobiological study of organs-on-a-chip/cells

References (examples) Haase and Kamm (2017), Chen et al. (2017), Curto et al. (2017), Varma et al. (2018), Scheinpflug et al. (2018), Kim et al. (2015), Bavli et al. (2016), Li et al. (2015), van Duinen et al. (2017), Jain et al. (2018), Maschmeyer et al. (2015a), and Lohasz et al. (2019) Huh et al. (2010), Benam et al. (2016), Stucki et al. (2015, 2018), Douville et al. (2010), Villenave et al. (2017) Scheinpflug et al. (2018), Li et al. (2010), Hoffmann et al. (2015), Meinert et al. (2017), and Rousselle et al. (2018) Park et al., 2012

Lind et al. (2017), Feinberg et al. (2007), and Grosberg et al. (2011) Svennersten et al. (2011) and Parate et al. (2017)

Curto et al. (2017)

Verhulsel et al. (2014), Gilbert et al. (2010), and Maginet al. (2016) Haase and Kamm (2017), Ahadian et al. (2018), Guenat and Berthiaume (2018), Kim et al. (2009), Kurth et al. (2012), Ergir et al. (2018), and McLean et al. (2018)

Stimulation and sensing

forces. To identify potential applications of mechanotransduction in OOC devices, this section focuses on the engineering details enabling the force transmission of shear forces, stretch, and compression, as well as force readout of tissue contraction.

Stimulation Shear stress The simplest way of introducing mechanical forces on cells is by applying fluid flow-induced shear forces, which represent, for example, blood flow through veins or interstitial flow through the bone matrix. The physical phenomena commence relatively simply with a mere fluid flow through a straight tube in which cells grow attached to the inner surface but can reach high complexity in cases of tube branching or even upon the addition of blood cells into the fluid. These processes are well described in a review by Sebastian and Dittrich (2018) and have direct implications for device design and experimental conditions. Moreover, fluid flow-induced shear forces are very often hard to exclude since continuous nutrient supply and waste removal within cell culture models typically depend on medium perfusion. Shear forces are commonly given in dynes per square centimeter, where 1 dyn=cm2 5 1  1025 N 5 1  1021 Pa:

Physiologically relevant shear forces for endothelial cell culture models range from about 10 to .50 dyn/cm2 (Papaioannou and Stefanadis, 2005; Kim et al., 2007); for bone osteoblasts, effects were reported between 8 and 30 dyn/cm2 in a very simplified cell model (Jeon and Jeong, 2012). Other tissues may have to be protected from shear forces to impede deviating effects from in vivo-like development. As shear forces can impinge already at significantly lower ranges for highly mechanosensitive cells—for instance, progenitor muscle cells have been shown to react at forces ,0.1 dyn/cm2 (Kurth et al., 2015) and MadinDarby canine kidney cells at 0.3 dyn/cm2 (Curto et al., 2017)—fluid-flow ratios and chip designs have to be considered carefully. For design considerations, we refer the reader to recent reviews of multiorgan systems (Wang et al., 2018b; Rogal et al., 2017; Ronaldson-Bouchard and Vunjak-Novakovic, 2018). Fluid flow-induced shear forces can be approximated according to the wall-shear rate model: τ 56

QUμ ; wUh2

ðrectangular channelÞ

τ 54

QUμ ; πUR3

ðcircular tubeÞ

where Q is the volumetric fluid-flow rate, μ is the dynamic viscosity of the medium, w and h are the channel width and height, respectively, and R is the inner radius of the tube (Squires and Quake, 2005; Efstathopoulos et al., 2008). The dimensions of the channel height or the inner radius affect the fluid

95

96

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.27 Vessel shear stress values in various types of vasculature measured in vivo. Reproduced with permission from Papaioannou, T.G., Stefanadis, C., 2005. Vascular wall shear stress: basic principles and methods. Hell. J. Cardiol. 46, 915.

mechanics most and thus determine whether the device is prone to inducing strong or weak shear forces. They are accordingly the most important geometrical constraints for device engineering. Notably, the application of constant shear forces may result in different outcomes than the application of pulsatile shear forces (Chin et al., 2011), which occur, for example, in the case of peristaltic pump-driven medium perfusion. In vivo-derived values for vessel shear rates are available in a review by Papaioannou and Stefanadis (Papaioannou and Stefanadis, 2005); see also Fig. 3.27. Such patient-derived information can be translated to in vivo-mimicking on-chip models to reinforce their medical relevance. In experimental terms, in vitro-induced shear force values and their effect on conventional 2D microfluidic cell culture are summarized in a review by Kim et al. (2007); see also Table 3.10. Since a concise and complete summary of shear rate-induced effects in OOC devices is still lacking in the literature, the interested reader is referred to Table 3.9 at the beginning of the “Mechanical” section; it provides references to scientific works and commercially available 3D culture models, each of which provide relevant flow rates for their devices to induce/circumvent shear forces. An emerging technology for 3D cell models is the direct printing of cells in a gel matrix. These printing processes, however, induce quite significant shear forces that may lead to necrosis and cell disruption, respectively (Zhou, 2017; Vijayavenkataraman et al., 2018). Studies on the long-term effect induced by these shear forces may yield valuable information with which to optimize this versatile

Stimulation and sensing

Table 3.10 Experimentally tested fluid flowinduced shear forces in microfluidic cell cultures. Cell type Vascular endothelial cells Hepatocytes Mouse embryonic stem cells Human endothelial cells Bovine aortic endothelial cells HeLa cells Chinese hamster ovary cells Human osteoblast-like bone cells Skeletal muscle progenitor cells (C2C12) MadinDarby canine kidney cells

Shear stress ðdyn=cm2 Þ

Reference

10‒100 ,2 6.5 4‒25 20 4‒25 4‒25 4‒25 8‒30 0.09‒0.94 0.3

Li et al. (2005) Powers et al. (2002) Fok and Zandstra (2005) Ranjan et al. (1996) Varma et al. (2018) Ranjan et al. (1996) Ranjan et al. (1996) Ranjan et al. (1996) Jeon and Jeong (2012) Kurth et al. (2015) Curto et al. (2017)

Data partly derived from Kim, L., Toh, Y.-C., Voldman, J., Yu, H., 2007. A practical guide to microfluidic perfusion culture of adherent mammalian cells. Lab Chip 7, 681694.

technology. New technologies for 3D cell culture model systems may not be the only factors influencing the experimental outcome (un)expectedly. For instance, a careful adaptation of shear forces is recommended for drug-testing models as mechanobiological input has been shown to affect drug delivery yields (Bhise et al., 2014). The diverse effects of fluid flow-induced shear forces on the cell models under investigation depict the complex interaction of biological systems with engineered environments and underscore the necessity of thorough planning and testing.

Stretch: tensile strain The application of mechanical stretch in cellular in vitro models commonly depends on flexible membranes that are either unidirectionally stretched by elongation of a particular membrane area or undergo a more complex deformation due to the application of either liquid or gas (positive or negative) pressure to a circular membrane from one side (Fig. 3.28). The first cell-stretch platforms were developed in the 1990s with the aim to induce uniaxial stretch in simpler 2D cell models, especially muscle. Muscle development inherently depends on cyclic stretching, since the mechanotransduction processes driving myogenesis are orchestrated by stretch-sensitive cation channels that upon activation induce changes in calcium homeostasis, which ultimately regulate cell fate (proliferation or differentiation) (Bassel-Duby and Olson, 2006). Although other tissues do not depend on this mechanical stimulus as strongly, their physiological and pathological states can still be influenced by tensile strain (Huh et al., 2010). The first cellstretching platforms comprised a silicon chamber that was elongated by stepping motors [e.g., Naruse et al. (1998)—a stretch chamber by Strex, Osaka, Japan]. To adapt this principle to OOC devices, the dimensions of the chambers were scaled

97

98

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.28 Organ-on-a-chip models that include mechanical stretch. (A) A PDMS-based chip design for an intestine model undergoing peristalsis-like motions, manufactured by Emulate (emulatebio.com): The gut epithelium is attached to a porous, flexible PDMS membrane in the middle of a microfluidic channel serving as the medium supply. The cell channel is flanked by two vacuum channels that, upon activation, exert cyclic membrane stretch. The medium perfusion-induced shear forces support the recreation of the gut microenvironment. (B) An alveolus-on-a-chip model system for parallelized testing, by AlveoliX (www.alveolix.com). Mechanical stretch of the cell-laden membrane (BM) representing the alveolar barrier consisting of epithelial type I (AT I) and type II (AT II) cells at the air interface and ECs at the liquid interface. The multilayered chip device represents exactly this mechanism, whereby the deformation of the lower microdiaphragm transfers the vacuum force to the upper cell-laden membrane during breathing mode. Medium exchange is enabled by membrane valves at the side of the flow chamber. BM, Basal membrane; ECs, endothelial cells; PDMS, polydimethylsiloxane. Adapted from (A) Villenave, R., et al., 2017. Human gut-on-a-chip supports polarized infection of coxsackie B1 virus in vitro. PLoS One 12, 2, e0169412; (B) Stucki, J.D., et al., 2018. Medium throughput breathing human primary cell alveolus-on-chip model. Sci. Rep. 8, 1, 14359.

down to smaller sizes and incorporated into closed chambers and channels. The membranes are still typically fabricated from silicon (mainly PDMS) and then sandwiched in between adjacent channels or culture compartments (Huh et al., 2013), (Stucki et al., 2015), (Quiro´s-Solano et al., 2016). To allow for

Stimulation and sensing

transmembrane cell-to-cell communication and effective exchange of molecules, the membranes can be perforated during fabrication. The primary focus of these OOC models is lung mimics (Huh et al., 2010, 2013; Benam et al., 2016; Stucki et al., 2015; Douville et al., 2010), but other tissue models are integrated into identical chip systems to study, for example, the intestine (Kim et al., 2012; Villenave et al., 2017). Applied strain forces are given in either linear, surface, circumferential, or radial strain. Linear strain—the predominant variant—is reportedly applied up to 50%, whereas most of the devices induce about 10% (Huh et al., 2013; Stucki et al., 2015; Douville et al., 2010; Villenave et al., 2017). Resulting surface strains are typically slightly higher—that is, up to 60%—and circumferential and radial strains are commonly applied at below 20% (Guenat and Berthiaume, 2018). Despite these differences in induced strain forces, stretching frequencies are relatively similar at around 0.2 Hz for all organ models and seldom exceed frequencies above 1 Hz. Most recent developments combining silicon membranes with microelectromechanical systems technology have, however, opened the way for highly miniaturized membranes that enable the generation of strain rates as high as 870 Hz allowing studies of even extreme physiological conditions such as blunt force trauma (Poulin, 2018). Since silicon-based polymer membranes can absorb hydrophobic molecules not only by surface interactions but also by the diffusion of molecules into the porous polymer, precautions must be taken to not alter concentrations of administered drugs and stimulants (Berthier et al., 2012; Wang et al., 2012; Wong and Ho, 2009). Alternative materials are, however, rarely used as stretchable membranes. For PDMS chips, low-pressure (oxygen) plasma activation and subsequent assembly of the membranes to the adjacent layer(s) results in reliable bonding (Huh et al., 2013; Wagner et al., 2013). In some protocols, the membranes are “glued” in between the adjacent layers, typically during the prepolymer curing process. It is strongly advised to extend such curing processes beyond the time specified in standard protocols (a minimum of 36 hours) to ensure the complete depletion of precursor molecules, which are highly cell-toxic (Kellogg et al., 2014; Regehr et al., 2009). Glues must be thoroughly validated for their biocompatibility. The clamping of membranes is an alternative option but is often prone to fluid leakage and air bubble generation, respectively.

Compression Although all cells within the human body regularly undergo compressive forces due to body movement and gravity, the development and maintenance of only a few tissues inherently depend on constant dynamic and cyclic mechanical loading. This force paradigm is the key stimulation driving bone remodeling processes (Tru¨ssel et al., 2012; Scheinpflug et al., 2018) as well as chondrogenesis (Huang et al., 2010). Some OOC platforms exist to mimic this early phase of chondrogenesis, but many of them lack the input of this mechanical load. Models that include mechanical input to bone cells within artificial scaffolds and chondrocytes have

99

100

CHAPTER 3 Organs-on-a-chip engineering

nevertheless been presented, though not in the typical OOC community and in comparably large dimensions (Li et al., 2010; Hoffmann et al., 2015; Meinert et al., 2017). Regarding bone, these models do, however, all fail to completely represent the cellular diversity, and thus the complex functions in interplay within in vivo bone tissue. Since the engineering of such a complex tissue still requires technological advancement in integrated scaffold formation, the models currently closest to in vivo conditions actually use in vivo-derived bone for subsequent on-chip culture and analyses (Torisawa et al., 2014). Existing solely in vitro microfluidic chip models, on the contrary, focus on more simplified bone models, very often studying bone marrow stem-cell differentiation or osteocyte orchestration, for example, but not comprising the complex architecture of the calcified bone matrix. Those models that also include dynamic mechanical loading use multilayered soft polymer chips with incorporated flexible membranes that can be pressed onto the cells (Park et al., 2012) and thereby are similar to the chip devices engineered to induce mechanical strain. Applied mechanical load is given in either pressure—for example, for membrane-loaded systems about 1 psi [equivalent to 0.69 N/cm2 (Park et al., 2012)]—or in the description of the experimental setup, such as distance of compression (in micro- or millimeters) or loading (in percentage) (Li et al., 2010; Hoffmann et al., 2015; Meinert et al., 2017).

Sensing To keep OOC models as simple as possible while optimizing robustness and reproducibility, force sensors have so far only rarely been integrated into these devices. As numerous immunocytochemical end-point assays exist, cell response to mechanical forces is often observed optically by means of fluorescent labels. Further options are cell contraction imaging and time-lapse monitoring of changes in cell morphology. A good overview of the specialized integration of molecule toolkits for mechanosensing and mechanoregulation in biological systems is given by Jacobs and Blank (2014). A special case is the monitoring of cell contraction, which primarily focuses on cardiac tissue. Lind et al. presented a thin-sheet signal transducer that, upon cell contraction, bends strongly (Fig. 3.29A) (Lind et al., 2017; Feinberg et al., 2007). Upon substrate sensor deformation, the change in sensor resistance directly correlates to the level of tissue contraction. The formed laminar cardiac tissues were able to generate stress in the range of 115 kPa. Other options for the monitoring of cell contractile forces have been developed, particularly for single-cell studies, but have not yet been incorporated in OOC devices, either due to their complex fabrication protocols or to their high optical transparency requirements. Examples for substrate deformation monitoring include soft polymer pillars that bend under cell contraction (Fig. 3.29B) (Mann et al., 2012; Fu et al., 2010; Lam et al., 2012) as well as the optical monitoring of cellinduced fluorescent bead displacements, which are incorporated within a 3D hydrogel of well-defined elasticity (Legant et al., 2010). Just recently, electrical sensors

Stimulation and sensing

FIGURE 3.29 (A) Sensing cardiomyocyte contraction: a laminar cardiomyocyte tissue bends a gauge wire during contraction. The change in sensor resistance correlates to the cantilever deflection. (B) Soft polymer pillars can measure cell contractile forces at high sensitivity, which can be tuned by changing the geometrical ratios. (C) Transepithelial electrical resistance changes following mechanical cell actuation by fluid flow-induced shear forces, as shown by Curto et al. (A) Reprinted by permission from Lind, J.U., et al., 2017. Instrumented cardiac microphysiological devices via multimaterial three-dimensional printing. Nat. Mater. 16, 303308; (B) Reprinted by permission from Fu, J., et al., 2010. Mechanical regulation of cell function with geometrically modulated elastomeric substrates. Nat. Methods 7, 733736; and (C) Adapted from Curto, V.F., et al., 2017. Organic transistor platform with integrated microfluidics for in-line multi-parametric in vitro cell monitoring. Microsyst. Nanoeng. 3, 17028.

were shown to measure changes in transepithelial electrical resistance that correlated to changes in fluid flow-induced shear forces, thereby broadening the scope of sensing mechanotransduction processes on-chip (Fig. 3.29C) (Curto et al., 2017).

Electrical Electrochemical sensors Electrochemical sensors combine biological selectivity and electroanalytical sensitivity within one device. They detect analytes of interest on a (bio-)

101

102

CHAPTER 3 Organs-on-a-chip engineering

functionalized layer of the sensor with the precision and the selectivity of biological recognition (Grieshaber et al., 2008); this catalytic or molecular recognition is then translated into a measurable electrical signal proportional to the concentration of the analyte (The´venot et al., 2001; Ronkainen et al., 2010). Because of their high sensitivity and versatility in application, electrochemical sensors are among the most used and available devices in biomolecular electronics. Here, we highlight electrochemical sensors with varying molecular recognition principles. For a concise overview of the sensors’ modes of action, fabrication techniques, and use in biological studies, please refer to the reviews by Zhang and Hoshino (2014), Bieg et al. (2017), or Primiceri et al. (2013).

Electrochemical enzyme biosensors Enzymes provide high specificity for their target molecule, which is typically bound to the enzymatic pocket. Leveraging this lock-and-key recognition for electrical transducers enables simple, fast, and very precise measurements (Ispas et al., 2012). In addition to substrate detection, enzymatic biosensors can also measure enzymatic inhibition. These inhibition-based enzymatic biosensors monitor the decrease in enzymatic activity caused by the analyte of interest (Amine et al., 2016). An overview of the different enzymatic mechanisms for analyte detection in enzyme-based biosensors is given in Fig. 3.30. Advances in the fields of microfabrication and miniaturization have created numerous new applications for enzyme-based electrochemical sensors in the biomedical and food industries (Ispas et al., 2012; Weltin et al., 2016; Monteiro and Almeida, 2019; Rotariu et al., 2016). The variety of strategies for enzymatic sensor modification allows for versatile use and the integration of the devices with liquid samples (Mross et al., 2015). Enzymatic transducers are widely leveraged

FIGURE 3.30 Mechanisms of enzymatic analyte detection in biosensors. Reproduced under a Creative Commons Attributions 4.0 International License from Asal, M., O¨zen, O¨., ˙ 2018. Recent developments in enzyme, DNA and immuno-based biosensors. ˘ I., Sahinler, ¸ M., Polatoglu, Sensors 18, 6, 1924 (Asal et al., 2018).

Stimulation and sensing

to monitor glucose and lactate metabolism in single-cell systems (Wang, 2008; Moser et al., 2002; Ges et al., 2008; Ges and Baudenbacher, 2010; Weltin et al., 2014), but also for flow through-based devices, in which the biosensors measure the liquid arriving downstream from the cell culture compartments (Misun et al., 2016; Moser et al., 2002; Eklund et al., 2006; Talaei et al., 2015; Boero et al., 2012; Frey et al., 2010; Sassa et al., 2008; Perdomo et al., 2000; Satoh et al., 2008; Dempsey et al., 1997; Brischwein et al., 2003). Applications of enzyme-based electrochemical biosensors for advanced cell culturing and OOC systems have generated significant interest in investigating microtissue metabolism in situ. Weltin et al. measured drug-induced damage to liver microtissues based by lactate secretion with a biosensor insert for standard well culture plates (Fig. 3.31A) (Weltin et al., 2017). Bavli et al. (2016) were able to maintain liver function on-chip for over a month and simultaneously recorded glucose and lactate concentrations with an integrated biosensor. Such real-time measurement has made it possible to predict mitochondrial stress, which is correlated with a shift toward anaerobic glycolysis. Work by Misun et al. (2016) integrated glass plugs containing electrochemical enzyme biosensors with dedicated hanging-drop networks to monitor 3D microtissues (Fig. 3.31B); that study measured glucose and lactate secretion with glucose and lactate oxidase to detect the metabolic activity of colon cancer microtissues in real time. A different approach is to continuously sample the perfusing medium from the cell culture chamber into a cartridge equipped with electrochemical sensors (Fig. 3.31C). In this way the still somewhat limited operational stability of enzymatic electrochemical biosensors can be overcome, as the individual sensors can be easily exchanged and required calibrations be run between measurements. This strategy has been used to quantitatively assess the amount of alanine aminotransferase secreted by a 3D liver model, which directly correlates to liver viability (La Cour et al., 2016).

Electrochemical immunosensors Antibodies and aptamers represent another class of molecular recognition proteins. The specific interaction of antibody or aptamer with the antigen can be leveraged for highly sensitive electrochemical biosensors by immobilizing antibodies, fragments of antibodies, or aptamers on the sensor surface (Ugo and Moretto, 2017; Piro and Reisberg, 2017). The versatility of antibodies and aptamers in detecting trace amounts of analytes—ranging from cells, viruses, and bacteria to small molecules and proteins—has generated considerable interest in their potential as biomarkers (Rezaei et al., 2016; Riahi et al., 2016). Electrochemical immunosensors permit numerous types of electrochemical methodologies, which are typically selected based on the sensitivity required for the desired immunoassay. For a comprehensive overview of electrochemical immunosensors, see Wen et al. (2017). Detection of liver-secreted biomarkers by an integrated immunosensor-on-achip has been demonstrated by Riahi et al. (2016). By continuously monitoring the medium of the OOC system, secreted biomarkers—albumin and

103

104

CHAPTER 3 Organs-on-a-chip engineering

FIGURE 3.31 Applications of enzyme-based biosensors. (A) Microsensor setup that can be dipped into 96-well cell culture plates to measure 3D liver microtissues. The lactate biosensor operates with a working electrode functionalized with lactate oxidase and a blank electrode to reduce unspecific background. Alternatively, oxygen is measured using a diffusion-limiting membrane with the same electrode layout. (B) Lactate and glucose biosensors on glass plug-ins are integrated into a microfluidic hanging-drop network to measure the metabolic activity of cancer microtissues. (C) Placing the electrochemical biosensors downstream of the cell culture enables the addition or change of specific cartridges on demand. 3D, Three-dimensional. (A) Reprinted from Weltin, A., Hammer, S., Noor, F., Kaminski, Y., Kieninger, J., Urban, G.A., 2017. Accessing 3D microtissue metabolism: lactate and oxygen monitoring in hepatocyte spheroids. Biosens. Bioelectron. 87, 941948 with permission from Elsevier; (B) Reproduced under a Creative Commons Attributions 4.0 International License from Misun, P.M., Rothe, J., Schmid, Y.R.F., Hierlemann, A., Frey, O., 2016. Multi-analyte biosensor interface for real-time monitoring of 3D microtissue spheroids in hanging-drop networks. Microsyst. Nanoeng. 2, 16022; (C) Reprinted from La Cour, J.B., Generelli, S., Barbe, L., Guenat, O.T., 2016. Low-cost disposable ALT electrochemical microsensors for in-vitro hepatotoxic assessment. Sens. Actuators, B Chem., 228, 360365 with permission from Elsevier.

transferrin—could be quantified in real time over 5 days. Using electrodes coated with aptamers, Shin et al. (2016) measured the response of cardiac microtissues to drugs. The sensor unit was connected to a culturing chamber and monitored creatine kinase that is secreted upon cardiac injury. Furthermore, label-free in situ biomarker measurements using an electrochemical immunosensor within an OOC system containing liver, heart, and cancer cells have been

Stimulation and sensing

demonstrated by Zhang et al. (2017). The setup consists of various culture and sensor units connected by tubing. This modular approach allows for sensor washing and regeneration steps but renders device operation rather complex because of the required valves and pumps. Integrating electrochemical immunosensors into OOC platforms enables the on-line readout of biomarkers [e.g., cytokines (Chen et al., 2015; Liu et al., 2010) and proteins (Arroyo-Curra´s et al., 2017; Chikkaveeraiah et al., 2012)], providing a tool with which to continuously inspect advanced cell cultures for viability and function in the OOC system. This aspect becomes particularly relevant, if dynamic drug-induced responses of tissue models are investigated.

pH- and ion-sensitive sensors Electrochemical sensors equipped with electronically conductive oxides—for example, iridium, platinum, and ruthenium oxides—can sense the potentiometric difference between sensor and reference electrode to measure pH (Olthuis et al., 1990; Fog and Buck, 1984; Kurzweil, 2009; Anderson et al., 2016; Shaegh et al., 2016). These sensors can also be combined with other sensors, such as those monitoring oxygen and CO2 (Arquint et al., 1994). pH sensors can track the metabolism of cells through extracellular changes in acidity (Thedinga et al., 2007), which has been demonstrated by Ges et al. (2008) in monitoring cardiac metabolism by integrated pH sensors-on-a-chip. Ion-sensitive field-effect transistors (ISFETs) can also monitor pH and can be used to sense redox potential, lactate, and glucose, as well as specific ions at low concentrations in real time. The ion concentration in the solution of interest defines the current through the transistor. Like most ion sensors, ISFETs are potentiometers, meaning that the electrical current at the solid/liquid interface changes with varying ion concentrations. This electrical potential difference is measured according to the Nernst equation. Nano ISFETs in particular have emerged as a promising and versatile sensing approach (Pachauri and Ingebrandt, 2016), and numerous setups have demonstrated applications for pH sensing (Cui et al., 2001; Knopfmacher et al., 2010), chemical sensing (Sudho¨lter et al., 1989), and label-free biosensing (Gao et al., 2012; Stern et al., 2007). Owing to the advantages of direct, ultrasensitive, and label-free detection, ISFETs have also been successfully used as nucleic acid sensors (Duan et al., 2012) and to monitor the kinetics of receptor binding (Wipf et al., 2016) and the intracellular recording of action potentials (Duan et al., 2012). Relatively inexpensive, multiparametric, and downscalable, ISFETs are thus good candidates for integration into OOC devices.

Electrochemical oxygen sensors Oxygen sensing can be electrochemically achieved through the equilibrium potential on solid electrolyteelectrode cells and the Nernst equation. The measurements are logarithmically dependent on the partial pressure of oxygen (Ramamoorthy et al., 2003). Kieninger et al. developed a transparent sensor that they integrated onto the bottom of a conventional cell culture flask (Kieninger et al., 2014); this sensing cell

105

106

CHAPTER 3 Organs-on-a-chip engineering

culture flask system enabled pericellular oxygen sensing in real time. Other applications of oxygen biosensors measure cellular oxygen consumption rates in a multiparametric silicon chip, monitoring cellular microphysiological patterns on single chips for several days (Brischwein et al., 2003).

Electrical biosensors Electrical impedance spectroscopy Electrical impedance spectroscopy (EIS) provides a method to quantitatively analyze biological systems ranging from single cells to complex 3D microtissues. The advantage of EIS lies in its label-free and nondestructive mode of action, allowing for real-time measurements to retrieve characteristic information, such as cell counts, tissue size, and intracellular features (Foster and Schwan, 1989). Although the measurement of biological systems and even single cells has been employed for almost a century (Am and Cole, 1937; Fricke and Morse, 1925), microfluidic technology has enabled the integration of EIS into perfused microsystems, and this, in turn, permits new ways to investigate biological samples. Flow-through measurements of cells passing over a set of electrodes enable particle counting and distinguish between individual cell types (Sun and Morgan, 2010). These impedance cytometers identify cell size and volume at lower frequencies and characterize cell membrane conductivity and intracellular features at higher frequencies (Sun and Morgan, 2010; Schade-Kampmann et al., 2008). Label-free measuring of biological systems is leveraged to count or characterize, for example, bacteria (Bernabini et al., 2011; DeBlois and Wesley, 1977), viruses (DeBlois and Wesley, 1977), and pollen (Zhang et al., 2005). In addition, cellular systems ranging from yeast to cancer and stem cells have also been investigated with multifrequency EIS approaches (Holmes et al., 2009; Hassan and Bashir, 2014; Cheung et al., 2005; Holmes and Morgan, 2010; Song et al., 2013; Han et al., 2007; Chen et al., 2011b; Haandbæk et al., 2016). Furthermore, microfluidic systems with immobilized or trapped cells have been developed to measure impedimetric parameters of single cells (Malleo et al., 2010; Jang and Wang, 2007; Zhu et al., 2014, 2015). A subtype of EIS, electrical cell-substrate impedance sensing (ECIS), is used to detect the proliferation and viability of attached or spread cells (Wegener et al., 2000). The approach measures the growth and subsequent spreading of the cells on an electrode-patterned surface. The impedance spectra over the sensor array changes by cell movement or total occupied regions of the sensor. ECIS is a wellestablished technology for many applications, ranging from cancer spreading to cytotoxicity and environmental diagnostics (Harman et al., 2016; Peters et al., 2015; Curtis et al., 2009; Xiao and Luong, 2005; Xing et al., 2006). In addition to ECIS, multielectrode arrays can be used in a similar fashion (Obien et al., 2017) as well as for neurobiological applications (Malerba et al., 2018) but are outside the scope of this section.

Stimulation and sensing

Recent advances have made it possible to use EIS even for more complex cell systems such as microtissues. Label-free and nondestructive monitoring of advanced in vitro models with EIS offers a distinct advantage over other methods. Combined with integration into OOC systems, EIS offers drug-effect measurement in real time. Liver microtissues have been monitored during photodynamic treatment in a dedicated EIS device using two needle electrodes (Molckovsky and Wilson, 2001). Using EIS in hydrodynamic traps, Thielecke et al. investigated drug-induced changes in the morphology of spherical microtissues (Thielecke et al., 2001). Kloß et al. (2008) successfully identified apoptosis of multiple microtissues in microcavities in a higher throughput manner using a fourelectrode setup per well (Fig. 3.32A). True integration into OOC devices was achieved by developing an EIS sensor plug for hanging-drop networks (Schmid et al., 2016). This setup made it possible to monitor the growth of cancer

FIGURE 3.32 Applications of EIS for 3D cell culture or organ-on-a-chip devices. (A) Microcavity array for the simultaneous culture of nine human melanoma microtissues. A setup of four electrodes makes it possible to monitor drug-induced apoptosis on-chip using EIS. (B) Plug-in EIS sensors make it possible to track the beating of cardiac microtissues in a hanging-drop network. (C) Multiplexed tilting chip that enables the parallel culture and EIS readout of 15 microtissues. Drug-induced toxicity is measured by rolling the microtissue over the electrodes and assessing the size of the microtissue. 2D, Two-dimensional; 3D, three-dimensional; EIS, electrical impendence spectroscopy. Adapted with permission from Bu¨rgel, S.C., Diener, L., Frey, O., Kim, J.-Y., Hierlemann, A., 2016. Automated, multiplexed electrical impedance spectroscopy platform for continuous monitoring of microtissue spheroids. Anal. Chem. 88, 22, 1087610883. © 2016, American Chemical Society.

107

108

CHAPTER 3 Organs-on-a-chip engineering

microtissues and to measure the beating of cardiac microtissues hanging underneath the EIS sensor plug (Fig. 3.32B). Similar to EIS of single cells that pass over electrodes, a gravity-driven tilting platform enabled cancer microtissues to roll over a pair of electrodes in a study of drug-induced effects (Fig. 3.32C) (Bu¨rgel et al., 2016). By multiplexing the assay on a microscopy slide, 15 parallel experiments can be conducted simultaneously to test drug-induced toxicity in real time by means of EIS-measured microtissue size.

Transepithelial electrical resistance sensors The epithelial barrier function is a crucial aspect of multiple organ function and health. The gut, lung, kidney, liver, and blood-brain barrier all rely on epithelia layers to maintain homeostasis, ensure nutrient transport, and filter out harmful pathogens. These organs are widely studied in vitro, and there is a need for both mimicking and probing the biological and morphological properties of those barriers. Transepithelial electrical resistance is a key measure of the permeability of a cell layer to electrolytes present in the medium. Transepithelial electrical resistance is commonly employed in combination with filter membrane inserts by measuring an electric impedance spectrum over a cell barrier grown on top of the filter (Srinivasan et al., 2015). The transepithelial electrical resistance value of a cell barrier is acquired by applying alternating current at one or more frequencies across the cell layer. The associated impedance is measured in ohms and normalized to Ohm  cm2 by multiplying by the membrane area. Multiple apparatuses dedicated to transepithelial electrical resistance measurements in filter membrane systems are commercially available. These systems generally rely on either handheld electrodes inserted on both sides of the insert (EVOM ohmmeter, World Precision Instruments, Sarasota, FL, United States) or use dedicated multiwell holders with integrated electrodes and the ability to measure impedance at multiple frequency points (cellZscope, Mu¨nster, Germany; xCELLigence, ACEA Biosciences, San Diego, CA, United States). The downside of this approach is that the filter membrane has a strong influence on the measurement, which should be compensated. Furthermore, these systems are largely static, while the trend in cell culture is gradually shifting toward perfused systems. Lastly, filter membrane approaches typically have a relatively large surface, requiring large amounts of starting material or long culture times if they are to reach full confluency. In the past few years, impedance measurement systems integrated into microfluidic systems have been reported to bridge the gap between the utility of electrical sensing and the complexity of OOC models. The first approach involves the patterning of electrodes inside a microfluidic cell culture system. In 2017, Ingber et al. (Wyss Institute, Harvard University) reported a PDMS-based microfluidic system with integrated planar electrodes for impedance measurements (Henry et al., 2017). Other groups developed PDMS chips embedded with conductive biocompatible wires that carry the current close to the cell compartments (Griep et al., 2013; Odijk et al., 2015; van der Helm et al., 2016, 2019). These platforms have been used in combination with impedance spectroscopy to

Stimulation and sensing

characterize perfused, membrane-supported models of tissues such as the gut, lung, and blood-brain barrier. Most other reported solutions integrate either inserts linked to perfused microfluidic channels (Alexander et al., 2018b; Maschmeyer et al., 2015b; Zeller et al., 2017) or microfluidic devices in which cell layers are cultured on built-in porous membranes (Shah et al., 2016; Wang et al., 2017) and make use of a single-frequency voltohmmeter for transepithelial electrical resistance measurements. Typical drawbacks of OOC and other microfluidic cell culture systems include the level of standardization, throughput, and compatibility with standard laboratory equipment. To overcome these limitations, Mimetas developed the OrganoPlate (Fig. 3.33), an OOC platform composed of up to 96 microfluidic chips integrated into a 384-well plate. Each chip can be perfused through leveling between reservoirs, which is made continuous by changing the inclination angle at given time intervals using a modified rocker. The platform has been demonstrated for toxicological and drug-transport studies on barrier tissues such as the gut (Trietsch et al., 2017), kidney (Vormann et al., 2018; Vriend et al., 2018), blood-brain barrier, and vasculature (van Duinen et al., 2017, 2018; Wevers et al., 2018), as well as for studying cancer (Moisan et al., 2018; Lanz et al., 2017). Thus far, readouts on the OrganoPlate have largely been optical, including a range of fluorescent and luminescent assays. Recently, Mimetas unveiled an

FIGURE 3.33 Apparatus for on-chip transepithelial electrical resistance measurement in the OrganoPlate (Mimetas). System impedance is characterized by the channel resistance in series with the extracellular matrix-supported cell layer. The electrical connection is made by dipping conductive electrodes into the entry wells of each chip (A1, A3, B1, and B3 in the schematic). Courtesy of Mimetas.

109

110

CHAPTER 3 Organs-on-a-chip engineering

automated transepithelial electrical resistance measurement device dedicated to the OrganoPlate platform to allow for electrical measurements. This commercially available apparatus is connected to the microfluidic chips by inserting a lid adapter with an integrated electrode array into the access wells of the OrganoPlate. The OrganoTEER performs impedance readouts at multiple frequencies to record the electrical parameters of a tubule for up to 120 tissues in parallel. Further development of the transepithelial electrical resistance measurement system for OOC technologies is crucial for the industrial adoption of OOC as a preclinical development tool. Challenges such as scalability, throughput, and the cross-platform normalization of impedance values will need to be resolved before widespread adoption of these systems is possible.

Commercial sensors Outside university laboratories, several small- and medium-sized enterprises are developing—often in close collaboration with academic groups—integrated commercial biosensors for cell culture systems (Table 3.11). Commercially available biosensors are used for clinical, food, environment, and biothreat applications [for a review of the application of such commercial biosensors, see Bahadir and Sezgintu¨rk (2015)]; few sensors, however, are or can be integrated into lab-on-achip applications. Since the sensor part of these devices is often the central focus, academic laboratories employ rapid prototyping (e.g., PDMS) to engineer the fluid part around the sensor. Commercial entities, in contrast, can develop sensors that are integrated into more robust devices, for example, by injection molding. These methods require a high initial investment for master fabrication but later enable the mass production of sensor chips with a very high reproducibility. In addition, ad- and absorption effects at the media-chip interface, as seen for PDMS (Wong and Ho, 2009), can be reduced significantly. Table 3.11 Examples of commercially available organ-on-a-chip sensors. Manufacturer

Analyte

Sensor type

References

C-CIT Sensors AG

Lactate and glucose

Electrochemistry

MicruX Technologies

Antigens via immobilized antibodies Endothelial proliferation, barrier function, and motility Oxygen

Impedimetric immunosensor

Spichiger and Spichiger-Keller (2011) Ravalli et al. (2016)

Cell-substrate impedance sensing

Szulcek et al. (2014)

Quenching of luminescence

Santoro et al. (2012)

Ibidi GmbH

PreSens Precision Sensing GmbH

(Continued)

References

Table 3.11 Examples of commercially available organ-on-a-chip sensors. Continued Manufacturer

Analyte

Sensor type

References

Cellasys GmbH

Metabolic and morphological components

Alexander et al. (2018a)

Mimetas

Barrier function

SiMPLINext SA Colibri Photonics GmbH Surflay Nanotec GmbH

Barrier function

Impedance, extracellular acidification (pH), oxygen consumption, and temperature Transepithelial electrical resistance Transepithelial electrical resistance Fluorescence lifetime and/or fluorescence intensity

TissUse GmbH Micronit Micro Technologies

3Brain Axion BioSystems

Oxygen

pH, temperature, oxygen, molecular components Cell viability

Fluorescence lifetime and/or fluorescence intensity

Fluorescent optical fiber

Peptide, a protein, and an antibody detection assay Neuronal activity Electrophysiology

Complementary metal oxide semiconductor-based single-photon counting optical sensor CMOS electrophysiology Microelectrode array

www.mimetas.com www.simplinext. com Sonntag et al. (2016) and Bavli et al. (2016) Massing et al. (2016) and Olszyna et al. (2019) Sergachev et al. (2013) Van Dorst et al. (2016)

Lonardoni (2017) www. axionbiosystems. com

References Agarwal, A., Goss, J.A., Cho, A., McCain, M.L., Parker, K.K., 2013. Microfluidic heart on a chip for higher throughput pharmacological studies. Lab Chip 13 (18), 35993608. Ahadian, S., et al., 2018. Organ-on-a-chip platforms: a convergence of advanced materials, cells, and microscale technologies. Adv. Healthc. Mater. 7 (2), 153. Alexander, F., Eggert, S., Wiest, J., 2018a. A novel lab-on-a-chip platform for spheroid metabolism monitoring. Cytotechnology 70 (1), 375386. Alexander, F.A., Eggert, S., Wiest, J., Wiest, J., 2018b. Skin-on-a-chip: transepithelial electrical resistance and extracellular acidification measurements through an automated airliquid interface. Genes (Basel) 9 (2). Available from: https://doi.org/10.3390/ genes9020114. Alford, P.W., Nesmith, A.P., Seywerd, J.N., Grosberg, A., Parker, K.K., 2011. Vascular smooth muscle contractility depends on cell shape. Integr. Biol. 3 (11), 10631070.

111

112

CHAPTER 3 Organs-on-a-chip engineering

Alveolix, 2019. “Technology”. [Online]. Available from: ,http://www.alveolix.com/ technology/. (accessed 03.02.19.) Am, H.J.C., Cole, K.S., 1937. Transverse electric impedance of Nitella. J. Gen. Physiol. 21, 189201. Amine, A., Arduini, F., Moscone, D., Palleschi, G., 2016. Recent advances in biosensors based on enzyme inhibition. Biosens. Bioelectron. 76, 180194. Anderson, J.R., et al., 2000. Fabrication of topologically complex three-dimensional microfluidic systems in PDMS by rapid prototyping. Anal. Chem. 72 (14), 31583164. Anderson, M., Moshnikova, A., Engelman, D.M., Reshetnyak, Y.K., Andreev, O.A., 2016. Probe for the measurement of cell surface pH in vivo and ex vivo. Proc. Natl. Acad. Sci. U.S.A. 113 (29), 81778181. Arquint, P., Koudelka-Hep, M., van der Schoot, B.H., van der Wal, P., de Rooij, N.F., 1994. Micromachined analyzers on a silicon chip. Clin. Chem. 40 (9), 18051809. Arroyo-Curra´s, N., Somerson, J., Vieira, P.A., Ploense, K.L., Kippin, T.E., Plaxco, K.W., 2017. Real-time measurement of small molecules directly in awake, ambulatory animals. Proc. Natl. Acad. Sci. U.S.A. 114 (4), 645650. ¨ zen, O ¨ ., Sahinler, Asal, M., O ¸ M., Polato˘glu, ˙I., 2018. Recent developments in enzyme, DNA and immuno-based biosensors. Sensors 18 (6), 1924. Ashauer, M., Glosch, H., Hedrich, F., Hey, N., Sandmaier, H., Lang, W., 1999. Thermal flow sensor for liquids and gases based on combinations of two principles. Sens. Actuators, A: Phys. 73 (12), 713. Attia, U.M., Marson, S., Alcock, J.R., 2009. Micro-injection moulding of polymer microfluidic devices. Microfluid. Nanofluid. 7, 128. Au, A.K., Lai, H., Utela, B.R., Folch, A., 2011. Microvalves and micropumps for BioMEMS. Micromachines 2, 179220. Avci, H., Gu¨zel, F.D., Erol, S., Akpek, A., 2018. Recent advances in organ-on-a-chip technologies and future challenges: a review. Turkish J. Chem. 42 (3), 587610. Ayyapan, M., Uttamchand, N.K., Rajan, R.A.A., 2016. Mechanical and wear properties of copper-lead alloy prepared by powder metallurgy processing technique. J. Chem. Technol. Metall. 51 (6), 726734. Bahadir, E.B., Sezgintu¨rk, M.K., 2015. Applications of commercial biosensors in clinical, food, environmental, and biothreat/biowarfare analyses. Anal. Biochem. 478, 107120. Baker, B.M., Chen, C.S., 2012. Deconstructing the third dimension  how 3D culture microenvironments alter cellular cues. J. Cell Sci. 125, 30153024. Bassel-Duby, R., Olson, E.N., 2006. Signaling pathways in skeletal muscle remodeling. Annu. Rev. Biochem. 75 (1), 1937. Bavli, D., et al., 2016. Real-time monitoring of metabolic function in liver-on-chip microdevices tracks the dynamics of mitochondrial dysfunction. Proc. Natl. Acad. Sci. U.S. A. 113 (16), E2231E2240. Becker, H., Heim, U., 2000. Hot embossing as a method for the fabrication of polymer high aspect ratio structures. Sens. Actuators 83, 130135. Becker, H., Locascio, L.E., 2002. Polymer microfluidic devices. Talanta 56 (2), 267287. Benam, K.H., et al., 2015. Engineered in vitro disease models. Annu. Rev. Pathol. Mech. Dis. 10 (1), 195262. Benam, K.H., et al., 2016. Small airway-on-a-chip enables analysis of human lung inflammation and drug responses in vitro. Nat. Methods 13, 151157.

References

Bernabini, C., Holmes, D., Morgan, H., 2011. Micro-impedance cytometry for detection and analysis of micron-sized particles and bacteria. Lab Chip 11 (3), 407412. Bertassoni, L.E., et al., 2014. Hydrogel bioprinted microchannel networks for vascularization of tissue engineering constructs. Lab Chip 14 (13), 22022211. Berthier, E., Young, E.W.K., Beebe, D., 2012. Engineers are from PDMS-land, biologists are from polystyrenia. Lab Chip 12 (7), 12241237. Bhatia, S.N., Ingber, D.E., 2014. Microfluidic organs-on-chips. Nat. Biotechnol. 32 (8), 760772. Bhise, N.S., Ribas, J., Manoharan, V., Shrike, Y., Polini, A., 2014. Organ-on-a-chip platforms for studying drug delivery systems. J. Control. Release 190, 8293. Bieg, C., Fuchsberger, K., Stelzle, M., 2017. Introduction to polymer-based solid-contact ion-selective electrodes—basic concepts, practical considerations, and current research topics. Anal. Bioanal. Chem. 409 (1), 4561. Boero, C., Olivo, J., De Micheli, G., Carrara, S., 2012. New approaches for carbon nanotubes-based biosensors and their application to cell culture monitoring. IEEE Trans. Biomed. Circuits Syst. 6 (5), 479485. Bondot, T., Petit, N., Ritty, S., Kieffer, C.-A., 2012. Flow-rate calibration and control in a microfluidic device. In: Application EP-2720103-A1 Borenstein, J.T., Terai, H., King, K.R., Weinberg, E.J., Kaazempur-Mofrad, M.R., Vacanti, J.P., 2002. Microfabrication technology for vascularized tissue engineering. Biomed. Microdevices 4 (3), 167175. Boyd-Moss, M., Baratchi, S., Di Venere, M., Khoshmanesh, K., 2016. Self-contained microfluidic systems: a review. Lab Chip 16 (17), 31773192. Brennan, M.D., Rexius-Hall, M.L., Elgass, L.J., Eddington, D.T., 2014. Oxygen control with microfluidics. Lab Chip 14, 43054318. Brischwein, M., Motrescu, E.R., Cabala, E., Otto, A.M., Grothe, H., Wolf, B., 2003. Functional cellular assays with multiparametric silicon sensor chips. Lab Chip 3 (4), 234240. Bronkhorst, 2019. Mini CORI-FLOW datasheet M12M14. [Online]. Available from: ,https://www.bronkhorst.com/products/liquid-flow/mini-cori-flow/m14/. (accessed 30.01.19.) Bu¨rgel, S.C., Diener, L., Frey, O., Kim, J.-Y., Hierlemann, A., 2016. Automated, multiplexed electrical impedance spectroscopy platform for continuous monitoring of microtissue spheroids. Anal. Chem. 88 (22), 1087610883. Byun, C.K., Abi-Samra, K., Cho, Y.K., Takayama, S., 2014. Pumps for microfluidic cell culture. Electrophoresis 35, 245257. CSEM, 2010. “Disposable Pump” for Microfluidic Systems CSEM, 2010. TURBISC Pump—Miniaturized Pulsation-Free, Bi-directional Pump for Liquid Handling Caplin, J.D., Granados, N.G., James, M.R., Montazami, R., Hashemi, N., 2015. Microfluidic organ-on-a-chip technology for advancement of drug development and toxicology. Adv. Healthc. Mater. 4 (10), 14261450. Capulli, A.K., et al., 2014. Approaching the in vitro clinical trial: engineering organs on chips. Lab Chip 14 (17), 31813186. Carpi, F., Smela, E., 2009. Biomedical Applications of Electroactive Polymer Actuators. John Wiley & Sons.

113

114

CHAPTER 3 Organs-on-a-chip engineering

Cetoni, 2019. Syringe stirrer neMIX. [Online]. Available from: ,https://www.cetoni.com/ products/syringe-stirrer-nemix/. (accessed 03.02.19.) Chauvet, A., Tibiletti, T., Caffarri, S., Chergui, M., 2014. A microfluidic flow-cell for the study of the ultrafast dynamics of biological systems. Rev. Sci. Instrum. 85 (10), 103118. Chen, Y., Zhang, L., Chen, G., 2008a. Fabrication, modification, and application of poly (methyl methacrylate) microfluidic chips. Electrophoresis 29, 18011814. Chen, J.M., Huang, P.-C., Lin, M.-G., 2008b. Analysis and experiment of capillary valves for microfluidics on a rotating disk. Microfluid. Nanofluid. 4 (5), 427437. Chen, C.F., Liu, J., Chang, C.C., DeVoe, D.L., 2009. High-pressure on-chip mechanical valves for thermoplastic microfluidic devices. Lab Chip 9 (24), 35113516. Chen, S.Y.C., Hung, P.J., Lee, P.J., 2011a. Microfluidic array for three-dimensional perfusion culture of human mammary epithelial cells. Biomed. Microdevices 13 (4), 753758. Chen, J., et al., 2011b. Classification of cell types using a microfluidic device for mechanical and electrical measurement on single cells. Lab Chip 11 (18), 31743181. Chen, P., Huang, N.-T., Chung, M.-T., Cornell, T.T., Kurabayashi, K., 2015. Label-free cytokine micro- and nano-biosensing towards personalized medicine of systemic inflammatory disorders. Adv. Drug Deliv. Rev. 95, 90103. Chen, Y., et al., 2017. A microfluidic circulatory system integrated with capillary-assisted pressure sensors. Lab Chip 17 (4), 653662. Cheung, K., Gawad, S., Renaud, P., 2005. Impedance spectroscopy flow cytometry: onchip label-free cell differentiation. Cytometry A 65 (2), 124132. Chikkaveeraiah, B.V., Bhirde, A.A., Morgan, N.Y., Eden, H.S., Chen, X., 2012. Electrochemical immunosensors for detection of cancer protein biomarkers. ACS Nano 6 (8), 65466561. Chin, L.K., Yu, J.Q., Fu, Y., Yu, T., Liu, A.Q., Luo, K.Q., 2011. Production of reactive oxygen species in endothelial cells under different pulsatile shear stresses and glucose concentrations. Lab Chip 11 (11), 18561863. Cho, H., Kim, H.-Y., Kang, J.Y., Kim, T.S., 2007. How the capillary burst microvalve works. J. Colloid Interface Sci. 306 (2), 379385. Chung, K., Lee, H., Lu, H., 2009. Multiplex pressure measurement in microsystems using volume displacement of particle suspensions. Lab Chip 9 (23), 3345. Clapham, D.E., 2003. TRP channels as cellular sensors. Nature 426, 517524. Coppeta, J.R., et al., 2017. A portable and reconfigurable multi-organ platform for drug development with onboard microfluidic flow control. Lab Chip 17 (1), 134144. Cui, Y., Wei, Q., Park, H., Lieber, C.M., 2001. Nanowire nanosensors for highly sensitive and selective detection of biological and chemical species. Science 293 (5533), 12891292 (80-.). Curtis, T.M., et al., 2009. A portable cell-based impedance sensor for toxicity testing of drinking water. Lab Chip 9 (15), 21762183. Curto, V.F., et al., 2017. Organic transistor platform with integrated microfluidics for inline multi-parametric in vitro cell monitoring. Microsyst. Nanoeng. 3, 17028. Darby, S., et al., 2010. A metering rotary nanopump for microfluidic systems. Lab Chip 10 (23), 32183226. De Jong, J., Lammertink, R.G.H., Wessling, M., 2006. Membranes and microfluidics: a review. Lab Chip 6 (9), 11251139.

References

DeBlois, R.W., Wesley, R.K., 1977. Sizes and concentrations of several type C oncornaviruses and bacteriophage T2 by the resistive-pulse technique. J. Virol. 23 (2), p. 227 LP-233. Dempsey, E., et al., 1997. Design and development of a miniaturised total chemical analysis system for on-line lactate and glucose monitoring in biological samples. Anal. Chim. Acta 346 (3), 341349. Discher, D.E., Janmey, P., Wang, Y.L., 2005. Tissue cells feel and respond to the stiffness of their substrate. Science 310 (5751), 11391143 (80-.). Douville, N.J., et al., 2010. Combination of fluid and solid mechanical stresses contribute to cell death and detachment in a microfluidic alveolar model. Lab Chip 11, 609619. Duan, X., Li, Y., Rajan, N.K., Routenberg, D.A., Modis, Y., Reed, M.A., 2012. Quantification of the affinities and kinetics of protein interactions using silicon nanowire biosensors. Nat. Nanotechnol. 7 (6), 401407. Duan, X., et al., 2012. Intracellular recordings of action potentials by an extracellular nanoscale field-effect transistor. Nat. Nanotechnol. 7 (3), 174179. Dubeau-Larame´e, G., Rivie`re, C., Jean, I., Mermut, O., Cohen, L.Y., 2014. Microflow1, a sheathless fiber-optic flow cytometry biomedical platform: demonstration onboard the international space station. Cytometry A 85 (4), 322331. van Duinen, V., Zhu, D., Ramakers, C., van Zonneveld, A.J., Vulto, P., Hankemeier, T., 2018. Perfused 3D angiogenic sprouting in a high-throughput in vitro platform. Angiogenesis 19. Edington, C.D., et al., 2018. Interconnected microphysiological systems for quantitative biology and pharmacology studies. Sci. Rep. 8, 4530. Efstathopoulos, E.P., Patatoukas, G., Pantos, I., Benekos, O., Katritsis, D., Kelekis, N.L., 2008. Wall shear stress calculation in ascending aorta using phase contrast magnetic resonance imaging. Investigating effective ways to calculate it in clinical practice. Phys. Med. 24, 175181. Eklund, S.E., Snider, R.M., Wikswo, J., Baudenbacher, F., Prokop, A., Cliffel, D.E., 2006. Multianalyte microphysiometry as a tool in metabolomics and systems biology. J. Electroanal. Chem. 587 (2), 333339. ElveFlow, 2019a. Bubble Trap. [Online]. Available from: ,https://www.elveflow.com/ microfluidic-tutorials/microfluidic-reviews-and-tutorials/microfluidic-fittings-and-tubing-resources/why-and-how-to-use-a-microfluidic-bubble-trap/. (accessed 03.02.19.) Elveflow, 2019b. MLS—microfluidic bubble detector. [Online]. Available from: ,https:// www.elveflow.com/microfluidic-flow-control-products/microfluidic-flow-control-module/ microfluidic-liquid-sensor/. (accessed 03.02.19.) Emulate, 2019. Technology. [Online]. Available from: ,https://emulatebio.com/organs-onchips/. (accessed 03.02.19.) Ergir, E., Bachmann, B., Redl, H., Forte, G., Ertl, P., 2018. Small force, big impact: next generation organ-on-a-chip systems incorporating biomechanical cues. Front. Physiol. 9, 18. no. October. Esch, M.B., et al., 2016. HHS Public Access 15 (10), 22692277. Eyer, K., et al., 2017. Single-cell deep phenotyping of IgG-secreting cells for highresolution immune monitoring. Nat. Biotechnol. 35, 977982. Faley, S., et al., 2008. Microfluidic platform for real-time signaling analysis of multiple single T cells in parallel. Lab Chip 8 (10), 17001712. Faustino, V., Catarino, S.O., Lima, R., Minas, G., 2016. Biomedical microfluidic devices by using low-cost fabrication techniques: a review. J. Biomech. 49 (11), 22802292.

115

116

CHAPTER 3 Organs-on-a-chip engineering

Feaver, R.E., Gelfand, B.D., Blackman, B.R., 2013. Human haemodynamic frequency harmonics regulate the inflammatory phenotype of vascular endothelial cells. Nat. Commun. 4, 15111525. Feinberg, A.W., Feigel, A., Shevkoplyas, S.S., Sheehy, S., Whitesides, G.M., Parker, K.K., 2007. Muscular thin films for building actuators and powering devices. Science 317 (5843), 13661370. FibreCellSystems, 2019. Pumps. [Online]. Available from: ,https://www.fibercellsystems. com/products/pumps/. (accessed 03.02.19.) Fluigent, 2019a. Pressure-based flow control LINEUP series. [Online]. Available from: ,https://www.fluigent.com/product/microfluidic-components/lineup-series/. (accessed 30.01.19.) Fluigent, 2019b. Bubble trap. [Online]. Available from: ,https://www.fluigent.com/product/microfluidic-components/bubble-trap/. (accessed 03.02.19.) Fog, A., Buck, R.P., 1984. Electronic semiconducting oxides as pH sensors. Sens. Actuators 5 (2), 137146. Fok, E.Y.L., Zandstra, P.W., 2005. Shear-controlled single-step mouse embryonic stem cell expansion and embryoid bodybased differentiation. Stem Cells 23 (9), 13331342. Formulatrix, 2019a. Valve control. [Online]. Available from: ,http://help.formulatrix.com/ mantis/4.0/Default.htm#Designing_Dispenses/Concepts/Valve_Control.htm. (accessed 03.02.19.) Formulatrix, 2019b. Microfluidic pumps. [Online]. Available from: ,https://formulatrix. com/microfluidic-pumps/. (accessed 03.02.19.) Foster, K.R., Schwan, H.P., 1989. Dielectric properties of tissues and biological materials: a critical review. Crit. Rev. Biomed. Eng. 17 (1), 25104. Frey, O., Talaei, S., van der Wal, P.D., Koudelka-Hep, M., de Rooij, N.F., 2010. Continuous-flow multi-analyte biosensor cartridge with controllable linear response range. Lab Chip 10 (17), 22262234. Fricke, H., Morse, S., 1925. The electric resistance and capacity of blood for frequencies between 800 and 4(1/2) million cycles. J. Gen. Physiol. 9 (2), 153167. Fu, J., et al., 2010. Mechanical regulation of cell function with geometrically modulated elastomeric substrates. Nat. Methods 7, 733736. Gao, A., et al., 2012. Enhanced sensing of nucleic acids with silicon nanowire field effect transistor biosensors. Nano Lett. 12 (10), 52625268. Ges, I.A., Baudenbacher, F., 2010. Enzyme-coated microelectrodes to monitor lactate production in a nanoliter microfluidic cell culture device. Biosens. Bioelectron. 26 (2), 828833. Ges, I.A., Dzhura, I.A., Baudenbacher, F.J., 2008. On-chip acidification rate measurements from single cardiac cells confined in sub-nanoliter volumes. Biomed. Microdevices 10 (3), 347354. Gilbert, P.M., et al., 2010. Substrate elasticity regulates skeletal muscle stem cell selfrenewal in culture. Science 329, 10781081 (80-.). Griep, L.M., et al., 2013. BBB on chip: microfluidic platform to mechanically and biochemically modulate blood-brain barrier function. Biomed. Microdevices 15 (1), 145150. Grieshaber, D., Mackenzie, R., Vo¨ro¨s, J., Reimhult, E., 2008. Electrochemical biosensors— sensor principles and architectures. Sensors 8 (3), 14001458.

References

Grist, S.M., Chrostowski, L., Cheung, K.C., 2010. Optical oxygen sensors for applications in microfluidic cell culture. Sensors (Switzerland) 10, 92869316. Grosberg, A., Alford, P.W., McCain, M.L., Parker, K.K., 2011. Ensembles of engineered cardiac tissues for physiological and pharmacological study: heart on a chip. Lab Chip 11, 41654173. Grover, W.H., Skelley, A.M., Liu, C.N., Lagally, E.T., Mathies, R.A., 2003. Monolithic membrane valves and diaphragm pumps for practical large-scale integration into glass microfluidic devices. Sens. Actuators, B: Chem. 89, 315323. Guckenberger, D.J., De Groot, T.E., Wan, A.M.D., Beebe, D.J., Young, E.W.K., 2015. Micromilling: a method for ultra-rapid prototyping of plastic microfluidic devices. Lab Chip 15 (11), 23642378. Guenat, O.T., Berthiaume, F., 2018. Incorporating mechanical strain in organs-on-a-chip: lung and skin. Biomicrofluidics 12, 042207. Haandbæk, N., Bu¨rgel, S.C., Rudolf, F., Heer, F., Hierlemann, A., 2016. Characterization of single yeast cell phenotypes using microfluidic impedance cytometry and optical imaging. ACS Sensors 1 (8), 10201027. Haase, K., Kamm, R.D., 2017. Advances in on-chip vascularization. Regen. Med. 12 (3), 285302. Hagmeyer, B., Schu¨tte, J., Bo¨ttger, J., Gebhardt, R., Stelzle, M., 2013. Tailoring microfluidic systems for organ-like cell culture applications using multiphysics simulations. SPIE MOEMS-MEMS 8615, 861509. Hamad, K., Kaseem, M., Ko, Y.G., Deri, F., 2014. Biodegradable polymer blends and composites: an overview. Polym. Sci. Ser. A 56 (6), 812829. Hamilton, 2019. Automated liquid handling. [Online]. Available from: ,https://www. hamiltoncompany.com/automated-liquid-handling. (accessed 03.02.19.) Han, A., Yang, L., Frazier, A.B., 2007. Quantification of the heterogeneity in breast cancer cell lines using whole-cell impedance spectroscopy. Clin. Cancer Res. 13 (1), 139143. Harman, R.M., Curtis, T.M., Argyle, D.J., Coonrod, S.A., Van de Walle, G.R., 2016. A comparative study on the In vitro effects of the DNA methyltransferase inhibitor 5azacytidine (5-AzaC) in breast/mammary cancer of different mammalian species. J. Mammary Gland Biol. Neoplasia 21 (12), 5166. Hassan, U., Bashir, R., 2014. Coincidence detection of heterogeneous cell populations from whole blood with coplanar electrodes in a microfluidic impedance cytometer. Lab Chip 14 (22), 43704381. van der Helm, M.W., et al., 2016. Direct quantification of transendothelial electrical resistance in organs-on-chips. Biosens. Bioelectron. 85, 924929. van der Helm, M.W., et al., 2019. Non-invasive sensing of transepithelial barrier function and tissue differentiation in organs-on-chips using impedance spectroscopy. Lab Chip. Available from: https://doi.org/10.1039/C8LC00129D. Henry, O.Y.F., Villenave, R., Cronce, M.J., Leineweber, W.D., Benz, M.A., Ingber, D.E., 2017. Organs-on-chips with integrated electrodes for trans-epithelial electrical resistance (TEER) measurements of human epithelial barrier function. Lab Chip 17 (13), 22642271. Hoffmann, W., Feliciano, S., Martin, I., de Wild, M., Wendt, D., 2015. Novel perfused compression bioreactor system as an in vitro model to investigate fracture healing. Front. Bioeng. Biotechnol. 3, 10.

117

118

CHAPTER 3 Organs-on-a-chip engineering

Holmes, D., Morgan, H., 2010. Single cell impedance cytometry for identification and counting of CD4 T-cells in human blood using impedance labels. Anal. Chem. 82 (4), 14551461. Holmes, D., et al., 2009. Leukocyte analysis and differentiation using high speed microfluidic single cell impedance cytometry. Lab Chip 9 (20), 28812889. Homan, K.A., et al., 2016. Bioprinting of 3D convoluted renal proximal tubules on perfusable chips. Sci. Rep. 6, 113. Hosokawa, K., Maeda, R., 2000. A pneumatically-actuated three-way microvalve fabricated with polydimethylsiloxane using the membrane transfer technique. J. Micromech. Microeng. 10, 415420. Huang, H., Kamm, R.D., Lee, R.T., 2004. Cell mechanics and mechanotransduction: pathways, probes, and physiology. Am. J. Physiol. Cell Physiol. 287 (1), C1C11. Huang, S.B., Wu, M.H., Cui, Z.F., Cui, Z., Lee, G.B., 2008. A membrane-based serpentine-shape pneumatic micropump with pumping performance modulated by fluidic resistance. J. Micromech. Microeng. 18 (12), 045008. Huang, A.H., Farrell, M.J., Mauck, R.L., 2010. Mechanics and mechanobiology of mesenchymal stem cell-based engineered cartilage. J. Biomech. 43, 128136. Huh, D., Matthews, B.D., Mammoto, A., Montoya-Zavala, M., Hsin, H.Y., Ingber, D.E., 2010. Reconstituting organ-level lung functions on a chip. Science 328 (5986), 1662. LP-1668. Huh, D., Hamilton, G.A., Ingber, D.E., 2011. From 3D cell culture to organs-on-chips. Trends Cell Biol. 21 (12), 745754. Huh, D., et al., 2013. Microfabrication of human organs-on-chips. Nat. Protoc. 8 (11), 21352157. Iliescu, C., Taylor, H., Avram, M., Miao, J., Franssila, S., 2012. A practical guide for the fabrication of microfluidic devices using glass and silicon. Biomicrofluidics 6, 116. InSphero, 2019. Technology. [Online]. Available from: ,https://insphero.com/science/ enabling-technology/. (accessed 03.02.19.) Ingber, D.E., 2003. Mechanobiology and diseases of mechanotransduction. Ann. Med. 35, 564577. Irimia, D., Toner, M., 2006. Cell handling using microstructured membranes. Lab Chip 6, 345352. Ispas, C.R., Crivat, G., Andreescu, S., 2012. Review: recent developments in enzymebased biosensors for biomedical analysis. Anal. Lett. 45 (23), 168186. Jaalouk, D.E., Lammerding, J., 2009. Mechanotransduction gone awry. Nat. Rev. Mol. Cell Biol. 10, 6373. Jacobs, M.J., Blank, K., 2014. Joining forces: integrating the mechanical and optical single molecule toolkits. Chem. Sci. 5, 16801697. Jacobson, S.C., Ermakov, S.V., Ramsey, J.M., 1999. Minimizing the number of voltage sources and fluid reservoirs for electrokinetic valving in microfluidic devices. Anal. Chem. 71, 32733276. Jain, A., et al., 2018. Primary human lung alveolus-on-a-chip model of intravascular thrombosis for assessment of therapeutics. Clin. Pharmacol. Ther. 103, 332340. Jang, K.-J., et al., 2013. Human kidney proximal tubule-on-a-chip for drug transport and nephrotoxicity assessment. Integr. Biol. 5 (9), 11191129. Jang, L.S., Wang, M.H., 2007. Microfluidic device for cell capture and impedance measurement. Biomed. Microdevices 9 (5), 737743. Jenke, C., et al., 2017. The combination of micro diaphragm pumps and flow sensors for single stroke based liquid flow control. Sensors 17 (4), 755.

References

Jeon, J.H., Jeong, O.C., 2012. Effect of shear stress magnitude on intracellular calcium expression in bone cells. Microelectron. Eng. 97, 329332. Jeong, O.C., Konishi, S., 2008. Fabrication of a peristaltic micro pump with novel cascaded actuators. J. Micromech. Microeng. 18, 025022. Johnson, B.N., et al., 2016. 3D printed nervous system on a chip. Lab Chip 16 (8), 13931400. Johnson, I., Spence, M. (Eds.), 2010. Molecular Probes Handbook, A Guide to Fluorescent Probes and Labeling Technologies. Life Technologies. Juncker, D., et al., 2002. Autonomous microfluidic capillary system. Anal. Chem. 74, 61396144. Kaigala, G.V., Hoang, V.N., Backhouse, C.J., 2008. Electrically controlled microvalves to integrate microchip polymerase chain reaction and capillary electrophoresis. Lab Chip 8, 10711078. Kaihara, S., et al., 2000. Silicon micromachining to tissue engineer branched vascular channels for liver fabrication. Tissue Eng. 6 (2), 105117. Kartalov, E.P., Maltezos, G., Anderson, W.F., Taylor, C.R., Scherer, A., 2007. Electrical microfluidic pressure gauge for elastomer microelectromechanical systems. J. Appl. Phys. 102 (8), 84909849094. Kasendra, M., et al., 2018. Development of a primary human small intestine-on-a-chip using biopsy-derived organoids. Sci. Rep. 8 (1), 2871. Kellogg, R.A., Go´mez-Sjo¨berg, R., Leyrat, A.A., Tay, S., 2014. High-throughput microfluidic single-cell analysis pipeline for studies of signaling dynamics. Nat. Protoc. 9 (7), 17131726. Kieninger, J., Aravindalochanan, K., Sandvik, J.A., Pettersen, E.O., Urban, G.A., 2014. Pericellular oxygen monitoring with integrated sensor chips for reproducible cell culture experiments. Cell Prolif. 47 (2), 180188. Kieninger, J., Weltin, A., Flamm, H., Urban, G.A., 2018. Microsensor systems for cell metabolism-from 2D culture to organ-on-chip. Lab Chip 18 (9), 12741291. Kim, J., et al., 2006. Photopolymerized check valve and its integration into a pneumatic pumping system for biocompatible sample delivery. Lab Chip 6, 10911094. Kim, L., Toh, Y.-C., Voldman, J., Yu, H., 2007. A practical guide to microfluidic perfusion culture of adherent mammalian cells. Lab Chip 7, 681694. Kim, D.-H., Wong, P.K., Park, J., Levchenko, A., Sun, Y., 2009. Microengineered platforms for cell mechanobiology. Annu. Rev. Biomed. Eng. 11, 203233. Kim, H.J., Huh, D., Hamilton, G., Ingber, D.E., 2012. Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow. Lab Chip 12, 21652174. Kim, J.-Y., et al., 2015. 3D spherical microtissues and microfluidic technology for multitissue experiments and analysis. J. Biotechnol. 205, 2435. King, K.R., Wang, C.C.J., Kaazempur-Mofrad, M.R., Vacanti, J.P., Borenstein, J.T., 2004. Biodegradable microfluids. Adv. Mater. 16 (22), 20072012. Kloß, D., et al., 2008. Microcavity array (MCA)-based biosensor chip for functional drug screening of 3D tissue models. Biosens. Bioelectron. 23 (10), 14731480. Knopfmacher, O., et al., 2010. Nernst limit in dual-gated Si-nanowire FET sensors. Nano Lett. 10 (6), 22682274. Krishnan, R., et al., 2011. Substrate stiffening promotes endothelial monolayer disruption through enhanced physical forces. Am. J. Physiol. Cell. Physiol. 300 (1), C146C154. Kuo, J.T.W., Yu, L., Meng, E., Kuo, J.T.W., Yu, L., Meng, E., 2012. Micromachined thermal flow sensors—a review. Micromachines 3 (3), 550573.

119

120

CHAPTER 3 Organs-on-a-chip engineering

Kuo, A.P., Bhattacharjee, N., Lee, Y.-S., Castro, K., Kim, Y.T., Folch, A., 2019. Highprecision stereolithography of biomicrofluidic devices. Adv. Mater. Technol. 1800395, 1800395. Kurth, F., Eyer, K., Franco-Obrego´n, A., Dittrich, P.S., 2012. A new mechanobiological era: Microfluidic pathways to apply and sense forces at the cellular level. Curr. Opin. Chem. Biol. 16, 400408. Kurth, F., Franco-Obrego´n, A., Casarosa, M., Ku¨ster, S.K., Wuertz-Kozak, K., Dittrich, P. S., 2015. Transient receptor potential vanilloid 2-mediated shear-stress responses in C2C12 myoblasts are regulated by serum and extracellular matrix. FASEB J. 29 (11), 47264737. Kurth, F., Dittrich, P.S., Walde, P., Seebach, D., 2018. Influence of the membrane dye R18 and of DMSO on cell penetration of guanidinium-rich peptides. Chem. Biodivers. 15, e1800302. Kurzweil, P., 2009. Metal oxides and ion-exchanging surfaces as pH sensors in liquids: state-of-the-art and outlook. Sensors (Basel) 9 (6), 49554985. La Cour, J.B., Generelli, S., Barbe, L., Guenat, O.T., 2016. Low-cost disposable ALT electrochemical microsensors for in-vitro hepatotoxic assessment. Sens. Actuators, B: Chem. 228, 360365. Lai, H., Folch, A., 2011. Design and dynamic characterization of ‘single-stroke’ peristaltic PDMS micropumps. Lab Chip 11, 336342. Lakowicz, J.R., 1999. Principles of Fluorescence Spectroscopy, Second ed. Kluwer Academic/Plenum Publishers, New York. Lam, R.H.W., Sun, Y., Chen, W., Fu, J., 2012. Elastomeric microposts integrated into microfluidics for flow-mediated endothelial mechanotransduction analysis. Lab Chip 12 (10), 1865. Lanz, H.L., et al., 2017. Therapy response testing of breast cancer in a 3D high-throughput perfused microfluidic platform. BMC Cancer 17 (1), 709. Lee, D.E., Soper, S., Wang, W.J., 2008. Design and fabrication of an electrochemically actuated microvalve. Microsyst. Technol. 14, 17511756. Lee, H., Cho, D.W., 2016. One-step fabrication of an organ-on-a-chip with spatial heterogeneity using a 3D bioprinting technology. Lab Chip 16 (14), 26182625. Lee, S.H., Sung, J.H., 2018. Organ-on-a-chip technology for reproducing multiorgan physiology. Adv. Healthc. Mater. 7 (2), 1700419. Lee, S.J., Chan, J.C.Y., Maung, K.J., Rezler, E., Sundararajan, N., 2007. Characterization of laterally deformable elastomer membranes for microfluidics. J. Micromech. Microeng. 17, 843851. Legant, W.R., Miller, J.S., Blakely, B.L., Cohen, D.M., Genin, G.M., Chen, C.S., 2010. Measurement of mechanical tractions exerted by cells in three-dimensional matrices. Nat. Methods 7, 969971. Li, X., Tian, T., 2018. Phytochemical characterization of Mentha spicata L. under differential dried-conditions and associated nephrotoxicity screening of main compound with organ-on-a-chip. Front. Pharmacol. 9, 110. no. September. Li, Y.-S.J., Haga, J.H., Chien, S., 2005. Molecular basis of the effects of shear stress on vascular endothelial cells. J. Biomech. 38 (10), 19491971. Li, Z., Yao, S.-J., Alini, M., Stoddart, M.J., 2010. Chondrogenesis of human bone marrow mesenchymal stem cells in fibrinpolyurethane composites is modulated by frequency and amplitude of dynamic compression and shear stress. Tissue Eng., A 16, 575584.

References

Li, X., Xu, S., He, P., Liu, Y., 2015. In vitro recapitulation of functional microvessels for the study of endothelial shear response, nitric oxide and [Ca21]i. PLoS One 10 (5), 119. Lind, J.U., et al., 2017. Instrumented cardiac microphysiological devices via multimaterial three-dimensional printing. Nat. Mater. 16, 303308. Lindken, R., Rossi, M., Große, S., Westerweel, J., 2009. Micro-particle image velocimetry (μPIV): recent developments, applications, and guidelines. Lab Chip 9 (17), 2551. Liu, Y., Tuleouva, N., Ramanculov, E., Revzin, A., 2010. Aptamer-based electrochemical biosensor for interferon gamma detection. Anal. Chem. 82 (19), 81318136. Liu, M.-C., et al., 2013. Electrofluidic pressure sensor embedded microfluidic device: a study of endothelial cells under hydrostatic pressure and shear stress combinations. Lab Chip 13 (9), 1743. Lohasz, C., Rousset, N., Renggli, K., Hierlemann, A., Frey, O., 2019. Scalable microfluidic platform for flexible configuration of and experiments with microtissue multiorgan models. SLAS Technol. 24 (1), 7995. Lonardoni, D., Amin, H., Di Marco, S., Maccione, A., Berdondini, L., Nieus, T., 2017. Recurrently connected and localized neuronal communities initiate coordinated spontaneous activity in neuronal networks. PLoS Computational Biology 13, e1005672. Loskill, P., Marcus, S.G., Mathur, A., Reese, W.M., Healy, K.E., 2015. μOrgano: a Lego®like plug & play system for modular multi-organ-chips. PLoS One 10 (10), e0139587. Loskill, P., et al., 2017. WAT-on-a-chip: a physiologically relevant microfluidic system incorporating white adipose tissue. Lab Chip 17 (9), 16451654. Ma, X., et al., 2018. 3D bioprinting of functional tissue models for personalized drug screening and in vitro disease modeling. Adv. Drug Deliv. Rev. 132, 235251. Magin, C.M., Alge, D.L., Anseth, K.S., 2016. Bio-inspired 3D microenvironments: a new dimension in tissue engineering. Biomed. Mater. 11 (2), 022001. Malda, J., et al., 2013. 25th anniversary article: engineering hydrogels for biofabrication. Adv. Mater. 25 (36), 50115028. Malerba, M., Amin, H., Angotzi, G.N., Maccione, A., Berdondini, L., 2018. In: Ertl, P., Rothbauer, M. (Eds.), Fabrication of Multielectrode Arrays for Neurobiology Applications BT - Cell-Based Microarrays: Methods and Protocols. Springer New York, New York, pp. 147157. Malkoc, V., 2018. Challenges and the future of 3D bioprinting. J. Biomed. Imaging Bioeng 2 (1), 6465. Malleo, D., Nevill, J.T., Lee, L.P., Morgan, H., 2010. Continuous differential impedance spectroscopy of single cells. Microfluid. Nanofluid. 9 (23), 191198. Mann, J.M., Lam, R.H.W., Weng, S., Sun, Y., Fu, J., 2012. A silicone-based stretchable micropost array membrane for monitoring live-cell subcellular cytoskeletal response. Lab Chip 12 (4), 731740. Maschmeyer, I., et al., 2015a. A four-organ-chip for interconnected long-term co-culture of human intestine, liver, skin and kidney equivalents. Lab Chip 15 (12), 26882699. Maschmeyer, I., et al., 2015b. Chip-based human liverintestine and liverskin cocultures  a first step toward systemic repeated dose substance testing in vitro. Eur. J. Pharm. Biopharm. 95, 7787. Massing, J., Kaden, D., Ka¨hler, C.J., Cierpka, C., 2016. Luminescent twocolor tracer particles for simultaneous velocity and temperature measurements in microfluidics. Meas. Sci. Technol. 27, 015301.

121

122

CHAPTER 3 Organs-on-a-chip engineering

Materne, E.-M., et al., 2015. The multi-organ chip—a microfluidic platform for long-term multi-tissue coculture. J. Vis. Exp. (98), e52526. Mathur, A., Roy, S.S., Tweedie, M., Mukhopadhyay, S., Mitra, S.K., Mclaughlin, J.A., 2009. Characterisation of PMMA microfluidic channels and devices fabricated by hot embossing and sealed by direct bonding. Curr. Appl. Phys. 9 (6), 11991202. Mazzocchi, A.R., Rajan, S.A.P., Votanopoulos, K.I., Hall, A.R., Skardal, A., 2018. In vitro patient-derived 3D mesothelioma tumor organoids facilitate patient-centric therapeutic screening. Sci. Rep. 8 (1), 112. McLean, I.C., Schwerdtfeger, L.A., Tobet, S.A., Henry, C.S., 2018. Powering ex vivo tissue models in microfluidic systems. Lab Chip 18 (10), 13991410. Mcdonald, J.C., Duffy, D.C., Anderson, J.R., Chiu, D.T., 2000. Fabrication of microfluidic systems in PDMS. Electrophoreses 21, 2740. Meinert, C., Schrobback, K., Hutmacher, D.W., Klein, T.J., 2017. A novel bioreactor system for biaxial mechanical loading enhances the properties of tissue-engineered human cartilage. Sci. Rep. 7, 16997. Merceron, T.K., Murphy, S.V., 2015. Hydrogels for 3D Bioprinting Applications. Elsevier Inc. Mills, C.A., et al., 2006. Transparent micro- and nanopatterned poly(lactic acid) for biomedical applications. J. Biomed. Mater. Res., A 76 (4), 781787. Mimetas, 2019. Technology. [Online]. Available from: ,https://mimetas.com/page/technology. (accessed 03.02.19.) Misun, P.M., Rothe, J., Schmid, Y.R.F., Hierlemann, A., Frey, O., 2016. Multi-analyte biosensor interface for real-time monitoring of 3D microtissue spheroids in hanging-drop networks. Microsyst. Nanoeng. 2, 16022. Modena, M.M., Chawla, K., Misun, P.M., Hierlemann, A., 2018. Smart cell culture systems: integration of sensors and actuators into microphysiological systems. ACS Chem. Biol. 13 (7), 17671784. Moisan, A., et al., 2018. Mechanistic investigations of diarrhea toxicity induced by antiHER2/3 combination therapy. Mol. Cancer Ther. 17 (7), 14641474. Molckovsky, A., Wilson, B.C., 2001. Monitoring of cell and tissue responses to photodynamic therapy by electrical impedance spectroscopy. Phys. Med. Biol. 46 (4), 9831002. Monteiro, T., Almeida, M.G., 2019. Electrochemical enzyme biosensors revisited: old solutions for new problems. Crit. Rev. Anal. Chem. 49, 4466. Moraes, C., Mehta, G., Lesher-Perez, S.C., Takayama, S., 2012. Organs-on-a-chip: a focus on compartmentalized microdevices. Ann. Biomed. Eng. 40 (6), 12111227. Moser, I., Jobst, G., Urban, G.A., 2002. Biosensor arrays for simultaneous measurement of glucose, lactate, glutamate, and glutamine. Biosens. Bioelectron. 17 (4), 297302. Moya, A., et al., 2018. Online oxygen monitoring using integrated inkjet-printed sensors in a liver-on-a-chip system. Lab Chip 18 (14), 20232035. Mross, S., Pierrat, S., Zimmermann, T., Kraft, M., 2015. Microfluidic enzymatic biosensing systems: a review. Biosens. Bioelectron. 70, 376391. Murphy, S.V., Atala, A., 2014. 3D bioprinting of tissues and organs. Nat. Biotechnol. 32 (8), 773785. Naruse, K., Yamada, T., Sai, X.R., Hamaguchi, M., Sokabe, M., 1998. Pp125FAK is required for stretch dependent morphological response of endothelial cells. Oncogene 17 (4), 455463.

References

Nguyen, N.-T., Truong, T.-Q., 2005. Flow rate measurement in microfluidics using optical sensors. In: First International Conference on Sensing Technology. pp. 2831 Nguyen, N.-T., Wereley, S.T., 2002. Fundamentals and Applications of Microfluidics. Artech House. Obien, M.E.J., Gong, W., Frey, U., Bakkum, D.J., 2017. CMOS-Based High-Density Microelectrode Arrays: Technology and Applications. Springer, Singapore, pp. 339. Odijk, M., et al., 2015. Measuring direct current trans-epithelial electrical resistance in organ-on-a-chip microsystems. Lab Chip 15 (3). Oh, K.W., Ahn, C.H., 2006. A review of microvalves. J. Micromech. Microeng. 16, R13R39. Oleaga, C., et al., 2016. Multi-Organ toxicity demonstration in a functional human in vitro system composed of four organs. Sci. Rep. 6, 20030. ¨ zu¨m, C., Da¨hne, L., 2019. Label-free bioanalysis based Olszyna, M., Debrassi, A., U on low-Q whispering gallery modes: rapid preparation of microsensors by means of layer-by-layer technology. Adv. Funct. Mater. 29 (2), 1805998. Olthuis, W., Robben, M.A.M., Bergveld, P., Bos, M., Van Der Linden, W.E., 1990. pH sensor properties of electrochemically grown iridium oxide. Sens. Actuators, B: Chem. 2, 247256. Ong, L.J.Y., Islam, A., DasGupta, R., Iyer, N.G., Leo, H.L., Toh, Y.C., 2017. A 3D printed microfluidic perfusion device for multicellular spheroid cultures. Biofabrication 9, 045005. Oomen, P.E., Skolimowski, M.D., Verpoorte, E., 2016. Implementing oxygen control in chip-based cell and tissue culture systems. Lab Chip 16, 33943414. Oskooei, A., Gu¨nther, A., 2015. Bubble pump: scalable strategy for in-plane liquid routing. Lab Chip 15 (13), 28422853. Pachauri, V., Ingebrandt, S., 2016. Biologically sensitive field-effect transistors: from ISFETs to NanoFETs. Essays Biochem. 60 (1), 8190. Papaioannou, T.G., Stefanadis, C., 2005. Vascular wall shear stress: basic principles and methods. Hell. J. Cardiol. 46, 915. Parate, D., et al., 2017. Enhancement of mesenchymal stem cell chondrogenesis with shortterm low intensity pulsed electromagnetic fields. Sci. Rep. 7, 9421. Park, S.H., Sim, W.Y., Min, B.H., Yang, S.S., Khademhosseini, A., Kaplan, D.L., 2012. Chip-based comparison of the osteogenesis of human bone marrow- and adipose tissue-derived mesenchymal stem cells under mechanical stimulation. PLoS One 7, e44689. Park, S.-H.L.D., Lim, J., Park, J.Y., 2015. Concise review: stem cell microenvironment on a chip: current technologies for tissue engineering and stem cell biology. Stem Cells Transl. Med. 4, 13521368. Parker, 2015. Electroactive Polymer (EAP) Technology—BUL 3800-EAP Overview Pemble, C.M., Towe, B.C., 1999. A miniature shape memory alloy pinch valve. Sens. Actuat. A 77, 145148. Perdomo, J., Hinkers, H., Sundermeier, C., Seifert, W., Martı´nez Morell, O., Knoll, M., 2000. Miniaturized real-time monitoring system for L-lactate and glucose using microfabricated multi-enzyme sensors. Biosens. Bioelectron. 15 (910), 515522. Peters, M.F., Lamore, S.D., Guo, L., Scott, C.W., Kolaja, K.L., 2015. Human stem cellderived cardiomyocytes in cellular impedance assays: bringing cardiotoxicity screening to the front line. Cardiovasc. Toxicol. 15 (2), 127139.

123

124

CHAPTER 3 Organs-on-a-chip engineering

Pilarski, P.M., Adamia, S., Backhouse, C.J., 2005. An adaptable microvalving system for on-chip polymerase chain reactions. J. Immunol. Meth. 305, 4858. Piro, B., Reisberg, S., 2017. Recent advances in electrochemical immunosensors. Sensors 17, 794. Pocock, K.J., et al., 2016. Low-temperature bonding process for the fabrication of hybrid glassmembrane organ-on-a-chip devices. J. Micro/Nanolithogr., MEMS, MOEMS 15 (4), 044502. Polinkovsky, M., Gutierrez, E., Levchenko, A., Groisman, A., 2009. Fine temporal control of the medium gas content and acidity and on-chip generation of series of oxygen concentrations for cell cultures. Lab Chip 9 (8), 10731084. Poulin, A., Imboden, M., Sorba, F., Grazioli, S., Martin-Olmos, C., Rosset, S., et al., 2018. An ultra-fast mechanically active cell culture substrate. Sci. Rep. 8, 9895. Pound, P., Ebrahim, S., Sandercock, P., Bracken, M.B., Roberts, I., 2004. Where is the evidence that animal research benefits humans? BMJ 328 (7438), p. 514 LP-517. Powers, M.J., et al., 2002. A microfabricated array bioreactor for perfused 3D liver culture. Biotechnol. Bioeng. 78 (3), 257269. Primiceri, E., Chiriaco`, M.S., Rinaldi, R., Maruccio, G., 2013. Cell chips as new tools for cell biology  results, perspectives and opportunities. Lab Chip 13 (19), 3789. Quiro´s-Solano, W.F., Gaio, N., Silvestri, C., Pandraud, G., Sarro, P.M., 2016. PEDOT: PSS: a conductive and flexible polymer for sensor integration in organ-on-chip platforms. Procedia Eng. 168, 11841187. Ramamoorthy, R., Dutta, P.K., Akbar, S.A., 2003. Oxygen sensors: materials, methods, designs and applicationse. J. Mater. Sci. 38 (21), 42714282. Ranjan, V., Waterbury, R., Xiao, Z., Diamond, S.L., 1996. Fluid shear stress induction of the transcriptional activator c-fos in human and bovine endothelial cells, HeLa, and Chinese hamster ovary cells. Biotechnol. Bioeng. 49 (4), 383390. Ravalli, A., Lozzi, L., Marrazza, G., 2016. Micro-flow immunosensor based on thin-film interdigitated gold array microelectrodes for cancer biomarker detection. Curr. Drug Deliv. 13 (3), 400408. Regehr, K.J., et al., 2009. Biological implications of polydimethylsiloxane-based microfluidic cell culture. Lab Chip 9, 21322139. Rennert, K., et al., 2015. A microfluidically perfused three dimensional human liver model. Biomaterials 71, 119131. Rezaei, B., Ghani, M., Shoushtari, A.M., Rabiee, M., 2016. Electrochemical biosensors based on nanofibres for cardiac biomarker detection: a comprehensive review. Biosens. Bioelectron. 78, 513523. Rhodin, J.A.G., 2014. Architecture of the vessel wall. Compr. Physiol. Riahi, R., et al., 2016. Automated microfluidic platform of bead-based electrochemical immunosensor integrated with bioreactor for continual monitoring of cell secreted biomarkers. Sci. Rep. 6 (1), 24598. Richter, M., Wackerle, M., Woias, P., Hillerich, B., 1999. novel flow sensor with high time resolution based on differential pressure. In: Technical Digest. IEEE International MEMS 99 Conference. 12th IEEE International Conference on Micro Electro Mechanical Systems (Cat. No. 99CH36291). pp. 118123 Rodrı`guez-Villarreal, A.I., Arundell, M., Carmona, M., Samitier, J., 2010. High flow rate microfluidic device for blood plasma separation using a range of temperatures. Lab Chip 10, 211219.

References

Rogal, J., Probst, C., Loskill, P., 2017. Integration concepts for multi-organ chips: how to maintain flexibility?!. Futur. Sci. OA 3 (2), FSO180. Ronaldson-Bouchard, K., Vunjak-Novakovic, G., 2018. Organs-on-a-chip: a fast track for engineered human tissues in drug development. Cell Stem Cell 22 (3), 310324. Ronkainen, N.J., Halsall, H.B., Heineman, W.R., 2010. Electrochemical biosensors. Chem. Soc. Rev. 39 (5), 17471763. Rotariu, L., Lagarde, F., Jaffrezic-Renault, N., Bala, C., 2016. Electrochemical biosensors for fast detection of food contaminants  trends and perspective. TrAC, Trends Anal. Chem. 79, 8087. Rothbauer, M., Zirath, H., Ertl, P., 2018. Recent advances in microfluidic technologies for cell-to-cell interaction studies. Lab Chip 18 (2), 249270. Rousselle, P., Braye, F., Dayan, G., 2018. Re-epithelialization of adult skin wounds: cellular mechanisms and therapeutic strategies. Adv. Drug Deliv. Rev. Saggiomo, V., Velders, A.H., 2015. Simple 3D printed scaffold-removal method for the fabrication of intricate microfluidic devices. Adv. Sci. 2 (9), 15. Sajay, B.N.G., Yin, C.S., Ramadan, Q., 2017. Optimization of micro-fabricated porous membranes for intestinal epithelial cell culture and in vitro modeling of the human intestinal barrier. J. Micromech. Microeng. 27, 124004. Santoro, R., Krause, C., Martin, I., Wendt, D., 2012. On-line monitoring of oxygen as a non-destructive method to quantify cells in engineered 3D tissue constructs. J. Tissue Eng. Regen. Med. 6 (9), 696701. dos Santos, V., Brandalise, R.N., Savaris, M., 2017. Eng. Biomater. Sarkar, S., Dadhania, M., Rourke, P., Desai, T.A., Wong, J.Y., 2005. Vascular tissue engineering: microtextured scaffold templates to control organization of vascular smooth muscle cells and extracellular matrix. Acta Biomater. 1 (1), 93100. Sartorius, 2019. Cubis MSA225S datasheet. [Online]. Available from: ,http://balancecanada. com/Documents/MSA225S-000-DU_000.pdf. (accessed 30.01.19.) Sassa, F., Morimoto, K., Satoh, W., Suzuki, H., 2008. Electrochemical techniques for microfluidic applications. Electrophoresis 29 (9), 17871800. Satoh, W., Hosono, H., Yokomaku, H., Morimoto, K., Upadhyay, S., Suzuki, H., 2008. Integrated electrochemical analysis system with microfluidic and sensing functions. Sensors 8 (2), 11111127. Schade-Kampmann, G., Huwiler, A., Hebeisen, M., Hessler, T., Di Berardino, M., 2008. On-chip non-invasive and label-free cell discrimination by impedance spectroscopy. Cell Prolif. 41 (5), 830840. Schasfoort, R.B.M., Schlautmann, S., Hendrikse, L., van den Berg, A., 1999. Field-effect flow control for microfabricated fluidic networks. Science 286, 942945 (80-.). Scheinpflug, J., et al., 2018. Journey into bone models: a review. Genes (Basel) 9 (5), 247. Schimek, K., et al., 2013. Integrating biological vasculature into a multi-organ-chip microsystem. Lab Chip 13 (18), 3588. Schmid, Y.R.F., Bu¨rgel, S.C., Misun, P.M., Hierlemann, A., Frey, O., 2016. Electrical electrical impedance spectroscopy for microtissue spheroid analysis in hanging-drop networks. ACS Sens. 1 (8), 10281035. Sebastian, B., Dittrich, P.S., 2018. Microfluidics to mimic blood flow in health and disease. Annu. Rev. Fluid Mech. 50, 483504.

125

126

CHAPTER 3 Organs-on-a-chip engineering

Sensirion, 2019. LPG10 liquid flow sensor. [Online]. Available from: ,https://www.sensirion. com/en/about-us/company/technology/cmosens-technology-for-liquid-flow/. (accessed 30.01.19.) Sergachev, I., Rusanov, A., Trushkin, E., Sakharov, D., Marx, U., Tonevitsky, A., 2013. Fluorescent optical fiber sensors for cell viability monitoring. Analyst 138 (14), 4066. Shaegh, S.A.M., et al., 2016. A microfluidic optical platform for real-time monitoring of pH and oxygen in microfluidic bioreactors and organ-on-chip devices. Biomicrofluidics 10, 044111. Shah, P., et al., 2016. A microfluidics-based in vitro model of the gastrointestinal humanmicrobe interface. Nat. Commun. 7 (1), 11535. Shin, S.R., et al., 2016. Aptamer-based microfluidic electrochemical biosensor for monitoring cell-secreted trace cardiac biomarkers. Anal. Chem. 88 (20), 1001910027. Shirure, V.S., George, S.C., 2017. Design considerations to minimize the impact of drug absorption in polymer-based organ-on-a-chip platforms. Lab Chip 17, 681690. Skardal, A., Shupe, T., Atala, A., 2016. Organoid-on-a-chip and body-on-a-chip systems for drug screening and disease modeling. Drug Discov. Today 21 (9), 13991411. Song, H., et al., 2013. A microfluidic impedance flow cytometer for identification of differentiation state of stem cells. Lab Chip 13 (12), 2300. Sonntag, F., et al., 2010. Design and prototyping of a chip-based multi-micro-organoid culture system for substance testing, predictive to human (substance) exposure. J. Biotechnol. 148 (1), 7075. Sonntag, F., Gru¨nzner, S., Schmieder, F., Busek, M., Klotzbach, U., Franke, V., 2015. Multilayer based lab-on-a-chip-systems for substance testing. Proc. SPIE 9351, 93510C. Sonntag, F., et al., 2016. Universal lab-on-a-chip platform for complex, perfused 3D cell cultures. In: Proc. of SPIE, vol. 9705. p. 970516 Spichiger, S., Spichiger-Keller, U.E., 2011. New single-use sensors for online measurement of glucose and lactate: the answer to the PAT initiative. Single-Use Technology in Biopharmaceutical Manufacture. John Wiley & Sons, Inc, pp. 295299. Squires, T.M., Quake, S.R., 2005. Microfluidics: fluid physics at the nanoliter scale. Rev. Mod. Phys. 77 (3), 9771026. Srinivasan, B., Kolli, A.R., Esch, M.B., Abaci, H.E., Shuler, M.L., Hickman, J.J., 2015. TEER measurement techniques for in vitro barrier model systems. J. Lab. Autom 20 (2), 107126. Stern, E., et al., 2007. Label-free immunodetection with CMOS-compatible semiconducting nanowires. Nature 445 (7127), 519522. Stucki, A.O., et al., 2015. A lung-on-a-chip array with an integrated bio-inspired respiration mechanism. Lab Chip 15, 13021310. Stucki, J.D., et al., 2018. Medium throughput breathing human primary cell alveolus-onchip model. Sci. Rep. 8 (1), 14359. Studer, V., Hang, G., Pandolfi, A., Ortiz, M., Anderson, W.F., Quake, S.R., 2004. Scaling properties of a low-actuation pressure microfluidic valve. J. Appl. Phys. 95, 393398. Su, X., Young, E.W.K., Underkofler, H.A.S., Kamp, T.J., January, C.T., Beebe, D.J., 2011. Microfluidic cell culture and its application in high-throughput drug screening: cardiotoxicity assay for hERG channels. J. Biomol. Screen. 16 (1), 101111. Sudho¨lter, E.J.R., van der Wal, P.D., Skowronska-Ptasinska, M., van den Berg, A., Reinhoudt, D.N., 1989. Ion-sensing using chemically-modified ISFETs. Sens. Actuators 17 (12), 189194.

References

Sun, J., Wang, Y., Qian, Z., Hu, C., 2011. An approach to architecture 3D scaffold with interconnective microchannel networks inducing angiogenesis for tissue engineering. J. Mater. Sci. Mater. Med. 22, 25652571. Sun, T., Morgan, H., 2010. Single-cell microfluidic impedance cytometry: a review. Microfluid. Nanofluid. 8, 423443. Sundararajan, N., Kim, D., Berlin, A.A., 2005. Microfluidic operations using deformable polymer membranes fabricated by single layer soft lithography. Lab Chip 5, 350354. Svennersten, K., Berggren, M., Richter-Dahlfors, A., Jager, E.W.H., 2011. Mechanical stimulation of epithelial cells using polypyrrole microactuators. Lab Chip 11, 32873293. Szulcek, R., Bogaard, H.J., van Nieuw Amerongen, G.P., 2014. Electric cell-substrate impedance sensing for the quantification of endothelial proliferation, barrier function, and motility. J. Vis. Exp. 85, e51300. Talaei, S., van der Wal, P.D., Ahmed, S., Liley, M., de Rooij, N.F., 2015. Enzyme SU-8 microreactors: simple tools for cell-culture monitoring. Microfluid. Nanofluid. 19 (2), 351361. Tanaka, Y., Fujikawa, T., Kazoe, Y., Kitamori, T., 2013. An active valve incorporated into a microchip using a high strain electroactive polymer. Sens. Actuators, B: Chem. 184, 163169. Tecan, 2019. Liquid handling and automation. [Online]. Available from: ,https://lifesciences.tecan.com/products/liquid_handling_and_automation. (accessed 03.02.19.) Temiz, Y., Lovchik, R.D., Kaigala, G.V., Delamarche, E., 2015. Lab-on-a-chip devices: How to close and plug the lab? Microelectron. Eng. 132, 156175. Thedinga, E., et al., 2007. Online monitoring of cell metabolism for studying pharmacodynamic effects. Toxicol. Appl. Pharmacol. 220 (1), 3344. The´venot, D.R., Toth, K., Durst, R.A., Wilson, G.S., 2001. Electrochemical biosensors: recommended definitions and classification. Biosens. Bioelectron. 16 (12), 121131. Thielecke, H., Mack, A., Robitzki, A., 2001. A multicellular spheroid-based sensor for anti-cancer therapeutics. Biosens. Bioelectron. 16 (45), 261269. Thomassen, 2018. Organ on Chip in Development (ORCHID) TissUse, 2019. Technology. [Online]. Available from: ,https://www.tissuse.com/en/ technology/. (accessed 03.02.19.) Torisawa, Y., et al., 2014. Bone marrow-on-a-chip replicates hematopoietic niche physiology in vitro. Nat. Methods 11, 663671. Tracey, M.C., Johnston, I.D., Davis, J.B., Tan, C.K.L., 2006. Dual independent displacement-amplified micropumps with a single actuator. J. Micromech. Microeng. 16, 14441452. Trietsch, S.J., Israe¨ls, G.D., Joore, J., Hankemeier, T., Vulto, P., 2013. Microfluidic titer plate for stratified 3D cell culture. Lab Chip 13 (18), 3548. Trietsch, S.J., et al., 2017. Membrane-free culture and real-time barrier integrity assessment of perfused intestinal epithelium tubes. Nat. Commun. 8 (1), 262. Tru¨ssel, A., Mu¨ller, R., Webster, D., 2012. Toward mechanical systems biology in bone. Ann. Biomed. Eng. 40 (11), 24752487. Tsamandouras, N., Chen, W.L.K., Edington, C.D., Stokes, C.L., Griffith, L.G., Cirit, M., 2017. Integrated gut and liver microphysiological systems for quantitative in vitro pharmacokinetic studies. AAPS J. 19 (5), 14991512. Tsao, C.W., 2016. Polymer microfluidics: simple, low-cost fabrication process bridging academic lab research to commercialized production. Micromachines 7 (12), 225.

127

128

CHAPTER 3 Organs-on-a-chip engineering

Ugo, P., Moretto, L.M., 2017. Electrochemical immunosensors and aptasensors. Chemosensors 5 (2), 13. Unger, M.A., Chou, H.P., Thorsen, T., Scherer, A., Quake, S.R., 2000. Monolithic microfabricated valves and pumps by multilayer soft lithography. Science 288, 113116 (80-.). Van Dorst, B., et al., 2016. Integration of an optical CMOS sensor with a microfluidic channel allows a sensitive readout for biological assays in point-of-care tests. Biosens. Bioelectron. 78, 126131. van Duinen, V., et al., 2017. 96 perfusable blood vessels to study vascular permeability in vitro. Sci. Rep. 7, 18071. Varma, S., Fendyur, A., Box, A., Voldman, J., 2017. Multiplexed cell-based sensors for assessing the impact of engineered systems and methods on cell health. Anal. Chem. 89, 46634670. Varma, S., Garcia-Cardena, G., Voldman, J., 2018. Unraveling endothelial cell phenotypic regulation by spatial hemodynamic flows with microfluidics. In: Proc. MicroTAS 2018. pp. 109112 Verhulsel, M., Vignes, M., Descroix, S., Malaquin, L., Vignjevic, D.M., Viovy, J.L., 2014. A review of microfabrication and hydrogel engineering for micro-organs on chips. Biomaterials 35 (6), 18161832. Verma, R., Adhikary, R.R., Banerjee, R., 2016. Smart material platforms for miniaturized devices: implications in disease models and diagnostics. Lab Chip 16 (11), 19781992. Vijayavenkataraman, S., Yan, W.C., Lu, W.F., Wang, C.H., Fuh, J.Y.H., 2018. 3D bioprinting of tissues and organs for regenerative medicine. Adv. Drug Deliv. Rev. 132, 296332. Villenave, R., et al., 2017. Human gut-on-a-chip supports polarized infection of coxsackie B1 virus in vitro. PLoS One 12 (2), e0169412. Vogel, V., Sheetz, M., 2006. Local force and geometry sensing regulate cell functions. Nat. Rev. Mol. Cell Biol. 7 (4), 265275. Vormann, M.K., et al., 2018. Nephrotoxicity and kidney transport assessment on 3D perfused proximal tubules. AAPS J. 20 (5), 90. Vriend, J., et al., 2018. Screening of drug-transporter interactions in a 3D microfluidic renal proximal tubule on a chip. AAPS J. 20 (5), 87. Wagner, I., et al., 2013. A dynamic multi-organ-chip for long-term cultivation and substance testing proven by 3D human liver and skin tissue co-culture. Lab Chip 13 (18), 35383547. Walker, G.M., Beebe, D.J., 2002. A passive pumping method for microfluidic devices. Lab Chip 2 (3), 131. Wang, C.H., Lee, G.B., 2006. Pneumatically driven peristaltic micropumps utilizing serpentine-shape channels. J. Micromech. Microeng. 16, 341348. Wang, J., 2008. Electrochemical glucose biosensors. Chem. Rev. 108 (2), 814825. Wang, Y.N., Fu, L.M., 2018. Micropumps and biomedical applications  a review. Microelectron. Eng. 195, 121138. Wang, Y.I., Shuler, M.L., 2018. UniChip enables long-term recirculating unidirectional perfusion with gravity-driven flow for microphysiological systems. Lab Chip 18 (17), 25632574. Wang, L., et al., 2009. Polydimethylsiloxane-integratable micropressure sensor for microfluidic chips. Biomicrofluidics 3 (3), 034105.

References

Wang, J.D., Douville, N.J., Takayama, S., ElSayed, M., 2012. Quantitative analysis of molecular absorption into PDMS microfluidic channels. Ann. Biomed. Eng. 40, 18621873. Wang, X., Phan, D.T.T., Zhao, D., George, S.C., Hughes, C.C.W., Lee, A.P., 2016. An onchip microfluidic pressure regulator that facilitates reproducible loading of cells and hydrogels into microphysiological system platforms. Lab Chip 16 (5), 868876. Wang, Y.I., Abaci, H.E., Shuler, M.L., 2017. Microfluidic blood-brain barrier model provides in vivo-like barrier properties for drug permeability screening. Biotechnol. Bioeng. 114 (1), 184194. Wang, T., Chen, J., Zhou, T., Song, L., 2018a. Fabricating microstructures on glass for microfluidic chips by glass molding process. Micromachines 9 (6), 115. Wang, Y.I., Carmona, C., Hickman, J.J., Shuler, M.L., 2018b. Multiorgan microphysiological systems for drug development: strategies, advances, and challenges. Adv. Healthc. Mater. 7 (2), 129. Wegener, J., Keese, C.R., Giaever, I., 2000. Electric cell-substrate impedance sensing (ECIS) as a noninvasive means to monitor the kinetics of cell spreading to artificial surfaces. Exp. Cell Res. 259 (1), 158166. Weibel, D.B., Kruithof, M., Potenta, S., Sia, S.K., Lee, A., Whitesides, G.M., 2005. Torque-actuated valves for microfluidics. Anal. Chem. 77, 47264733. Weibel, D.B., Siegel, A.C., Lee, A., George, A.H., Whitesides, G.M., 2007. Pumping fluids in microfluidic systems using the elastic deformation of poly(dimethylsiloxane). Lab Chip 7, 18321836. Weltin, A., et al., 2014. Cell culture monitoring for drug screening and cancer research: a transparent, microfluidic, multi-sensor microsystem. Lab Chip 14 (1), 138146. Weltin, A., Kieninger, J., Urban, G.A., 2016. Microfabricated, amperometric, enzymebased biosensors for in vivo applications. Anal. Bioanal. Chem. 408 (17), 45034521. Weltin, A., Hammer, S., Noor, F., Kaminski, Y., Kieninger, J., Urban, G.A., 2017. Accessing 3D microtissue metabolism: lactate and oxygen monitoring in hepatocyte spheroids. Biosens. Bioelectron. 87, 941948. Wen, W., Yan, X., Zhu, C., Du, D., Lin, Y., 2017. Recent advances in electrochemical immunosensors. Anal. Chem. 89 (1), 138156. Wevers, N.R., et al., 2018. A perfused human bloodbrain barrier on-a-chip for highthroughput assessment of barrier function and antibody transport. Fluids Barriers CNS 15 (1), 23. van der Wijngaart, W., Chugh, D., Man, E., Melin, J., Stemme, G., 2007. A lowtemperature thermopneumatic actuation principle for gas bubble microvalves. J. Microelectromech. Syst. 16, 765774. Wikswo, J.P., et al., 2013. Engineering challenges for instrumenting and controlling integrated organ-on-chip systems. IEEE Trans. Biomed. Eng. 60 (3), 682690. Williams, D.F., 2008. On the mechanisms of biocompatibility. Biomaterials 29 (20), 29412953. Wipf, M., et al., 2016. Label-Free FimH protein interaction analysis using silicon nanoribbon bioFETs. ACS Sens. 1 (6), 781788. Wong, I., Ho, C.-M., 2009. Surface molecular property modifications for poly(dimethylsiloxane) (PDMS) based microfluidic devices. Microfluid. Nanofluid. 7 (3), 291306. Wrasse, A.N., Bertoldi, D., Morales, R.E.M., Jose da Silva, M., 2017. Two-phase flow rate measurement using a capacitive sensor and a Venturi meter. In: 2017 IEEE Sensors. pp. 13

129

130

CHAPTER 3 Organs-on-a-chip engineering

Xiao, C., Luong, J.H.T., 2005. Assessment of cytotoxicity by emerging impedance spectroscopy. Toxicol. Appl. Pharmacol. 206 (2), 102112. Xiao, S., et al., 2017. A microfluidic culture model of the human reproductive tract and 28-day menstrual cycle. Nat. Commun. 8, 14584. Xing, J.Z., Zhu, L., Gabos, S., Xie, L., 2006. Microelectronic cell sensor assay for detection of cytotoxicity and prediction of acute toxicity. Toxicol. Vitr. 20 (6), 9951004. Yang, B.Z., Lin, Q., 2007. A latchable microvalve using phase change of paraffin wax. Sens. Actuators, A: Phys. 134, 194200. Yang, Q., Lian, Q., Xu, F., 2017. Perspective: fabrication of integrated organ-on-a-chip via bioprinting. Biomicrofluidics 11, 031301. Yavuz, C., Oliaei, S.N.B., Cetin, B., Yesil-Celiktas, O., 2016. Sterilization of PMMA microfluidic chips by various techniques and investigation of material characteristics. J. Supercrit. Fluids 107, 114121. Yi, H.-G., Lee, H., Cho, D.-W., 2017. 3D printing of organs-on-chips. Bioengineering 4, 10. Yoo, J.C., Choi, Y.J., Kang, C.J., Kim, Y.S., 2007. A novel polydimethylsiloxane microfluidic system including thermopneumatic-actuated micropump and paraffin-actuated microvalve A 2007. Sens. Actuators 139, 216220. Yum, K., Hong, S.G., Healy, K.E., Lee, L.P., 2014. Physiologically relevant organs on chips. Biotechnol. J. 9 (1), 1627. Zeller, P., et al., 2017. Hepatocytes cocultured with Sertoli cells in bioreactor favors Sertoli barrier tightness in rat. J. Appl. Toxicol. 37 (3), 287295. Zengerle, R., Geiger, W., Richter, M., Ulrich, J., Kluge, S., Richter, A., 1995. Transient measurements on miniaturized diaphragm pumps in microfluid systems. Sens. Actuators, A: Phys. 47 (13), 557561. Zhang, J.X.J., Hoshino, K., 2014. Electrical transducers. In: Zhang, J.X.J., Hoshino, K. (Eds.), Molecular Sensors and Nanodevices. Elsevier, Oxford, pp. 169232. Zhang, Z., Zhe, J., Chandra, S., Hu, J., 2005. An electronic pollen detection method using Coulter counting principle. Atmos. Environ. 39 (30), 54465453. Zhang, Y.S., et al., 2017. Multisensor-integrated organs-on-chips platform for automated and continual in situ monitoring of organoid behaviors. Proc. Natl. Acad. Sci. U.S.A. 114 (12), E2293E2302. Zhang, B., Korolj, A., Lai, B.F.L., Radisic, M., 2018. Advances in organ-on-a-chip engineering. Nat. Rev. Mater. 3 (8), 257278. Zheng, F., Fu, F., Cheng, Y., Wang, C., Zhao, Y., Gu, Z., 2016. Organ-on-a-chip systems: microengineering to biomimic living systems. Small 12 (17), 22532282. Zhou, Y., 2017. The recent development and applications of fluidic channels by 3D printing. J. Biomed. Sci. 24, 80. Zhu, Z., Frey, O., Franke, F., Haandbæk, N., Hierlemann, A., 2014. Real-time monitoring of immobilized single yeast cells through multifrequency electrical impedance spectroscopy. Anal. Bioanal. Chem. 406, 70157025. Zhu, Z., Frey, O., Haandbaek, N., Franke, F., Rudolf, F., Hierlemann, A., 2015. Time-lapse electrical impedance spectroscopy for monitoring the cell cycle of single immobilized S. pombe cells. Sci. Rep. 5, 17180. no. October.

CHAPTER

Lung-on-a-chip platforms for modeling disease pathogenesis

4

Alessandra Dellaquila1,2, Emma K. Thome´e1,3, Alexander H. McMillan1,4 and Sasha Cai Lesher-Pe´rez1 1

Elvesys Microfluidic Innovation Center, Paris, France Biomolecular Photonics, Department of Physics, University of Bielefeld, Bielefeld, Germany 3 University of Strasbourg, Strasbourg, France 4 Department of Microbial and Molecular Systems, Centre for Surface Chemistry and Catalysis (COK), KU Leuven, Leuven, Belgium 2

Introduction The global prevalence and major burden of lung disease on human health and economy highlight the importance of understanding lung pathology to advance treatment. In the ongoing endeavor to improve the tools and methodologies in the field, we can now create artificial cellular microenvironments that closely mimic those found in the human body. This will potentially revolutionize the way in which human biology is investigated, in the form of organ-on-a-chip (OOC) devices. The novel capabilities of microengineering systems can be leveraged to mimic lung functions, which involve a complex and hierarchical milieu of fluid and solid mechanical stresses. These platforms are termed lung-on-a-chip (LOC) and will be the focus of this chapter. We provide an overview of lung physiology and common pathologies, along with the conventional in vitro approaches to study them. We then introduce LOC development and illustrate the uses of these devices and the effects they have had on lung pathology research. An outline of the current state of LOC devices is given alongside foreseen improvements that we deem critical to the future success of LOC devices as effective research tools.

Anatomy and physiology of the respiratory system The respiratory system supplies oxygen and removes carbon dioxide via gas exchange across the capillary alveolar interface. The respiratory system also maintains blood pH, filters xenobiotics, and serves as a blood reservoir within the pulmonary vasculature. The system is composed of nose, pharynx, larynx, trachea, bronchi, and lungs. The trachea bifurcates into two primary bronchi and Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00004-8 © 2020 Elsevier Inc. All rights reserved.

133

134

CHAPTER 4 Lung-on-a-chip platforms

FIGURE 4.1 (A) Simplified schematic of the anatomy of the respiratory system; (B) respiratory airways include the bronchioles and the alveoli; (C) physiology of gas exchange at the alveolar capillary interface; and (D) a schematic model of distribution of the main cell types of the respiratory organs.

continues to branch into lobar and segmental bronchi, bronchioles, and terminal bronchioles (Fig. 4.1A). The terminal bronchioles in the lungs are cone-shaped, spongy organs located in the thoracic cavity, further separate into respiratory bronchioles, alveolar ducts, and alveolar sacs (Fig. 4.1B). Alveoli are grape-like structures with a diameter of approximately 250 μm. Gas exchange with the blood occurs through the alveolar membrane, at the alveolar capillary interface (Fig. 4.1C). The adult human lung has around 300 million alveoli, with a surface area of up to 100 m2 (Anjali and Mahto, 2016; Fonseca et al., 2017; Siegel et al., 2017; Wang et al., 2016). The airways are coated by epithelium, which appears as a pseudostratified, ciliated layer across the trachea, and the bronchi, where it is mainly composed of basal cells, ciliated cells, and goblet cells, which are responsible for mucus secretion, with a protective function. In the bronchioles the epithelium is thinner and

Introduction

cuboidal and is composed of club secretory cells and few ciliated cells; club cells regulate immune and metabolic activities by secreting surfactants and antimicrobial proteins. The alveolus membrane is layered with two types of pneumocytes, type I (ATI) and type II (ATII) alveolar epithelial cells. Type I cells are flat, squamous cells involved in gas exchange and immune response and compose 90% of the alveolar surface, while type II cells are cuboidal and secrete surfactants to reduce the alveolar surface tension (Anjali and Mahto, 2016; Siegel et al., 2017; Wang et al., 2016) (Fig. 4.1D).

The burden of respiratory diseases: classification, impact on global health, and statistics Respiratory diseases mainly involve airways, lung tissue, and pulmonary circulation and are commonly classified as (1) obstructive, characterized by narrowed airways with formation of mucus/liquid plugs, (2) restrictive, when the lung volume is reduced, or (3) infectious, caused by microorganisms (Anjali and Mahto, 2016). According to the World Health Organization, respiratory diseases are the major cause of death and disability in the world, especially chronic obstructive pulmonary disease (COPD), asthma, acute lower respiratory tract infection and pneumonia, tuberculosis, and lung cancer (Forum of International Respiratory Societies and European Respiratory Society, 2017; Wang et al., 2016). COPD and asthma are the two main chronic obstructive diseases and affect 200 and 334 million individuals worldwide, respectively. COPD is the third leading cause of death, with 3 million death per year. Asthma affects 14% of children, representing the leading chronic childhood disease. Mortality from lower respiratory tract infection (more than 4 million individuals annually) exceeds mortality from the human immunodeficiency virus, tuberculosis and malaria combined and is a leading cause of death in children aged under 5 years (Forum of International Respiratory Societies and European Respiratory Society, 2017). Tuberculosis is the most widespread infectious disease worldwide, with more than 10 million new cases and almost 2 million deaths each year (Fonseca et al., 2017). Lung cancer is estimated to be the leading cause of cancer deaths (Siegel et al., 2017), with 1.6 million deaths per year. The social and economic burden of respiratory is high, and the costs of COPD and asthma are estimated to exceed h30 and h15 billion per annum, respectively, in the European Union alone (Blume and Davies, 2013). Another respiratory disease is cystic fibrosis (CF), a multiorgan pathology caused by a mutation in the CF transmembrane conductance regulator (CFTR) gene that affects more than 70,000 individuals worldwide. This mutation reduces ion transport, causing decreased mucociliary clearance, chronic inflammation, and bacterial colonization at the pulmonary level, with the lung symptoms being the most serious (Lavelle et al., 2016; Wang et al., 2014). The establishment and improvement of study models in research are thus necessary for a better understanding of lung pathophysiology and developing new therapies.

135

136

CHAPTER 4 Lung-on-a-chip platforms

In vitro and in vivo models of respiratory disease pathogenesis Building appropriate research models is fundamental to investigate the mechanisms of human lung development and pathophysiology. An ideal lung model should hierarchically reproduce morphological and functional aspects. At the cellular level, it should include various human cell types from lung tissue (alveolar and bronchial epithelial cells), vasculature (vascular endothelial cells), immune cells, and neural tissue (nerve cells) to recreate cell cell interactions and the interplay between cells and the surrounding three-dimensional (3D) microenvironment. At the tissue level, the physiological mechanisms should provide cyclic stretching to mimic the mechanical stimulation via respiration and the controlled flow of both air and blood typical of the respiratory airways (Bajaj et al., 2016; Miller and Spence, 2017). At the system level, the components of the respiratory tract (bronchial, alveolar, etc.) should be integrated for a complete mimicking of the respiratory apparatus. In vivo, ex vivo, and in vitro lung models will be described in this chapter, addressing the main advantages and drawbacks of each approach.

The in vivo approach: animal models of lung diseases Because of ethical considerations, the use of human models is severely restricted to minimal or no-detriment studies, such as analyses of the immune response (e.g., skin allergy test) or physical outputs such as breathing peak flow rate and heart rate (e.g., exercise), as well as to the study of tissue biopsies (Blume and Davies, 2013; Gordon et al., 2015). Experimental animal models have thus provided most of the current knowledge of lung pathophysiology and are the cornerstone of developing new drugs (Blume and Davies, 2013; Miller and Spence, 2017). Several animal species have been used as lung disease models of asthma, acute respiratory distress syndrome (ARDS), COPD, CF, tuberculosis, and lung cancer; these include small animal models (mice, rats, guinea pigs, hamsters), rabbits, dogs, sheep, and nonhuman primate models (NHPs) (Fonseca et al., 2017; Fricker et al., 2014; Han et al., 2018; Rosen et al., 2018) (Table 4.1). NHPs are considered the best model because their lungs mirror the anatomical, physiological, and immune features of the human lung (Fonseca et al., 2017; Miller et al., 2017), but costs and ethical issues strictly limit their use, especially for largescale experiments. Owing to economics and ease of scientific manipulation, small animals and specifically mice are the most commonly used model (Fonseca et al., 2017; Marque´s-Garcı´a and Marcos-Vadillo, 2016; Meurs et al., 2008). Mice models have primarily been “humanized” through grafting of human cells and tissues (Calderon et al., 2013; Ito et al., 2012) or transgenic, knockout, or knockin approaches (Pe´rez-Rial et al., 2015) to recreate particular human disease features. Once established, these humanized models are further developed to achieve the appropriate pathophysiological conditions or progression (Table 4.2).

Table 4.1 List of the main aspects achieved/recreated by means of animal (in vivo) models of the common lung pathologies. Pathology

Aspects achieved/mimicked using animal models

ARDS

• Analysis of the mechanisms of AFC reduction and pulmonary edema (Huppert and Matthay, 2017)

• Altered alveolar ion transport and decreased epithelium permeability causes AFC reduction in mice exposed to influenza virus (Chen et al., 2004)

• MSC therapy to decrease Escherichia coli pneumonia injury (Devaney et al., 2015; Gupta et al., 2012) Asthma

• Asthma development and exacerbation mechanisms induced by viruses



CF

• • •



• Lung cancer



• • COPD





• Tuberculosis







by means of sensitized animal models (Han et al., 2018; Mullane and Williams, 2014) The target of the T-helper type 2 cells responsible for driving allergic asthma mechanism (Holt et al., 1999; Zosky and Sly, 2007) Role of airway remodeling in chronic and acute AHR (Meurs et al., 2008) CFTR mouse models recreated the altered nasal epithelium typical of human CF (Lavelle et al., 2016; McCarron et al., 2018); CFTR 2 / 2 mutant pigs were used to model electrolyte transport defects and mimic the impaired bacteria clearance mechanisms (Chen et al., 2010; Ostedgaard et al., 2011; Stoltz et al., 2010); CFTR 2 / 2 ferrets were used to recreate mucus plugging and defective ion transport of the tracheal epithelium (cAMP-dependent chloride transport) (Fisher et al., 2013; Sun et al., 2010) Test of therapeutics [e.g., pH modification for acidic CF tissues (Alaiwa et al., 2016) and gene therapy (Cmielewski et al., 2014; Cooney et al., 2016)] Preclinical models for testing chemotherapeutics, drug combinations (Sandler et al., 2006), and chemopreventive therapies (Herzog et al., 1997; Wang et al., 2009; You and Bergman, 1998) Xenograft models as tool for personalized medicine and cancer recurrence prediction (Dong et al., 2010; John et al., 2011; Kellar et al., 2015) GEMMs to assess carcinogenesis, cancer prevention (Kellar), and metastatic mechanisms (Gazdar et al., 2016, 2015) Determine the role of immune cells (macrophages, neutrophils, NKs, T cells) by means of CS-induced models (Beckett et al., 2013; Dhami et al., 2000; Eppert et al., 2013; Fricker et al., 2014; Motz et al., 2010) Defining the involvement of proteins (chemokines, cytokines), enzymes (protease), and oxidative stress in COPD pathogenesis (Foronjy and D’Armiento, 2006; Fricker et al., 2014) Create models of bacterial and viral exacerbation (Gaschler et al., 2009; Pérez-Rial et al., 2015; Zhou et al., 2013) Models of latent TB, TB granuloma formation (guinea pigs, rabbits) (Fonseca et al., 2017; Gupta and Katoch, 2005; Orme and Basaraba, 2014), and necrosis (Kramnik and Beamer, 2016) Modeling HIV/TB coinfection to test new therapeutics for HIV prophylaxis and treatment (Diedrich and Flynn, 2011; Pawlowski et al., 2012) Models for testing anti-TB vaccines and drugs (Cardona and Williams, 2017; Zhan et al., 2017)

AFC, alveolar fluid clearance; AHR, airway hyperresponsiveness; ARDS, acute respiratory distress syndrome; CF, cystic fibrosis; CFTR, cystic fibrosis transmembrane conductance regulator; COPD, chronic obstructive pulmonary disease; CS, cigarette smoke; GEMM, genetically engineered mouse model; HIV, human immunodeficiency virus; MSC, mesenchymal stem (stromal) cell; NK, natural killer; TB, tuberculosis.

138

CHAPTER 4 Lung-on-a-chip platforms

Table 4.2 List of lung in vitro models and the main outcomes achieved using each of them. Model

Aspects achieved

2D cultures

• Study of ciliary structure and ciliary dysfunction (Pifferi et al.,

ALI cultures

• Used for clinical diagnostics in TB (acid fast smear assays) • Showing mucociliary differentiation of tracheal epithelial cells

2009; Rutland and Cole, 1980)

(Whitcutt et al., 1988)

• Study of pulmonary toxicity: air pollutants (Upadhyay and

• • Organoids



• • •

Biological (decellularized) scaffolds

• •

• •



Bioengineered scaffolds

• • •

Palmberg, 2018), cigarette smoke (Li, 2016; Thorne and Adamson, 2013), infectious agents (Miller and Spence, 2017) Epithelium and ECM remodeling in asthma (Benam et al., 2015) Cellular effect of CFTR mutation in CF (Miller and Spence, 2017) Progenitor and PSC cells can generate functional proximal and distal airway epithelium (Ghaedi et al., 2014; Gotoh et al., 2014; Konishi et al., 2016) Mimicry of the human fetal lung development and tumorigenesis (Kaisani et al., 2014; Miller et al., 2018) Screening of proteins responsible for goblet cells metaplasiaa in diseased lung (Danahay et al., 2015) Analysis of lung stem cells self-renewal mechanisms and interactions with the niche (Dye et al., 2016; Kretzschmar and Clevers, 2016; Zepp et al., 2017) Ventilated scaffolds enabled pneumocytes differentiation (Doryab et al., 2016; Inanlou and Kablar, 2005) Effective stem-cell differentiation compared to other matrixbased models (Cortiella et al., 2010; Hoganson et al., 2014; Price et al., 2010) Large-scale production of alveolar epithelium (Ghaedi et al., 2014) Chronic lung disease (COPD, IPF) acellular scaffolds as models of disease pathomechanics (Booth et al., 2012; Gilpin and Wagner, 2018; Sava et al., 2017) Use of airway scaffolds (trachea) for clinical transplants (Delaere et al., 2018; Laurance, 2010; Macchiarini et al., 2008) Formation of alveolar-like structures (Doryab et al., 2016; Douglas et al., 1976) Alveolar and vascular regeneration (Shigemura et al., 2006) Tissue-engineered trachea and bronchus (Nichols et al., 2017)

2D, Two-dimensional; ALI, air liquid interface; CF, cystic fibrosis; CFTR, cystic fibrosis transmembrane conductance regulator; COPD, chronic obstructive pulmonary disease; ECM, extracellular matrix; IPF, idiopathic pulmonary fibrosis; PSC, pluripotent stem; TB, tuberculosis. a Goblet cells metaplasia is responsible for mucus hypersecretion and airway obstruction, features typical of lung diseases such as COPD, chronic asthma, and CF (Nadkarni).

In vitro and in vivo models of respiratory disease pathogenesis

In vivo models of the main lung diseases Exposing mice to cigarette smoke for 3 6 months is a method of reproducing smoke-induced COPD hallmarks of airway remodeling, inflammation, and emphysema (Beckett et al., 2013; Fricker et al., 2014; Pe´rez-Rial et al., 2015). The cigarette smoke induced COPD mouse model is also used in combination with bacterial or viral infection to create models of COPD exacerbation and with hypoxia conditions or growth factor inhibitors to induce pulmonary hypertension and emphysema in severe COPD models. Mice sensitized to allergens such as ovalbumin, house dust mite, or Aspergillus fumigatus are commonly used to model asthma-related exacerbations and airway hyperresponsiveness, typical of chronic allergic asthma (Benam et al., 2015; Han et al., 2018; Meurs et al., 2008; Mullane and Williams, 2014). Murine models of influenza pneumonia have contributed to elucidating the mechanism of reduction of the alveolar fluid clearance typical of ARDS, showing the reduced activity of alveolar ion channels and the altered permeability of alveolar epithelium (Huppert and Matthay, 2017). Furthermore, mice infected by Mycobacterium tuberculosis have been used to study the mechanisms of tuberculosis dormancy (Alnimr, 2015), granuloma formation (Orme and Basaraba, 2014), and tuberculosis necrosis (Kramnik and Beamer, 2016). CFTR-knockout mice have been used as CF models, as well as under bacteria-challenged conditions, achieved by inoculating mice with Pseudomonas aeruginosa. However, this model has demonstrated limited validity because of an inability to spontaneously develop the pathological features of CF. The absence of characteristic lung inflammation and mucus plugs resulted in the development of more accurate CF ferret and pig models (Lavelle et al., 2016; McCarron et al., 2018; Rosen et al., 2018; Wang et al., 2014). Human tumor xenografts in mice are the most commonly used lung cancer models (Kellar et al., 2015; Ruggeri et al., 2014) in research and in preclinical pharmacology studies. Genetically engineered mouse models also represent a valuable tool in studying lung tumorigenesis and cancer progression. By mimicking the tumor microenvironment, these models can provide more reliable outcomes for drug toxicity tests (Gazdar et al., 2016, 2015).

Limitations of animal models Despite the contribution of animal models to understanding disease mechanisms, animals and humans are substantially different (Miller and Spence 2017): murine airways differ from human airways in size, airway branching patterns, and anatomy of upper airways (Hofmann et al., 1989; Jong and de Maina, 2010a; Miller and Spence, 2017). Furthermore, human and murine lung development occurs on different time scales (Snoeck, 2015) and via differing molecular pathways and self-renewal of progenitor cells (Nikoli´c et al., 2018). Mice and humans share several cell types, but these differ in location, distribution, quantity, morphology, and function, effectively creating gaps in recapitulating specific

139

140

CHAPTER 4 Lung-on-a-chip platforms

pathophysiological features (Mullane and Williams, 2014; Nichols et al., 2013). More critically, animals do not necessarily develop human pathologies such as asthma or tuberculosis (Fonseca et al., 2017; Zosky and Sly, 2007); thus in vivo models typically represent limited features of the disease. Moreover, respiratory and heartbeat rates are substantially higher in rodents, resulting in faster physiologic and metabolic processes, a difference that should be taken into account in pharmacokinetic studies (Jong and de Maina, 2010a; Perinel et al., 2017). Thus correlation of test results across species and translation to human clinical trials is still challenging and raises the issue of the validity of in vivo research (Fonseca et al., 2017). Because of these drawbacks, more than 80% of drugs that pass the preclinical animal testing stage in mice fail in clinical trials, with a substantial loss of time and resources invested, underscoring the need for less expensive, more representative, and higher throughput models (Miller and Spence, 2017; Mullane and Williams, 2014; Perrin, 2014).

Ex vivo approach The use of ex vivo models allows targeted studies of airway and parenchyma that cannot be ethically conducted in vivo while still providing a 3D tissue architecture (Blume and Davies, 2013; Gordon et al., 2015). Biopsy samples obtained from diseased and healthy (control) patients by bronchoscopy have been used to investigate the pathophysiological mechanisms and cellular immune responses to allergic inflammation in asthmatic patients and to investigate human tissue responses during therapeutic testing and development (Bhowmick and GappaFahlenkamp, 2016; Blume and Davies, 2013). Human precision-cut lung slices have been shown to be useful models for immune responses to allergies, new drug formulation, toxicology assays, and tumor studies (Constant et al., 2015; Gordon et al., 2015). However, these models suffer from short-term viability and lack of reproducibility due to the source variability; in addition, the barrier properties can be compromised by direct exposure to stimulating, challenge agents applied to the tissue pre- or postsample acquisition (Blume and Davies, 2013).

Use of in vitro models Several cell types can be used to produce in vitro models; these can be classified into three main categories: 1. Primary human airway epithelial cells: Primary cells from bronchial brushing and biopsies or from commercial suppliers mimic differentiated aspects of the epithelium, exhibiting the normal pulmonary phenotype under healthy conditions. Normal human bronchial epithelial and small-airway epithelial cells are available for disease and pharmacological studies. Furthermore, immortalization protocols are now available to extend the lifespan of primary

In vitro and in vivo models of respiratory disease pathogenesis

cells without losing the differentiation capability (Gordon et al., 2015). Primary cells also offer the potential for personalized medicine studies, but obtaining patient cells can be difficult. 2. Cell lines: Immortalized cell lines are obtained from tumors or by viral transfection and are commonly used in research because of their ease of culturing, the high throughput of proliferation, and the longevity of the culture, which effectively leads to lower costs than those of in vivo studies (Gordon et al., 2015). Human bronchial epithelial cells, such as the adenocarcinoma line Calu-3, have been used to mimic central airways in transport studies, model barrier properties, and recreate disease and infection models of the bronchial epithelium. Peripheral airway models have been recreated mainly using the A549 alveolar adenocarcinoma cell line as a model of alveolar type II cells for drug formulation, transfection, and infection studies (Bhowmick and Gappa-Fahlenkamp, 2016; Gordon et al., 2015). The A549 line has also been used in studying non small cell lung carcinoma, which represents 80% 85% of lung cancer cases (Gazdar et al., 2010). 3. Stem cells: The controlled differentiation of human embryonic stem cells or induced pluripotent stem cells represents a powerful tool with the potential for large-scale production (Gordon et al., 2015). The process commonly used is directed differentiation (Dye et al., 2016a,b), which consists of providing biological cues to cells to drive their differentiation into specific cell types, but there is still a lack of standardized protocols for cell production and characterization. Directed differentiation enables the use of cells from patients genetically predisposed to pathologies and genetic modification of humaninduced pluripotent stem cells (Constant et al., 2015; Dye et al., 2016a,b). In vitro models range in their complexity and accuracy in recreating various aspects of the lungs. The simplest model is the two-dimensional (2D) monolayer media-submerged model, in which cells of the same type are cultured on a dish or a membrane and exposed to media on both the apical and basal sides (Bhowmick and Gappa-Fahlenkamp, 2016; Miller and Spence, 2017). The model is used in clinical diagnostics of ciliary dysfunction and tuberculosis (Miller et al., 2017; Rosenfeld et al., 2014). The lack of an in vivo like architecture in 2D in vitro models causes cellular dedifferentiation, and protein expression and cellular response only partially capture the underlying physiological mechanisms (Bhowmick and GappaFahlenkamp, 2016). 3D in vitro models reproduce a microenvironment that can mimic in vivo architecture and functional features. Cells seeded into 3D models have been shown to create functional tight junctions, with the formation of basement membranes, polarization, and a phenotype resembling that of in vivo conditions (Bhowmick and Gappa-Fahlenkamp, 2016; Miller and Spence, 2017). These platforms include air liquid interface (ALI) 3D organoids, tissue-engineered constructs, and LOC systems. Unlike the submerged model, the ALI model aims to mimic the airway microenvironment by seeding epithelial cells on a permeable

141

142

CHAPTER 4 Lung-on-a-chip platforms

membrane of a Transwell insert, with the apical surface in contact with the atmosphere (air) and the basal side exposed to culture medium (liquid) (Bhowmick and Gappa-Fahlenkamp, 2016; Miller and Spence, 2017; Whitcutt et al., 1988). ALI systems enable the evaluation of barrier properties and have recreated mucus production and epithelial cell responses under both health and pathology conditions (Benam et al., 2015; Gordon et al., 2015). Furthermore, in vivo like cell cell interaction can be recreated by using multicellular models, in which different cell types are cocultured to study their interplay under healthy and inflammatory conditions (Gordon et al., 2015; Upadhyay and Palmberg, 2018). Several commercial ALI models are available: CULTEX (Hannover, Germany) and VITROCELL (Waldkirch, Germany) are in vitro aerosol systems used to investigate cellular responses to particles and gases (Bhowmick and Gappa-Fahlenkamp, 2016; Upadhyay and Palmberg, 2018). EpiAirway (MatTek, Ashland, MA, United States) and MucilAir (Epithelix, Geneva, Switzerland) are ready-to-use ALI systems composed of primary cells from donors with different pathologies (e.g., asthma, COPD, CF) that can reproduce various anatomical sites (e.g. trachea, bronchi). The coculture of human fibroblasts enables the study of complex inflammatory mechanisms, such as in the fibrotic lung (EpiAirwayFT, MucilAir-HF). These commercial in vitro platforms can recreate in vivo barrier features, including cellular morphology, mucus production, and tight junctions, providing a sufficiently complex model to evaluate airway disease mechanisms and perform toxicology, drug delivery, and infection studies. Long-term and chronic studies have also been possible because of the long shelf life of these systems (Bhowmick and Gappa-Fahlenkamp, 2016; Gordon et al., 2015). OncoCilAir, a 3D in vitro lung cancer model, has been used in preclinical studies of anticancer therapies to investigate tumor stroma interactions (Constant et al., 2015; Mas et al., 2016). 3D organoids are self-assembled aggregates of various cell types grown in extracellular matrix-like gel (e.g., Matrigel) (Bhowmick and Gappa-Fahlenkamp, 2016; Miller and Spence, 2017). Organoids can be formed from primary stem cells or from pluripotent stem cells and can be assembled as tracheospheres, bronchospheres, or alveolospheres to recapitulate the features of lung compartments (Barkauskas et al., 2017; Gkatzis et al., 2018). Their ability to mimic cellular spatial organization and recreate specific organ functions have defined organoids as innovative platforms for modeling lung development, lung disease, and lung cancer (Nadkarni et al., 2016). Organoids have been used to clarify the role of growth factors and signaling pathways in branching morphogenesis (Gkatzis et al., 2018; Nadkarni et al., 2016) and that of genetic alterations and malignant transformation in lung tumorigenesis. Intestinal organoids have been developed successfully to evaluate CFTR mutations and drug formulations in CF treatment (Dekkers et al., 2013). Organoids from stem cells have been used to study pulmonary viral infections (Chen et al., 2017) and to elucidate the complex early-stage mechanisms typical of infectious diseases such as tuberculosis (Bielecka and Elkington, 2018; Fonseca et al., 2017).

Current lung-on-a-chip systems

Lung tissue engineering enables the development of lung substitutes for transplantation in patients with end-stage lung diseases and is a platform for drug screening and potentially personalized therapeutics (Doryab et al., 2016; Hoganson et al., 2014; Langer and Vacanti, 2016). Engineered scaffolds can integrate both vascular and airway structures into a complex 3D lung construct, reproducing the necessary gas exchange (Hoganson et al., 2014). Two parallel approaches are used: the first consists of engineering synthetic substrates for cell growth and differentiation using either biopolymers (e.g., collagen, Matrigel, gelatin) or biocompatible polymers [e.g., poly(DL-lactic acid), polyglycolic acid, poly2-hydroxyethyl methacrylate] and blends. The second involves the fabrication of decellularized biological scaffolds from which the cellular components of the organ or tissues are removed using detergents while preserving the anatomical structure and extracellular matrix (Doryab et al., 2016). The device is then recellularized with specific cell types, eventually with the aid of bioreactors (Doryab et al., 2016; Langer and Vacanti, 2016). This method demonstrates the feasibility of recreating functional lung constructs and is potentially applicable for clinical use (Nichols et al., 2017). While a core goal of lung tissue engineering is the ability to replace or restore functional lung tissue, scaffolds are a valuable platform for evaluating lung development, studying disease, and testing therapeutics. Studies using decellularized scaffolds have shown how extracellular matrix from diseased and aged donors influences cell attachment, proliferation, and survival and clarified the role that structural remodeling of the extracellular matrix plays in pathology initiation and progression (Gilpin et al., 2017; Tjin et al., 2017; Wagner et al., 2014). Despite their advantages over traditional culture models, organoids and lung tissue engineering approaches have some drawbacks, key among which is the difficulty to spatially pattern and organize different cell types, the controlled distribution of biochemical molecules, or the defined mechanical stimulation across the 3D tissue. The LOC technology overcomes these challenges using microfluidic devices that can mimic tissue stretching, physiological flow, and biochemical stimuli (Bajaj et al., 2016; Seo and Huh, 2019).

Current lung-on-a-chip systems The respiratory system exhibits dynamic behavior of cyclic motion physical stresses, as the lung is fundamentally a mechanical organ. With each breath, the lungs undergo a considerable change in volume and an approximate change in tissue length of 4% 25% (Gump et al., 2001). The resulting macroscale stress constitutes only one part of the complex physical forces (of varying magnitude, direction, and frequency) transmitted down to the microscale in the alveoli, where surface tension and fluid shear stresses from blood and interstitial flow begin to dominate (Fredberg and Kamm, 2006; Guenat and Berthiaume, 2018).

143

144

CHAPTER 4 Lung-on-a-chip platforms

While 3D lung cell culture platforms are significant improvements over their 2D counterparts, they still largely fail to take into account the mechanical cues that are vital to lung development and function, and consequently are inadequate lung analogs for disease modeling (Mammoto et al., 2013). Leveraging the advantages of microfluidics, OOC platforms have enabled researchers to develop more biomimetic systems that recreate both the biochemical and mechanical aspects of cellular microenvironments in human organ systems. A wide variety of organ functions can thus be mimicked. The versatility of microfluidic systems can be leveraged to model a wide variety of physiological lung conditions, in both healthy and diseased states, at various levels within the hierarchical lung structure. Moreover, by taking a modular approach, individual microfluidic units can be combined to study the complex interactions between the respiratory system’s distinct functional zones. Skolimowski et al. (2012a,b) demonstrated this concept with interconnected microfluidic compartments of media at varying oxygen levels to model the anaerobic sinus environment, the microaerobic tracheal and bronchial zone, and the highly aerobic alveolar zone (Fig. 4.2). These compartments were connected with a series of micropumps and valves, which enabled the study of antibiotic efficacy with changes in oxygen tension on bacteria associated with mortality in CF. This system serves as an example of the new capabilities that microfluidics offer to drive discoveries in lung pathology.

FIGURE 4.2 The modular microfluidic airway model. Each of the interconnected compartments is maintained at varying levels of oxygen concentration, corresponding to different zones throughout the human airway. Reprinted from Skolimowski, M., Weiss Nielsen, M., Abeille, F., et al., 2012. Modular microfluidic system as a model of cystic fibrosis airways. Biomicrofluidics 6 (3), 1 11, with the permission of AIP Publishing.

Current lung-on-a-chip systems

FIGURE 4.3 Mechanically active lung-on-a-chip device. The three-layer PDMS device consists of two central chambers separated by a thin porous membrane seeded with alveolar epithelial cells on one side and vascular endothelial cells on the other. An air liquid interface is established by flowing air in the upper, alveolar channel. As illustrated, vacuum can be applied to the two-side chambers in order to deform the elastic PDMS walls, and thus the membrane. This serves to unidirectionally stretch the cells and simulates the stretching that is undergone during the expansion and contraction of the alveoli in the human body. PDM, Spolydimethylsiloxane. From Huh, D., Matthews, B.D., Mammoto, A., Montoya-Zavala, M., Hsin, H.Y., Ingber, D.E., 2010. Reconstituting organ-level lung functions on a chip. Science 328, 1662 1668. doi:10.1126/ science.1188302. Reprinted with permission from AAAS.

In this section, we discuss from a structural and microfluidic standpoint, innovative and noteworthy LOC systems and how their designs mimic a variety of lung functions and conditions. This chapter considers OOC devices to be systems that can recreate organ-level function and response, most often via tissue tissue interfaces, unlike simpler, gel-based 3D cell cultures or single-tissue cultures.

Mechanically active alveolar capillary interface The pioneering work by Huh et al. (2010) in simulating the mechanical stresses on lung cells was the first to mimic the cyclic stretching of the alveolar membrane at an ALI in a microfluidic device. The multicompartmental polydimethylsiloxane (PDMS) device (Fig. 4.3) consisted of two primary media channels separated by a thin, extracellular matrix-coated porous membrane, one side of which was seeded with alveolar epithelial cells and the other with microvascular endothelial cells. After cells reached confluence, an ALI was established by flowing air through the epithelial, or alveolar, channel and cell culture media through the endothelial, or vascular, channel, recreating the ALI found at the alveolar capillary interface. To simulate the cyclic strain that this interface experiences in vivo, the group used an innovative actuation method employing lateral

145

146

CHAPTER 4 Lung-on-a-chip platforms

microfluidic channels, through which vacuum could be applied and subsequent unidirectional stretching of the elastomeric membrane could be achieved. With this setup, endothelial cell alignment responses mimicked those of actual endothelium and in vivo blood vessels (Iba and Sumpio, 1991; Thodeti et al., 2009), differing from cell behavior in 3D culture systems that do not incorporate an ALI or mechanical actuation (Pampaloni et al., 2007). The platform was then evaluated from an immunological standpoint to verify that the addition of cyclic mechanical stresses in an in vitro lung model could produce more complex and biomimetic organ-level responses. Pulmonary inflammation response was first demonstrated using a medium containing the proinflammatory cytokine tumor necrosis factor α (TNFα) into the vascular channel of the microfluidic device. The endothelial expression of intercellular adhesion molecule 1 (ICAM-1) and subsequent adhesion and transmembrane migration of fluorescently labeled neutrophils were monitored as a measure of the system’s inflammatory response. The inclusion of mechanical strain on the alveolar membrane produced no changes in immune response, but this was not the case when nanoparticles were used as the stimulant. Silica nanoparticles, often used in airborne particle toxicity studies (Lin et al., 2006; Napierska et al., 2009), were deposited as a thin liquid layer onto the epithelial side of the membrane and their absorption across the membrane increased in the presence of physiologically similar levels of membrane strain. The immune response, determined again by ICAM-1 expression, also increased. This response was similar to that observed in murine lungs exposed to silica nanoparticles and consistent with in vivo evidence, suggesting nanoparticle toxicity attributable to cross-membrane absorption (Nel et al., 2006). This is in contrast to the lower levels of absorption shown in static in vitro Transwell-based systems or the same PDMS device used without mechanical actuation. This biomimetic platform was the first demonstration of a mechanically active system generating more realistic in vitro lung environments and subsequent immune responses than conventional static systems and has since become a foundation in microfluidic configuration for further development of LOC devices. Some of these are discussed in detail in Organ-on-a-chip systems for modeling pathological conditions section.

Mimicking the pulmonary parenchymal environment Building on the microengineered device that Huh et al. pioneered, Stucki et al. aimed to reproduce an even more biomimetic alveolar strain regime on a chip (Stucki et al., 2015). Where many systems recreate strain in either a linear (Huh et al., 2012) or 2D manner (Tschumperlin and Margulies, 1998; Vlahakis et al., 1999), an alveolar sac expands and contracts in vivo in three dimensions, not unlike a balloon. While there is still much to be learned about the mechanobiology of the lung (Waters et al., 2012), it follows that the strain applied on in vitro systems should aim to closely replicate the 3D stretching occurring in vivo, and

Current lung-on-a-chip systems

FIGURE 4.4 (A) LOC system incorporating an artificial alveolar membrane (a) onto which epithelial and endothelial cells can be seeded, and that can be three-dimensionally stretched with mechanical actuation of the diaphragm-inspired lower membrane (b) by an electropneumatic pump. (B) Schematic of the device separated into a fluid part and a pneumatic part. The fluid part contains three cell culture wells (i) above porous, flexible membranes (ii). The basolateral chambers (iii), to be filled with cell culture media, can be found directly beneath the membranes. The pneumatic component features a microdiaphragm (iv) at the base of each of the basolateral chambers and directly adjacent to pneumatic microchannels (v) in order to apply the vacuum that causes their deformation. (C) Photograph of the device with colored fluid inserted into the basolateral channels for visualization (scale bar: 10 mm). LOC, Lung-on-a-chip. Reprinted (adapted) with permission from Stucki, A.O., Stucki, J.D., Hall, S.R.R., Felder, M., Mermoud, Y., Schmid, R.A., Geiser, T., Guenat, O.T., 2015. A lung-on-a-chip array with an integrated bio-inspired respiration mechanism. Lab Chip 15, 1302 1310. doi:10.1039/C4LC01252F

Published by The Royal Society of Chemistry.

studies have indeed shown that strain profiles affect cell responses (Berry et al., 2003; Deng et al., 2009; Gould et al., 2012; Park et al., 2004). This device (Fig. 4.4) employs two PDMS membranes to achieve 3D stretching. The first is a porous membrane that acts as an alveolar barrier, similar to the mechanically active alveolar capillary interface device described above, with epithelial and endothelial cells seeded onto the apical and basal sides of the membrane, respectively. The second membrane is situated below the cellularized membrane and a media compartment and serves as a diaphragm-inspired mechanical actuator. Supplying vacuum below this microdiaphragm deforms it, transmitting a negative pressure through the incompressible cell culture medium onto the artificial alveolar membrane. The resulting membrane deformation thus mimics the 3D stretching observed in the alveoli and can be precisely tuned to specific mechanical inputs such as strain and frequency. Another important objective of the work of Stucki et al. was to address the constraints of robustness and reproducibility of LOC-type devices, considerations often deemed vital to the eventual adoption of OOC platforms on a larger scale

147

148

CHAPTER 4 Lung-on-a-chip platforms

(Esch et al., 2015). From a biological standpoint, one issue of reproducibility arises during the cell-seeding stage of an OOC, where the limited control over cells, once they are within the microfluidic device, can often result in imprecise or nonhomogeneous cell distributions. Unlike most LOC devices, which cannot be disassembled after fabrication nondestructively, the LOC developed by Stucki et al. was composed of two principal parts that could be reversibly bonded to one another with sufficient applied pressure, allowing the device to be readily opened and closed more than once. Thus when the device was disassembled, both sides of the artificial alveolar membrane could be accessed, facilitating straightforward and controlled cell seeding via conventional pipetting methods. This approach improves LOC ease of use and cellular reproducibility and represents an important step toward wider scale use of LOC devices.

Simulating surface tension stresses One aspect of simulating the lung microenvironment that has been only briefly discussed is recreating the in vivo fluid mechanical stresses (as opposed to solid mechanical stresses that result from cyclic stretching, of which the previous two systems have been exemplars). Fluid stresses, or more specifically surface tension-related fluid stresses, are critical role lung diseases such as ventilatorinduced lung injury, ARDS, and neonatal respiratory distress syndrome, where the propagation of pathological air liquid menisci with each expansion of the alveoli imposes a variety of stresses on the underlying epithelial cells (Bilek et al., 2003). The difficulty in recreating the complex interplay of both the solid and fluid mechanical stresses is reflected in the fact that previous models investigating these types of lung diseases simulated either the mechanical stretching (Tschumperlin et al., 2000) or the surface tension-related effects of air liquid plugs (Huh et al., 2007) in isolation, but not simultaneously. Douville et al. were the first to bridge this gap with a LOC that allowed the study of the combined effects of the solid and fluid stresses observed in surface tension-related diseases (Douville et al., 2011). The PDMS device consisted of two chambers (termed alveolar and actuation chambers) separated by a thin membrane (Fig. 4.5). The upper (alveolar) chamber was filled with cell culture medium, so that human alveolar basal epithelial cells could be seeded onto the upper side of the membrane, and the lower (actuation) chamber was put under vacuum to stretch the membrane in a 3D manner, similar to the membrane deformation achieved by Stucki et al. After epithelial cell confluence was reached in the device’s horizontal configuration, the device could be positioned vertically to orient the liquid meniscus perpendicular to the cell layer. Upon stretching of the membrane, the meniscus could be propagated along the cell layer, exposing it to the fluid mechanical stresses from surface tension effects. With careful tuning of the vacuum applied to the actuation channel, the propagation speed and frequency can be controlled to model a range of stress scenarios. This combinatorial approach resulted in cell morphology and death that were significantly different

Current lung-on-a-chip systems

FIGURE 4.5 (A) “Alveoli-on-chip” microfluidic device composed of three principal parts: an upper piece with a semiopen cell culture chamber; a thin, porous membrane onto which epithelial cells are seeded; and a lower piece with the actuation chamber, to which vacuum can be applied to deform the membrane. Cell culturing is done with the device in its horizontal orientation (B), then it is rotated into its vertical orientation for study with a combination of solid and fluid mechanical stresses. (D) Cross-section schematic illustrating the device’s two orientations. In its vertical orientation deformation of the membrane leads to a propagation of the air liquid interface, exposing the adjacent cells to surface tension-related stresses. Reprinted with permission from Douville, N.J., Zamankhan, P., Tung, Y.C., Li, R., Vaughan, B.L., Tai, C.F., et al., 2011. Combination of fluid and solid mechanical stresses contribute to cell death and detachment in a microfluidic alveolar model. Lab Chip 11, 609 619, reproduced by permission of The Royal Society of Chemistry (http://www.rsc.org/).

149

150

CHAPTER 4 Lung-on-a-chip platforms

FIGURE 4.6 Organotypic LOC device featuring two media-filled endothelial-lined lumens and one central epithelial-lined lumens, where an ALI is present. The lumens run through a collagen and fibrinogen gel matrix, incorporating pulmonary fibroblasts. ALI, Air liquid interface; LOC, lung-on-a-chip. Reprinted (adapted) with permission from Barkal, L. J., et al., 2017. Microbial volatile communication in human organotypic lung models. Nat. Commun. 8 (1).

than when the system was used to model cyclic membrane stretching. The findings were in agreement with hypotheses drawn from clinical studies (Chu et al., 1967; Hirschl et al., 1998, 1996) that suggested that surface tension forces play a critical role in pathologies such as neonatal respiratory distress syndrome and ARDS and that modeling cyclic stretching alone is insufficient.

Complex organotypic cocultures More recently, Barkal et al. (2017) introduced an organotypic LOC model of the terminal bronchiole, whose biological complexity and innovative design features represent a step toward mimicking in vivo biology and a more user-friendly and versatile platform. The system (Fig. 4.6) is based on a fibroblast and collagen gel matrix structure that supports epithelial and endothelial cell monolayers. This type of matrix is rarely integrated into in vitro lung platforms, where most often a microporous polymer membrane is used without a fibroblast-integrated support matrix, whether in Transwell-based systems (Yamaya et al., 2002) or more recent LOC devices (Benam et al., 2016a,b). During the fabrication process, three PDMS rods are encased in the hydrogel matrix and, once removed, leave behind channels with circular cross-sections. These three parallel channels, one to model the bronchial airway and two to model lateral vascular capillaries, offer a biologically relevant geometry onto which epithelial and endothelial cells can be

Organ-on-a-chip systems for modeling pathological conditions

cultured and result in distinctly different cell behavior that contrasts with the traditionally flat cell monolayers used in other in vitro models (Bischel et al., 2014). Furthermore, the ability to control the shape and size of the lumina with precision allowed the team to create lumina that matched the average dimensions of the terminal bronchioles and neighboring vascular capillaries in the human body (Anderson and Foraker, 1962; Hansen and Ampaya, 1975). This microscale design and fabrication approach, along with the use of primary human cells, can create a truly organotypic device to mimic the complex immunoinflammatory environment present in the lung. Innovation in the physical design of a LOC chip device must also be matched by careful consideration of how it can be used simply and effectively to create a robust experimental platform. In the LOC of Barkal et al., each lumen can be accessed through pipette-compatible ports, allowing easy and specific insertion and removal of material for chemical analysis. This enables evaluation under two distinct pathological conditions: First, pathogens could be introduced into the bronchial lumen to investigate lung response to direct pathogen contact. Second, a complementary module was designed to facilitate host pathogen communication via volatile compounds, which has been shown to produce unique host infection responses (Briard et al., 2016; Koo et al., 2014). The external microbial culture module could be positioned over the LOC in a sealed dish and facilitate bacteria and fungi-derived volatile compound interactions with the bronchiole model, a largely unexplored phenomenon that has particular relevance in CF patients colonized by multiple microbial species (Amin et al., 2010). This type of modular approach will likely prove to be indispensable in expanding OOC capabilities and helps this device stand out as a comprehensive LOC.

Organ-on-a-chip systems for modeling pathological conditions LOC technology has enabled in vitro studies of lung pathology under dynamic, more in vivo like conditions. Microfluidic size scales aid the mimicking of ALI phenomena, rendering them ideal platforms for small-airway pathophysiological models. This technology is expected to contribute significantly to understanding the mechanisms, causes, and symptoms of pulmonary diseases and developing treatments. In this section, we organize systems by components of the airway hierarchy, primarily focusing on the ALI of small airways and alveoli. The majority of these systems use a cellularized membrane, similar to the work done by the Ingber group (Huh et al., 2010). However, we first highlight the precursor microfluidic systems that provided the framework for studying on-chip lung pathophysiologies.

151

152

CHAPTER 4 Lung-on-a-chip platforms

FIGURE 4.7 A microfluidic device recapitulating the formation of liquid plug formation and propagation in the small airways. Reprinted (adapted) from Huh, D., Fujioka, H., Tung, Y.-C., Futai, N., Paine, R., Grotberg, J.B., et al., 2007. Acoustically detectable cellular-level lung injury induced by fluid mechanical stresses in microfluidic airway systems. PNAS 104 (48), 18886 18891. © 2007 National Academy of Sciences, U.S.A.

Microfluidic precursors as lung pathology models: liquid plugs in small airways The Takayama group (Huh et al., 2007) (Fig. 4.7) was among the first to demonstrate an on-chip small-airway pathological condition, by replicating the formation and propagation of mucus plugs in the airway lumen. These plugs are similar to those that form in vivo and are associated with pulmonary diseases and conditions such as COPD, asthma, pulmonary edema, and bronchiolitis (Cassidy et al., 1999). Liquid plugs form abnormally following dysregulation of surfactant levels or a dysfunction of the surfactant itself, resulting in a viscous film coating of the small-airway epithelium that increases the probability of air liquid instabilities and plug generation. The plugs block the airways, inhibiting gas exchange and air flow. During inhalation, as the lung inflates, the liquid plug is driven further down along the airway lumen, until it eventually ruptures and reopens the airway. This rupturing causes a breathing sound termed crackles (Piirila¨ and Sovija¨rvi, 1995). Listening for crackles is a clinical detection method for associated pulmonary diseases. This microfluidic model showed that the mechanical forces resulting from plug rupturing cause damage to airway epithelial cells, and the on-chip plug rupture produced sounds to the clinical phenomenon. Of note, the formation of plugs is also observed during artificial ventilation of patients undergoing medical interventions, and this model system could recreate plug formation. The model was subsequently leveraged to systematically produce and monitor on-chip liquid

Organ-on-a-chip systems for modeling pathological conditions

plugs for the focused evaluation of plug characteristics and airway pressure changes. This was accomplished by integrating a liquid plug generator to introduce a controlled air liquid segmented flow in the airway lumen of the chip (Tavana et al., 2010). Furthermore, the addition of pulmonary surfactants to reduce surface tension in the liquid plugs resulted in less damage to the epithelial cells (Tavana et al., 2011). The application of surface phenomena captured epithelial responses and highlighted microfluidics as an important technology platform for LOC modeling.

Modeling lung inflammation, asthma, and chronic obstructive pulmonary disease in small-airway chips A more recent small-airway-on-a-chip device was developed to model human lung inflammatory diseases and identify new antiinflammatory therapies (Benam et al., 2016a,b) (Fig. 4.8). This system mimicked the effect of epithelial endothelial crosstalk on lung inflammation in a microfluidic device, as a platform for obtaining organ-level responses to pathological processes. A functional, healthy small airway was reconstructed on a microfluidic chip containing human airway epithelial cells and endothelial cells, using similar device architecture. Permeability measurements confirmed the barrier integrity of the cellularized membrane, while immunofluorescence confocal microscopy identified the mixed tissue composition (ciliated, goblet, club, and basal cells, representative of a healthy small-airway epithelium). The ciliary beating frequency on the apical surface of the endothelial cells was similar to that observed in healthy human cilia in vivo. An asthma-like state was induced in the chip by perfusing the cell culture medium with interleukin 13, a mediator of allergic asthma, which increased the number of goblet cells and secretion of proinflammatory cytokines in the vascular compartment. A decrease in ciliary beating frequency was also observed, mirroring the ciliary beating behavior in asthma patients (Thomas et al., 2010). This airway chip was also used to model inflammatory responses to pathogenic infections in the small airways that trigger exacerbation of asthma and COPD using a viral mimic, polycytidylic acid. This immunostimulant was perfused through the chip to induce a severe asthma exacerbation-like state, resulting in the secretion of several proinflammatory cytokines. Interestingly, a monoculture of epithelial cells seeded onto the chip under the same culture conditions and the same stimulus exhibited significantly less secretion of cytokines than the epithelial endothelial coculture. The additive effect on cytokine secretion in the cocultured LOC suggests a synergistic inflammatory dynamic between the endothelial and epithelial layers. The system was able to detect cytokine expression as a pathophysiological outcome and to recreate neutrophil recruitment when flowing neutrophils through the lower, “vascular” channel. This mirrors the initial adhesion and rolling of neutrophils underflow in microvessels when circulating

153

154

CHAPTER 4 Lung-on-a-chip platforms

FIGURE 4.8 A functional small airway was reconstructed on a microfluidic chip, containing differentiated human cells to model human lung inflammatory diseases, expressing actively beating cilia function. Reprinted by permission from Benam, K.H., Villenave, R., Lucchesi, C., Varone, A., Hubeau, C., Lee, H.-H., et al., 2016a. Small airway-on-a-chip enables analysis of human lung inflammation and drug responses in vitro. Nat. Methods 13, 151 157.

through inflamed living tissues (Lawrence and Springer, 1993). Proinflammatory responses were also produced in devices with epithelial cells harvested from COPD patients. A similar immunostimulation regime was performed with either polycytidylic acid or the bacterial-derived stimulant lipopolysaccharide endotoxin (LPS) to model the inflammatory effects of COPD exacerbation. Clinical disease features of COPD exacerbation, specifically upregulation of the proinflammatory cytokine macrophage colony stimulating factor secretion, were recaptured in chips seeded with cells from COPD patients but not in from healthy controls.

Organ-on-a-chip systems for modeling pathological conditions

The ability to model such functional responses suggested that the model could be used for drug discovery, and this hypothesis was confirmed by assessing whether drugs could reverse the interleukin 13-induced phenotype. A Janus kinase inhibitor restored the epithelium to normal conditions, suggesting that the drug could be used as an antiinflammatory and that the microfluidic model could be leveraged to evaluate therapeutic responses. This work established a comprehensive template for modeling on-chip lung inflammation, but also highlighted the utility of LOC devices for running parallel studies while modulating the cultured cells or stimulation regimes. This ability to parallelize on a relatively small footprint extends the functionality of air liquid in vitro culture platforms while adding control of flow and environmental elements, bridging the gap with the more dynamic environment of the lung. Cigarette smoke is known to cause damage to multiple organs in the body, especially the lungs (United States Surgeon General, 2014), but electronic cigarettes and other tobacco products have been less characterized. The effective modeling of lung inflammation in the airway-on-a-chip system was leveraged to study smoke-induced pathophysiology in vitro by connecting the device to a smoking machine (Benam et al., 2016a,b). The small-airway microdevice was seeded with human epithelium cells from either COPD patients or healthy individuals and exposed to smoke from various tobacco products. A microrespirator was integrated into the device to apply smoke directly into the airway of the chip, mimicking realistic inhalation of smoke by cyclically “breathing” in and out of the chip with representative frequency and volumes. Unlike previous in vitro models where the cells were submerged in medium containing diluted smoke extract, effectively skewing the impact of smoking conditions, this model exposed the cells to whole smoke. Key smoke-generated phenotypes of the epithelium, such as increased oxidative stress, were reproduced under the wholesmoke regimes. Furthermore, as the cells were exposed to high levels of smoke (nine cigarettes over a single 75-minute period), they exhibited certain features seen in chronic smokers, such as expression of most oxidation reduction genes. By implementing a pre-conditioned smoking regime, this model can rapidly create phenotypic disease features of chronic smokers, paving the way for future studies of cellular alterations associated with chronic smoking. A major advantage of this model is the ability to run side-by-side comparisons between cells from the same COPD patient, enabling a true-match comparison. This comparison, coupled with single-cell resolution, high-speed imaging, and subsequent statistical analysis of the apical ciliary beating, revealed that smoking had a heterogeneous impact on ciliary beating patterns: beating frequency was only reduced over some areas of the epithelial layer. In contrast, this true-match comparison did not find a significant alteration following exposure to electronic cigarettes. Overall, this demonstrates how the model can be used to account for interindividual differences. The limitation of this study was that the model was not used at its full potential, as only epithelial cells were exposed instead of an epithelial/endothelial coculture.

155

156

CHAPTER 4 Lung-on-a-chip platforms

Pulmonary edema and intravascular thrombosis modeled in alveolus-on-a-chip devices The mechanical activation of lung cells in LOC systems was most notably shown by Huh et al. to model the alveolar capillary interface (Huh et al., 2010). The coupling of mechanical forces generated from a repetitive breathing motion and the device architecture mimicked a lung epithelium-like barrier in which the more instructive disease process of inflammation could be evaluated. Building on this “breathing” LOC system modeled interleukin 2 induced pulmonary edema (Huh et al., 2012). Pulmonary edema is a complex condition in which vascular fluid accumulates in the airspace of the alveoli as a result of leakage from the pulmonary vessels (Miles, 1977). A less common form of pulmonary edema can develop as a chemotherapeutic side effect, when interleukin 2 is administered to patients with melanoma or renal cancers. Cells in the LOC device were mechanically stimulated by cyclic stretching to produce the appropriate barrier and physico-similar properties. Once achieved, clinical doses of interleukin 2 were introduced into the vascular channel, provoking responses similar to those observed in patients (Conant et al., 1989), including vascular leakage and fibrin deposition. Separately, the Ingber group stimulated inflammatory mechanisms and platelet activation in the pulmonary vascular endothelium to model thrombus formation within pulmonary vessels (Jain et al., 2018) using a two-chamber LOC platform similar to the classic model (Fig. 4.9) (Huh et al., 2010). In this modified version, the upper (air) chamber was seeded with primary human lung alveolar

FIGURE 4.9 Intravascular thrombosis modeled in alveolus-on-chip. (A) A structural sketch showing the ALI of the alveolus that was modeled in the device. (B) A principal sketch of the device architecture. (C) Illustration of the cellular compartments of the device showing the epithelial cells in the air compartment and the endothelial cell lining the walls of the liquid compartment, forming the “vascular tube.” ALI, Air liquid interface. Reprinted from Jain, A., Barrile, R., van der Meer, A., Mammoto, A., Mammoto, T., De Ceunynck, K., et al., 2018. Primary human lung alveolus-on-a-chip model of intravascular thrombosis for assessment of therapeutics. Clin. Pharmacol. Ther. 103, 332 340, with the permission of John Wiley and Sons.

Organ-on-a-chip systems for modeling pathological conditions

epithelial cells. A principal difference from the previous model was that the lower (vascular) compartment was lined on all four extracellular matrix-coated walls with human vascular endothelial cells to form a vascular tube, instead of being limited to the basal surface of the ALI (Jain et al., 2018). This enabled perfusion of whole blood through the vascular tube, mimicking healthy human microvessels, without spontaneous platelet aggregation and thrombus formation (Furie and Furie, 2008). The ability to perfuse whole blood in this model is an important step toward a more individualized assessment of disease features or drug responses using lung cells and blood from the same patient. This model system demonstrated three important aspects of disease modeling in OOC technology: in vivo like responses of intravascular thrombosis, organ-level contributions to inflammation-induced thrombosis, and evaluation of potential antithrombotic therapeutics. First, the model captured a prothrombotic disease state by stimulating the endothelialized channel with perfusion of the inflammatory cytokine TNFα. The endothelium responded in a concentration-dependent manner, with increasing TNFα levels corresponding to more inflammation in the endothelium. Concomitantly, increased platelet aggregation and subsequent thrombus formation occurred. Implementing a synthetic biology approach, this system decoupled the endothelial/epithelial cellular interaction, enabling the analysis of the independent and collective responses of the lung epithelium and endothelium to determine that the prothrombotic effects were escalated as a function of the epithelial endothelial coculture. LPS, a bacterial endotoxin known to induce pulmonary thrombosis, was introduced into the endothelial tube to study thrombus formation in a monoculture of endothelial cells and in an endothelial epithelial coculture. Surprisingly, LPS did not induce any thrombus formation when the endothelium was stimulated in the absence of the epithelium, but the same treatment in the epithelial endothelial coculture increased barrier permeability and platelet aggregation in the vessel. As these responses were only observed in the device with a more complete cellular milieu, it appeared that LPS-induced thrombosis occurred through indirect stimulation of the epithelium. These findings demonstrate one of the major advantages of OOC devices, as a similar study cannot be performed in vivo without the ability to control and manipulate cellular constituents individually. Another major advantage of this system is the ability to sample by collecting the outflow from the vascular channel and analyzing the cytokines produced within the LPS-stimulated alveolus-on-a-chip device. Effective sampling within LOC devices provides more depth in evaluating and modeling pathophysiology progression by measuring shifts in cytokines with functional changes. It is critical to note that when performing this thrombotic evaluation, the model did not mimic the breathing motion of the lungs, although the system can incorporate dynamic stimulation. The cyclic strain from breathing can affect lung pathology, and its incorporation in future prothrombotic studies will augment the physiological relevance of this model. Furthermore, this system has the potential to become a precursor test platform for evaluating thrombolytic or antithrombolytic therapeutics.

157

158

CHAPTER 4 Lung-on-a-chip platforms

Modeling cystic fibrosis in a microfluidic device The multiorgan nature of CF (Elborn, 2007) makes it difficult to develop comprehensive in vitro models of the disease, which is why animal models are commonly used to study CF development and progression. Microfluidic pulmonary models can mimic the accumulation of viscous mucus, trapping bacteria to result in recurrent infections and other complications. Skolimowski et al. used a twochamber microfluidic device and an alginate hydrogel to model the thick mucus layer in the bronchi of CF patients (Skolimowski et al., 2012a,b); the hydrogel was formed on top of a membrane by introducing sodium alginate into the top channel and flowing calcium chloride in the bottom channel. A human subbronchial gland cell line (Calu-3) was seeded on the bottom of the membrane in the lower compartment. To model bacterial infection, P. aeruginosa laboratory strain PAO1 bacteria were added into the top channel, to form a biofilm in the hydrogel. This model was used to study the efficacy of antibiotic treatment by introducing antibiotic drugs in relevant concentrations and cycles into the medium flowing in the bottom channel. The same group developed a modular multicompartment microfluidic system (described in Current lung-on-a-chip systems section) to model CF conditions in the various compartments of the airway system (Skolimowski et al., 2012a,b). The segmentation of the system into a pulmonarylike hierarchy, and the effective inoculation of the model with P. aeruginosa, served as a test platform of bacterial response to antibiotic treatments. Testing bacterial responses at different oxygen levels can be achieved in a more conventional system, but the footprint and scalability of an interconnected system adds a new aspect and provides a holistic approach that may identify feedback loops that would be nonexistent in singular testing modules using one oxygen concentration at a time. This interconnected modular approach could be used to understand metabolism, cytokine secretion, and other cellular- and tissue-level responses and could be extended into an integrated LOC module for pathophysiological study.

Lung tumor-on-a-chip devices As previously mentioned, lung cancer is estimated to be the most prevalent lethal cancer, making the search for suitable therapeutics against lung tumors one of the foremost challenges facing modern healthcare. 3D in vitro systems have allowed researchers to study lung tumors in a more complex microenvironment that resembles the physiological environment better than static 2D models (Asghar et al., 2015). Tumor-on-chip systems emerged over the past 10 years, aiming to model the local tumor environment during various stages of the cancer cascade. Traditionally, tumor-on-chip models have recreated the general tissue microenvironment in which tumors form, rather than an organ-specific environment. These models enable long-term studies, providing sufficient time for the slow growth of the on-chip tumor, which can take weeks to establish in vitro. In addition to gaining a general understanding of the disease complexity at the organ level, the

Improvements needed in lung-on-a-chip platforms

development of tumor-on-a-chip platforms is motivated by their potential in preclinical drug evaluation, with better control of individual parameters than animal models. Interfacing cancer progression with the microenvironmental cues of an organ system may provide new perspectives on progression mechanisms and as such, insight into therapeutic targets. One of the most sophisticated microfluidic orthotopic lung tumor models (Hassell et al., 2017) built on an alveolus-on-a-chip device (Jain et al., 2018) (Fig. 4.9, Section 4.3), demonstrated the influence of mechanical cues from the cyclical process of breathing on tumor growth and vascular invasion. The epithelial tissue layer in the model was injected with human non small cell lung cancer cells that express high levels of fluorescent proteins to optically monitor tumor growth. Upon subjecting the chip to cyclic mechanical stretching, cell proliferation was inhibited by 50% and the tumor cells grew centered in smaller areas over the epithelium instead of spreading over larger areas. This work suggested the potential of a positive feedback loop dependent on motion, in which a mechanical strain associated with unencumbered breathing decreases epithelial growth factor receptor (EGFR) phosphorylation, while the loss of motion from increased cell proliferation in the alveoli enhances tumor growth, with a concomitant increase in phosphorylated EGFR levels. This feedback loop additionally highlights the adverse effect that enhanced EGFR phosphorylation has on rociletinib treatment, underscoring the importance, superiority, and utility of models that exhibit physiological relevance (Jain et al., 2018). In addition to gaining a general understanding of the disease complexity of cancer at the organ level, tumor-on-a-chip devices have potential as preclinical screening platforms for drugs. Advances in the field of nanotechnology have provided a range of novel nanotherapeutics, and their safety and efficacy must be evaluated. A platform to assess drug sensitivity for the treatment of lung cancer was developed to mimic the 3D tumor microenvironment (Xu et al., 2013) (Fig. 4.10). Non-small-cell lung cancer cells and stromal cells were exposed onchip to chemotherapeutic drugs with different gradient concentrations to assay an appropriate concentration and single- and combined-drug chemotherapy regimes. By using patient-specific cells, this platform may personalize effective clinical cancer treatments.

Improvements needed in lung-on-a-chip platforms for disease modeling and lung regeneration Limitations of current organ-on-a-chip systems OOC approaches hold great promise for the future development of in vitro studies of lung pathophysiology. Microfluidic models increase our knowledge of pulmonary diseases by their ability to mimic dynamic, physiologically relevant features of the human lung and approach cell culture studies using synthetic biology

159

160

CHAPTER 4 Lung-on-a-chip platforms

FIGURE 4.10 A microfluidic platform to assess drug sensitivity for the treatment of lung cancer. The device contains parallel cell culture chambers with nonsmall cancer cells and stromal cells. Reprinted from Xu, Z., Gao, Y., Hao, Y., Li, E., Wang, Y., Zhang, J., et al., 2013. Application of a microfluidic chip-based 3D co-culture to test drug sensitivity for individualized treatment of lung cancer. Biomaterials 34, 4109 4117. © 2013, with permission from Elsevier.

methods. Some of the major accomplishments in disease modeling using LOC platforms have been described in this chapter. Another important accomplishment is the ability to perform a true-match comparison between the responses from normal cells and disease-exposed cells from the same patient, which is a step toward the development of individualized treatments. The development of these systems over the past decade illustrates the ongoing efforts of creating LOC platforms to model pulmonary diseases, predict clinical outcomes, and evaluate drug efficacy. However, technical challenges remain. Although microfluidic technology allows multiplexing and automation of cellular assays with high throughput, a general hurdle is to render the design useful, easy to handle, and compatible with the equipment available in biology laboratories. With increasing device complexity, design aspects and usability become important considerations, so as to avoid producing nontranslatable, nonadoptable technologies. Furthermore, with miniaturization, small variations can still exert a great impact, so high robustness and reproducibility are required of these devices. In this section, we discuss the limitations of the current state-of-the-art of LOC devices and the necessary improvements, followed by an outline of future directions of in vitro lung pathology modeling.

Improvements to organ-on-a-chip technology Proof-of-concept studies have demonstrated the usefulness of LOC technology, but further research is needed to develop even more sophisticated systems that better mimic human organ functions and expand the usability of the devices.

Improvements needed in lung-on-a-chip platforms

To date, LOC devices have not reached their full potential in terms of biological complexity. The current OOC platforms still lack high fidelity of cellular components and are oversimplified to only a few main cell types assumed to be sufficiently representative of the whole tissue (Takebe et al., 2017). Thus an important next step toward expanding the use of LOC devices should be the incorporation of other cell types, especially immune cells such as macrophages and lymphocytes, to study the crosstalk between cells. For example, in the breathing LOC model developed by Huh et al. (2012), the findings suggested that interleukin 2 induced toxicity may be mediated by lymphocytes or other specific immune cells. Even though the experiments were not performed, the device can incorporate immune cells to study their particular contribution to responses, a feature that is not possible in animal models. In addition, the use of human-induced pluripotent stem cells can recreate the complex lung tissue functionality and architecture by deriving various types of lung cells from both proximal and distal airways and by mimicking the branching morphogenesis process directly on a microfluidic platform (Dye et al., 2016a,b; Gordon et al., 2015). Moreover, even though a simplified vascular component has already been integrated into LOC platforms (Jain et al., 2018), the incorporation and innervation of mesenchymal and neural tissue are fundamental to fully recreate the in vivo microenvironment. There are also several engineering challenges that must be addressed in device complexity and material composition, to both improve these systems and increasing their accessibility for end-users. PDMS is currently the most widely used material in fabricating LOC devices, because of its ease of use, rapid fabrication process, optical clarity, and elasticity. However, as it is hydrophobic and gaspermeable, small hydrophobic molecules can be readily absorbed by PDMS, complicating toxicology studies (Gomez-Sjoberg et al., 2010). While the gas permeability of PDMS is beneficial for culturing cells under optimal gas conditions, the small scale and permeability can result in hyperoxia. Achieving a tissue- and disease-specific gas environment may require changing the culture conditions (i.e., incubator gas content) to match in vitro oxygen tension levels to that of in vivo tissue levels. Controlling specific oxygen environments increases the complexity of culturing these systems, as PDMS is too gas-permeable to allow the cultured tissue to consume oxygen through cellular respiration or to modulate oxygen tension levels. Other materials traditionally used for cell culture experiments, such as polystyrene, can overcome this problem, but they are much stiffer, with a Young’s modulus three to four orders of magnitude greater than that of PDMS, which far exceeds that of biological tissues. In addition, they cannot be stretched like PDMS to emulate breathing. The elasticity of PDMS membranes enables the application of mechanical strain on the cells using vacuum pressure applied to specific channels in the devices. Another critical drawback of PDMS is that it presents challenges in larger scale manufacturing, limiting the scale-up of production and distribution to enduser researchers and clinicians. Injection molding could enable high-throughput fabrication of PDMS devices. However, the process would be expensive to

161

162

CHAPTER 4 Lung-on-a-chip platforms

implement and difficult to make robust in order to control batch to batch variability based on the premixing and curing of PDMS. Unless there is a highly profitable application, it is unlikely that investments into mass production of PDMS devices will be made (Klapperich, 2009). Furthermore, the fabrication of ultrathin membranes to mimic the basement membrane of lung alveoli and bronchiolar epithelium remains an obstacle. This membrane, onto which cells are seeded, must be ultrathin to recreate appropriate cell communication through the basal membrane, concomitantly capturing in vivo like responses and functions. One solution for constructing such a membrane is to directly deposit it on the device. Yang et al. demonstrated the use of this technique to model an alveolar tumor and evaluate the anticancer drug gefitinib (Yang et al., 2018), using a biocompatible poly(lactic-co-glycolic acid) electrospun nanofiber membrane as cellular scaffold prepared directly on the PDMS layer. The scaffold provided a 3D environment for cell growth and the thickness was controllable to a few microns by modifying the spinning time. Although the microchip was cultured as a static system, the study provides a good example of how tissue engineering techniques can be integrated into OOC technology. Improving device materials or using a composite of various materials can capture microenvironmental parameters such as the integration of chemical signals or mechanical properties that are more in line with in vivo stiffness. This will benefit the OOC field by capturing more physiologically relevant cellular responses in these surrogate in vitro model systems. One of the greatest promises of microfluidic in vitro systems is the potential to integrate on-chip microsensors for real-time detection. This would enable control and manipulation of parameters during the cell assay. Electrochemical or optical sensors can be incorporated using techniques from microtechnology to monitor environmental conditions such as local temperature and oxygen levels (Murphy and Atala, 2016) or cell behavior and mechanical, electrical, and chemical stimuli. Quantitative information about the cell culture can then be obtained as a functional readout on cell behavior or to benchmark to other systems or other experiments. To date, LOC research has largely depended on fluorescence microscopy, which is a qualitative analytical method. The lack of quantitative standardized parameters is a general problem in the OOC field. So far, microsensors have not been widely integrated into LOC systems. In barrier systems, such as models of the lung alveolus, small airway, or blood brain barrier, it is imperative to monitor barrier formation and function for the formation of a confluent barrier. Quantitative information on the barrier can be obtained by measuring transepithelial electrical resistance (TEER). Proof-of-concept studies have measured TEER in LOC systems (Henry et al., 2017), but the technique has not been widely implemented, perhaps because the technology has not until recently been robust enough for accurate readouts. In addition, TEER measurements in ALI cultures are difficult, because of the need to have both channels filled with liquid on either side of the membrane. While this enables the TEER measurement, if cells are submerged in liquid for too long, the cells can become highly stressed and suffer from damages.

Improvements needed in lung-on-a-chip platforms

Other, better defined sensors for microscale integration include gas-based sensors for nitric oxide, oxygen, and carbon dioxide, and pH sensors. These environmental sensors can provide insight into the metabolic activity of the LOC and how it impacts the functional outcomes being evaluated. These sensors, along with TEER readings, are noninvasive and enable continuous nonendpoint analysis, enhancing the study of lung pathophysiology. The small dimensions of microfluidic devices are the main strength of the technology, offering lower reagent consumption and fast analysis. However, the importance of scaling and relative dimensions in biology has not been widely discussed in the LOC field (Bajaj et al., 2016). Several approaches have been suggested for scaling, including allometric scaling by the relative mass of organs and functional scaling by factors such as blood flow or metabolic rate. Developing allometrically scaled models of the lung on the microscale is physically challenging, as it is impossible to recreate realistic ratios of ALI surface area, dimensions, and blood volumes (Bajaj et al., 2016). Proportional scaling and relevant dimensions become increasingly important when developing multi-OOC or body-on-achip devices (Moraes et al., 2013) to mimic the interactions between organs or disease or drug response on a multiorgan level. Body-on-a-chip systems have already been used to model processes such as pharmacokinetic clearance in drugtoxicity studies (Prantil-Baun et al., 2018) and are discussed in further detail in the chapter entitled “Human body-on-a-chip systems.” Body-on-a-chip systems would be well-positioned to model diseases that simultaneously affect the lungs and other organs, such as CF or smoke-induced injuries; these multicompartmental systems could have a great impact on our understanding of disease mechanisms and their interplay. In addition, crosstalk between lung components or functions could theoretically be modeled using a multicompartmental system and appropriate scaling to capture the in vivo hierarchy. The LOC systems that have been developed to date are largely represented by barrier-function models rather than parenchymal tissue-function models. In addition to gas exchange, the lung is responsible for controlling pH balance in the blood, preventing embolisms, protecting against infection, enabling speech, and providing a blood reservoir. Future LOC systems should include parenchymal tissue functions modeled with spheroid or organoid cultures, as discussed in the following sections.

Innovative strategies and fabrication methods for improving standard microfluidic devices Organoids-on-a-chip and of the need for multiorgan platforms Researchers agree that the use of 3D models is fundamental to recapitulate human pathophysiology. Organoids have emerged as a promising tool for addressing the limitations of common 2D approaches (In vitro and in vivo models of respiratory

163

164

CHAPTER 4 Lung-on-a-chip platforms

disease pathogenesis section) (Devarasetty et al., 2018; Wilkinson et al., 2017). Using a synergistic approach, it is possible to combine OOC devices and organoids to enable a higher throughput readout for organoids and a more in vivo like cellular composition for OOC devices (Takebe et al., 2017). This integrated technology has been used to create multi-OOC platforms to study interorgan interactions (Skardal et al., 2016; Takebe et al., 2017): Skardal et al. fabricated a modular device composed of a microfluidic circulatory system that connected bioprinted lung, heart, and liver organoid microreactors to evaluate drug responses both in individual organoids and systemically (Skardal et al., 2017, 2015). The lung module was fabricated by culturing vascular endothelial cells, lung bronchial epithelial cells, and fibroblasts over a porous membrane, creating a layered 3D organoid. TEER and electrophysiological sensors were used to monitor the organoids and their CFTR activity. Similarly, the Advanced Tissue-Engineered Human Ectypal Network Analyzer program designed by the Los Alamos National Laboratory is a system composed of lung, heart, liver, and kidney organoids of millimetric size that uses standard clinical diagnostic tools (Skardal et al., 2019). Leveraging 3D organoid cultures in a dynamic controllable microenvironment will recreate structure function relationships with more complexity, such as organoid alveolar sacs or bronchospheres at the end of bronchiole-like microfluidic tubes. This combinatorial approach may close the hierarchical, geometric, and fundamentally biological gaps between in vitro and in vivo models and position the next generation of OOC systems.

Three-dimensional printing for microfabrication and bioprinting 3D printing methods are a valuable tool in overcoming the main limitations of conventional microfabrication techniques, which are time-consuming and require expensive equipment and special facilities (Amin et al., 2016; Bhattacharjee et al., 2016; Ho et al., 2015). This technology has been used to fabricate master molds, microfluidic devices, and microfluidic components for generating gradients, mixing, and sensing, for a wide range of biological and cellular applications (Amin et al., 2016; Knowlton et al., 2016b; Lee et al., 2016; Zhou, 2017). Unlike standard techniques in microfluidic device fabrication, 3D printing enables the production of complex 3D structures such as microvascular networks (Bertassoni et al., 2014; Hasan et al., 2014; Wu et al., 2010), which are fundamental to ensure vascularized tissues. While 3D printing can produce complex structures using plastic polymers, bioprinting enables high spatial control in positioning and patterning cells and biomaterials to form heterogeneous biological structures (Ho et al., 2015; Murphy and Atala, 2014; Verhulsel et al., 2014; Yi et al., 2017). Moreover, the ability to print a wide range of biomaterials as alternatives to PDMS circumvents the issue of PDMS absorption of small hydrophobic molecules (Halldorsson et al., 2015; Wang et al., 2012). Optimizing the biomaterials used for printing can help capture tissue- and organ-specific properties such as mechanical characteristics and chemical cues. The technology has been used to fabricate tissue constructs such as

Conclusions

bone (Young Park et al., 2015), liver (Lee et al., 2016), skin (Kim et al., 2017), kidney (Homan et al., 2016; Sochol et al., 2016), nervous system (Johnson et al., 2016), heart (Zhang et al., 2016), and lung. Park et al. harnessed 3D cell printing to fabricate a vascularized airway-on-a-chip using an extracellular matrix bioink encapsulating endothelial cells and fibroblasts and print a 3D vascular network integrated with an airway epithelium model (Park et al., 2018). The platform was used to investigate the role of cytokines in asthma and asthma exacerbation and closely recreated in vivo the pathophysiological mechanisms. Furthermore, 3D printing can also be used to fabricate both the microfluidic device and the OOC components (Knowlton et al., 2016a); Lee et al. produced an OOC device in a single-step process by printing poly(ε-caprolactone) as the platform material and encapsulated various cell types in an extracellular matrix-based hydrogel, eliminating the cell-seeding step. The 3D bioprinted device was used to create a liver-on-a-chip platform, exhibiting enhanced liver function and confirming the versatility and potential of rapid prototyping techniques (Lee and Cho, 2016). By providing a variety of rapid, scalable, and reproducible fabrication methods, 3D printing can improve LOC design and engineering, while bioprinting can enable the spatial control of a more complex cellular milieu native to the lung tissue, comprehensive of the mesenchymal and neural tissue components.

Conclusions Harnessing microfluidic systems for precision control and monitoring of cellular microenvironments is a logical and necessary progression in modeling diseases in vitro. While microfluidic devices can potentially revolutionize research in this field, there are numerous challenges that must be overcome before broad adoption and use of these devices, and eventual replacement of conventional disease modeling techniques can be expected. As discussed throughout this chapter, the challenges that LOC devices face generally fall into two broad categories: biological complexity and robust engineering design. It is the opinion of the authors that at the time of writing, the most significant factors limiting LOC devices, and where the most innovation is required, are engineering design challenges. Biomimetic models cannot truly proliferate until a robust engineering base is established. By focusing on creating highly reproducible and easy-to-use microfluidic systems, greater adoption and use of LOC devices can be realized and their capabilities pushed further. The next steps in LOC technology must involve a push for more interdisciplinary thinking and awareness, echoing the development of microfluidics in general, a field defined by a complex interplay of multidisciplinary foundations and technological requirements driven by members of the biology, materials science, chemistry, electronics, and engineering communities. Parallel advancements in organoid synthesis and 3D printing will have great implications for LOC devices,

165

166

CHAPTER 4 Lung-on-a-chip platforms

so breakthroughs will require creative amalgamations of discrete techniques and expertise. OOC devices and LOC systems, in particular, are largely still in their infancy, and much progress is required before they become a truly transformative technology. Although the techniques and execution are not yet optimized, the notable advantages of LOC platforms should be viewed as an inspiration for continued progress.

Acknowledgments This work was supported by the Marie Sklodowska-Curie Grant Agreements: No. 766181 project “DeLIVER” (author AD), No. 766007 project “MaMi” (author EKT), No. 722591 project “PHOTOTRAIN,” No. 753743 an Individual Fellowship (author SCLP), all from the European Union’s Horizon 2020 research and innovation program. The authors thank Julien Ridouard for producing original figures and schematics and Lisa Muiznieks, PhD, for insightful feedback and edits.

References Alaiwa, M.H.A., Launspach, J.L., Sheets, K.A., Rivera, J.A., Gansemer, N.D., Taft, P.J., et al., 2016. Repurposing tromethamine as inhaled therapy to treat CF airway disease. JCI Insight 1. Available from: https://doi.org/10.1172/jci.insight.87535. Alnimr, A.M., 2015. Dormancy models for Mycobacterium tuberculosis: a minireview. Braz. J. Microbiol. 46, 641 647. Available from: https://doi.org/10.1590/S1517838246320140507. Amin, R., Dupuis, A., Aaron, S.D., Ratjen, F., 2010. The effect of chronic infection with Aspergillus fumigatus on lung function and hospitalization in patients with cystic fibrosis. Chest 137, 171 176. Available from: https://doi.org/10.1378/chest.09-1103. Amin, R., Knowlton, S., Hart, A., Yenilmez, B., Ghaderinezhad, F., Katebifar, S., et al., 2016. 3D-printed microfluidic devices. Biofabrication 8, 022001. Available from: https://doi.org/10.1088/1758-5090/8/2/022001. Anderson, A.E., Foraker, A.G., 1962. Relative dimensions of bronchioles and parenchymal spaces in lungs from normal subjects and emphysematous patients. Am. J. Med. 32, 218 226. Available from: https://doi.org/10.1016/0002-9343(62)90291-7. Anjali, P.S., Mahto, S., 2016. Revisiting pulmonary diseases using microfluidic technology. Int. J. Biosens. Bioelectron. 1, 00001. Asghar, W., El Assal, R., Shafiee, H., Pitteri, S., Paulmurugan, R., Demirci, U., 2015. Engineering cancer microenvironments for in vitro 3-D tumor models. Mater. Today 18, 539 553. Available from: https://doi.org/10.1016/j.mattod.2015.05.002. Bajaj, P., Harris, J.F., Huang, J.-H., Nath, P., Iyer, R., 2016. Advances and challenges in recapitulating human pulmonary systems: at the cusp of biology and materials. ACS Biomater. Sci. Eng. 2, 473 488. Available from: https://doi.org/10.1021/ acsbiomaterials.5b00480. ´ lvarez-Garcia´, Y.R., Niu, M., Jime´nez-Torres, J.A., Barkal, L.J., Procknow, C.L., A Brockman-Schneider, R.A., et al., 2017. Microbial volatile communication in human

References

organotypic lung models. Nat. Commun. 8. Available from: https://doi.org/10.1038/ s41467-017-01985-4. Barkauskas, C.E., Chung, M.-I., Fioret, B., Gao, X., Katsura, H., Hogan, B.L.M., 2017. Lung organoids: current uses and future promise. Development 144, 986 997. Available from: https://doi.org/10.1242/dev.140103. Beckett, E.L., Stevens, R.L., Jarnicki, A.G., Kim, R.Y., Hanish, I., Hansbro, N.G., et al., 2013. A new short-term mouse model of chronic obstructive pulmonary disease identifies a role for mast cell tryptase in pathogenesis. J. Allergy Clin. Immunol. 131, 752 762. Available from: https://doi.org/10.1016/j.jaci.2012.11.053. Benam, K.H., Dauth, S., Hassell, B., Herland, A., Jain, A., Jang, K.-J., et al., 2015. Engineered in vitro disease models. Annu. Rev. Pathol. 10, 195 262. Available from: https://doi.org/10.1146/annurev-pathol-012414-040418. Benam, K.H., Villenave, R., Lucchesi, C., Varone, A., Hubeau, C., Lee, H.-H., et al., 2016a. Small airway-on-a-chip enables analysis of human lung inflammation and drug responses in vitro. Nat. Methods 13, 151 157. Available from: https://doi.org/10.1038/ nmeth.3697. Benam, K.H., Novak, R., Nawroth, J., Hirano-Kobayashi, M., Ferrante, T.C., Choe, Y., et al., 2016b. Matched-comparative modeling of normal and diseased human airway responses using a microengineered breathing lung chip. Cell Syst. 3, 456 466.e4. Available from: https://doi.org/10.1016/j.cels.2016.10.003. Berry, C., Cacou, C., Lee, D., Bader, D., Shelton, J., 2003. Dermal fibroblasts respond to mechanical conditioning in a strain profile dependent manner. Biorheology 40, 337 345. Bertassoni, L.E., Cecconi, M., Manoharan, V., Nikkhah, M., Hjortnaes, J., Cristino, A.L., et al., 2014. Hydrogel bioprinted microchannel networks for vascularization of tissue engineering constructs. Lab Chip 14, 2202 2211. Available from: https://doi.org/ 10.1039/C4LC00030G. Bhattacharjee, N., Urrios, A., Kang, S., Folch, A., 2016. The upcoming 3D-printing revolution in microfluidics. Lab Chip 16, 1720 1742. Available from: https://doi.org/ 10.1039/c6lc00163g. Bhowmick, R., Gappa-Fahlenkamp, H., 2016. Cells and culture systems used to model the small airway epithelium. Lung 194, 419 428. Available from: https://doi.org/10.1007/ s00408-016-9875-2. Bielecka, M.K., Elkington, P., 2018. Advanced cellular systems to study tuberculosis treatment. Curr. Opin. Pharmacol. 42, 16 21. Available from: https://doi.org/10.1016/j. coph.2018.06.005. Bilek, A.M., Dee, K.C., Gaver, D.P., 2003. Mechanisms of surface-tension-induced epithelial cell damage in a model of pulmonary airway reopening. J. Appl. Physiol. 94, 770 783. Available from: https://doi.org/10.1152/japplphysiol.00764.2002. Bischel, L.L., Sung, K.E., Jime´nez-Torres, J.A., Mader, B., Keely, P.J., Beebe, D.J., 2014. The importance of being a lumen. FASEB J. 28, 4583 4590. Available from: https:// doi.org/10.1096/fj.13-243733. Blume, C., Davies, D.E., 2013. In vitro and ex vivo models of human asthma. Eur. J. Pharm. Biopharm. 84, 394 400. Available from: https://doi.org/10.1016/j. ejpb.2012.12.014. Booth, A.J., Hadley, R., Cornett, A.M., Dreffs, A.A., Matthes, S.A., Tsui, J.L., et al., 2012. Acellular normal and fibrotic human lung matrices as a culture system for in vitro

167

168

CHAPTER 4 Lung-on-a-chip platforms

investigation. Am. J. Respir. Crit. Care Med. 186, 866 876. Available from: https:// doi.org/10.1164/rccm.201204-0754OC. Briard, B., Heddergott, C., Latge´, J.-P., 2016. Volatile compounds emitted by Pseudomonas aeruginosa stimulate growth of the fungal pathogen Aspergillus fumigatus. mBio 7, e00219. Available from: https://doi.org/10.1128/mBio.00219-16. Calderon, V.E., Valbuena, G., Goez, Y., Judy, B.M., Huante, M.B., Sutjita, P., et al., 2013. A humanized mouse model of tuberculosis. PLoS One 8, e63331. Available from: https://doi.org/10.1371/journal.pone.0063331. Cardona, P.-J., Williams, A., 2017. Experimental animal modelling for TB vaccine development. Int. J. Infect. Dis. 56, 268 273. Available from: https://doi.org/10.1016/j. ijid.2017.01.030. Cassidy, K.J., Halpern, D., Ressler, B.G., Grotberg, J.B., 1999. Surfactant effects in model airway closure experiments. J. Appl. Physiol. 87, 415 427. Available from: https://doi. org/10.1152/jappl.1999.87.1.415. Chen, X.-J., Seth, S., Yue, G., Kamat, P., Compans, R.W., Guidot, D., et al., 2004. Influenza virus inhibits ENaC and lung fluid clearance. Am. J. Physiol.-Lung Cell. Mol. Physiol. 287, L366 L373. Available from: https://doi.org/10.1152/ajplung.00011.2004. Chen, J.-H., Stoltz, D.A., Karp, P.H., Ernst, S.E., Pezzulo, A.A., Moninger, T.O., et al., 2010. Loss of anion transport without increased sodium absorption characterizes newborn porcine cystic fibrosis airway epithelia. Cell 143, 911 923. Available from: https://doi.org/10.1016/j.cell.2010.11.029. Chen, Y.-W., Huang, S.X., de Carvalho, A.L.R.T., Ho, S.-H., Islam, M.N., Volpi, S., et al., 2017. A three-dimensional model of human lung development and disease from pluripotent stem cells. Nat. Cell Biol. 19, 542 549. Available from: https://doi.org/10.1038/ ncb3510. Chu, J., Clements, J.A., Cotton, E.K., Klaus, M.H., Sweet, A.Y., Tooley, W.H., 1967. Neonatal pulmonary ischemia. I. Clinical and physiological studies. Pediatrics 40, 709. Cmielewski, P., Donnelley, M., Parsons, D.W., 2014. Long-term therapeutic and reporter gene expression in lentiviral vector treated cystic fibrosis mice. J. Gene Med. 16, 291 299. Available from: https://doi.org/10.1002/jgm.2778. Conant, E., Fox, K., Miller, W., 1989. Pulmonary edema as a complication of interleukin-2 therapy. Am. J. Roentgenol. 152, 749 752. Available from: https://doi.org/10.2214/ ajr.152.4.749. Constant, S., Huang, S., Wisniewski, L., Mas, C., 2015. Advanced human in vitro models for the discovery and development of lung cancer therapies. In: Omboon Vallisuta and Suleiman Olimat (Eds.), Drug Discovery and Development - From Molecules to Medicine, IntechOpen ,https://doi.org/10.5772/60606. Cooney, A.L., Abou Alaiwa, M.H., Shah, V.S., Bouzek, D.C., Stroik, M.R., Powers, L.S., et al., 2016. Lentiviral-mediated phenotypic correction of cystic fibrosis pigs. JCI Insight 1, e88730. 10.1172/jci.insight.88730. Cortiella, J., Niles, J., Cantu, A., Brettler, A., Pham, A., Vargas, G., et al., 2010. Influence of acellular natural lung matrix on murine embryonic stem cell differentiation and tissue formation. Tissue Eng., A 16, 2565 2580. Available from: https://doi.org/10.1089/ ten.tea.2009.0730. Danahay, H., Pessotti, A.D., Coote, J., Montgomery, B.E., Xia, D., Wilson, A., et al., 2015. Notch2 is required for inflammatory cytokine-driven goblet cell metaplasia in the lung. Cell Rep. 10, 239 252. Available from: https://doi.org/10.1016/j.celrep.2014.12.017.

References

Dekkers, J.F., Wiegerinck, C.L., de Jonge, H.R., Bronsveld, I., Janssens, H.M., de Winterde Groot, K.M., et al., 2013. A functional CFTR assay using primary cystic fibrosis intestinal organoids. Nat. Med. 19, 939 945. Available from: https://doi.org/10.1038/ nm.3201. Delaere, P., Van Raemdonck, D., Vranckx, J., 2018. Tracheal transplantation. Intensive Care Med. Available from: https://doi.org/10.1007/s00134-018-5445-9. Deng, L., Bosse, Y., Brown, N., Chin, L.Y.M., Connolly, S.C., Fairbank, N.J., et al., 2009. Stress and strain in the contractile and cytoskeletal filaments of airway smooth muscle. Pulm. Pharmacol. Ther. 22, 407 416. Available from: https://doi.org/10.1016/j. pupt.2009.04.008. Devaney, J., Horie, S., Masterson, C., Elliman, S., Barry, F., O’Brien, T., et al., 2015. Human mesenchymal stromal cells decrease the severity of acute lung injury induced by E. coli in the rat. Thorax 70, 625 635. Available from: https://doi.org/10.1136/thoraxjnl-2015-206813. Devarasetty, M., Mazzocchi, A.R., Skardal, A., 2018. Applications of bioengineered 3D tissue and tumor organoids in drug development and precision medicine: current and future. BioDrugs 32, 53 68. Available from: https://doi.org/10.1007/s40259-017-0258x. Dhami, R., Gilks, B., Xie, C., Zay, K., Wright, J.L., Churg, A., 2000. Acute cigarette smoke-induced connective tissue breakdown is mediated by neutrophils and prevented by alpha1-antitrypsin. Am. J. Respir. Cell Mol. Biol. 22, 244 252. Available from: https://doi.org/10.1165/ajrcmb.22.2.3809. Diedrich, C.R., Flynn, J.L., 2011. HIV-1/Mycobacterium tuberculosis coinfection immunology: how does HIV-1 exacerbate tuberculosis? Infect. Immun. 79, 1407 1417. Available from: https://doi.org/10.1128/IAI.01126-10. Dong, X., Guan, J., English, J.C., Flint, J., Yee, J., Evans, K., et al., 2010. Patientderived first generation xenografts of non small cell lung cancers: promising tools for predicting drug responses for personalized chemotherapy. Clin. Cancer Res. 1078 0432. Doryab, A., Amoabediny, G., Salehi-Najafabadi, A., 2016. Advances in pulmonary therapy and drug development: lung tissue engineering to lung-on-a-chip. Biotechnol. Adv. 34, 588 596. Available from: https://doi.org/10.1016/j.biotechadv.2016.02.006. Douglas, W.H.J., Moorman, G.W., Teel, R.W., 1976. The formation of histotypic structures from monodisperse fetal rat lung cells cultured on a three-dimensional substrate. In Vitro 12, 373 381. Available from: https://doi.org/10.1007/BF02796315. Douville, N.J., Zamankhan, P., Tung, Y.C., Li, R., Vaughan, B.L., Tai, C.F., et al., 2011. Combination of fluid and solid mechanical stresses contribute to cell death and detachment in a microfluidic alveolar model. Lab Chip 11, 609 619. Available from: https:// doi.org/10.1039/c0lc00251h. Dye, B.R., Dedhia, P.H., Miller, A.J., Nagy, M.S., White, E.S., Shea, L.D., et al., 2016a. A bioengineered niche promotes in vivo engraftment and maturation of pluripotent stem cell derived human lung organoids. eLife 5. Available from: https://doi.org/10.7554/ eLife.19732. Dye, B.R., Miller, A.J., Spence, J.R., 2016b. How to grow a lung: applying principles of developmental biology to generate lung lineages from human pluripotent stem cells. Curr. Pathobiol. Rep. 4, 47 57. Available from: https://doi.org/10.1007/s40139-0160102-x.

169

170

CHAPTER 4 Lung-on-a-chip platforms

Elborn, J., 2007. How can we prevent multisystem complications of cystic fibrosis? Semin. Respir. Crit. Care Med. 28, 303 311. Available from: https://doi.org/10.1055/s-2007981651. Eppert, B.L., Wortham, B.W., Flury, J.L., Borchers, M.T., 2013. Functional characterization of T cell populations in a mouse model of chronic obstructive pulmonary disease. J. Immunol. (Baltimore, MD) 1950 (190), 1331 1340. Available from: https://doi.org/ 10.4049/jimmunol.1202442. Esch, E.W., Bahinski, A., Huh, D., 2015. Organs-on-chips at the frontiers of drug discovery. Nat. Rev. Drug Discov. 14, 248 260. Available from: https://doi.org/10.1038/ nrd4539. Fisher, J.T., Tyler, S.R., Zhang, Y., Lee, B.J., Liu, X., Sun, X., et al., 2013. Bioelectric characterization of epithelia from neonatal CFTR knockout ferrets. Am. J. Respir. Cell Mol. Biol. 49, 837 844. Available from: https://doi.org/10.1165/rcmb.2012-0433OC. Fonseca, K.L., Rodrigues, P.N.S., Olsson, I.A.S., Saraiva, M., 2017. Experimental study of tuberculosis: From animal models to complex cell systems and organoids. PLoS Pathog. 13, e1006421. Available from: https://doi.org/10.1371/journal.ppat.1006421. Foronjy, R., D’Armiento, J., 2006. The effect of cigarette smoke derived oxidants on the inflammatory response of the lung. Clin. Appl. Immunol. Rev. 6, 53 72. Forum of International Respiratory Societies, 2017. The Global Impact of Respiratory Disease, Second Edition European Respiratory Society, Sheffield. Fredberg, J.J., Kamm, R.D., 2006. Stress transmission in the lung: pathways from organ to molecule. Annu. Rev. Physiol. 68, 507 541. Available from: https://doi.org/10.1146/ annurev.physiol.68.072304.114110. Fricker, M., Deane, A., Hansbro, P.M., 2014. Animal models of chronic obstructive pulmonary disease. Expert Opin. Drug Discov. 9, 629 645. Available from: https://doi.org/ 10.1517/17460441.2014.909805. Furie, B., Furie, B.C., 2008. Mechanisms of thrombus formation. N. Engl. J. Med. 359, 938 949. Available from: https://doi.org/10.1056/NEJMra0801082. Gaschler, G.J., Skrtic, M., Zavitz, C.C.J., Lindahl, M., Onnervik, P.-O., Murphy, T.F., et al., 2009. Bacteria challenge in smoke-exposed mice exacerbates inflammation and skews the inflammatory profile. Am. J. Respir. Crit. Care Med. 179, 666 675. Available from: https://doi.org/10.1164/rccm.200808-1306OC. Gazdar, A.F., Gao, B., Minna, J.D., 2010. Lung cancer cell lines: useless artifacts or invaluable tools for medical science? Lung Cancer (Amst. Neth.) 68, 309 318. Available from: https://doi.org/10.1016/j.lungcan.2009.12.005. Gazdar, A.F., Hirsch, F.R., Minna, J.D., 2016. From mice to men and back: an assessment of preclinical model systems for the study of lung cancers. J. Thorac. Oncol. 11, 287 299. Available from: https://doi.org/10.1016/j.jtho.2015.10.009. Gazdar, A.F., Savage, T.K., Johnson, J.E., Berns, A., Sage, J., Linnoila, R.I., et al., 2015. The comparative pathology of genetically engineered mouse models for neuroendocrine carcinomas of the lung. J. Thorac. Oncol. 10, 553 564. Available from: https://doi.org/ 10.1097/JTO.0000000000000459. Ghaedi, M., Mendez, J.J., Bove, P.F., Sivarapatna, A., Raredon, M.S.B., Niklason, L.E., 2014. Alveolar epithelial differentiation of human induced pluripotent stem cells in a rotating bioreactor. Biomaterials 35, 699 710. Available from: https://doi.org/10.1016/ j.biomaterials.2013.10.018.

References

Gilpin, S.E., Li, Q., Evangelista-Leite, D., Ren, X., Reinhardt, D.P., Frey, B.L., et al., 2017. Fibrillin-2 and Tenascin-C bridge the age gap in lung epithelial regeneration. https://doi.org/10.1016/j. Biomaterials 140, 212 219. Available from: biomaterials.2017.06.027. Gilpin, S.E., Wagner, D.E., 2018. Acellular human lung scaffolds to model lung disease and tissue regeneration. Eur. Respir. Rev. 27. Available from: https://doi.org/10.1183/ 16000617.0021-2018. Gkatzis, K., Taghizadeh, S., Huh, D., Stainier, D.Y.R., Bellusci, S., 2018. Use of threedimensional organoids and lung-on-a-chip methods to study lung development, regeneration and disease. Eur. Respir. J. 52. Available from: https://doi.org/10.1183/ 13993003.00876-2018. Gomez-Sjoberg, R., Leyrat, A.A., Houseman, B.T., Shokat, K., Quake, S.R., 2010. Biocompatibility and reduced drug absorption of sol 2 gel-treated poly(dimethyl siloxane) for microfluidic cell culture applications. Anal. Chem. 82, 8954 8960. Available from: https://doi.org/10.1021/ac101870s. Gordon, S., Daneshian, M., Bouwstra, J., Caloni, F., Constant, S., Davies, D.E., et al., 2015. Non-animal models of epithelial barriers (skin, intestine and lung) in research, industrial applications and regulatory toxicology. ALTEX 32, 327 378. Available from: https://doi.org/10.14573/altex.1510051. Gotoh, S., Ito, I., Nagasaki, T., Yamamoto, Y., Konishi, S., Korogi, Y., et al., 2014. Generation of alveolar epithelial spheroids via isolated progenitor cells from human pluripotent stem cells. Stem Cell Rep. 3, 394 403. Available from: https://doi.org/ 10.1016/j.stemcr.2014.07.005. Gould, R.A., Chin, K., Santisakultarm, T.P., Dropkin, A., Richards, J.M., Schaffer, C.B., et al., 2012. Cyclic strain anisotropy regulates valvular interstitial cell phenotype and tissue remodeling in three-dimensional culture. Acta Biomater. 8, 1710 1719. Available from: https://doi.org/10.1016/j.actbio.2012.01.006. Guenat, O.T., Berthiaume, F., 2018. Incorporating mechanical strain in organs-on-a-chip: lung and skin. Biomicrofluidics 12. Available from: https://doi.org/10.1063/1.5024895. Gump, A., Haughney, L., Fredberg, J., 2001. Relaxation of activated airway smooth muscle: relative potency of isoproterenol vs. tidal stretch. J. Appl. Physiol. 90, 2306 2310. Available from: https://doi.org/10.1152/jappl.2001.90.6.2306. Gupta, N., Krasnodembskaya, A., Kapetanaki, M., Mouded, M., Tan, X., Serikov, V., et al., 2012. Mesenchymal stem cells enhance survival and bacterial clearance in murine Escherichia coli pneumonia. Thorax 67, 533 539. Available from: https://doi. org/10.1136/thoraxjnl-2011-201176. Gupta, U., Katoch, V., 2005. Animal models of tuberculosis. Tuberculosis 85, 277 293. Halldorsson, S., Lucumi, E., Go´mez-Sjo¨berg, R., Fleming, R.M.T., 2015. Advantages and challenges of microfluidic cell culture in polydimethylsiloxane devices. Biosens. Bioelectron. 63, 218 231. Available from: https://doi.org/10.1016/j. bios.2014.07.029. Han, M., Rajput, C., Ishikawa, T., Jarman, C.R., Lee, J., Hershenson, M.B., 2018. Small animal models of respiratory viral infection related to asthma. Viruses 10. Available from: https://doi.org/10.3390/v10120682. Hansen, J.E., Ampaya, E.P., 1975. Human air space shapes, sizes, areas, and volumes. J. Appl. Physiol. 38, 990 995. Available from: https://doi.org/10.1152/ jappl.1975.38.6.990.

171

172

CHAPTER 4 Lung-on-a-chip platforms

Hasan, A., Paul, A., Vrana, N.E., Zhao, X., Memic, A., Hwang, Y.-S., et al., 2014. Microfluidic techniques for development of 3D vascularized tissue. Biomaterials 35, 7308 7325. Available from: https://doi.org/10.1016/j.biomaterials.2014.04.091. Hassell, B.A., Goyal, G., Lee, E., Sontheimer-Phelps, A., Levy, O., Chen, C.S., et al., 2017. Human organ chip models recapitulate orthotopic lung cancer growth, therapeutic responses, and tumor dormancy in vitro. Cell Rep. 21, 508 516. Available from: https://doi.org/10.1016/j.celrep.2017.09.043. Henry, O.Y.F., Villenave, R., Cronce, M.J., Leineweber, W.D., Benz, M.A., Ingber, D.E., 2017. Organs-on-chips with integrated electrodes for trans-epithelial electrical resistance (TEER) measurements of human epithelial barrier function. Lab Chip 17, 2264 2271. Available from: https://doi.org/10.1039/C7LC00155J. Herzog, C.R., Lubet, R.A., You, M., 1997. Genetic alterations in mouse lung tumors: Implications for cancer chemoprevention. J. Cell. Biochem. 67, 49 63. Available https://doi.org/10.1002/(SICI)1097-4644(1997)28/29 1 , 49::AID-JCB6 . 3.0. from: CO;2-W. Hirschl, R.B., Conrad, S., Kaiser, R., Zwischenberger, J.B., Bartlett, R.H., Booth, F., et al., 1998. Partial liquid ventilation in adult patients with ARDS: a multicenter phase I-II trial. Adult PLV Study Group. Ann. Surg. 228, 692 700. Hirschl, R.B., Tooley, R., Parent, A., Johnson, K., Bartlett, R.H., 1996. Evaluation of gas exchange, pulmonary compliance, and lung injury during total and partial liquid ventilation in the acute respiratory distress syndrome. Crit. Care Med. 24, 1001 1008. Ho, C.M.B., Ng, S.H., Li, K.H.H., Yoon, Y.-J., 2015. 3D printed microfluidics for biological applications. Lab Chip 15, 3627 3637. Available from: https://doi.org/10.1039/ C5LC00685F. Hofmann, W., Koblinger, L., Martonen, T.B., 1989. Structural differences between human and rat lungs: implications for Monte Carlo modeling of aerosol deposition. Health Phys. 57 (Suppl 1), 46 47. 41 46; discussion. Hoganson, D.M., Bassett, E.K., Vacanti, J.P., 2014. Lung tissue engineering. Front. Biosci. (Landmark Ed.) 19, 1227 1239. Holt, P.G., Macaubas, C., Stumbles, P.A., Sly, P.D., 1999. The role of allergy in the development of asthma. Nature 402, 12 17. Available from: https://doi.org/10.1038/ 35037009. Homan, K.A., Kolesky, D.B., Skylar-Scott, M.A., Herrmann, J., Obuobi, H., Moisan, A., et al., 2016. Bioprinting of 3D convoluted renal proximal tubules on perfusable chips. Sci. Rep. 6, 34845. Available from: https://doi.org/10.1038/srep34845. Huh, D., Fujioka, H., Tung, Y.-C., Futai, N., Paine, R., Grotberg, J.B., et al., 2007. Acoustically detectable cellular-level lung injury induced by fluid mechanical stresses in microfluidic airway systems. Proc. Natl. Acad. Sci. 104, 18886 18891. Available from: https://doi.org/10.1073/pnas.0610868104. Huh, D., Matthews, B.D., Mammoto, A., Montoya-Zavala, M., Hsin, H.Y., Ingber, D.E., 2010. Reconstituting organ-level lung functions on a chip. Science 328, 1662 1668. Available from: https://doi.org/10.1126/science.1188302. Huh, D., Leslie, D.C., Matthews, B.D., Fraser, J.P., Jurek, S., Hamilton, G.A., et al., 2012. A human disease model of drug toxicity induced pulmonary edema in a lung-on-achip microdevice. Sci. Transl. Med. 4, 159ra147. Available from: https://doi.org/ 10.1126/scitranslmed.3004249.

References

Huppert, L.A., Matthay, M.A., 2017. Alveolar fluid clearance in pathologically relevant conditions: in vitro and in vivo models of acute respiratory distress syndrome. Front. Immunol. 8, 371. Available from: https://doi.org/10.3389/fimmu.2017.00371. Iba, T., Sumpio, B.E., 1991. Morphological response of human endothelial cells subjected to cyclic strain in vitro. Microvasc. Res. 42, 245 254. Inanlou, M.R., Kablar, B., 2005. Contractile activity of skeletal musculature involved in breathing is essential for normal lung cell differentiation, as revealed in Myf5 2 / 2 : MyoD 2 / 2 embryos. Dev. Dyn. 233, 772 782. Available from: https://doi.org/ 10.1002/dvdy.20381. Ito, R., Takahashi, T., Katano, I., Ito, M., 2012. Current advances in humanized mouse models. Cell. Mol. Immunol. 9, 208 214. Available from: https://doi.org/10.1038/ cmi.2012.2. Jain, A., Barrile, R., van der Meer, A., Mammoto, A., Mammoto, T., De Ceunynck, K., et al., 2018. Primary human lung alveolus-on-a-chip model of intravascular thrombosis for assessment of therapeutics. Clin. Pharmacol. Ther. 103, 332 340. Available from: https://doi.org/10.1002/cpt.742. John, T., Kohler, D., Pintilie, M., Yanagawa, N., Pham, N.-A., Li, M., et al., 2011. The ability to form primary tumor xenografts is predictive of increased risk of disease recurrence in early-stage non small cell lung cancer. Clin. Cancer Res. 17, 134 141. Johnson, B.N., Lancaster, K.Z., Hogue, I.B., Meng, F., Kong, Y.L., Enquist, L.W., et al., 2016. 3D Printed Nervous System on a Chip. Lab Chip. 16 (8), 1393 1400. Available from: https://doi.org/10.1039/c5lc01270h. Jong, M., de, Maina, T., 2010a. Of mice and humans: are they the same?—Implications in cancer translational research. J. Nucl. Med. 51, 501 504. Available from: https://doi. org/10.2967/jnumed.109.065706. Kaisani, A., Delgado, O., Fasciani, G., Kim, S.B., Wright, W.E., Minna, J.D., et al., 2014. Branching morphogenesis of immortalized human bronchial epithelial cells in threedimensional culture. Differentiation 87, 119 126. Available from: https://doi.org/ 10.1016/j.diff.2014.02.003. Kellar, A., Egan, C., Morris, D., 2015. Preclinical murine models for lung cancer: clinical trial applications. BioMed Res. Int. 2015, 621324. Available from: https://doi.org/ 10.1155/2015/621324. Kim, B.S., Lee, J.-S., Gao, G., Cho, D.-W., 2017. Direct 3D cell-printing of human skin with functional transwell system. Biofabrication 9, 025034. Available from: https://doi. org/10.1088/1758-5090/aa71c8. Klapperich, C.M., 2009. Microfluidic diagnostics: time for industry standards. Expert Rev. Med. Devices 6 (3), 211 213. Available from: https://doi.org/10.1586/erd.09.11. Knowlton, S., Yenilmez, B., Tasoglu, S., 2016a. Towards single-step biofabrication of organs on a chip via 3D printing. Trends Biotechnol. 34, 685 688Special Issue: Biofabrication. Available from: https://doi.org/10.1016/j.tibtech.2016.06.005. Knowlton, S., Yu, C.H., Ersoy, F., Emadi, S., Khademhosseini, A., Tasoglu, S., 2016b. 3Dprinted microfluidic chips with patterned, cell-laden hydrogel constructs. Biofabrication 8, 025019. Available from: https://doi.org/10.1088/1758-5090/8/2/025019. Konishi, S., Gotoh, S., Tateishi, K., Yamamoto, Y., Korogi, Y., Nagasaki, T., et al., 2016. Directed induction of functional multi-ciliated cells in proximal airway epithelial spheroids from human pluripotent stem cells. Stem Cell Rep. 6, 18 25. Available from: https://doi.org/10.1016/j.stemcr.2015.11.010.

173

174

CHAPTER 4 Lung-on-a-chip platforms

Koo, S., Thomas, H.R., Daniels, S.D., Lynch, R.C., Fortier, S.M., Shea, M.M., et al., 2014. A breath fungal secondary metabolite signature to diagnose invasive aspergillosis. Clin. Infect. Dis. 59, 1733 1740. Available from: https://doi.org/10.1093/cid/ciu725. Kramnik, I., Beamer, G., 2016. Mouse models of human TB pathology: roles in the analysis of necrosis and the development of host-directed therapies. Semin. Immunopathol. 38, 221 237. Available from: https://doi.org/10.1007/s00281-015-0538-9. Kretzschmar, K., Clevers, H., 2016. Organoids: modeling development and the stem cell niche in a dish. Dev. Cell 38, 590 600. Available from: https://doi.org/10.1016/j. devcel.2016.08.014. Langer, R., Vacanti, J., 2016. Advances in tissue engineering. J. Pediatr. Surg. 51, 8 12. Available from: https://doi.org/10.1016/j.jpedsurg.2015.10.022. Laurance, J., 2010. British boy receives trachea transplant built with his own stem cells. BMJ 340, c1633. Available from: https://doi.org/10.1136/bmj.c1633. Lavelle, G.M., White, M.M., Browne, N., McElvaney, N.G., Reeves, E.P., 2016. Animal models of cystic fibrosis pathology: phenotypic parallels and divergences. BioMed Res. Int. 2016, 5258727. Available from: https://doi.org/10.1155/2016/5258727. Lawrence, M.B., Springer, T.A., 1993. Neutrophils roll on E-selectin. J. Immunol. 151, 6338. Lee, H., Cho, D.-W., 2016. One-step fabrication of an organ-on-a-chip with spatial heterogeneity using a 3D bioprinting technology. Lab Chip 16, 2618 2625. Available from: https://doi.org/10.1039/C6LC00450D. Lee, J.W., Choi, Y.-J., Yong, W.-J., Pati, F., Shim, J.-H., Kang, K.S., et al., 2016. Development of a 3D cell printed construct considering angiogenesis for liver tissue engineering. Biofabrication 8, 015007. Available from: https://doi.org/10.1088/17585090/8/1/015007. Li, X., 2016. In vitro toxicity testing of cigarette smoke based on the air-liquid interface exposure: a review. Toxicol. In Vitro 36, 105 113. Available from: https://doi.org/ 10.1016/j.tiv.2016.07.019. Lin, W., Huang, Y., Zhou, X.-D., Ma, Y., 2006. In vitro toxicity of silica nanoparticles in human lung cancer cells. Toxicol. Appl. Pharmacol. 217, 252 259. Available from: https://doi.org/10.1016/j.taap.2006.10.004. Macchiarini, P., Jungebluth, P., Go, T., Asnaghi, M.A., Rees, L.E., Cogan, T.A., et al., 2008. Clinical transplantation of a tissue-engineered airway. Lancet 372, 2023 2030. Available from: https://doi.org/10.1016/S0140-6736(08)61598-6. Mammoto, T., Mammoto, A., Ingber, D.E., 2013. Mechanobiology and developmental control. Annu. Rev. Cell Dev. Biol. 29, 27 61. Available from: https://doi.org/10.1146/ annurev-cellbio-101512-122340. Marque´s-Garcı´a, F., Marcos-Vadillo, E., 2016. Review of mouse models applied to the study of asthma. Methods Mol. Biol. (Clifton, NJ) 1434, 213 222. Available from: https://doi.org/10.1007/978-1-4939-3652-6_15. Mas, C., Boda, B., Caul Futy, M., Huang, S., Wisniewski, L., Constant, S., 2016. Establishment of a tumour-stroma airway model (OncoCilAir) to accelerate the development of human therapies against lung cancer. Altern. Lab. Anim. ATLA 44, 479 485. McCarron, A., Donnelley, M., Parsons, D., 2018. Airway disease phenotypes in animal models of cystic fibrosis. Respir. Res. 19, 54. Available from: https://doi.org/10.1186/ s12931-018-0750-y.

References

Meurs, H., Gosens, R., Zaagsma, J., 2008. Airway hyperresponsiveness in asthma: lessons from in vitro model systems and animal models. Eur. Respir. J. 32, 487 502. Miles, W.A., 1977. Pulmonary edema: an anatomic, pathophysiologic, and roentgenologic analysis. J. Natl. Med. Assoc. 69, 4. Miller, A.J., Spence, J.R., 2017. In vitro models to study human lung development, disease and homeostasis. Physiology 32, 246 260. Miller, L.A., Royer, C.M., Pinkerton, K.E., Schelegle, E.S., 2017. Nonhuman primate models of respiratory disease: past, present, and future. ILAR J. 58, 269 280. Available from: https://doi.org/10.1093/ilar/ilx030. Miller, A.J., Hill, D.R., Nagy, M.S., Aoki, Y., Dye, B.R., Chin, A.M., et al., 2018. In vitro induction and in vivo engraftment of lung bud tip progenitor cells derived from human pluripotent stem cells. Stem Cell Rep. 10, 101 119. Available from: https://doi.org/ 10.1016/j.stemcr.2017.11.012. Moraes, C., Labuz, J.M., Leung, B.M., Inoue, M., Chun, T.-H., Takayama, S., 2013. On being the right size: scaling effects in designing a human-on-a-chip. Integr. Biol. 5, 1149. Available from: https://doi.org/10.1039/c3ib40040a. Motz, G.T., Eppert, B.L., Wortham, B.W., Amos-Kroohs, R.M., Flury, J.L., Wesselkamper, S.C., et al., 2010. Chronic cigarette smoke exposure primes NK cell activation in a mouse model of chronic obstructive pulmonary disease. J. Immunol. (Baltimore, MD) 1950 (184), 4460 4469. Available from: https://doi.org/10.4049/ jimmunol.0903654. Mullane, K., Williams, M., 2014. Animal models of asthma: reprise or reboot? Biochem. Pharmacol. 87, 131 139. Available from: https://doi.org/10.1016/j.bcp.2013.06.026. Murphy, S.V., Atala, A., 2014. 3D bioprinting of tissues and organs. Nat. Biotechnol. 32, 773 785. Available from: https://doi.org/10.1038/nbt.2958. Murphy, S.V., Atala, A., 2016. Regenerative Medicine Technology: On-a-Chip Applications for Disease Modeling, Drug Discovery and Personalized Medicine, Gene and Cell Therapy. CRC Press. Nadkarni, R.R., Abed, S., Draper, J.S., 2016. Organoids as a model system for studying human lung development and disease. Biochem. Biophys. Res. Commun. 473, 675 682. Available from: https://doi.org/10.1016/j.bbrc.2015.12.091. Napierska, D., Thomassen, L.C.J., Rabolli, V., Lison, D., Gonzalez, L., Kirsch-Volders, M., et al., 2009. Size-dependent cytotoxicity of monodisperse silica nanoparticles in human endothelial cells. Small 5, 846 853. Available from: https://doi.org/10.1002/ smll.200800461. Nel, A., Xia, T., Ma¨dler, L., Li, N., 2006. Toxic potential of materials at the nanolevel. Science 311, 622 627. Available from: https://doi.org/10.1126/science.1114397. Nichols, J.E., Francesca, S.L., Vega, S.P., Niles, J.A., Argueta, L.B., Riddle, M., et al., 2017. Giving new life to old lungs: methods to produce and assess whole human paediatric bioengineered lungs. J. Tissue Eng. Regen. Med. 11, 2136 2152. Available from: https://doi.org/10.1002/term.2113. Nichols, J.E., Niles, J.A., Vega, S.P., Cortiella, J., 2013. Novel in vitro respiratory models to study lung development, physiology, pathology and toxicology. Stem Cell Res. Ther. 4, S7. Available from: https://doi.org/10.1186/scrt368. Nikoli´c, M.Z., Sun, D., Rawlins, E.L., 2018. Human lung development: recent progress and new challenges. Development 145, dev163485. Available from: https://doi.org/ 10.1242/dev.163485.

175

176

CHAPTER 4 Lung-on-a-chip platforms

Orme, I.M., Basaraba, R.J., 2014. The formation of the granuloma in tuberculosis infection. Semin. Immunol. 26, 601 609. Available from: https://doi.org/10.1016/j. smim.2014.09.009. Ostedgaard, L.S., Meyerholz, D.K., Chen, J.-H., Pezzulo, A.A., Karp, P.H., Rokhlina, T., et al., 2011. The ΔF508 mutation causes CFTR misprocessing and cystic fibrosis like disease in pigs. Sci. Transl. Med. 3, 74ra24. Available from: https://doi.org/10.1126/ scitranslmed.3001868. Pampaloni, F., Reynaud, E.G., Stelzer, E.H.K., 2007. The third dimension bridges the gap between cell culture and live tissue. Nat. Rev. Mol. Cell Biol. 8, 839 845. Available from: https://doi.org/10.1038/nrm2236. Park, J.S., Chu, J.S.F., Cheng, C., Chen, F., Chen, D., Li, S., 2004. Differential effects of equiaxial and uniaxial strain on mesenchymal stem cells. Biotechnol. Bioeng. 88, 359 368. Available from: https://doi.org/10.1002/bit.20250. Park, J.Y., Ryu, H., Lee, B., Ha, D.-H., Ahn, M., Kim, S., et al., 2018. Development of a functional airway-on-a-chip by 3D cell printing. Biofabrication 11, 015002. Available from: https://doi.org/10.1088/1758-5090/aae545. Pawlowski, A., Jansson, M., Sko¨ld, M., Rottenberg, M.E., Ka¨llenius, G., 2012. Tuberculosis and HIV co-infection. PLoS Pathog. 8, e1002464. Available from: https:// doi.org/10.1371/journal.ppat.1002464. ´ ., Peces-Barba, G., 2015. Animal models of chronic Pe´rez-Rial, S., Giro´n-Martı´nez, A obstructive pulmonary disease. Arch. Bronconeumol. 51, 121 127. Available from: https://doi.org/10.1016/j.arbres.2014.06.016. Perinel, S., Pourchez, J., Leclerc, L., Avet, J., Durand, M., Pre´voˆt, N., et al., 2017. Development of an ex vivo human-porcine respiratory model for preclinical studies. Sci. Rep. 7, 43121. Available from: https://doi.org/10.1038/srep43121. Perrin, S., 2014. Preclinical research: make mouse studies work. Nature 507, 423 425. Pifferi, M., Montemurro, F., Cangiotti, A.M., Ragazzo, V., Cicco, M.D., Vinci, B., et al., 2009. Simplified cell culture method for the diagnosis of atypical primary ciliary dyskinesia. Thorax 64, 1077 1081. Available from: https://doi.org/10.1136/ thx.2008.110940. Piirila¨, P., Sovija¨rvi, A.R.A., 1995. Crackles: recording, analysis and clinical significance. Eur. Respir. J. 8, 2139 2148. Available from: https://doi.org/10.1183/ 09031936.95.08122139. Prantil-Baun, R., Novak, R., Das, D., Somayaji, M.R., Przekwas, A., Ingber, D.E., 2018. Physiologically based pharmacokinetic and pharmacodynamic analysis enabled by microfluidically linked organs-on-chips. Annu. Rev. Pharmacol. Toxicol. 58, 37 64. Available from: https://doi.org/10.1146/annurev-pharmtox-010716-104748. Price, A.P., England, K.A., Matson, A.M., Blazar, B.R., Panoskaltsis-Mortari, A., 2010. Development of a decellularized lung bioreactor system for bioengineering the lung: the matrix reloaded. Tissue Eng., A 16, 2581 2591. Available from: https://doi.org/ 10.1089/ten.tea.2009.0659. Rosen, B.H., Chanson, M., Gawenis, L.R., Liu, J., Sofoluwe, A., Zoso, A., et al., 2018. Animal and model systems for studying cystic fibrosis. J. Cyst. Fibros. 17, S28 S34. Available from: https://doi.org/10.1016/j.jcf.2017.09.001. Rosenfeld, L., Cheng, Y., Rao, J., Tang, S.K., 2014. Rapid detection of tuberculosis using droplet-based microfluidics. Presented at the Microfluidics, BioMEMS, and Medical Microsystems XII, International Society for Optics and Photonics, p. 897611.

References

Ruggeri, B.A., Camp, F., Miknyoczki, S., 2014. Animal models of disease: pre-clinical animal models of cancer and their applications and utility in drug discovery. Biochem. Pharmacol. 87, 150 161. Rutland, J., Cole, P., 1980. Non-invasive sampling of nasal cilia for measurement of beat frequency and study of ultrastructure. Lancet 316, 564 565. Available from: https:// doi.org/10.1016/S0140-6736(80)91995-9. Sandler, A., Gray, R., Perry, M.C., Brahmer, J., Schiller, J.H., Dowlati, A., et al., 2006. Paclitaxel carboplatin alone or with bevacizumab for non small-cell lung cancer. N. Engl. J. Med. 355, 2542 2550. Sava, P., Ramanathan, A., Dobronyi, A., Peng, X., Sun, H., Ledesma-Mendoza, A., et al., 2017. Human pericytes adopt myofibroblast properties in the microenvironment of the IPF lung. JCI Insight 2. ,https://doi.org/10.1172/jci.insight.96352. Seo, J., Huh, D., 2019. Chapter 7 Microphysiological models of human organs: a case study on microengineered lung-on-a-chip systems. In: Borenstein, J.T., Tandon, V., Tao, S.L., Charest, J.L. (Eds.), Microfluidic Cell Culture Systems, second ed. In: Micro and Nano Technologies. Elsevier 187 208.https://doi.org/10.1016/B978-0-12813671-3.00007-4 Shigemura, N., Okumura, M., Mizuno, S., Imanishi, Y., Matsuyama, A., Shiono, H., et al., 2006. Lung tissue engineering technique with adipose stromal cells improves surgical outcome for pulmonary emphysema. Am. J. Respir. Crit. Care Med. 174, 1199 1205. Available from: https://doi.org/10.1164/rccm.200603-406OC. Siegel, R.L., Miller, K.D., Jemal, A., 2017. Cancer statistics, 2017. CA Cancer J. Clin. 67, 7 30. Skardal, A., Devarasetty, M., Kang, H.-W., Mead, I., Bishop, C., Shupe, T., et al., 2015. A hydrogel bioink toolkit for mimicking native tissue biochemical and mechanical properties in bioprinted tissue constructs. Acta Biomater. 25, 24 34. Available from: https://doi.org/10.1016/j.actbio.2015.07.030. Skardal, A., Shupe, T., Atala, A., 2016. Organoid-on-a-chip and body-on-a-chip systems for drug screening and disease modeling. Drug Discov. Today 21, 1399 1411,Pathophysiology models for drug discovery. Available from: https://doi.org/10.1016/j.drudis.2016.07.003. Skardal, A., Murphy, S.V., Devarasetty, M., Mead, I., Kang, H.-W., Seol, Y.-J., et al., 2017. Multi-tissue interactions in an integrated three-tissue organ-on-a-chip platform. Sci. Rep. 7, 8837. Available from: https://doi.org/10.1038/s41598-017-08879-x. Skardal, A., Shupe, T., Atala, A., 2019. Chapter 44 Body-on-a-chip: regenerative medicine for personalized medicine. In: Atala, A., Lanza, R., Mikos, A.G., Nerem, R. (Eds.), Principles of Regenerative Medicine, Third Edition Academic Press, Boston, MA, pp. 769 786. Available from: https://doi.org/10.1016/B978-0-12-809880-6.00044-8. Skolimowski, M., Abeille, F., Nielsen, M.-W., Lopacinska, J.D., Molin, S., Taboryski, R., et al., 2012a. Microfluidic model of cystic fibrosis bronchi. In: Fifteenth International Conference on Miniaturized Systems for Chemistry and Life Science, vol. 3. 2011, Curran Associates, Inc., Seattle, WA, pp. 2 118 Skolimowski, M., Weiss Nielsen, M., Abeille, F., Skafte-Pedersen, P., Sabourin, D., Fercher, A., et al., 2012b. Modular microfluidic system as a model of cystic fibrosis airways. Biomicrofluidics 6, 1 11. Available from: https://doi.org/10.1063/1.4742911. Snoeck, H.-W., 2015. Modeling human lung development and disease using pluripotent stem cells. Development 142, 13 16. Available from: https://doi.org/10.1242/ dev.115469.

177

178

CHAPTER 4 Lung-on-a-chip platforms

Sochol, R.D., Gupta, N.R., Bonventre, J.V., 2016. A role for 3D printing in kidney-on-achip platforms. Curr. Transplant. Rep. 3, 82 92. Available from: https://doi.org/ 10.1007/s40472-016-0085-x. Stoltz, D.A., Meyerholz, D.K., Pezzulo, A.A., Ramachandran, S., Rogan, M.P., Davis, G.J., et al., 2010. Cystic fibrosis pigs develop lung disease and exhibit defective bacterial eradication at birth. Sci. Transl. Med. 2, 29ra31. Available from: https://doi.org/ 10.1126/scitranslmed.3000928. Stucki, A.O., Stucki, J.D., Hall, S.R.R., Felder, M., Mermoud, Y., Schmid, R.A., et al., 2015. A lung-on-a-chip array with an integrated bio-inspired respiration mechanism. Lab. Chip 15, 1302 1310. Available from: https://doi.org/10.1039/c4lc01252f. Sun, X., Sui, H., Fisher, J.T., Yan, Z., Liu, X., Cho, H.-J., et al., 2010. Disease phenotype of a ferret CFTR-knockout model of cystic fibrosis. J. Clin. Invest. 120, 3149 3160. Available from: https://doi.org/10.1172/JCI43052. Takebe, T., Zhang, B., Radisic, M., 2017. Synergistic engineering: organoids meet organson-a-chip. Cell Stem Cell 21, 297 300. Available from: https://doi.org/10.1016/j. stem.2017.08.016. Tavana, H., Kuo, C.-H., Lee, Q.Y., Mosadegh, B., Huh, D., Christensen, P.J., et al., 2010. Dynamics of liquid plugs of buffer and surfactant solutions in a micro-engineered pulmonary airway model. Langmuir 26, 3744 3752. Available from: https://doi.org/ 10.1021/la903038a. Tavana, H., Zamankhan, P., Christensen, P.J., Grotberg, J.B., Takayama, S., 2011. Epithelium damage and protection during reopening of occluded airways in a physiologic microfluidic pulmonary airway model. Biomed. Microdevices 13, 731 742. Available from: https://doi.org/10.1007/s10544-011-9543-5. Thodeti, C.K., Matthews, B., Ravi, A., Mammoto, A., Ghosh, K., Bracha, A.L., et al., 2009. TRPV4 channels mediate cyclic strain induced endothelial cell reorientation through integrin-to-integrin signaling. Circ. Res. 104, 1123 1130. Available from: https://doi.org/10.1161/CIRCRESAHA.108.192930. Thomas, B., Rutman, A., Hirst, R.A., Haldar, P., Wardlaw, A.J., Bankart, J., et al., 2010. Ciliary dysfunction and ultrastructural abnormalities are features of severe asthma. J. Allergy Clin. Immunol. 126. Available from: https://doi.org/10.1016/j. jaci.2010.05.046722-729.e2. Thorne, D., Adamson, J., 2013. A review of in vitro cigarette smoke exposure systems. Exp. Toxicol. Pathol. 65, 1183 1193. Available from: https://doi.org/10.1016/j. etp.2013.06.001. Tjin, G., White, E.S., Faiz, A., Sicard, D., Tschumperlin, D.J., Mahar, A., et al., 2017. Lysyl oxidases regulate fibrillar collagen remodelling in idiopathic pulmonary fibrosis. Dis. Model. Mech. 10, 1301 1312. Available from: https://doi.org/10.1242/dmm.030114. Tschumperlin, D.J., Margulies, S.S., 1998. Equibiaxial deformation-induced injury of alveolar epithelial cells in vitro. Am. J. Physiol. 275, L1173 L1183. Available from: https://doi.org/10.1152/ajplung.1998.275.6.L1173. Tschumperlin, D.J., Oswari, J., Margulies, A.S.S., 2000. Deformation-induced injury of alveolar epithelial cells. Am. J. Respir. Crit. Care Med. 162, 357 362. Available from: https://doi.org/10.1164/ajrccm.162.2.9807003. United States Surgeon General, 2014. The Health Consequences of Smoking --50 Years of progress: A Report of the Surgeon General: (510072014-001). https://doi.org/10.1037/ e510072014-001

References

Upadhyay, S., Palmberg, L., 2018. Air-liquid interface: relevant in vitro models for investigating air pollutant-induced pulmonary toxicity. Toxicol. Sci. 164, 21 30. Verhulsel, M., Vignes, M., Descroix, S., Malaquin, L., Vignjevic, D.M., Viovy, J.-L., 2014. A review of microfabrication and hydrogel engineering for micro-organs on chips. Biomaterials 35, 1816 1832. Available from: https://doi.org/10.1016/j. biomaterials.2013.11.021. Vlahakis, N.E., Schroeder, M.A., Limper, A.H., Hubmayr, R.D., 1999. Stretch induces cytokine release by alveolar epithelial cells in vitro. Am. J. Physiol.-Lung Cell. Mol. Physiol. 277, L167 L173. Available from: https://doi.org/10.1152/ajplung.1999.277.1. L167. Wagner, D.E., Bonenfant, N.R., Parsons, C., Sokocevic, D., Brooks, E.M., Borg, Z.D., et al., 2014. Comparative decellularization and recellularization of normal versus emphysematous human lungs. Biomaterials 35, 3281. Available from: https://doi.org/ 10.1016/j.biomaterials.2013.12.103. Wang, H., Naghavi, M., Allen, C., Barber, R.M., Bhutta, Z.A., Carter, A., et al., 2016. Global, regional, and national life expectancy, all-cause mortality, and cause-specific mortality for 249 causes of death, 1980 2015: a systematic analysis for the Global Burden of Disease Study 2015. The Lancet 388, 1459 1544. Available from: https:// doi.org/10.1016/S0140-6736(16)31012-1. Wang, J.D., Douville, N.J., Takayama, S., ElSayed, M., 2012. Quantitative analysis of molecular absorption into PDMS microfluidic channels. Ann. Biomed. Eng. 40, 1862 1873. Available from: https://doi.org/10.1007/s10439-012-0562-z. Wang, Y., Wrennall, J.A., Cai, Z., Li, H., Sheppard, D.N., 2014. Understanding how cystic fibrosis mutations disrupt CFTR function: from single molecules to animal models. Int. J. Biochem. Cell Biol. 52, 47 57. Available from: https://doi.org/10.1016/j. biocel.2014.04.001. Wang, Y., Zhang, Z., Garbow, J.R., Rowland, D.J., Lubet, R.A., Sit, D., et al., 2009. Chemoprevention of lung squamous cell carcinoma in mice by a mixture of Chinese herbs. Cancer Prev. Res. (Phila. Pa.) 2, 634 640. Available from: https://doi.org/ 10.1158/1940-6207.CAPR-09-0052. Waters, C.M., Roan, E., Navajas, D., 2012. Mechanobiology in lung epithelial cells: measurements, perturbations, and responses. Compr. Physiol. 2, 1 29. Available from: https://doi.org/10.1002/cphy.c100090. Whitcutt, M.J., Adler, K.B., Wu, R., 1988. A biphasic chamber system for maintaining polarity of differentiation of culture respiratory tract epithelial cells. In Vitro Cell. Dev. Biol. 24, 420 428. Available from: https://doi.org/10.1007/BF02628493. Wilkinson, D.C., Alva-Ornelas, J.A., Sucre, J.M.S., Vijayaraj, P., Durra, A., Richardson, W., et al., 2017. Development of a three-dimensional bioengineering technology to generate lung tissue for personalized disease modeling. Stem Cells Transl. Med. 6, 622 633. Available from: https://doi.org/10.5966/sctm.2016-0192. Wu, W., Hansen, C.J., Arago´n, A.M., Geubelle, P.H., White, S.R., Lewis, J.A., 2010. Direct-write assembly of biomimetic microvascular networks for efficient fluid transport. Soft Matter 6, 739 742. Available from: https://doi.org/10.1039/B918436H. Xu, Z., Gao, Y., Hao, Y., Li, E., Wang, Y., Zhang, J., et al., 2013. Application of a microfluidic chip-based 3D co-culture to test drug sensitivity for individualized treatment of lung cancer. Biomaterials 34, 4109 4117. Available from: https://doi.org/10.1016/j. biomaterials.2013.02.045.

179

180

CHAPTER 4 Lung-on-a-chip platforms

Yamaya, M., Hosoda, M., Suzuki, T., Yamada, N., Sasaki, H., 2002. Human airway epithelial cell culture. Methods Mol. Biol. (Clifton, NJ) 188, 7 16. Available from: https:// doi.org/10.1385/1-59259-185-X:07. Yang, X., Li, K., Zhang, X., Liu, C., Guo, B., Wen, W., et al., 2018. Nanofiber membrane supported lung-on-a-chip microdevice for anti-cancer drug testing. Lab Chip 18, 486 495. Available from: https://doi.org/10.1039/C7LC01224A. Yi, H.-G., Lee, H., Cho, D.-W., 2017. 3D printing of organs-on-chips. Bioengineering (Basel Switz.) 4. Available from: https://doi.org/10.3390/ bioengineering4010010. You, M., Bergman, G., 1998. Preclinical and clinical models of lung cancer chemoprevention. Hematol. Oncol. Clin. North Am. 12, 1037 1053. Available from: https://doi.org/ 10.1016/S0889-8588(05)70040-X. Young Park, J., Shim, J.-H., Choi, S.-A., Jang, J., Kim, M., Hwa Lee, S., et al., 2015. 3D printing technology to control BMP-2 and VEGF delivery spatially and temporally to promote large-volume bone regeneration. J. Mater. Chem. B 3, 5415 5425. Available from: https://doi.org/10.1039/C5TB00637F. Zepp, J.A., Zacharias, W.J., Frank, D.B., Cavanaugh, C.A., Zhou, S., Morley, M.P., et al., 2017. Distinct mesenchymal lineages and niches promote epithelial self-renewal and myofibrogenesis in the lung. Cell 170, 1134-1148.e10. Available from: https://doi.org/ 10.1016/j.cell.2017.07.034. Zhan, L., Tang, J., Sun, M., Qin, C., 2017. Animal models for tuberculosis in translational and precision medicine. Front. Microbiol. 8, 717. Available from: https://doi.org/ 10.3389/fmicb.2017.00717. Zhang, Y.S., Arneri, A., Bersini, S., Shin, S.-R., Zhu, K., Goli-Malekabadi, Z., et al., 2016. Bioprinting 3D microfibrous scaffolds for engineering endothelialized myocardium and heart-on-a-chip. Biomaterials 110, 45 59. Available from: https://doi.org/10.1016/j. biomaterials.2016.09.003. Zhou, Y., 2017. The recent development and applications of fluidic channels by 3D printing. J. Biomed. Sci. 24, 80. Available from: https://doi.org/10.1186/s12929-017-0384-2. Zhou, Y., Kang, M.-J., Jha, B.K., Silverman, R.H., Lee, C.G., Elias, J.A., 2013. Role of ribonuclease L in viral pathogen-associated molecular pattern/influenza virus and cigarette smoke-induced inflammation and remodeling. J. Immunol. (Baltimore, MD) 1950 (191), 2637 2646. Available from: https://doi.org/10.4049/jimmunol.1300082. Zosky, G.R., Sly, P.D., 2007. Animal models of asthma. Clin. Exp. Allergy 37, 973 988. Available from: https://doi.org/10.1111/j.1365-2222.2007.02740.x.

CHAPTER

Requirements for designing organ-on-a-chip platforms to model the pathogenesis of liver disease

5 Seiichi Ishida

Division of Pharmacology, National Institute of Health Sciences, Kawasaki, Japan

Introduction The liver is a multifunction organ with a unique physiological structure, which makes difficult to reproduce its physiological functions in vitro. For example, hepatocytes, the main parenchymal cells in the liver, lose their functions quickly after the cell preparation from excised liver (Hewitt et al., 2007; Khetani and Bhatia, 2008). Such dysfunction of hepatocytes in conventional in vitro cell culture makes it difficult to utilize hepatocytes for the evaluation of drug-induced liver injury, which is the most common cause of acute liver insult. The limitation of conventional in vitro cell culture system raises unmet needs in pharmaceutical industry. These are summarized as follows from the aspects of pharmacokinetic evaluation and toxicity evaluation. Pharmacokinetic side includes two points: one is long-term maintenance of metabolic and transport activity and the other is evaluation of biliary excretion and sinusoidal backflux. Toxicity assessment includes several issues: toxicity caused by metabolites, long-term repeated exposure, toxicity due to cholestasis, liver fibrosis, and toxicity due to immune response. Various culture techniques have been developed to solve these unmet needs by reproducing complex organs in an in vitro culture device, and the liver is no exception (LeCluyse et al., 2012; Underhill and Khetani, 2018). Organ-on-a-chip is one of such devices, which is enable cells derived from human organs to be cultured in small chambers connected by medium flow in a microfabricated channel (Fig. 5.1). These devices have been developed over the past two decades as “micro cell culture analogs” (Viravaidya et al., 2004). Various platforms (Zhang et al., 2018) and culturing methods (Kimura et al., 2018; Tetsuka et al., 2017) have been developed to date. Platforms that mimic the functions of organs in vitro are also termed microphysiological systems (MPS). In the pharmaceutical industry, body-on-a-chip or MPS is expected as an assay system for early prediction of drug-induced liver injury and drug development for the treatment of liver disease, because such culture system enables long-term culture and fat Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00005-X © 2020 Elsevier Inc. All rights reserved.

181

182

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

t Bile duc

d flow Main bloo Target organ

lism

Metabo

In

Out

l vein Por ta

FIGURE 5.1 Conceptual image of organ(s)-on-a-chip.

accumulation of hepatocytes by devising the culture medium and cell adhesion surface, and because there is a possibility to restore the bile excretion capacity of hepatocytes by improving cell cell adhesion. In addition, by coculturing with nonparenchymal cells such as stellate cells or Kupffer cells, it is possible to reproduce liver fibrosis and hepatotoxicity due to immune response. In order to construct such a culture system, it is important to utilize not only human cryo-preserved hepatocytes that have been conventionally used but also newly developed cell resources and to understand a culture substrate that provides a scaffold to which cells adhere. Only by combining them it becomes possible to reproduce the complex histology of the liver in vitro for constructing in vitro hepatic cell culture systems. Thus the chapter first outlines hepatic cell types and their functions. Next, fundamental technology considerations for constructing the culture unit are discussed, with a focus on drug-induced liver injury and a model of hepatic fibrosis. In the final part of the chapter, we will outline a model connecting a liver MPS and a small-intestine MPS as an example of an organs-on-a-chip platform.

Liver function and structure Liver function The liver is the largest internal organ in the human body, and at 1.0 1.5 kg, accounts for approximately 2% of total body weight (LeCluyse et al., 2012). The liver has multiple physiological functions, including the storage of glycogen as energy source, which is synthesized from glucose, fatty acids, and amino acids, and the synthesis of glutamine, bile acids, cholesterol and lipids, urea, albumin, and blood clotting factors (Jungermann and Kietzmann, 1996; van Grunsven, 2017). The liver processes foreign substances that flow into the hepatic portal vein after absorption by digestive organs or as detoxification products of xenobiotics (Jungermann and Kietzmann, 1996; van Grunsven, 2017), maintaining the internal environment of the body.

Liver function and structure

Hepatic lobules and sinusoids Liver tissue sections exhibit characteristic structural units termed hepatic lobules, which are organized into hexagonal columns that are 1 2 mm in diameter (LeCluyse et al., 2012). A portal triad consisting of the hepatic portal vein, hepatic artery, and bile duct is located at the apex of the hexagon, and the central vein penetrates the center of the hexagonal column. In a healthy individual at rest, around 1 L of blood passes through these structures every minute (van Grunsven, 2017) (Fig. 5.2A). Blood from the hepatic portal vein in the hepatic lobule flows into liver-specific capillaries known as sinusoids that flow toward the central vein and outward into systemic circulation. Sinusoids are composed of highly permeable sinusoidal endothelial cells surrounded by hepatocytes, the main parenchymal cell in the liver. A gap known as the space of Disse lies between the sinusoidal endothelial cells and parenchymal cells and contains various nonparenchymal cells (Fig. 5.2B). Hepatic parenchymal cells account for

(A) Central vein Hepatic artery Bile duct Portal vein

Sinusoid

Space of Disse

(B)

Sinusoid

Portal traid

Hepatocyte

Kupffercell

Hepatic artery

Central vein

Portal vein Bile duct

Stellate cell

Bile canaliculi

Endothelial cell

FIGURE 5.2 Liver structure. (A) Hepatic lobule. Hepatic lobules are organized into 1-2mm hexagonal columns. A portal triad (portal vein, hepatic artery, and bile duct) is located at the apex of the hexagon. The central vein penetrates the center of the hexagonal column. (B) Hepatic sinusoid. Hepatic sinusoids consist of highly permeable sinusoidal endothelial cells that are surrounded by hepatocytes, the main parenchymal cell in the liver.

183

184

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

approximately 70% of the total cell count in the liver and carry out many of the functions of this organ (Gebhardt, 1992), while nonparenchymal cells such as hepatic stellate cells are involved in immune responses and liver fibrosis.

Cellular components of the sinusoids Substances in the blood that flows through the sinusoids are taken up by the parenchymal cells that line the sinusoids and undergo biotransformation into substrates for various biochemical reactions. A fraction of the generated substance returns to the sinusoidal tract where it is stored in hepatocytes. Another fraction can be excreted into the bile canaliculi that lie between the hepatocytes for discharge into the intestinal tract along with bile. The ability of hepatocytes to take up and excrete substances gives these cells a unique polarity (Benedicto et al., 2012). While many cell types have apical and basolateral surfaces, with tight junctions at cell cell contacts (Fig. 5.3A), hepatocytes have two tight junctions at cell cell contacts and small gaps at the apical surface between the tight junctions that form the bile canaliculi (Fig. 5.3B). In hepatocytes, substances taken into the cell can be exported via two regulated excretion pathways: efflux into the sinusoid and biliary excretion into the bile canaliculi. The active uptake of highly hydrophilic substances and active excretion from cells are mediated by transporters with defined intracellular localizations to maintain hepatocyte polarity (Chu et al., 2017). Hepatocytes are involved in the biotransformation of xenobiotics, which flow into the liver from the stomach, duodenum, gall bladder, pancreas, spleen, and small intestine. Some of these substances are highly toxic and must be detoxified. Highly hydrophilic compounds cannot pass through cell membranes composed of lipids by passive diffusion, unlike hydrophobic compounds, which can diffuse across cell membranes into cells. Hepatocytes express a series of drug-metabolizing enzymes (A)

(B)

Apical surface

Tight junction

Apical surface

Tight junction

Basal membrane Basal surface

Basolateral surface

Bile canaliculi

FIGURE 5.3 A comparative scheme of (A) epithelial polarization and (B) hepatocyte polarization. Most epithelial cells have an apical domain at the cell apex. In contrast, the apical surface of hepatocytes lies between the two tight junctions at cell cell contacts and forms small gaps termed bile canaliculi into which bile is secreted.

Liver function and structure

that include oxidases and conjugating enzymes to increase the water solubility of lipophilic exogenous compounds and enable their excretion. Oxidation, reduction, and hydrolysis by phase I enzymes introduce a highly reactive functional group or polar moiety to lipophilic compounds; the reaction products then undergo conjugation by phase II enzymes of various groups, including glutathione, sulfate, glycine, and glucuronic acid, which exhibit a large molecular weight and low reactivity to promote conversion into bulky and highly polar compounds, completing the detoxification process. Importantly, biotransformation reactions associated with elimination of xenobiotics can be species-dependent, and the substrate specificity of these enzymatic reactions may differ between humans and model animals (Sharer et al., 1995; Turpeinen et al., 2007). Sinusoidal endothelial cells are vascular endothelial cells that line the sinusoid lumen and represent 15% 20% of liver cells (Gebhardt, 1992). Oxygen-poor and nutrient-rich portal vein blood and oxygen-rich arterial blood from the hepatic artery flow into the sinusoid and back into the bloodstream through the central vein. Around 80% of the blood flowing into sinusoids is supplied by the hepatic portal vein, and the remaining 20% is supplied by the hepatic artery (Jungermann and Kietzmann, 1996). Approximately 15 25 hepatic parenchymal cells line a sinusoid. Sinusoidal endothelial cells lack a basement membrane and are the most permeable endothelial cells in the body, owing to the presence of pores termed fenestrae (Wisse et al., 1985). Fenestrae are 50 200 nm in diameter and allow the passage of substances from the sinusoidal blood flow that do not exceed the pore size (Braet and Wisse, 2002). Hepatic stellate cells are nonparenchymal cells found in the space of Disse that lies between sinusoids and hepatocytes. These cells are similar in shape to stretched dendrites and account for 5% 8% of the cells in the liver (Gebhardt, 1992; Braet and Wisse, 2002).

Zonation Observation of sinusoids along the path of blood flow shows that the sinusoid cell environment is not homogeneous (Gebhardt, 1992; Kietzmann, 2017). Blood from the hepatic portal vein and hepatic artery flows mainly to the central vein while exchanging substances with hepatocytes on both sides of the sinusoid. The oxygen concentration gradient influences uptake of substances by the sinusoid. The partial pressure of oxygen is 60 65 mmHg in periportal blood but drops to 30 35 mmHg for perivenous blood (Jungermann and Kietzmann, 2000). Likewise, nutrients flowing from the portal vein and waste products from sinusoid cells generate concentration gradients along the sinusoids, affecting cell functions such as carbohydrate, amino acid, and lipid metabolism (e.g., oxidative phosphorylation and glycolysis), drug metabolism, protein synthesis, and the size and number of fenestrae in sinusoidal endothelial cells (Fig. 5.4) (van Grunsven, 2017; Kietzmann, 2017). Xenobiotic metabolism is also affected, resulting in a site-specific induction of hepatic toxicity, which is discussed in the “Site-specific

185

186

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

Zone 1

Zone 2

Zone 3

Hepatic artery

Blood flow

Portal vein

Central vein

oxygen

Gluconeogenesis Ureagenesis Fenestrae size

Glycolysis Bile acid synthesis Metabolism Fenestrae number

FIGURE 5.4 Zonation of the hepatic sinusoid, the oxygen gradient, and functional zonation. Sinusoids can be divided into the periportal region (zone 1), midlobular region (zone 2), and pericentral region (zone 3). Cell functions alter with blood flow from the periportal region to the pericentral region, owing to the concentration gradient of oxygen.

hepatic toxicity in the liver (liver zonation)” section. Alterations in cell functions following fluctuations in the concentration gradient of substances suggest a threezone division of sinusoids: periportal region (zone 1), midlobular region (zone 2), and pericentral region (zone 3) (Rappaport, 1977). Oxygen is thought to be involved in the formation of these zones, because some gene expressions are thought to be controlled by oxygen tensions in primary hepatocytes (Jungermann and Kietzmann, 1996, 2000; Soto-Gutierrez et al., 2017; Gebhardt and Matz-Soja, 2014).

Drug-induced liver injury Drug-induced liver injury is the most common cause of acute liver insult and can arise from toxicity that is not detected during the drug-development process (Lee, 2003). Accurate prediction of drug-induced liver toxicity is difficult with currently available in vitro/in vivo evaluation systems. This section outlines the limits of current nonclinical and clinical trials and expectations for in vitro cell-based assays, summarizes the characteristics of hepatocytes used in drug-induced liver injury tests, and considers how to incorporate MPS techniques as a culture method to increase predictability of drug-induced liver injury in vitro.

Drug-induced liver injury

Limitations of conventional nonclinical studies and clinical trials In nonclinical studies, hepatic injury by drug candidate compounds is commonly assessed using animals, although the accuracy of these predictions is not high (Olson et al., 2000). Biotransformation by drug-metabolizing enzymes in the liver may actually exert toxicity. Moreover, the drug-metabolizing enzymes are known to exhibit species-specific differences (Zhang et al., 2000), resulting in metabolite profiles that differ between experimental animals and humans. Several drugs have been withdrawn from the market following hepatic injury observed at the clinical or postmarketing stage that was not detected during animal testing (Vernetti et al., 2017). In contrast, some drugs that induce hepatic injury in animals do not cause liver damage in humans (Sistare et al., 2016), owing to interspecies variations in the metabolism and excretion of chemical compounds. This highlights the difficulty in making accurate predictions about drug safety from animal data alone and the need for in vitro human cell-based assays. Inaccurate safety predictions caused by the limitations of clinical testing are also a matter for concern; clinical trials typically recruit middle-aged and relatively healthy individuals except in terms of the target disease. Clinical trials have been described as too few, too narrow, too median-aged, too simple, too brief (Rogers, 1987). Insufficient numbers of subjects in clinical trials, age group bias, and simplicity of the medication protocol are common problems in clinical trial design that cause drug-induced liver damage to be overlooked, particularly for fragile populations, such as children or the elderly (Anderson, 2002). In addition, multiple diseases necessitate a multiple-drug regimen (polypharmacy), which may lead to drug-drug interactions and unexpected liver injury (Guthrie et al., 2015). Some drug combinations are tested in clinical trials, but it is impossible to comprehensively assess potential drug combinations even in nonclinical animal studies. This emphasizes the need for in vitro cell-based assays to evaluate the hepatic toxicity of drug candidates.

In vitro cell-based assay Ideally, in vitro cell-based assays would reproduce in vivo evaluation tests. These assays have been traditionally carried out using established cell lines under twodimensional culture conditions, but various cellular resources and sophisticated culture methods that better mimic the physiological environment of the liver have become available in recent years. MPS are cell cultures that mimic the biological environment, which cannot be reproduced in conventional two-dimensional cultures that often lack particular factors. This would enable the assessment of long-term, repeated-dose toxicity, hepatic injury from cholestasis, and liver-lobule site-specific toxicity, which the pharmaceutical industry cites as unmet needs that cannot be satisfied by conventional culture methods. The combination of newly developed hepatocyte resources and MPS is expected to fulfill these gaps.

187

188

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

Hepatocytes Hepatocytes are essential for evaluating drug-induced liver injury (Figs. 5.4 and 5.5). The pharmaceutical industry has been using primary hepatocytes prepared from human liver tissue since the 1990s (Houston, 1994). Currently, both the US Food and Drug Administration (http://www.fda.gov/downloads/Drugs/ GuidanceComplianceRegulatoryInformation/Guidances/ucm292362.pdf) and the European Medicines Agency (http://www.ema.europa.eu/docs/en_GB/document_library/Scientific_guideline/2012/07/WC500129606.pdf) recommend primary human hepatocytes as the gold standard for use in in vitro test systems to study cytochrome induction. The use of primary hepatocytes in in vitro testing involves several issues, including interdonor variations in hepatocyte function (Shimada et al., 1994; Roymans et al., 2005), the limited numbers of hepatocytes that can be obtained from a single donor, and the loss of the drug metabolism-related properties of primary hepatocytes upon culture (Boess et al., 2003; Kidambi et al., 2009; Bell et al., 2018; Rodrı´guez-Antona et al., 2002; Guguen-Guillouzo and Guillouzo, 2010). Therefore the HepG2 (Aden et al., 1979) and Huh7 (Nakabayashi et al., 1982) hepatocarcinoma cell lines are widely used in conventional in vitro

FIGURE 5.5 Hepatocytes of various origins. (A) Cryopreserved hepatocytes from a single donor. (B) Cryopreserved hepatocytes from a pool of donors. (C) Hepatocytes from human-liver chimeric mice (PXB-cells). (D) HepaRG cells. (E) HepatoCells. (F) ARE hepatocytes. (G I) Hepatocyte-like cells derived from induced pluripotent stem cells. Cells were cultured in the manufacturer’s recommended medium for several days after thawing the frozen cell stock at 37 C in a humidified atmosphere containing 5% CO2. ARE, AssayReady Expanded.

Drug-induced liver injury

cell-based assays. Although these cells can be cultured in a relatively inexpensive medium to obtain large number of uniform cells, the activity of enzymes involved in drug metabolism, a characteristic function of hepatocytes, is often lower than that of primary hepatocytes (Ren et al., 2018). To compensate for the low drugmetabolizing enzyme activity of HepG2 cells, attempts have been made to forcibly express relevant enzymes via plasmid- or adenovirus-mediated gene transfer (Yoshikawa et al., 2009). Since plasmids and adenoviruses have a limited number of genes that can be introduced into the cell at a given time, techniques using artificial chromosomes have been developed to allow transfer and stable expression of multiple genes, although these approaches are not yet widely used (Satoh et al., 2017). Recent reports described enhanced expression of drug-metabolizing enzymes following HepG2 cell culturing with low concentrations of the DNA methylation inhibitor 5-azacytidine for several weeks (Gailhouste et al., 2017). The gene encoding the detoxification enzyme cytochrome P450 3A4 (CYP3A4) has a CpG-rich region in the promoter region that, in cultured cells such as HepG2, is hypermethylated relative to of its levels in human primary hepatocytes. Thus treatment of HepG2 cells with 5-azacytidine could promote demethylation of the promoter region that increases CYP3A4 expression, restoring hepatocyte function to a more physiological level. HepG2 cells can display strain-specific characteristics (Hewitt and Hewitt, 2004) and varying behavior under different cell culture conditions (Doostdar et al., 1988) and periods (Wilkening and Bader, 2003) or depending on their origin (Hewitt and Hewitt, 2004), leading to interlaboratory variations. When constructing an in vitro cell-based assay, attention should be paid to this variability, as well as to differing enzyme activity, lot-to-lot variations, and the lower functional activity of hepatocyte cell lines. The HepaRG cell line was developed to address these constraints (Gripon et al., 2002). HepaRG cells are differentiated hepatic cells that are derived from hepatocyte progenitor cells and can be passaged multiple times. However, if contact inhibition is maintained without subculturing, the cells will spontaneously differentiate into liver parenchyma and biliary cells. The addition of 2% dimethyl sulfoxide to the culture medium promotes expression and activity of many drug-metabolizing enzymes to levels that are comparable to those in human hepatocytes (Gerets et al., 2012; Jennen et al., 2010; Aninat et al., 2006; Josse´ et al., 2008; Anthe´rieu et al., 2010). Relative to primary hepatocytes, HepaRG cells have similar polarity (Le Vee et al., 2013; Szabo et al., 2013) and transporter expression, with the exception of the organic anion transporting polypeptide 1B3 (OATP1B3) (Aninat et al., 2006; Anthe´rieu et al., 2010; Le Vee et al., 2006, 2013). These cells maintain stable activity for up to 4 weeks (Josse´ et al., 2008; Klein et al., 2014) and thus can be used in assays that require long-term culture. However, these cells do have some characteristics that should be considered. First, although most drug-metabolizing enzymes are expressed in HepaRG cells at typical levels, the expression of CYP3A4 is particularly high, whereas that of CYP2D6 is low relative to primary cells (Kanebratt and Andersson, 2008). In addition, as mentioned earlier,

189

190

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

OATP1B3 expression is low, as is that of the bile salt export pump (BSEP). When these drug-metabolizing enzymes and transporters contribute to the pharmacokinetics of a given compound, the actual pharmacokinetics may differ from that measured in assays, skewing the pharmacokinetic prediction of drugcandidate compounds. Second, since there is only one line of HepaRG cells, the cells must be sustainably maintained for use in in vitro cell-based assays. However, maintenance should be conducted carefully, to avoid changing cellular traits by multiple passages. Addition of 2% dimethyl sulfoxide to the culture medium enables continuous maintenance of the high drug-metabolizing enzyme activity that is the hallmark of HepaRG. In cocultures, high concentrations of dimethyl sulfoxide in the medium may affect cellular proliferation and characteristics. As the differentiation of HepaRG cells is reversible (Legendre et al., 2009), when differentiated HepaRG cells are cultured in a dimethyl sulfoxide-free medium, they enter a cell proliferation cycle, and the expression of drugmetabolizing enzymes and other proteins decreases (Legendre et al., 2009). Therefore the need to promote cell proliferation that can be achieved by the addition of dimethyl sulfoxide must be balanced against the need to maintain high hepatocyte activity. Multiple attempts have been made to immortalize human hepatocytes over the long term (Schippers et al., 1997; Tsuruga et al., 2008), but immortalization can lead to decreased metabolic activity, perhaps through the acquired proliferative activity (Nyberg et al., 1994; Guillouzo et al., 2007). Systems that regulate cell proliferation have been developed that separate the proliferative phase from the culture period; during the latter, relatively high metabolic activity is stably maintained (Zuo et al., 2017; Levy et al., 2015). The Corning HepatoCell line was established by introducing the simian virus 40 large T antigen into the parental cell line and screening and selecting cells with high cytochrome function. To induce differentiation to a mature hepatocyte phenotype prior to cryopreservation, the immortalizing gene was removed from the cells (Zuo et al., 2017). Assay-Ready Expanded hepatocytes are primary human hepatocytes transduced with the E6 and E7 human papillomavirus genes that can be expanded in the presence of oncostatin M. The removal of oncostatin M induces differentiation of the cells into metabolically functional, polarized hepatocytes that have functional bile canaliculi (Levy et al., 2015). These methods address proliferation limitations in primary hepatocytes by enabling expansion of immortalized cells. A system in which human hepatocytes are transplanted into mice to develop a humanized liver in the murine body has been established (Dandri et al., 2001; Mercer et al., 2001). PXB-mice are chimeric mice developed by transplanting human hepatocytes into severe combined immunodeficiency mice with liver failure induced by hepatocyte-targeted expression of the urokinase-type plasminogen activator gene. Hepatic replacement in PXB-mice is more than 80% (Tateno et al., 2004). PXB-mice are a useful humanized animal model (Watari et al., 2018) and a good resource for preparing PXB hepatocytes (Tateno et al., 2013; Watari et al., 2018). Hu-Liver cells can be isolated from chimeric mice that have

Drug-induced liver injury

a humanized liver after transplantation of human hepatocytes into mouse albumin enhancer/promoter-driven herpes simplex virus type 1 thymidine kinase transgenic NOG mice (Hasegawa et al., 2011). Hu-Liver cells display similar characteristics to primary human hepatocytes (Uehara et al., 2019). Contamination by mouse hepatocytes hindered the practical use of both systems early in development, but subsequent improvements allowed these systems to be used in assays to assess drug metabolism by humanized hepatocytes. Transplanted human hepatocytes can be expanded approximately 1000-fold in humanized liver from these chimeric mice. Thus passage of hepatocytes through chimeric mice allows the generation of sufficient numbers of hepatocytes from a single lot, increasing experimental reproducibility. Hepatocytes derived from humanized-liver chimeric mice can be prepared without freezing, and cell isolation can be regularly planned with conveniently stable cell distribution. In addition, cell preparation procedures can commence immediately after removal of the liver, enabling convenient control of cells, unlike isolation of hepatocytes from liver tissue samples obtained during surgery. Hepatocytes derived from humanized liver chimeric mice have cellular functions, such as the activities of drug-metabolizing enzymes, which are equivalent to those of human hepatocytes (Uehara et al., 2019). In addition, these cells can be cultured for extended periods while maintaining or activating drug metabolic activity (Watari et al., 2018). The differentiation of hepatocytes from pluripotent stem cells has ethical implications, whereas differentiation from somatic stem cells poses challenges associated with cell specificity (Snykers et al., 2009). Human- induced pluripotent stem cells (iPSC) are widely used to establish pluripotent stem cells from human somatic cells (Takahashi et al., 2007), and can be established from relatively abundant sources such as blood cells and skin fibroblasts. The resulting stem cells will reflect donor-specific differences in genetic profiles and disease status (Raab et al., 2014). Treatment of iPSC cells with differentiation-inducing factors promotes stepwise differentiation into definitive endoderm cells, hepatoblast-like cells, and finally hepatocyte-like cells (Takayama et al., 2017) (Fig. 5.6). However, hepatocyte-like cells induced to differentiate by such methods are

bFGF Activin A

iPS cells

BMP4 FGF4

HGF HGF Oncostatin M EGF

Definitive Hepatoblast endoderm cells like cells

Hepatocytelike cells

FIGURE 5.6 Schematic procedure for differentiation of human iPSC into hepatocyte-like cells. Treatment of iPSC cells with differentiation-inducing factors promotes stepwise differentiation into definitive endoderm cells, hepatoblast-like cells, and hepatocyte-like cells. iPSC, Induced pluripotent stem cells.

191

192

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

immature and have characteristics that more closely resemble those of fetal or neonatal hepatocytes, such as higher expression levels of certain genes than in adult human hepatocytes (Schwartz et al., 2014). Therefore when comparison is made under the same conditions as human hepatocytes, gene characteristics to fetal hepatocytes are expressed sometimes high, whereas genes expressed in human adult hepatocytes are relatively low. Some phase II enzymes and transporters are expressed at levels equivalent to those of human adult hepatocytes, so iPSCs can still represent an important cellular resource in in vitro cell-based assays (Fig. 5.7).

Microphysiological systems for drug-induced liver injury In vitro cell-based assays to predict drug-induced liver injury by chemical compounds have long been performed, but conventional test systems of cultured human hepatocytes often exhibit insufficient predictive value. MPS has been developed to meet the unmet needs that the pharmaceutical industry faces in in vitro assay systems. In this section, we outline how MPS can be utilized to solve these issues by focusing on four different applications: long-term repeateddose toxicity tests, assessment of toxicity by cholestasis, site-specific hepatic toxicity in the liver (liver zonation), and idiosyncratic drug-induced liver injury. Liver fibrosis, in which the liver stellate cells are involved, is described in the “Liver fibrosis” section.

Long-term repeated-dose toxicity tests To construct an in vitro cell-based assay that corresponds to an in vivo repeateddose toxicity test, cultured hepatocytes must remain functional for long periods, which can be difficult to achieve using conventional culture methods (den Braver-Sewradj et al., 2016; Griffin and Houston, 2005). Repeated-dose toxicity studies are conducted not to observe acute cytotoxicity from high concentrations of compounds but to observe accumulated cytotoxicity by continued exposure to low concentrations of compounds over a prolonged period. Thus to mimic in vivo experiments and to obtain accurate repeated-dose toxicity results, cultured hepatocytes used in in vitro assays must maintain their cell activity for at least 2 4 weeks. Maintenance of drug-metabolizing activity is particularly important, since toxicity and liver injury may arise from metabolites produced in the body rather than the administered compound (Thompson et al., 2016). Thus high metabolic activity is required to produce sufficient quantities of a metabolite (at least 10% of the parent compound) to assess the cellular effects of a metabolite. Furthermore, many drug candidates have reduced dosing requirements and therefore may exhibit slow clearance rates (Di and Obach, 2015). The cells must be exposed to these compounds for long periods so that the effect of metabolites can be observed. However, primary human hepatocytes cannot be cultivated over long periods of time (den Braver-Sewradj et al., 2016; Griffin and Houston, 2005), and hepatic drug metabolic activity resembling that of the in vivo situation often

FIGURE 5.7 Comparison of gene expression in human cryopreserved hepatocytes and human iPSC-derived hepatocytes. RNA was labeled and hybridized to measure relative gene expression by Genopal gene expression macroarray (Mitsubishi Chemical, Tokyo). iPSC, Induced pluripotent stem cells.

194

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

diminishes after several days of two-dimensional culturing (Boess et al., 2003; Kidambi et al., 2009; Bell et al., 2018, Watari et al., 2018; Rodrı´guez-Antona et al., 2002; Guillouzo et al., 2007). Therefore cells and methods suitable for long-term culturing have been developed. MPS can serve as a test platform for repeated-dose toxicity tests and circumvents issues associated with conventional two-dimensional culture systems. As described earlier, HepaRG cells tolerate cultivation for 4 weeks while maintaining activity similar to that of human primary hepatocytes even after induction of differentiation (Josse´ et al., 2008; Klein et al., 2014). These cells maintain high drug metabolic activity and are thus suitable for use in assays requiring long-term culture, since the toxicity of both the parent compound and its metabolites can be tested (Lu¨bberstedt et al., 2011). However, differences in the activity of drug-metabolizing enzymes and transporter expression, together with the availability of only one suitable cell line, limit the use of MPS to test liver toxicity. Toxicity tests that reflect interindividual differences are valuable, but again, limited by constraints on long-term culture and preservation of activity by primary human hepatocytes using conventional culture methods. Medium is an important basis of cell culturing. Cellartis Power Primary HEP Medium (TaKaRa Bio, Kusatsu, Japan) was developed to enable long-term culture and has been shown to maintain the characteristic hexagonal shape of human hepatocytes after 28 days of culturing (Fig. 5.8, right), unlike conventional hepatocyte maintenance medium (Fig. 5.8, left). Likewise, iPSC-derived hepatocyte-like cells can be cultured for extended periods in Cellartis Enhanced hiPS-HEP Long-Term Maintenance Medium (TaKaRa Bio), maintaining their metabolic activity. This culture method also enhances formation of capillary bile ducts (Fig. 5.9A) and enables intracellular accumulation of test compounds (Fig. 5.9B).

FIGURE 5.8 Long-term maintenance of hepatocyte cultures. Human cryopreserved hepatocytes were cultured in standard medium (left) and in Cellartis Power Primary HEP Medium (TaKaRa Bio, Kusatsu, Japan; right) for 28 days. Cell morphology was observed with a phasecontrast microscope.

Drug-induced liver injury

FIGURE 5.9 Long-term culture of human iPSC-derived hepatocytes. Human iPSC-derived hepatocytes were cultured in Cellartis Power Primary HEP Medium (TaKaRa Bio, Kusatsu, Japan) for 25 days. (A) Observation scheme of bile canaliculi formation. Fluorescein diacetate was taken up into cells and metabolized to fluorescein, then secreted into the bile canaliculi by the MRP2 transporter. (B) Cells were observed with a phase-contrast microscope (left) and with a fluorescence microscope (right). iPSC, Induced pluripotent stem cells.

Coculture of human hepatocytes with stromal cells stabilizes the cellular functions of hepatocytes (Bhatia et al., 1999; Guillouzo, 1998). To construct an MPS for coculturing hepatocytes, dot-shaped micropatterns are formed on the surface of the culture dish using microfabrication technology, and hepatocytes are then seeded in the dotted region coated with collagen and surrounded by 3T3-J2 murine embryonic fibroblasts. This method enables long-term culture of hepatocytes and maintains cell function (Khetani and Bhatia, 2008; Khetani et al., 2013; Wang et al., 2010). Since hepatocyte function can be maintained for 7 days without changing the medium in the micropatterned coculture, accurate evaluation of drug metabolite clearance is possible (Wang et al., 2010). Coculture of hepatocyte spheroids with 3T3-Swiss Albino mouse embryo cells as feeder cells on Cell-able plates (Transparent Inc., Inzai, Japan) was also shown to maintain drugmetabolizing activity of human liver parenchymal cells over 21 days (Ohkura et al., 2014). With this culture method, half-maximal inhibitory concentrations of fialuridine, a compound known to exert toxicity through cellular accumulation of metabolic products, were shown to decrease from 18.1 μM after 7 days of culturing to 0.9 μM after 21 days, similar to the values reported in clinical studies (Ogihara et al., 2015). Thus coculture is a useful approach for constructing MPS that enable long-term exposure to assess drug-induced liver injury.

195

196

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

The extracellular matrix (ECM) comprises several biomaterials, including collagens, proteoglycans, and glycosaminoglycans (Rojkind and Ponce-Noyola, 1982), and controls the dynamic biochemical, physicochemical, and mechanostructural environment to regulate various cell behaviors such as adhesion, proliferation, and differentiation (Lane et al., 2014; Watt and Huck, 2013; Kim et al., 2012). Addition of ECM constituents to hepatocyte cultures substantially improves hepatocyte functions (Kocarek et al., 1993; Moghe et al., 1996; Musat et al., 1993), as has been shown by the similarity of actin filament organization of rat primary hepatocytes cultured in the presence of ECM to that of intact liver (Musat et al., 1993). Various polymer-based biomaterials have been developed to better reflect the cellular microenvironment in in vitro cultures (Kim et al., 2012). One such biomaterial, collagen, has also been used as a substrate to mimic the growth environment of native ECM (Takezawa et al., 2007a,b). Sandwiching and culturing hepatocytes between gelled layers of collagen preserve cell morphology, increase cell survival, and retain hepatocyte function across extended culture periods (Dunn et al., 1991, 1992). Matrigel extracted from Engelbreth-Holm-Swarm mouse sarcoma tissue that is rich in ECM proteins, such as laminin, type IV collagen, heparin sulfate proteoglycan, and numerous growth factors, has also been used as an ECM constituent instead of collagen. Collagen vitrigel membranes, generated as transparent membranes by vitrification of type I collagen gel, are easy to handle, have high mechanical strength, and promote cell adhesion (Takezawa et al., 2004). Hepatocytes derived from humanliver chimeric mice can be maintained on collagen vitrigel membranes for 2 3 weeks, while maintaining expression and activity of drug-metabolizing enzymes, unlike hepatocytes in conventional two-dimensional cultures (Watari et al., 2018). These examples demonstrate the value of ECM components as culture substrates when constructing MPS that support maintenance of cell activity over extended culture periods.

Assessment of toxicity by cholestasis Drug-induced liver injury can be classified as hepatocellular, mixed, or cholestatic injury (Be´nichou, 1990). Thus in vitro evaluation systems are also needed for cholestatic-type liver injury, particularly since cholestasis may represent nearly 50% of total liver injuries (Bjo¨rnsson and Olsson, 2005). Drug-induced cholestatic injury is thought to arise from inhibition of bile acid excretion from hepatocytes and subsequent bile acid accumulation following treatment with a hepatotoxic compound (Stieger et al., 2000; Fattinger et al., 2001; Byrne et al., 2002; Kostrubsky et al., 2006). Predictions of cholestatic-type drug liver injury were typically based on the results of membrane vesicle assays, in which the BSEP is embedded in membranes in a cell-free in vitro system (Morgan et al., 2010, 2013; Dawson et al., 2012; Pedersen et al., 2013; Ko¨ck et al., 2014). However, membrane vesicle assays lack key components of drug metabolism systems, particularly the bile acid excretion transporter, and thus likely do not provide an accurate reflection of cholestatic-type drug-induced liver injury (Dawson et al., 2012). Moreover,

Drug-induced liver injury

cholestatic liver injury exhibits species-specific differences because of varying bile acid composition in rodents and humans (Marion et al., 2012), requiring in vitro cell-based assays that use human hepatocytes. Several structural considerations must be taken into account when developing assays to assess the potential of a compound to induce cholestatic injury. Since bile acids are excreted mainly in the bile canaliculi, these structures must form efficiently on the apical surface of the cell between the two tight junctions that are formed between adjacent hepatocytes. Transporters responsible for the excretion of bile acids and xenobiotics must also be properly oriented with respect to this apical plane. However, when the hepatocytes are isolated from liver tissue, the orientation of the transporter to the bile canaliculus is lost and biliary excretion disappears (Bow et al., 2008); sandwich cultures of hepatocytes restore the polarity and excretion ability of transporters (Swift et al., 2010; Pfeifer et al., 2013). This shows that using an appropriate ECM as a culture substrate is important in constructing MPS to evaluate cholestasis-type drug injury.

Site-specific hepatic toxicity in the liver (liver zonation) Many compounds are known to induce local injury within the liver (Gebhardt, 1992), most notably acetaminophen. Nearly half of all cases of drug-induced liver damage are attributable to acetaminophen overdose (Yoon et al., 2016). In therapeutic doses, acetaminophen is converted into a glucuronide conjugate of low toxicity around the portal region (Araya et al., 1986) or to a sulfate conjugate in the central vein region (Araya et al., 1986; Pang et al., 1988). High doses increase the concentration of active metabolites, leading to cell death in the vicinity of the central vein (Bartolone et al., 1989; Birge et al., 1990; Bruno et al., 1991; Placke et al., 1987). This site-specific acetaminophen toxicity is thought to relate to CYP2E1 expression by hepatocytes located around the central vein (Anundi et al., 1993). Oxygen is an important factor in controlling zone-dependent function and gene expression in hepatocytes (Jungermann and Kietzmann, 2000) and must also be taken into account in the development of MPS (Soto-Gutierrez et al., 2017). Allen and Bhatia observed an oxygen concentration gradient in microfluidic bioreactors and showed that acetaminophen-induced liver injury is caused by low oxygen partial pressure corresponding to that seen in the perivenous zone 3 (Allen and Bhatia, 2003; Allen et al., 2005). Recently, studies using a liver acinus MPS reported that the cell culture environments of zones 1 and 3 can be reproduced in small spaces by regulating the flow rate of the culture medium, and that CYP2E1 expression levels vary by local oxygen concentrations and that expression levels that mimic those in zone-specific liver disorders can be achieved (Lee-Montiel et al., 2017). Zonation was reconstituted in vitro using two chambers with different medium reflux rates, and a vascularized liver acinus MPS reconstructed continuous physiological zonation in one reactor (Li et al., 2018). As oxygen governs position-specific hepatocyte function and gene expression in the liver, it is critical to developing MPS that reproduce site-specific functions of hepatic lobules.

197

198

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

Idiosyncratic drug-induced liver injury Idiosyncratic drug-induced liver injury is a toxicity that appears in a small number of populations independent to the drug dose or route, and to the patient genetic or ethnical back ground, and is observed when the drug is administered to a large number of patients after marketing. The involvement of the immune system is suggested because the time to the onset of toxicity often takes several weeks or more from the start of dosing (Kaplowitz, 2005). It is desirable to be able to predict at an early stage of drug development, as it is a serious toxicity that appears for the first time after marketing and causes the withdrawal of the drug from the market. Due to the long time to the onset of toxicity and the involvement of the immune system, coculture with nonparenchymal cells involved in the immune response along with hepatic parenchymal cells is considered necessary to evaluate such toxicity. Researchers from Organovo and Roche have used the bioprinter to reconstruct liver tissue like structures that have been successfully used to find drug-induced liver damage not found in standard preclinical models (Nguyen et al., 2016). They used primary human parenchymal (hepatocyte) and nonparenchymal (endothelial and hepatic stellate) cells obtained from patients to create a three-dimensional liver-like tissue model by bioprinting technology. When this model was exposed to trovafloxacin, toxicity was observed at clinically relevant doses. Trovafloxacin is a drug that has not been found to cause severe liver damage in conventional nonclinical studies and has been withdrawn from the market due to deaths of patients taking it after it has been marketed (Lazarczyk et al., 2001). Also, Verbetti et al. have detected the toxicity of trovafloxacin by using a human, 3D, microfluidic, four-cell, sequentially layered, self-assembly liver model (Vernetti et al., 2016). They cocultured primary human hepatocytes along with human endothelial (EA.hy926), immune (U937), and stellate (LX-2) cells in physiological ratios. As a result, they have succeeded in observing cytotoxicity that may be related to the immune system. Both reports constructed liver-MPS by coculturing hepatocytes with liver nonparenchymal cells and succeeded in reproducing cytokine mediated immune responses that occur in the actual liver. This result is considered to lead to the construction of histopathology on MPS in the future.

Liver fibrosis Hepatic fibrosis can be caused by infection with the hepatitis B or C virus, alcohol intake, and diseases associated with lifestyle, such as nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (Trautwein et al., 2015). Liver endothelial cells, hepatic stellate cells, and Kupffer cells are all involved in liver fibrosis that arises in response to impaired hepatic repair mechanisms and accumulation of excess matrix proteins (Bataller and Brenner, 2005). Hepatic stellate cells are the main source of myofibroblasts that supply the

Liver fibrosis

ECM (Mederacke et al., 2013; Brenner et al., 2012; Kisseleva, 2017); in healthy liver, quiescent hepatic stellate cells store vitamin A, but following liver injury and especially chronic injury, quiescent hepatic stellate cells transdifferentiate and are activated, lipid droplets containing vitamin A disappear, and the cells acquire a fibrogenic, myofibroblast-like phenotype (Geerts, 2001; Guo and Friedman, 2007). Concomitantly, the expression of smooth-muscle actin increases and secretion of ECM components is enhanced (Friedman, 2008; Geerts, 2001). During these events, the ECM composition in hepatic stellate cells remodels from a collagen type IV laminin-rich, low-density basement membrane-like structure into a collagen type I and III-rich fibrillary matrix (Verbruggen et al., 1986; Shiratori et al., 1987; Geerts et al., 1989; van Oortmarssen et al., 1990; Hayasaka et al., 1991). Hepatic stellate cell activation results in changes to ECM stiffness: a positive feedback loop that involves several factors, such as Yes-associated protein (a transcriptional coactivator and effector of the mechanosensitive Hippo pathway) and bromodomain-containing protein 4, connects activation of hepatic stellate cells and enhanced production of ECM, leading to a hardening of the environment surrounding the cells (Zhubanchaliyev et al., 2016).

Hepatic stellate cells Hepatic fibrosis studies require human hepatic stellate cells, which can be isolated from human tissues and are also available as established cell lines. Human hepatic stellate cells prepared from liver samples can withstand two to five passages, and cryopreservation of cultured cells is also possible. Passaged cells are already in an activated state, so quiescent hepatic stellate cells must be derived from primary cells. Since isolation of quiescent hepatic stellate cells can be challenging, several quiescent hepatic stellate cells cell lines were developed. The LI90 cell line was established from a human hepatic mesenchymal cell tumor (Murakami et al., 1995) and retains the capacity to store vitamin A even though LI90 cells display an activated stellate-like morphology and expression of the smooth-muscle actin activation marker and production of ECM constituents. The TWNT-4 cell line was established by introducing human telomerase reverse transcriptase into LI90 cells (Shibata et al., 2003). LX-1 cells are an immortalized cell line produced by introducing SV40 T antigen into human hepatic stellate cells, and the LX-2 line was established by subculturing human hepatic stellate cells in low-serum medium (Xu et al., 2005).

Microphysiological systems for evaluating fibrosis Regardless of which cell line is used in MPS, regulation of the hepatic stellate cell activation state is critical for examining whether a given drug candidate induces hepatic fibrosis. As mentioned earlier, the rigidity of the ECM that is in contact with cells is one factor that regulates activation. Hepatic stellate cells grown on polyacrylamide gels of varying rigidity have been shown to be activated

199

200

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

FIGURE 5.10 Effect of cultivating LI90 cells on VECELL inserts (Cosmo Bio, Tokyo, Japan) on actin filaments. LI90 cells were cultured on a standard culture plate (left) or on a VECELL insert (right) for 7 days (Horiuchi et al., 2018). Actin filaments were stained with phalloidin labeled with Alexa Fluor 594 and observed by fluorescence microscopy. Red indicates actin filaments.

on hard substrates (12 kPa) but to maintain quiescence on soft substrates (0.4 kPa) (Li et al., 2007; Olsen et al., 2011), corresponding to the rigidity of normal rat liver tissue (0.3 0.6 kPa) and that of cirrhotic liver tissue (3 12 kPa). In contrast, the surface of a typical culture dish is around 100,000 kPa in rigidity and does not mimic physiological conditions (van Grunsven, 2017). Activated hepatic stellate cells exhibit spheroid-like morphology and are deactivated when cultured on Matrigel (Shimada et al., 2010); when activated hepatic stellate cells are seeded and cultured on a cell culture scaffold composed of an expanded polytetrafluoroethylene mesh coated with collagen type I (VECELL inserts, Cosmo Bio, Tokyo, Japan), the cells show a spheroid-like morphology similar to that seen with Matrigel culturing and remain in a deactivated state (Fig. 5.10) (Horiuchi et al., 2018). These findings show that MPS for evaluating liver fibrosis should provide an ECM that is as close as possible to the physiological environment of the cell.

Organs-on-a-chip This section discusses considerations for multiorgan MPS modeling the liver and small intestine.

Gut liver axis The gut liver axis is the portal vein connection of the liver and the small intestine and is a means of communication between these two organs (Fig. 5.11, left) (Vajro et al., 2013; Fukui, 2015; Scarpellini et al., 2014). NAFLD is a fatty liver condition that can be diagnosed by histology or imaging and is not caused by

Organs-on-a-chip

Kupffer cells Stellate cells Release of inflammatory cytokines

Cholic acid (CA) Primary bile acid Conjugation

Liver injury

PRRs PAMPs Leaky gut

Taurocholic acid (TCA) Deconjugation

Dysbiosis Dehydroxylation

High - fat diet Alcohol, Stress

Barrier function Reabsorption

Symbiosis

Gut–liver axis

Deoxycholic acid (DCA ) Secondary bile acid

Enterohepatic circulation

FIGURE 5.11 The gut liver axis and enterohepatic circulation (Ishida., 2018). Left: Gut liver axis. Intestine and liver actions combine to generate biological responses. Right: Enterohepatic circulation of bile acids. Primary bile acids are secreted into the duodenum, deconjugated by intestinal bacteria, dehydroxylated, converted into secondary bile acids, and reabsorbed in the small intestine. Various drugs have been shown to enter enterohepatic circulation.

alcohol consumption (Hashimoto et al., 2013). The pathological mechanism of NAFLD is similar to that of alcoholic steatohepatitis, where the barrier function of the intestinal tract is compromised by excessive alcohol intake (leaky gut), enhancing intestinal tract permeability and enabling factors released from intestinal bacteria to flow into the liver via the portal vein. These microparticles, termed pathogen-associated molecular patterns (PAMPs), are recognized by surfaceexpressed pattern recognition receptors on Kupffer cells and hepatic stellate cells. Stimulation of pattern recognition receptors on these cells releases inflammatory cytokines. In NAFLD patients, intestinal flora dysbiosis is another cause of altered intestinal tract permeability (Miele et al., 2009). As with alcoholic steatohepatitis, nutritional imbalances can affect intestinal flora, compromising the intestinal barrier function and increasing PAMP delivery to the liver via the portal vein. Disruptions to intestinal flora can contribute to the onset of NAFLD and be caused not only by diet but also by drugs. Increased intestinal permeability following drug-induced alterations in intestinal flora can enable both PAMPs and

201

202

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

the administered drug to flow into the liver via the portal vein, sometimes simultaneously. Roth et al. reported that hepatocellular damage from aflatoxin B1 was exacerbated by concomitant treatment with lipopolysaccharide, which is one of PAMPs (Roth et al., 2003). When intestinal barriers are compromised by lifestyle factors such as alcohol and a high-fat diet, PAMPs flow into the liver from the portal vein and induce an inflammatory state, rendering the liver more susceptible to damage from medication. Multiorgan MPS assays can assess the effect of drugs on multiple organs that share physiological linkages.

Enterohepatic circulation The small intestine absorbs nutrients, medicines, and xenobiotics that flow into the liver through the portal vein (Gao et al., 2014; Fagerholm, 2008; Roberts et al., 2002). Following hepatic metabolism, these substances can be excreted as bile and return to the small intestine through the bile duct. Bile acids are known as enterohepatic circulating substances (Fig. 5.11, right) (de Aguiar Vallim et al., 2013), and primary bile acids such as cholic acid and ketodeoxycholic acid are synthesized through hydroxylation of cholesterol in the liver. The primary bile acids are converted to taurocholic acid and glycocholic acid via a conjugation reaction to add a taurine or glycine moiety, respectively. Conjugated bile acids accumulate in the gallbladder and are secreted into the duodenum as needed. After reaching the small intestine, conjugated bile acids are deconjugated by bile acid hydrolase expressed by intestinal bacteria and then undergo 7-dehydroxylation by bacterial enzymes to form the secondary bile acids deoxycholic acid from cholic acid and lithocholic acid from chenodeoxycholic acid. These secondary bile acids are reabsorbed in the small intestine and return to the liver from the portal vein. Such enterohepatic circulation is repeated six to nine times a day. Drugs also undergo enterohepatic circulation. In general, glucuronide conjugates are excreted in bile but cannot be easily resorbed by the small intestine because of the high polarity of the molecules. However, when these compounds are hydrolyzed by β-glucuronidase from intestinal bacteria to restore the parent compounds, the lipid solubility increases, along with rates of intestinal absorption, so that they are easily absorbed from the intestinal tract and cause enterohepatic circulation. Digoxin, digitoxin, indomethacin, and morphine are all examples of drugs that undergo enterohepatic circulation. Metabolism of the anticancer drug irinotecan by carboxylesterase produces the active metabolite SN-38 that inhibits type I DNA topoisomerase activity. SN-38 is converted to a glucuronic acid conjugate (SN-38G) by UDP-glucuronosyltransferase 1A1 and excreted in bile. SN-38G is converted back to SN-38 by β-glucuronidase from intestinal bacteria and enters enterohepatic circulation (Kuhn, 1998). These drugs have a long half-life, because they enter the enterohepatic circulation; changes in the abundance and composition of intestinal flora that occur following, for example,

References

administration of antibiotics, will affect this metabolic cycle, influencing the efficacy and side effects of medications. To reproduce intestinal circulation in in vitro assays, two techniques must be developed. First, bile must be recovered from hepatocytes. As summarized in the section on biliary excretion, capillary bile duct formation can be promoted by certain culture conditions, but methods to recover the bile that accumulates in the bile canaliculi while avoiding dilution are still lacking. Second, intestinal epithelial cells must be cocultured with intestinal flora. In vivo, intestinal epithelial cells are supplied with oxygen from the blood and exist in an aerobic state, but the intestinal tract is anaerobic. Thus an MPS would need to cultivate intestinal epithelial cells and intestinal bacterial flora under their respective aerobic and anaerobic culture conditions.

Conclusion This chapter discussed drug-induced liver injury and hepatic fibrosis, outlining the needs of MPS to study liver disease. Constructing a system that mimics the functions and structure of the liver in vitro is complex, even though human hepatocytes are relatively easy to obtain. Multiple technical challenges must be overcome, including stable long-term culture conditions and reconstruction of intercellular factors and subsequent formation of bile ducts. While MPS engineering technologies have advanced, integration with culture technology is lagging, and this gap must be bridged before reconstruction of hepatic physiological functions in vitro is possible.

References Aden, D.P., Fogel, A., Plotkin, S., Damjanov, I., Knowles, B.B., 1979. Controlled synthesis of HBsAg in a differentiated human liver carcinoma-derived cell line. Nature 282, 615 616. de Aguiar Vallim, T.Q., Tarling, E.J., Edwards, P.A., 2013. Pleiotropic roles of bile acids in metabolism. Cell Metab. 17, 657 669. Allen, J.W., Bhatia, S.N., 2003. Formation of steady-state oxygen gradients in vitro: application to liver zonation. Biotechnol. Bioeng. 82, 253 262. Allen, J.W., Khetani, S.R., Bhatia, S.N., 2005. In vitro zonation and toxicity in a hepatocyte bioreactor. Toxicol. Sci. 84, 110 119. Anderson, G.D., 2002. Children versus adults: pharmacokinetic and adverse-effect differences. Epilepsia 43 (Suppl 3), 53 59. Aninat, C., Piton, A., Glaise, D., Le Charpentier, T., Langoue¨t, S., Morel, F., et al., 2006. Expression of cytochromes P450, conjugating enzymes and nuclear receptors in human hepatoma HepaRG cells. Drug Metab. Dispos. 34, 75 83.

203

204

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

Anthe´rieu, S., Chesne´, C., Li, R., Camus, S., Lahoz, A., Picazo, L., et al., 2010. Stable expression, activity, and inducibility of cytochromes P450 in differentiated HepaRG cells. Drug Metab. Dispos. 38, 516 525. Anundi, I., La¨hteenma¨ki, T., Rundgren, M., Moldeus, P., Lindros, K.O., 1993. Zonation of acetaminophen metabolism and cytochrome P450 2E1-mediated toxicity studied in isolated periportal and perivenous hepatocytes. Biochem. Pharmacol. 45, 1251 1259. Araya, H., Mizuma, T., Horie, T., Hayashi, M., Awazu, S., 1986. Heterogeneous distribution of the conjugation activity of acetaminophen and p-nitrophenol in isolated rat liver cells. J. Pharmacobiodyn. 9, 218 222. Bartolone, J.B., Cohen, S.D., Khairallah, E.A., 1989. Immunohistochemical localization of acetaminophen-bound liver proteins. Fundam. Appl. Toxicol. 13, 859 862. Bataller, R., Brenner, D.A., 2005. Liver fibrosis. J. Clin. Invest. 115, 209 218. Bell, C.C., Dankers, A.C.A., Lauschke, V.M., Sison-Young, R., Jenkins, R., Rowe, C., et al., 2018. Comparison of hepatic 2D sandwich cultures and 3D spheroids for longterm toxicity applications: a multicenter study. Toxicol. Sci. 162, 655 666. Benedicto, I., Molina-Jime´nez, F., Garcı´a-Buey, L., Gondar, V., Lo´pez-Cabrera, M., Moreno-Otero, R., et al., 2012. Role of tight junctions in hepatitis C virus infection. Rev. Esp. Enferm. Dig. 104, 255 263. Be´nichou, C., 1990. Criteria of drug-induced liver disorders. Report of an international consensus meeting. J. Hepatol. 11, 272 276. Bhatia, S.N., Balis, U.J., Yarmush, M.L., Toner, M., 1999. Effect of cell-cell interactions in preservation of cellular phenotype: cocultivation of hepatocytes and nonparenchymal cells. FASEB J. 13, 1883 1900. Birge, R.B., Bartolone, J.B., Hart, S.G., Nishanian, E.V., Tyson, C.A., Khairallah, E.A., et al., 1990. Acetaminophen hepatotoxicity: correspondence of selective protein arylation in human and mouse liver in vitro, in culture, and in vivo. Toxicol. Appl. Pharmacol. 105, 472 482. Bjo¨rnsson, E., Olsson, R., 2005. Outcome and prognostic markers in severe drug-induced liver disease. Hepatology 42, 481 489. Boess, F., Kamber, M., Romer, S., Gasser, R., Muller, D., Albertini, S., et al., 2003. Gene expression in two hepatic cell lines, cultured primary hepatocytes, and liver slices compared to the in vivo liver gene expression in rats: possible implications for toxicogenomics use of in vitro systems. Toxicol. Sci. 73, 386 402. Bow, D.A., Perry, J.L., Miller, D.S., Pritchard, J.B., Brouwer, K.L., 2008. Localization of P-gp (Abcb1) and Mrp2 (Abcc2) in freshly isolated rat hepatocytes. Drug Metab. Dispos. 36, 198 202. Braet, F., Wisse, E., 2002. Structural and functional aspects of liver sinusoidal endothelial cell fenestrae: a review. Comp. Hepatol. 1, 1. den Braver-Sewradj, S.P., den Braver, M.W., Vermeulen, N.P., Commandeur, J.N., Richert, L., Vos, J.C., 2016. Inter-donor variability of phase I/phase II metabolism of three reference drugs in cryopreserved primary human hepatocytes in suspension and monolayer. Toxicol. In Vitro 33, 71 79. Brenner, D.A., Kisseleva, T., Scholten, D., Paik, Y.H., Iwaisako, K., Inokuchi, S., et al., 2012. Origin of myofibroblasts in liver fibrosis. Fibrog. Tissue Repair 5 (Suppl 1), S17. Bruno, M.K., Cohen, S.D., Khairallah, E.A., 1991. Selective alterations in the profiles of newly synthesized proteins by acetaminophen (APAP) and its dimethylated analogues: relationship to oxidative stress. Adv. Exp. Med. Biol. 283, 257 260.

References

Byrne, J.A., Strautnieks, S.S., Mieli-Vergani, G., Higgins, C.F., Linton, K.J., Thompson, R. J., 2002. The human bile salt export pump: characterization of substrate specificity and identification of inhibitors. Gastroenterology 123, 1649 1658. Chu, X., Chan, G.H., Evers, R., 2017. Identification of endogenous biomarkers to predict the propensity of drug candidates to cause hepatic or renal transporter-mediated drugdrug interactions. J. Pharm. Sci. 106, 2357 2367. Dandri, M., Burda, M.R., To¨ro¨k, E., Pollok, J.M., Iwanska, A., Sommer, G., et al., 2001. Repopulation of mouse liver with human hepatocytes and in vivo infection with hepatitis B virus. Hepatology 33, 981 988. Dawson, S., Stahl, S., Paul, N., Barber, J., Kenna, J.G., 2012. In vitro inhibition of the bile salt export pump correlates with risk of cholestatic drug-induced liver injury in humans. Drug Metab. Dispos. 40, 130 138. Di, L., Obach, R.S., 2015. Addressing the challenges of low clearance in drug research. AAPS J. 17, 352 357. Doostdar, H., Duthie, S.J., Burke, M.D., Melvin, W.T., Grant, M.H., 1988. The influence of culture medium composition on drug metabolising enzyme activities of the human liver derived Hep G2 cell line. FEBS Lett. 241, 15 18. Dunn, J.C., Tompkins, R.G., Yarmush, M.L., 1991. Long-term in vitro function of adult hepatocytes in a collagen sandwich configuration. Biotechnol. Prog. 7, 237 245. Dunn, J.C., Tompkins, R.G., Yarmush, M.L., 1992. Hepatocytes in collagen sandwich: evidence for transcriptional and translational regulation. J. Cell Biol. 116, 1043 1053. Fagerholm, U., 2008. Prediction of human pharmacokinetics-biliary and intestinal clearance and enterohepatic circulation. J. Pharm. Pharmacol. 60, 535 542. Fattinger, K., Funk, C., Pantze, M., Weber, C., Reichen, J., Stieger, B., et al., 2001. The endothelin antagonist bosentan inhibits the canalicular bile salt export pump: a potential mechanism for hepatic adverse reactions. Clin. Pharmacol. Ther. 69, 223 231. Friedman, S.L., 2008. Hepatic stellate cells: protean, multifunctional, and enigmatic cells of the liver. Physiol. Rev. 88, 125 172. Fukui, H., 2015. Gut-liver axis in liver cirrhosis: How to manage leaky gut and endotoxemia. World J. Hepatol. 7, 425 442. Gailhouste, L., Liew, L.C., Yasukawa, K., Hagiwara, K., Iwazaki, N., Yamada, Y., et al., 2017. Epigenetic reprogramming of human hepatoma cells: a low-cost option for drug metabolism assessment. Cell. Mol. Gastroenterol. Hepatol. 5, 454 457. Gao, Y., Shao, J., Jiang, Z., Chen, J., Gu, S., Yu, S., et al., 2014. Drug enterohepatic circulation and disposition: constituents of systems pharmacokinetics. Drug Discov. Today 19, 326 340. Gebhardt, R., 1992. Metabolic zonation of the liver: regulation and implications for liver function. Pharmacol. Ther. 53, 275 354. Gebhardt, R., Matz-Soja, M., 2014. Liver zonation: Novel aspects of its regulation and its impact on homeostasis. World J. Gastroenterol. 20, 8491 8504. Geerts, A., 2001. History, heterogeneity, developmental biology, and functions of quiescent hepatic stellate cells. Semin. Liver Dis. 21, 311 335. Geerts, A., Vrijsen, R., Rauterberg, J., Burt, A., Schellinck, P., Wisse, E., 1989. In vitro differentiation of fat-storing cells parallels marked increase of collagen synthesis and secretion. J. Hepatol. 9, 59 68.

205

206

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

Gerets, H.H., Tilmant, K., Gerin, B., Chanteux, H., Depelchin, B.O., Dhalluin, S., et al., 2012. Characterization of primary human hepatocytes, HepG2 cells, and HepaRG cells at the mRNA level and CYP activity in response to inducers and their predictivity for the detection of human hepatotoxins. Cell Biol. Toxicol. 28, 69 87. Griffin, S.J., Houston, J.B., 2005. Prediction of in vitro intrinsic clearance from hepatocytes: comparison of suspensions and monolayer cultures. Drug Metab. Dispos. 33, 115 120. Gripon, P., Rumin, S., Urban, S., Le Seyec, J., Glaise, D., Cannie, I., et al., 2002. Infection of a human hepatoma cell line by hepatitis B virus. Proc. Natl. Acad. Sci. U.S.A. 99, 15655 15660. van Grunsven, L.A., 2017. 3D in vitro models of liver fibrosis. Adv. Drug Deliv. Rev. 121, 133 146. Guguen-Guillouzo, C., Guillouzo, A., 2010. General review on in vitro hepatocyte models and their applications. Methods Mol. Biol. 640, 1 40. Guillouzo, A., 1998. Liver cell models in in vitro toxicology. Environ. Health Perspect. 106 (Suppl 2), 511 532. Guillouzo, A., Corlu, A., Aninat, C., Glaise, D., Morel, F., Guguen-Guillouzo, C., 2007. The human hepatoma HepaRG cells: a highly differentiated model for studies of liver metabolism and toxicity of xenobiotics. Chem. Biol. Interact. 168, 66 73. Guo, J., Friedman, S.L., 2007. Hepatic fibrogenesis. Semin. Liver Dis. 27, 413 426. Guthrie, B., Makubate, B., Hernandez-Santiago, V., Dreischulte, T., 2015. The rising tide of polypharmacy and drug-drug interactions: population database analysis 1995-2010. BMC Med. 13, 74. Hasegawa, M., Kawai, K., Mitsui, T., Taniguchi, K., Monnai, M., Wakui, M., et al., 2011. The reconstituted ‘humanized liver’ in TK-NOG mice is mature and functional. Biochem. Biophys. Res. Commun. 405, 405 410. Hashimoto, E., Taniai, M., Tokushige, K., 2013. Characteristics and diagnosis of NAFLD/ NASH. J. Gastroenterol. Hepatol. 28 (Suppl 4), 64 70. Hayasaka, A., Koch, J., Schuppan, D., Maddrey, W.C., Hahn, E.G., 1991. The serum concentrations of the aminoterminal propeptide of procollagen type III and the hepatic content of mRNA for the alpha 1 chain of procollagen type III in carbon tetrachlorideinduced rat liver fibrogenesis. J. Hepatol. 13, 328 338. Hewitt, N.J., Hewitt, P., 2004. Phase I and II enzyme characterization of two sources of HepG2 cell lines. Xenobiotica 34, 243 256. Hewitt, N.J., Lecho´n, M.J., Houston, J.B., Hallifax, D., Brown, H.S., Maurel, P., et al., 2007. Primary hepatocytes: current understanding of the regulation of metabolic enzymes and transporter proteins, and pharmaceutical practice for the use of hepatocytes in metabolism, enzyme induction, transporter, clearance, and hepatotoxicity studies. Drug Metab. Rev. 39, 159 234. Horiuchi, S., Kuroda, Y., Fujii, R., Kim, S.R., Ishida, S., 2018. Deactivation of hepatic stellate cells by culturing on VECELL inserts. AATEX 23, 53 62. Houston, J.B., 1994. Utility of in vitro drug metabolism data in predicting in vivo metabolic clearance. Biochem. Pharmacol. 47, 1469 1479. Ishida, S., 2018. Organs-on-a-chip: current applications and consideration points for in vitro ADME-Tox studies. Drug Metab. Pharmacokinet. 33, 49 54. Jennen, D.G., Magkoufopoulou, C., Ketelslegers, H.B., van Herwijnen, M.H., Kleinjans, J. C., van Delft, J.H., 2010. Comparison of HepG2 and HepaRG by whole-genome gene

References

expression analysis for the purpose of chemical hazard identification. Toxicol. Sci. 115, 66 79. Josse´, R., Aninat, C., Glaise, D., Dumont, J., Fessard, V., Morel, F., et al., 2008. Longterm functional stability of human HepaRG hepatocytes and use for chronic toxicity and genotoxicity studies. Drug Metab. Dispos. 36, 1111 1118. Jungermann, K., Kietzmann, T., 1996. Zonation of parenchymal and nonparenchymal metabolism in liver. Annu. Rev. Nutr. 16, 179 203. Jungermann, K., Kietzmann, T., 2000. Oxygen: modulator of metabolic zonation and disease of the liver. Hepatology 31, 255 260. Kanebratt, KP., Andersson, TB., 2008. Evaluation of HepaRG cells as an in vitro model for human drug metabolism studies. Drug Metab Dispos 36, 1444 1452. Kaplowitz, N., 2005. Idiosyncratic drug hepatotoxicity. Nat. Rev. Drug Discov. 4, 489 499. Khetani, S.R., Bhatia, S.N., 2008. Microscale culture of human liver cells for drug development. Nat. Biotechnol. 26, 120 126. Khetani, S.R., Kanchagar, C., Ukairo, O., Krzyzewski, S., Moore, A., Shi, J., et al., 2013. Use of micropatterned cocultures to detect compounds that cause drug-induced liver injury in humans. Toxicol. Sci. 132, 107 117. Kidambi, S., Yarmush, R.S., Novik, E., Chao, P., Yarmush, M.L., Nahmias, Y., 2009. Oxygen-mediated enhancement of primary hepatocyte metabolism, functional polarization, gene expression, and drug clearance. Proc. Natl. Acad. Sci. U.S.A. 106, 15714 15719. Kietzmann, T., 2017. Metabolic zonation of the liver: the oxygen gradient revisited. Redox Biol. 11, 622 630. Kim, D.H., Provenzano, P.P., Smith, C.L., Levchenko, A., 2012. Matrix nanotopography as a regulator of cell function. J. Cell Biol. 197, 351 360. Kimura, H., Sakai, Y., Fujii, T., 2018. Organ/body-on-a-chip based on microfluidic technology for drug discovery. Drug Metab. Pharmacokinet. 33, 43 48. Kisseleva, T., 2017. The origin of fibrogenic myofibroblasts in fibrotic liver. Hepatology 65, 1039 1043. Klein, S., Mueller, D., Schevchenko, V., Noor, F., 2014. Long-term maintenance of HepaRG cells in serum-free conditions and application in a repeated dose study. J. Appl. Toxicol. 34, 1078 1086. Kocarek, T.A., Schuetz, E.G., Guzelian, P.S., 1993. Expression of multiple forms of cytochrome P450 mRNAs in primary cultures of rat hepatocytes maintained on matrigel. Mol. Pharmacol. 43, 328 334. Ko¨ck, K., Ferslew, B.C., Netterberg, I., Yang, K., Urban, T.J., Swaan, P.W., et al., 2014. Risk factors for development of cholestatic drug-induced liver injury: inhibition of hepatic basolateral bile acid transporters multidrug resistance-associated proteins 3 and 4. Drug Metab. Dispos. 42, 665 674. Kostrubsky, S.E., Strom, S.C., Kalgutkar, A.S., Kulkarni, S., Atherton, J., Mireles, R., et al., 2006. Inhibition of hepatobiliary transport as a predictive method for clinical hepatotoxicity of nefazodone. Toxicol. Sci. 90, 451 459. Kuhn, J.G., 1998. Pharmacology of irinotecan. Oncology (Williston Park) 12, 39 42. Lane, S.W., Williams, D.A., Watt, F.M., 2014. Modulating the stem cell niche for tissue regeneration. Nat. Biotechnol. 32, 795 803. Lazarczyk, D.A., Goldstein, N.S., Gordon, S.C., 2001. Trovafloxacin hepatotoxicity. Dig. Dis. Sci. 46, 925 926.

207

208

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

LeCluyse, E.L., Witek, R.P., Andersen, M.E., Powers, M.J., 2012. Organotypic liver culture models: meeting current challenges in toxicity testing. Crit. Rev. Toxicol. 42, 501 548. Lee, W.M., 2003. Drug-induced hepatotoxicity. N. Engl. J. Med. 349, 474 485. Lee-Montiel, F.T., George, S.M., Gough, A.H., Sharma, A.D., Wu, J., DeBiasio, R., et al., 2017. Control of oxygen tension recapitulates zone-specific functions in human liver microphysiology systems. Exp. Biol. Med. (Maywood) 242, 1617 1632. Legendre, C., Hori, T., Loyer, P., Aninat, C., Ishida, S., Glaise, D., et al., 2009. Drugmetabolising enzymes are down-regulated by hypoxia in differentiated human hepatoma HepaRG cells: HIF-1alpha involvement in CYP3A4 repression. Eur. J. Cancer 45, 2882 2892. Le Vee, M., Jigorel, E., Glaise, D., Gripon, P., Guguen-Guillouzo, C., Fardel, O., 2006. Functional expression of sinusoidal and canalicular hepatic drug transporters in the differentiated human hepatoma HepaRG cell line. Eur. J. Pharm. Sci. 28, 109 117. Le Vee, M., Noel, G., Jouan, E., Stieger, B., Fardel, O., 2013. Polarized expression of drug transporters in differentiated human hepatoma HepaRG cells. Toxicol. In Vitro 27, 1979 1986. Levy, G., Bomze, D., Heinz, S., Ramachandran, S.D., Noerenberg, A., Cohen, M., et al., 2015. Long-term culture and expansion of primary human hepatocytes. Nat. Biotechnol. 33, 1264 1271. Li, X., George, S.M., Vernetti, L., Gough, A.H., Taylor, D.L., 2018. A glass-based, continuously zonated and vascularized human liver acinus microphysiological system (vLAMPS) designed for experimental modeling of diseases and ADME/TOX. Lab Chip 18, 2614 2631. Li, Z., Dranoff, J.A., Chan, E.P., Uemura, M., Se´vigny, J., Wells, R.G., 2007. Transforming growth factor-beta and substrate stiffness regulate portal fibroblast activation in culture. Hepatology 46, 1246 1256. Lu¨bberstedt, M., Mu¨ller-Vieira, U., Mayer, M., Biemel, K.M., Kno¨spel, F., Knobeloch, D., et al., 2011. HepaRG human hepatic cell line utility as a surrogate for primary human hepatocytes in drug metabolism assessment in vitro. J. Pharmacol. Toxicol. Methods 63, 59 68. Marion, T.L., Perry, C.H., St Claire 3rd, R.L., Brouwer, K.L., 2012. Endogenous bile acid disposition in rat and human sandwich-cultured hepatocytes. Toxicol. Appl. Pharmacol. 261, 1 9. Mederacke, I., Hsu, C.C., Troeger, J.S., Huebener, P., Mu, X., Dapito, D.H., et al., 2013. Fate tracing reveals hepatic stellate cells as dominant contributors to liver fibrosis independent of its aetiology. Nat. Commun. 4, 2823. Mercer, D.F., Schiller, D.E., Elliott, J.F., Douglas, D.N., Hao, C., Rinfret, A., et al., 2001. Hepatitis C virus replication in mice with chimeric human livers. Nat. Med. 7, 927 933. Miele, L., Valenza, V., La Torre, G., Montalto, M., Cammarota, G., Ricci, R., et al., 2009. Increased intestinal permeability and tight junction alterations in nonalcoholic fatty liver disease. Hepatology 49, 1877 1887. Moghe, P.V., Berthiaume, F., Ezzell, R.M., Toner, M., Tompkins, R.G., Yarmush, M.L., 1996. Culture matrix configuration and composition in the maintenance of hepatocyte polarity and function. Biomaterials 17, 373 385.

References

Morgan, R.E., Trauner, M., van Staden, C.J., Lee, P.H., Ramachandran, B., Eschenberg, M., et al., 2010. Interference with bile salt export pump function is a susceptibility factor for human liver injury in drug development. Toxicol. Sci. 118, 485 500. Morgan, R.E., van Staden, C.J., Chen, Y., Kalyanaraman, N., Kalanzi, J., Dunn 2nd, R.T., et al., 2013. A multifactorial approach to hepatobiliary transporter assessment enables improved therapeutic compound development. Toxicol. Sci. 136, 216 241. Murakami, K., Abe, T., Miyazawa, M., Yamaguchi, M., Masuda, T., Matsuura, T., et al., 1995. Establishment of a new human cell line, LI90, exhibiting characteristics of hepatic Ito (fat-storing) cells. Lab Invest. 72, 731 739. Musat, A.I., Sattler, C.A., Sattler, G.L., Pitot, H.C., 1993. Reestablishment of cell polarity of rat hepatocytes in primary culture. Hepatology 18, 198 205. Nakabayashi, H., Taketa, K., Miyano, K., Yamane, T., Sato, J., 1982. Growth of human hepatoma cells lines with differentiated functions in chemically defined medium. Cancer Res. 42, 3858 3863. Nguyen, D.G., Funk, J., Robbins, J.B., Crogan-Grundy, C., Presnell, S.C., Singer, T., et al., 2016. Bioprinted 3D primary liver tissues allow assessment of organ-level response to clinical drug induced toxicity in vitro. PLoS One 11, e0158674. Nyberg, S.L., Remmel, R.P., Mann, H.J., Peshwa, M.V., Hu, W.S., Cerra, F.B., 1994. Primary hepatocytes outperform Hep G2 cells as the source of biotransformation functions in a bioartificial liver. Ann. Surg. 220, 59 67. Ogihara, T., Iwai, H., Inoue, Y., Katagi, J., Matsumoto, N., Motoi-Ohtsuji, M., et al., 2015. Utility of human hepatocyte spheroids for evaluation of hepatotoxicity. Fund. Toxicol. Sci. 2, 41 48. Ohkura, T., Ohta, K., Nagao, T., Kusumoto, K., Koeda, A., Ueda, T., et al., 2014. Evaluation of human hepatocytes cultured by three-dimensional spheroid systems for drug metabolism. Drug Metab. Pharmacokinet. 29, 373 378. Olsen, A.L., Bloomer, S.A., Chan, E.P., Gac¸a, M.D., Georges, P.C., Sackey, B., et al., 2011. Hepatic stellate cells require a stiff environment for myofibroblastic differentiation. Am. J. Physiol. Gastrointest. Liver Physiol. 301, G110 G118. Olson, H., Betton, G., Robinson, D., Thomas, K., Monro, A., Kolaja, G., et al., 2000. Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul. Toxicol. Pharmacol. 32, 56 67. van Oortmarssen, G.J., Habbema, J.D., van der Maas, P.J., de Koning, H.J., Collette, H.J., Verbeek, A.L., et al., 1990. A model for breast cancer screening. Cancer 66, 1601 1612. Pang, K.S., Cherry, W.F., Accaputo, J., Schwab, A.J., Goresky, C.A., 1988. Combined hepatic arterial-portal venous and hepatic arterial-hepatic venous perfusions to probe the abundance of drug metabolizing activities: perihepatic venous O-deethylation activity for phenacetin and periportal sulfation activity for acetaminophen in the oncethrough rat liver preparation. J. Pharmacol. Exp. Ther. 247, 690 700. Pedersen, J.M., Matsson, P., Bergstro¨m, C.A., Hoogstraate, J., Nore´n, A., LeCluyse, E.L., et al., 2013. Early identification of clinically relevant drug interactions with the human bile salt export pump (BSEP/ABCB11). Toxicol. Sci. 136, 328 343. Pfeifer, N.D., Goss, S.L., Swift, B., Ghibellini, G., Ivanovic, M., Heizer, W.D., et al., 2013. Effect of ritonavir on (99m)technetium-mebrofenin disposition in humans: a semi-PBPK modeling and in vitro approach to predict transporter-mediated DDIs. CPT Pharmacometrics Syst. Pharmacol. 2, e20.

209

210

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

Placke, M.E., Ginsberg, G.L., Wyand, D.S., Cohen, S.D., 1987. Ultrastructural changes during acute acetaminophen-induced hepatotoxicity in the mouse: a time and dose study. Toxicol. Pathol. 15, 431 438. Raab, S., Klingenstein, M., Liebau, S., Linta, L., 2014. A comparative view on human somatic cell sources for iPSC generation. Stem Cells Int. 2014, 768391. Rappaport, A.M., 1977. Microcirculatory units in the mammalian liver. Their arterial and portal components. Bibl. Anat. 16 Pt 2, 116 120. Ren, Z., Chen, S., Ning, B., Guo, L., 2018. Use of liver-derived cell lines for the study of drug-induced liver injury. In: Chen, M., Will, Y. (Eds.), Drug-Induced Liver Toxicity, Methods in Pharmacology and Toxicology. Springer Nature. Roberts, M.S., Magnusson, B.M., Burczynski, F.J., Weiss, M., 2002. Enterohepatic circulation: physiological, pharmacokinetic and clinical implications. Clin. Pharmacokinet. 41, 751 790. Rodrı´guez-Antona, C., Donato, M.T., Boobis, A., Edwards, R.J., Watts, P.S., Castell, J.V., et al., 2002. Cytochrome P450 expression in human hepatocytes and hepatoma cell lines: molecular mechanisms that determine lower expression in cultured cells. Xenobiotica 32, 505 520. Rogers, A.S., 1987. Adverse drug events: identification and attribution. Drug Intell. Clin. Pharm. 21, 915 920. Rojkind, M., Ponce-Noyola, P., 1982. The extracellular matrix of the liver. Collagen Relat. Res. 2, 151 175. Roth, R.A., Luyendyk, J.P., Maddox, J.F., Ganey, P.E., 2003. Inflammation and drug idiosyncrasy—is there a connection? J. Pharmacol. Exp. Ther. 307, 1 8. Roymans, D., Annaert, P., Van Houdt, J., Weygers, A., Noukens, J., Sensenhauser, C., et al., 2005. Expression and induction potential of cytochromes P450 in human cryopreserved hepatocytes. Drug Metab. Dispos. 33, 1004 1016. Satoh, D., Iwado, S., Abe, S., Kazuki, K., Wakuri, S., Oshimura, M., et al., 2017. Establishment of a novel hepatocyte model that expresses four cytochrome P450 genes stably via mammalian-derived artificial chromosome for pharmacokinetics and toxicity studies. PLoS One 12, e0187072. Scarpellini, E., Lupo, M., Iegri, C., Gasbarrini, A., De Santis, A., Tack, J., 2014. Intestinal permeability in non-alcoholic fatty liver disease: the gut-liver axis. Rev. Recent Clin. Trials 9, 141 147. Schippers, I.J., Moshage, H., Roelofsen, H., Mu¨ller, M., Heymans, H.S., Ruiters, M., et al., 1997. Immortalized human hepatocytes as a tool for the study of hepatocytic (de-)differentiation. Cell Biol. Toxicol. 13, 375 386. Schwartz, R.E., Fleming, H.E., Khetani, S.R., Bhatia, S.N., 2014. Pluripotent stem cellderived hepatocyte-like cells. Biotechnol. Adv. 32, 504 513. Sharer, J.E., Shipley, L.A., Vandenbranden, M.R., Binkley, S.N., Wrighton, S.A., 1995. Comparisons of phase I and phase II in vitro hepatic enzyme activities of human, dog, rhesus monkey, and cynomolgus monkey. Drug Metab. Dispos. 23, 1231 1241. Shibata, N., Watanabe, T., Okitsu, T., Sakaguchi, M., Takesue, M., Kunieda, T., et al., 2003. Establishment of an immortalized human hepatic stellate cell line to develop antifibrotic therapies. Cell Transplant. 12, 499 507. Shimada, H., Ochi, T., Imasato, A., Morizane, Y., Hori, M., Ozaki, H., et al., 2010. Gene expression profiling and functional assays of activated hepatic stellate cells suggest that myocardin has a role in activation. Liver Int. 30, 42 54.

References

Shimada, T., Yamazaki, H., Mimura, M., Inui, Y., Guengerich, F.P., 1994. Interindividual variations in human liver cytochrome P-450 enzymes involved in the oxidation of drugs, carcinogens and toxic chemicals: studies with liver microsomes of 30 Japanese and 30 Caucasians. J. Pharmacol. Exp. Ther. 270, 414 423. Shiratori, Y., Ichida, T., Geerts, A., Wisse, E., 1987. Modulation of collagen synthesis by fat-storing cells, isolated from CCl4- or vitamin A-treated rats. Dig. Dis. Sci. 32, 1281 1289. Sistare, F.D., Mattes, W.B., LeCluyse, E.L., 2016. The promise of new technologies to reduce, refine, or replace animal use while reducing risks of drug induced liver injury in pharmaceutical development. ILAR J. 57, 186 211. Snykers, S., De Kock, J., Rogiers, V., Vanhaecke, T., 2009. In vitro differentiation of embryonic and adult stem cells into hepatocytes: state of the art. Stem Cells 27, 577 605. Soto-Gutierrez, A., Gough, A., Vernetti, L.A., Taylor, D.L., Monga, S.P., 2017. Preclinical and clinical investigations of metabolic zonation in liver diseases: the potential of microphysiology systems. Exp. Biol. Med. (Maywood) 242, 1605 1616. Stieger, B., Fattinger, K., Madon, J., Kullak-Ublick, G.A., Meier, P.J., 2000. Drug- and estrogen-induced cholestasis through inhibition of the hepatocellular bile salt export pump (BSEP) of rat liver. Gastroenterology 118, 422 430. Swift, B., Pfeifer, N.D., Brouwer, K.L., 2010. Sandwich-cultured hepatocytes: an in vitro model to evaluate hepatobiliary transporter-based drug interactions and hepatotoxicity. Drug Metab. Rev. 42, 446 471. Szabo, M., Veres, Z., Baranyai, Z., Jakab, F., Jemnitz, K., 2013. Comparison of human hepatoma HepaRG cells with human and rat hepatocytes in uptake transport assays in order to predict a risk of drug induced hepatotoxicity. PLoS One 8, e59432. Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., et al., 2007. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 131, 861 872. Takayama, K., Akita, N., Mimura, N., Akahira, R., Taniguchi, Y., Ikeda, M., et al., 2017. Generation of safe and therapeutically effective human induced pluripotent stem cellderived hepatocyte-like cells for regenerative medicine. Hepatol. Commun. 1, 1058 1069. Takezawa, T., Ozaki, K., Nitani, A., Takabayashi, C., Shimo-Oka, T., 2004. Collagen vitrigel: a novel scaffold that can facilitate a three-dimensional culture for reconstructing organoids. Cell Transplant. 13, 463 473. Takezawa, T., Takeuchi, T., Nitani, A., Takayama, Y., Kino-Oka, M., Taya, M., et al., 2007a. Collagen vitrigel membrane useful for paracrine assays in vitro and drug delivery systems in vivo. J. Biotechnol. 131, 76 83. Takezawa, T., Nitani, A., Shimo-Oka, T., Takayama, Y., 2007b. A protein-permeable scaffold of a collagen vitrigel membrane useful for reconstructing crosstalk models between two different cell types. Cells Tissues Organs 185, 237 241. Tateno, C., Yoshizane, Y., Saito, N., Kataoka, M., Utoh, R., Yamasaki, C., et al., 2004. Near completely humanized liver in mice shows human-type metabolic responses to drugs. Am. J. Pathol. 165, 901 912. Tateno, C., Miya, F., Wake, K., Kataoka, M., Ishida, Y., Yamasaki, C., et al., 2013. Morphological and microarray analyses of human hepatocytes from xenogeneic host livers. Lab Invest. 93, 54 71.

211

212

CHAPTER 5 Requirements for designing organ-on-a-chip platforms

Tetsuka, K., Ohbuchi, M., Tabata, K., 2017. Recent progress in hepatocyte culture models and their application to the assessment of drug metabolism, transport, and toxicity in drug discovery: the value of tissue engineering for the successful development of a microphysiological system. J. Pharm. Sci. 106, 2302 2311. Thompson, R.A., Isin, E.M., Ogese, M.O., Mettetal, J.T., Williams, D.P., 2016. Reactive metabolites: current and emerging risk and hazard assessments. Chem. Res. Toxicol. 29, 505 533. Trautwein, C., Friedman, S.L., Schuppan, D., Pinzani, M., 2015. Hepatic fibrosis: concept to treatment. J. Hepatol. 62, S15 S24. Tsuruga, Y., Kiyono, T., Matsushita, M., Takahashi, T., Kasai, H., Todo, S., 2008. Establishment of immortalized human hepatocytes by introduction of HPV16 E6/E7 and hTERT as cell sources for liver cell-based therapy. Cell Transplant. 17, 1083 1094. Turpeinen, M., Ghiciuc, C., Opritoui, M., Tursas, L., Pelkonen, O., Pasanen, M., 2007. Predictive value of animal models for human cytochrome P450 (CYP)-mediated metabolism: a comparative study in vitro. Xenobiotica 37, 1367 1377. Uehara, S., Higuchi, Y., Yoneda, N., Yamazaki, H., Suemizu, H., 2019. Expression and inducibility of cytochrome P450s in human hepatocytes isolated from chimeric mice with humanised livers. Xenobiotica 49, 678 687. Underhill, G.H., Khetani, S.R., 2018. Advances in engineered human liver platforms for drug metabolism studies. Drug Metab. Dispos. 46, 1626 1637. Vajro, P., Paolella, G., Fasano, A., 2013. Microbiota and gut-liver axis: their influences on obesity and obesity-related liver disease. J. Pediatr. Gastroenterol. Nutr. 56, 461 468. Verbruggen, L.A., Geerts, A., Wisse, E., Lapiere, C., 1986. Serum and synovial fluid antibodies to collagen in rheumatic diseases: a review. Clin Rheumatol. 5, 440 444. Vernetti, L.A., Senutovitch, N., Boltz, R., DeBiasio, R., Shun, T.Y., Gough, A., et al., 2016. A human liver microphysiology platform for investigating physiology, drug safety, and disease models. Exp. Biol. Med. (Maywood) 241, 101 114. Vernetti, L.A., Vogt, A., Gough, A., Taylor, D.L., 2017. Evolution of experimental models of the liver to predict human drug hepatotoxicity and efficacy. Clin. Liver Dis. 21, 197 214. Viravaidya, K., Sin, A., Shuler, M.L., 2004. Development of a microscale cell culture analog to probe naphthalene toxicity. Biotechnol. Prog. 20, 316 323. Wang, W.W., Khetani, S.R., Krzyzewski, S., Duignan, D.B., Obach, R.S., 2010. Assessment of a micropatterned hepatocyte coculture system to generate major human excretory and circulating drug metabolites. Drug Metab. Dispos. 38, 1900 1905. Watari, R., Kakiki, M., Oshikata, A., Takezawa, T., Yamasaki, C., Ishida, Y., et al., 2018. A long-term culture system based on a collagen vitrigel membrane chamber that supports liver-specific functions of hepatocytes isolated from mice with humanized livers. J. Toxicol. Sci. 43, 521 529. Watt, F.M., Huck, W.T., 2013. Role of the extracellular matrix in regulating stem cell fate. Nat. Rev. Mol. Cell Biol. 14, 467 473. Wilkening, S., Bader, A., 2003. Influence of culture time on the expression of drugmetabolizing enzymes in primary human hepatocytes and hepatoma cell line HepG2. J. Biochem. Mol. Toxicol. 17, 207 213.

References

Wisse, E., De Zanger, R.B., Charels, K., Van Der Smissen, P., McCuskey, R.S., 1985. The liver sieve: considerations concerning the structure and function of endothelial fenestrae, the sinusoidal wall and the space of Disse. Hepatology 5, 683 692. Xu, L., Hui, A.Y., Albanis, E., Arthur, M.J., O’Byrne, S.M., Blaner, W.S., et al., 2005. Human hepatic stellate cell lines, LX-1 and LX-2: new tools for analysis of hepatic fibrosis. Gut 54, 142 151. Yoon, E., Babar, A., Choudhary, M., Kutner, M., Pyrsopoulos, N., 2016. Acetaminopheninduced hepatotoxicity: a comprehensive update. J. Clin. Transl. Hepatol. 4, 131 142. Yoshikawa, Y., Hosomi, H., Fukami, T., Nakajima, M., Yokoi, T., 2009. Establishment of knockdown of superoxide dismutase 2 and expression of CYP3A4 cell system to evaluate drug-induced cytotoxicity. Toxicol. In Vitro 23, 1179 1187. Zhang, Z., Schwartz, S., Wagner, L., Miller, W., 2000. A greedy algorithm for aligning DNA sequences. J. Comput. Biol. 7, 203 214. Zhang, B., Korolj, A., Fook, B., Lai, L., Radisic, M., 2018. Advances in organ-on-a-chip engineering. Nat. Rev. Mater. 3, 257 278. Zhubanchaliyev, A., Temirbekuly, A., Kongrtay, K., Wanshura, L.C., Kunz, J., 2016. Targeting mechanotransduction at the transcriptional level: YAP and BRD4 are novel therapeutic targets for the reversal of liver fibrosis. Front. Pharmacol. 7, 462. Zuo, R., Li, F., Parikh, S., Cao, L., Cooper, K.L., Hong, Y., et al., 2017. Evaluation of a novel renewable hepatic cell model for prediction of clinical CYP3A4 induction using a correlation-based relative induction score approach. Drug Metab. Dispos. 45, 198 207.

213

CHAPTER

Brain-on-a-chip systems for modeling disease pathogenesis

6

Alexander P. Haring1,2 and Blake N. Johnson1,2,3 1

Department of Industrial and Systems Engineering, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States 2 Macromolecules Innovation Institute, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States 3 School of Neuroscience, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States

Introduction Need for microphysiological brains-on-chips An organ-on-a-chip is defined by the National Institutes of Health as a microsystem that represents units of human organs, such as the lung, liver, or heart, modeling both structure and function. While the term originally referred to microfluidic cell culture platforms of adjacent fluidic channels that often contained cells growing on semipermeable plastic membranes, modern organ-on-a-chip technologies now use biofabrication techniques that provide simultaneous fabrication of microfluidic channels and integration with biology. Regardless of the fabrication approach, the primary purpose of any organ-on-a-chip is to reproduce higherorder pathophysiological responses in humans, which is challenging to achieve using reductionist approaches (e.g., monolayer cultures) or small-animal models. Thus organ-on-a-chip platforms have been used for fundamental research on disease pathogeneses and applications in drug discovery (Haring et al., 2017a; Yi et al., 2015). Over the past decade, organ-on-a-chip platforms were developed for multiple organs, including lung, heart, liver, and kidney (Huh et al., 2010; Grosberg et al., 2011; Bavli et al., 2016; Jang et al., 2013), and researchers are now examining the integration of multiple organs on a chip (Jie et al., 2017). While a detailed discussion of such models is beyond the scope of this chapter, microphysiological neural systems-on-chips (NSCs), also commonly termed brains-on-chips (BOCs) or spinal cords-on-chips (SOCs), represent an emerging

Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00006-1 © 2020 Elsevier Inc. All rights reserved.

215

216

CHAPTER 6 Brain-on-a-chip systems

critical domain within this research field because of the disproportionate adverse effect of neurological diseases and disorders on society. Neurological diseases and disorders account for 8 of the 10 most disabling diseases, as defined by the World Health Organization (Menken et al., 2000). The development of microphysiological systems enables analyses of these diseases and disorders that are difficult to achieve with traditional cell culture methods or animal models (Huh et al., 2010). In addition, organs-on-chips are compatible with human cells, providing a potential preclinical screening alternative to animal models (Mak et al., 2014). Importantly, NSCs fabricated using human cells, which can be patient-derived, reduce translational challenges that arise from pathophysiological variations among species (Hackam and Redelmeier, 2006). The histological, electrophysiological, and bioanalytical techniques available for monitoring animals and human subjects during clinical trials limit the sampling frequency (i.e., the rate at which measurements can be collected) because of their invasiveness or requirement of sacrificial samples. These constraints limit the ability to model the dynamics associated with higher-order processes in the nervous system (e.g., neuroregenerative or neurodegenerative). Integrating sensors and NSCs can assist in understanding the dynamic nature of neuropathophysiological processes with near real-time resolution.

Prevalent diseases and disorders Neurological diseases are highly prevalent in today’s society, with 6.4 million people in the United States (US) currently affected by a neurological disease or disorder (Kochanek et al., 2016; N. C. f. H. Statistics, 2016). In addition to genetic and infectious causes, injury-induced disorders are common. It is estimated that up to 5.3 million people are currently suffering from traumatic brain injury, and 1.7 million people will experience a traumatic brain injury each year in the US alone (Ma et al., 2014). Approximately 20,000 spinal cord injuries and 200,000 peripheral nerve injuries occur annually in the US (Kehoe et al., 2012). While surgical advances have been made that improve patient outcomes to minor nerve injuries, severe nerve injuries are challenging to regenerate (Houschyar et al., 2016). Two of the most prevalent and severe problems are stroke and Alzheimer’s disease, which are the third and sixth leading causes of death in the US, respectively (Kochanek et al., 2016). Currently, there are no available treatments that slow the progression of Alzheimer’s disease, as the underlying mechanism is not understood; however, research suggests that amyloid β is associated with some of the neurodegenerative symptoms (Folch et al., 2016; Szeto and Lewis, 2016). Parkinson’s disease is another major neurological disease that affects over 10 million people worldwide and causes progressive deterioration of motor function with no cure at present (Haring et al., 2017a; Szeto and Lewis, 2016). The poor translation between animal models and human trials (Francardo, 2018) and difficulty in delivering drugs across the blood brain barrier (BBB) (Tonda-Turo et al., 2018) are two translational barriers associated with the development of therapies for Parkinson’s disease and Alzheimer’s disease. Brain

Introduction

tumors affected over 26,000 Americans in 2017, and the most common brain tumor, glioblastoma multiforme, has no known cure; its mean survival time is 10 18 months (Preusser et al., 2011; American Society of Clinical Oncology, 2017). With increased life expectancy worldwide and decreasing fertility rates in many countries, an aging population is expected to exhibit even larger numbers of these diseases, which disproportionately affect the elderly. Alzheimer’s disease alone is expected to affect 16 million Americans by the year 2050 (Haring et al., 2017a). Consequently, there is a need to create NSCs that focus specifically on the brain, spinal cord, and peripheral nervous system, the interfaces thereof, and systems that involve all components. However, NSC design and fabrication approaches vary widely, depending on application.

Design and biofabrication approaches for systems-on-chips While multiple designs for BOCs, SOCs, and NSCs are available, they can be broadly categorized into three groups: microfluidic-based systems, hydrogelbased systems, and compartmentalized static fluid systems, or combinations thereof. While brain and spinal cord slice cultures represent gold standard NSCs for in vitro studies, the scope of this chapter is limited to a discussion of NSCs that could potentially be fabricated using patient-derived cells. Biofabrication processes such as three-dimensional (3D) bioprinting have enabled the creation of hydrogel-based systems in which cells exhibit biomimetic 3D cell cell and cell matrix interactions that also contain microfluidic channels (Kolesky et al., 2014). Each design has advantages and disadvantages in particular applications, as summarized in Table 6.1. For example, microfluidic systems are commonly used for modeling the BBB because of the convenience of seeding astrocytes and endothelial cells in a multilayer configuration that mimics the BBB. Microfluidic systems are also attractive for studying organ response to mechanical stimuli, owing to compatibility with pneumatic actuation approaches (Huh et al., 2010). Alternatively, hydrogel-based systems have shown promise for modeling pathogenesis and regeneration in the brain and spinal cord because of their mimicry of native cell cell and cell matrix interactions from both a chemical and mechanical perspective (Lozano et al., 2015; Haring et al., 2019). Our group recently created highly processable, bioinspired, 3D-printable bioinks for fabricating freestanding glial and neural tissues that exhibit moduli ranging from 6 to 13 kPa (Haring et al., 2019). In addition to considering the anatomy of the neural system of interest (e.g., brain, spinal cord, peripheral nervous system, or a combination thereof), the NSC design should be influenced by the need for integrating drugdelivery systems (e.g., polyester microspheres), as well as sensors and actuators (e.g., optoelectronics), because of the role of chemotaxis and electrical phenomena in many higher-order pathophysiological diseases. These approaches have resulted in the creation of BOCs that exhibit a variety of designs. In particular, multiple studieson BBB models and models for neurite outgrowth have been reported.

217

218

CHAPTER 6 Brain-on-a-chip systems

Table 6.1 Summary of the advantages and disadvantages associated with different microphysiological neural system designs.

System type

Dimension of cell cell and cell matrix interaction

Microfluidic

2D

Compartmentalized

2D or 3D

Spatially patterned

2D

Hydrogel

3D

Mimics extracellular matrix, provides realistic cell cell and cell matrix interactions, potential for highly biomimetic systems

Spheroid

3D

Provides realistic 3D cell cell interactions, extracellular matrix is native to cells

3D, Three-dimensional.

Advantages

Disadvantages

Dynamic flow, dynamic control of chemical environment, stable for long-term experiments, ideal for modeling vascular systems Mimics heterogeneous systems, control of interaction and mass transport across compartments, ideal for studying neurite outgrowth, transport, and interaction Mimics heterogeneous environment, ideal for controlling differentiation or migration

Does not mimic 3D cell cell and cell matrix interactions

Restricted to 2D culture under static conditions unless combined with microfluidics or hydrogel systems

Static fluid environment, does not mimic 3D cell cell and cell matrix interactions Limited in scale without vascularization, challenging to create heterogeneous systems, challenging to characterize and monitor cell behavior using traditional techniques Systems are homogeneous, limited ability to incorporate additional functionality

State-of-the-art brains-on-chips

State-of-the-art brains-on-chips Microfluidic models An important area of research in pharmacological treatment of neurological diseases and disorders, the BBB has been modeled extensively with microfluidic BOCs. BOCs for BBB applications are typicallymicrofluidic-based constructs that facilitate modeling of the flow conditions in the BBB (Griep et al., 2013; Booth and Kim, 2012; Prabhakarpandian et al., 2013; Wevers et al., 2018). For example, as shown in Fig. 6.1, Booth and Kim created a BOC for BBB research using a polydimethylsiloxane (PDMS) two-channel microfluidic device that contained a semipermeable membrane coated with bEnd.3 brain endothelial cells on one side and C8-D1A astrocytes on the other (Booth and Kim, 2012). Transendothelial electrical resistance (TEER) assays were conducted by placing electrodes on both sides of the membrane, and the tests indicated tight junctions and a stable BBB model (Booth and Kim, 2012). Permeability was assessed by flowing various molecular-weight fluorescein isothiocyanate (FITC)-dextrans through the endothelial cell microfluidic channel and measuring the FITC-dextran concentration on the astrocyte side of the membrane (Booth and Kim, 2012). The results indicated permeability similar to that of previous BBB models and suggested that the system could be used for drug-delivery studies. In another BBB model, Griep et al. (2013) used a similar microfluidic design with a transwell membrane coated with hCMEC/D3 human BBB endothelial cells and functionalized with TEER electrodes to investigate the effects of shear stress and chemical stress on BBB function; increasing the flow rate, and therefore shear stress, increased the barrier tightness (Griep et al., 2013). The addition of tumor necrosis factor α resulted in decreased barrier tightness (Griep et al., 2013). Another microfluidic BBB model was developed by Prabhakarpandian et al. (2013), which used microgaps to separate the two channels: one containing a brain endothelial cell line (RBE4) and the other containing astrocyte-conditioned media. FITCdextran permeation studies and the presence of tight junction proteins suggested a stable and functional BBB model (Prabhakarpandian et al., 2013). As discussed earlier, the creation of 3D hydrogel systems that contain integrated microfluidic channels provides a new avenue for constructing biomimetic BOCs, SOCs, and NSCs. Recent efforts using microfluidic-based BBB models have evolved from incorporating extracellular matrix (ECM) replicating hydrogels instead of transwell membranes or filters (Wevers et al., 2018; Campisi et al., 2018).

Compartmentalized neuronal models Compartmentalized neuronal models are systems that contain multiple culture regions connected by microchannels, and are often used to study the development, behavior, and diseases associated with neural networks (Ionescu et al., 2016; Fantuzzo et al., 2018; Coquinco and Cynader, 2015). While the presence of microchannels facilitates the incorporation of fluid metering, compartmentalized neuronal models are often static fluid systems, which offers advantages for establishing controlled, biomimetic 3D spatiotemporal gradients of molecular species. Most

219

220

CHAPTER 6 Brain-on-a-chip systems

FIGURE 6.1 State-of-the-art brain-on-a-chip devices. (A) A multiple-organ system mimicking the conditions of the intestine, liver, and a brain tumor; the system biomimetically conditions drugs prior to exposure to glioma for improved drug screening (Jie et al., 2017). (B) A blood brain barrier device on a chip consisting of two channels separated by a porous membrane with cocultured brain endothelial cells and astrocytes. TEER electrodes were incorporated on both sides of the membrane to monitor tight junctions (Booth and Kim, 2012). (C) A spheroid-based brain-on-a-chip device that introduces interstitial flow and amyloid β to model Alzheimer’s disease. (D) A compartmentalized brain-on-a-chip device for modeling 3D neuronal networks by growing cortical neurons into an aligned extracellular matrix (Bang et al., 2015). PDMS, Polydimethylsiloxane; TEER, transendothelial electrical resistance. (A) © 2017 The Royal Society of Chemistry; (B) © 2012 The Royal Society of Chemistry; (C) © 2014 The Royal Society of Chemistry; and (D) © 2015 John Wiley and Sons.

commonly, these systems are surface-based devices that utilize cell-seeding approaches; however, some designs also incorporate hydrogel scaffolds as substrates for growth or chamber separation. The microchannels are typically parallel to facilitate the growth of aligned axonal networks between chambers. Similar to microfluidic systems, compartmentalized neuronal chambers are often fabricated from PDMS using soft lithography. Coquinco and Cynader developed a

State-of-the-art brains-on-chips

three-compartment design that isolated neuronal cell bodies but allowed for axonal interaction to study activity-dependent synaptic competition, an aspect underlying brain development (Coquinco and Cynader, 2015; Coquinco et al., 2014). This device allowed neurons in one compartment to interact and form synapses with neurons in both a control compartment and a test compartment in which the neuronal bodies were subject to activity enhancers or inhibitors (e.g., tetrodotoxin or potassium chloride) (Coquinco and Cynader, 2015; Coquinco et al., 2014). Johnson et al. (2016) created a 3D-printed NSC based on a compartmentalized neuronal chamber design that involved interaction between cells found in the central and peripheral nervous system and reported that Schwann cells participated in axon-tocell viral transmission and may affect infection by the pseudorabies virus. In an effort to mimic the development and behavior of neural networks in a realistic brain environment, Bang et al. (2015) developed a compartmentalized device in which axons grew through microchannels filled with Matrigel, as shown in Fig. 6.1. This allowed the production and visualization of 3D neural circuits for fundamental neuroscience and neurodegenerative therapy research (Bang et al., 2015). Thus, compartmentalized systems are particularly promising for modeling neurodegenerative diseases associated with axonal transport and synapse formation.

Hydrogel-based models Hydrogel-based BOCs, SOCs, and NSCs are another 3D culture platform that offers a highly biomimetic microenvironment. In addition to allowing cells to interact, migrate, and propagate in 3D, hydrogel systems also exhibit transport properties that mimic in vivo conditions (Huh et al., 2011). Lozano et al. (2015) 3D-printed a gellan gum hydrogel culture that mimicked the layered structure of the brain and demonstrated that cortical cells developed rapidly into neural networks in this environment. Gu et al. (2016) 3D-printed human neural stem cells in a polysaccharide hydrogel; the stem cells differentiated into neurons and neuroglia and then formed neural networks that exhibited spontaneous activity. A concentric cylinder design presented by Chwalek et al. (2015) was composed of a silk outer layer seeded with primary cortical neurons and a cell-free collagen hydrogel inner layer; neuronal projections using this design were found to penetrate into the inner layer, and cultures were maintained for several weeks (Chwalek et al., 2015). Maclean et al. (2018) recently reported that by using a programmable Fmoc chloroformate ester, astrocytes in self-assembled peptidebased hydrogel exhibited behavior similar to that observed in vivo. These hydrogel systems present a path toward realistic brain models, but challenges such as transport of nutrients, oxygen, and waste products must be overcome to achieve long-term cultures of macroscopic neural or innervated tissue constructs.

Spheroid models Spheroids are a type of scaffold-free 3D cell culture model composed of a spherical aggregation of cells (Sutherland, 1988). The advantage of spheroids over 2D cultures is the ability to create more realistic cell cell interactions that are found in vivo (Sutherland, 1988; Fennema et al., 2013). Spheroid cultures have been

221

222

CHAPTER 6 Brain-on-a-chip systems

incorporated into microfluidic NSCs to establish improved platforms for disease modeling and drug discovery. A BOC model developed by Park et al. (2015) showed that flowing media over neurospheroids at similar flow rates to those present in brain interstitial fluid resulted in more complex and robust neural networks. Amyloid β, a peptide involved in plaque formation in Alzheimer’s disease, was then applied to the neurospheroids, resulting in cell death and destruction of neural networks as shown in Fig. 6.1 (Park et al., 2015). Jorfi et al. (2018) developed neurospheroids from genetically-engineered stem cells with familial Alzheimer’s disease mutations, which accumulated two of the major markers of Alzheimer’s disease, amyloid β and phosphorylated tau protein. Fan et al. (2016) created a BOC for tumor research from an array of glioblastoma spheroids and used it to test the efficacy of a combinational therapy of irinotecan and pitavastatin. The ability to create a biomimetic microenvironment using patient-derived cells and large arrays for high-throughput testing makes spheroid models useful in certain BOC and SOC applications. Additional functionalities, such as drug release systems and bio-electronic interfaces, can also be integrated with BOCs to improve biomimicry of the model and facilitate sensing and stimulation.

Higher-order system-on-a-chip functionality Spatiotemporal control of chemoattractants, drugs, and other biologics in the microenvironment via integrated drug release systems Spatiotemporal control of soluble molecular species in the microenvironment is an important aspect that underlies neurophysiological and neuropathophysiological processes driven by cell migration and cell differentiation. It is also a critical design consideration in drug screening applications. Growth factor gradients can be established by several different techniques, which can be classified as flow- or static fluid based. The drug release systems associated with these techniques include microfluidic streams that provide concentration boundary conditions, loaded hydrogels, and loaded microspheres, each of which establish a unique diffusive gradient. Stimuli-responsive controlled release has also been examined (Gupta et al., 2015), including through 3D printing, a useful technique in creating novel, controlled drug-release systems with multicomponent programmable spatiotemporal release profiles (Haring et al., 2018). Contact printing can facilitate the fabrication of adsorbed molecular gradients on the surfaces of cell culture substrates. Hynd et al. (2007) used contact printing to pattern biotinylated ECM proteins (fibronectin and laminin) and the laminin peptide biotin-isoleucine-lysinevaline-alanine-valine (IKVAV) onto a hydrogel surface and reported that LRM55 astroglioma and primary rat neurons selectively adhered to the patterned areas. As discussed, microfluidic systems have also been used to establish controlled concentration gradients and offer advantages such as ease of changing the gradient parameters (e.g., distribution and concentration) during testing (Fan et al., 2016;

Higher-order system-on-a-chip functionality

Lee et al., 2011). In addition, microfluidic systems allow for biological conditioning or modification of chemical stimulants prior to application. For example, see Fig. 6.1, which describes a brain-tumor-on-a-chip design presented by Jie et al. (2017). Drugs screened against U251 glioma cells were first passed through a channel containing a Caco-2 human colorectal epithelial culture to mimic the intestine, followed by a chamber containing a HepG2 culture mimicking the liver, before entering the chamber that contained the glioma. This is a promising design for pharmaceutical applications because it provides an in vitro model for physiological drug metabolism prior to reaching the target organ (Jie et al., 2017). In another microfluidic system, a biomimetic heterogeneous synapse density distribution in cortical neurons was achieved by creating a stepwise gradient in the concentrations of the neurotrophic factors B27 and NGF (Kunze et al., 2011). This chemical treatment resulted in layered differences in synapse densities in vitro, which is important for creating realistic BOCs (Kunze et al., 2011). Exposing BOCs, SOCs, and NSCs to controlled distributions of molecular species provides a critical aspect of modeling higher-order responses in the nervous system and developing new therapies for neurological diseases and disorders.

Integration of bioelectronic interfaces with microphysiological neural systems for in situ sensing and stimulation capabilities In the majority of current neural microphysiological systems, data are acquired through conventional characterization approaches, such as microscopy, and via offline bioassays, such as enzyme-linked immunosorbent assay and polymerase chain reaction. While microscopy provides real-time monitoring, the output image data are high-dimensional in nature and consist of an intensity matrix associated with each color in the image, requiring software to facilitate analysis of spatiotemporal data. In addition, while such bioassays are robust, they are methodsbased approaches and thus do not provide real-time monitoring capability. Alternatively, sensors can facilitate the collection of low-dimensional, real-time data from BOCs, SOCs, and NSCs, provided they are integrated with microphysiological systems (e.g., 3D bioprinted mini-tissues). While "instrumented mini-tissues" based on the integration of sensors with microphysiological systems are still emerging, sensors are now being integrated into microfluidic-based NSCs. For example, in BBB models, TEER electrodes have been leveraged to quantify the formation, stability, and morphology of the BBB (Griep et al., 2013; Booth and Kim, 2012; Helms et al., 2016). TEER sensing is conducted by placing electrodes (typically chopstick or cup electrodes) on either side of the BBB model and measuring the resistance across the barrier (Helms et al., 2016); resistance increases as endothelial cells grow and form tight junctions. Destabilizing the BBB model results in decreased resistance, providing a useful approach for measuring conformational and morphological changes in the BBB in real time, as shown in Fig. 6.2 (Booth and Kim, 2012; Helms et al., 2016). Coronado-Vela´zquez et al. (2018) used a TEER-functionalized BBB model

223

224

CHAPTER 6 Brain-on-a-chip systems

to study the underlying biological mechanisms of primary amebic meningoencephalitis, a deadly disease caused by bacterial infection that disrupts intercellular junctions in the BBB. Microelectrode arrays have also been used to facilitate realtime monitoring of neuronal electrical activity (Huang et al., 2012; Obien et al., 2015), but they have not yet been applied to BOC, SOC, and NSC systems with higher-order structures. For example, Charkhkar et al. (2015) cultured primary mouse neurons that formed a spontaneously active network on a microelectrode array; modulation of this culture with oligomeric amyloid β resulted in decreased network activity, and application of Alzheimer’s disease therapeutics memantine

FIGURE 6.2 Sensing and the bioelectronic interface for brain-on-a-chip devices. (A) Real-time data acquired with TEER electrodes across a blood brain barrier model, demonstrating the degradative effects of Naegleria fowleri that cause primary amebic meningoencephalitis (Coronado-Vela´zquez et al., 2018). (B) A substrate-integrated microelectrode array for detecting activity in neuronal cultures used for investigating the neurotoxic effects of amyloid β in an Alzheimer’s disease model (Charkhkar et al., 2015). (C) A “bioelectronic neural pixel” that allows simultaneous chemical stimulation and electrical sensing at the single-neuron level. The reservoir contains positively charged compounds that are transported by the CEM through the PEDOT:PSS recording electrode to the neuron (Jonsson et al., 2016). (D) Human neuronal stem cells showed enhanced differentiation on a graphene substrate that also provided electrical coupling (Park et al., 2011). CEM, Cation exchange membrane; TEER, transendothelial electrical resistance. (A) © 2018 John Wiley and Sons; (B) © 2015 Elsevier; and (D) © 2011 John Wiley and Sons.

Higher-order system-on-a-chip functionality

and methylene blue resulted in a rebound in network activity, suggesting that this microelectrode array-based system-on-a-chip could be used as a drug-screening device for Alzheimer’s disease therapeutics (Charkhkar et al., 2015). While electrical sensors have facilitated the monitoring of conformational changes between cell assemblies and neuronal electrical activity in real time, it is also critical to measure other parameters associated with NSCs in real time to develop a mechanistic understanding of disease pathophysiology, particularly the temporal concentration of molecular species in the system. Biosensors are device-based bioassays that facilitate real-time detection, identification, and quantification of biologics (e.g., proteins, DNA, RNA, and cells). Biosensors are also compatible with label-free and sample preparation-free protocols, making them highly attractive platforms for in situ monitoring of biological systems and processes. A biosensor is composed of three primary components: a transducer, a receptor, and a readout system. A detailed review of biosensors is beyond the scope of this chapter and can be found elsewhere, but here we discuss the considerations underlying design of an effective biosensor for application in BOCs, SOCs, and NSCs. A biosensor for use in NSCs should exhibit rapid responses after expression, so as to not require additional considerations, such as time delays in data interpretation, sensitivity in hydrogel matrices, and spatial resolution. While the measurement must also be robust and the potential for falsepositive and -negative responses must be minimized, efforts should be taken to integrate controls within the measurement approach, for example, in the form of reference sensors. Sensors should quantify the level of a given biological target in the system; NSC constructs should also contain hierarchical sensor architectures that facilitate the simultaneous monitoring of molecular expression levels and higherorder functional properties, such as mechanical properties (e.g., viscoelastic). Thus sensor development efforts for BOCs, SOCs, and NSCs should emphasize versatile sensing formats, such as those with integrated physical property sensing, chemical sensing, and biosensing. In addition to millimeter- and micrometer-scale sensors developed from top-down fabrication approaches, molecular-scale sensors, such as plasmonic nanoparticles, are also promising tools for real-time monitoring of gene expression in BOCs, SOCs, and NSCs. We recently showed that it is possible to 3D-print hydrogel constructs that contain spatially programmed distributions of up to eight different types of plasmonic nanoparticles, demonstrating their potential in mini-tissue-based sensing applications (Haring et al., 2017b).

Molecular-level interface between electronic materials and neural tissue toward constructing next-generation neural interfaces and bionic neural systems Given the importance of monitoring and modulating neuronal electrical activity using electrical sensors and monitoring gene expression in neuronal and glial cells using biosensors, it is of interest to examine approaches for interfacing electronic

225

226

CHAPTER 6 Brain-on-a-chip systems

materials with biologics found in the nervous system (e.g., ECM domains, cell membranes, and intracellular domains). It is also of interest that the approaches for bioelectronic interface are selective to facilitate directed assembly and interface with molecular targets such as particular ECM moieties or surface receptor molecules. In addition to the need for molecular-scale integration of electronic materials and biomaterials, there is a need to consider the extent of mechanical matching between the integrated materials (Banerjee et al., 2009; Saha et al., 2008). Thus polymer-based electronics are attractive candidate materials for neural interfaces and bioelectronic therapies, because of their relatively closer elastic modulus to neural tissue than that of metals (kPa vs GPa). These constraints and requirements of an ideal bioelectronic neural interface approach suggest promise for molecular-scale directed selfassembly processes using polymer-based nanotechnology. Many of the strides taken in improving this bionic interface in the past two decades have been made through material science. For example, Kim et al. (2013) synthesized graphene oxide coated gold nanoparticles that enabled in situ monitoring of neural stem-cell differentiation through simultaneous electrochemical measurements and surface-enhanced Raman spectroscopy. Park et al. (2011) found that a graphene substrate yielded enhanced differentiation of human neural stem cells and provided good electrical coupling with the differentiated neurons for electrical stimulation. Kam et al. (2008) developed carbon nanotubes coated with the ECM protein laminin that supported the differentiation of neural stem cells into neurons and performed well as neural electrodes for stimulation. Shown in Fig. 6.2, a new capability for simultaneous electrical and chemical stimulation interface, called a neural pixel, was developed by Jonsson et al. (2016), who used electrically and ionically conductive polymers to develop electrodes that could deliver chemical stimulation and conduct electrical signals simultaneously at the single-cell level in hippocampal samples. Murbach et al. recently demonstrated the ability to polymerize polymer electrodes in situ as a new approach for creating highly compliant injectable neural interfaces for bioelectronic therapies (Tong et al., 2018; Murbach et al., 2018). Importantly, the design of these state-of-theart BOCs and the ability to integrate higher-order functionalities are constrained by the manufacturing approach utilized.

Manufacturing approaches Mold lithography Mold lithography has emerged as a useful technique for biological research; highresolution microscale molds are used to create both microfluidic devices and 3D hydrogel tissue constructs (Madou, 2011). While lithography processes often vary depending on the material removal step, here we discuss the creation of microfluidics and micromolds using SU-8 mold lithography, as it provides a relatively low-cost approach for creating high-quality devices (Lorenz et al., 1997). The first

Manufacturing approaches

step in the process is cleaning the surface of the wafer with piranha solution (a mixture of sulfuric acid and hydrogen peroxide). The wafer is then coated with a photosensitive chemical, SU-8 photoresist, and baked to solidify. A photomask or laser is used to selectively expose the surface, and then the wafer is baked again. Next, the wafer is submerged in a development solution that dissolves the areas of SU-8 not exposed to ultraviolet radiation. A final baking step can be added to improve the durability of the mold. The resultant mold is commonly termed the “master,” and the negative image of the desired system is etched into it. PDMS is typically the soft polymer used for making the negative impression of the mold because of its biocompatibility and ease of use. PDMS is available as a liquid-phase linear polymer with a crosslinking agent that can be mixed in prior to molding and exhibits a moderate potlife. PDMS can be cured at room temperature or with moderate heating and yields an elastomeric solid conforming to the master, which can be removed and used without additional chemical treatment. Soft lithography is the workhorse manufacturing process for microfluidic fabrication and is also used to create soft compartmentalized systems and stamps for contact printing (Sackmann et al., 2014). More complex microfluidic designs can be printed in multiple layers and affixed together with additional PDMS as a glue (Sackmann et al., 2014).

Contact printing To create biomimetic models of organs and diseases, it is imperative to control the spatial distribution of diffusive molecular species in 3D, such as native biomolecules (e.g., chemoattractants) or exogenous species (e.g., pharmaceuticals) (Bittner et al., 2018). Contact printing is a common approach for spatially distributing molecular species across planar cell culture substrates. A PDMS stamp is first produced using soft lithography. Next, the stamp is immersed in a solution that contains the target molecular species to facilitate their adsorption to the surface of the PDMS stamp as a coating (Hynd et al., 2007). The stamp is then held in mechanical contact with the desired cell culture substrate (e.g., a Petri dish) to transfer the molecular species to the substrate. This process can also be repeated using the same stamp or a different stamp to increase the concentration of transferred molecular species or to create substrates that incorporate multiple molecular species. The PDMS stamp is frequently plasma treated to create a hydrophilic surface prior to contact printing for improved results with aqueous solutions (Bodas and Khan-Malek, 2006).

Hydrogel casting Hydrogels are used for scaffolds and ECM replicants in 3D cell cultures for tissue engineering and study of microphysiological systems (Tibbitt and Anseth, 2009). Hydrogel geometry can be controlled through molding or 3D printing. Molding lowers requirements on rheological properties and associated equipment, while 3D printing offers improved control over spatial composition, geometry, and internal void space. Hydrogel casting requires a mold, or master, which can be

227

228

CHAPTER 6 Brain-on-a-chip systems

produced through lithography, 3D printing, or machining (Ryan et al., 2015). In the hydrogel casting process, an aqueous-phase polymer or prepolymer is poured into a mold and cured; the curing mechanisms vary by hydrogel system, and curing is often initiated by exposure to ultraviolet light or the addition of a curing agent. Cells can be mixed into the aqueous hydrogel precursor prior to curing if the curing process is not too harsh or can be applied to the surface of the hydrogel following curing.

Three-dimensional printing 3D printing is an emerging manufacturing process for fabrication of microphysiological systems, with the promise of creating complex geometries with spatially controlled composition and voids (Haring et al., 2017b, 2018). For NSCs, inkjet or microextrusion printing systems are typically used because of their ability to produce soft cell-laden tissue constructs using multiple types of printable matrices (Haring et al., 2017a; Murphy and Atala, 2014). In both systems, a robot controls the spatial position of the print head as material is deposited and assembled in a layer-by-layer process into a 3D construct (Murphy and Atala, 2014; Wong and Hernandez, 2012). In inkjet printing, the printing material, known as an ink, is expelled as droplets by a microactuator (often piezoelectric or thermal) (Murphy and Atala, 2014). In microextrusion printing the ink is extruded as a continuous filament from a syringe (often pneumatically or via screw-based extruders) (Murphy and Atala, 2014). The entire system is controlled by a machine code program that contains the tool path data as well as commands for tool triggering (dispensing in 3D printing). Robotic control in material assembly also improves the reproducibility of assembled devices relative to hand-made devices (Wong and Hernandez, 2012; Cesewski et al., 2018). The printed material can be either structural components such as silicone and polycaprolactone or other biocompatible polymers (e.g., hydrogels). Inkjet and microextrusion 3D printers can also accommodate multiple print heads for multi-material manufacturing, such as structural components and hydrogels in a single program, allowing for the fabrication of complex systems such as structural tissues (Murphy and Atala, 2014; Kang et al., 2016). Microextrusion 3D printing systems have been shown to incorporate robotic-assisted pick-and-place capability for integrating nonprinted high-quality electronic materials (e.g., diced chips) within 3D-printed soft polymer microfluidic networks (Cesewski et al., 2018). Thus 3D printing is poised to facilitate the fabrication of “instrumented microphysiological neural systems” (Table 6.2).

Future outlook Although the previous sections suggest a promising future for BOCs, SOCs, and NSCs, various challenges remain as potential barriers to progress. First, it has been recognized that 3D cell culture techniques, such as spheroids and hydrogel

Future outlook

Table 6.2 Common manufacturing processes for brains-on-chips and the associated types of microphysiological systems produced. Manufacturing process Mold lithography Contact printing Hydrogel casting 3D printing

Types of systems produced Microfluidic, spheroid, and compartmentalized systems Spatially patterned systems 3D cell culture systems Microfluidic, compartmentalized, spatially patterned, and 3D cell culture systems

3D, Three-dimensional.

cultures, must be implemented to create truly biomimetic systems-on-chips (Huh et al., 2011). This identifies an area requiring further investigation in microfluidic-based platforms. Second, one of the challenges of using 3D cell culture models is mass-transport limitations. While recent work has provided a method for developing vascularized thick tissues (Kolesky et al., 2016), further work is needed to incorporate them into 3D microphysiological neuronal and glial tissues. Third, NSCs require increased complexity and heterogeneity of cocultures. Although many BOCs, SOCs, and NSCs involve only one and at most three cell types and a single ECM mimic, native tissues consist of multiple cell types, often clustered in a specialized cocultured microenvironment (Jie et al., 2017). For example, the local composition of ECM, specifically the integrity of the perivascular space, has been identified to control neuronal activity (Watkins et al., 2014). Fourth, while multiple new materials have been developed for improved neural electrical interfacing (Park et al., 2011; Kam et al., 2008), their potential has not yet been fully realized through implementation in BOCs. Further research is required in the area of directed self-assembly processes for creating bioelectronic neural interfaces. An additional challenge for disease modeling using NSCs is achieving the throughput necessary for pharmaceutical drug screening. Wevers et al. (2018) addressed this by using a commercial tissue culture plate, OrganoPlate, as the substrate, which allowed 96 BBB models to be tested simultaneously. Finally, approaches for creating instrumented BOCs, such as computer-aided manufacturing processes, are essential to understanding the dynamic nature of disease pathophysiology in the nervous system. The development of these complex heterogeneous systems with higher-order functionalities could yield in vitro models that accurately mimic higher-order physiological and pathophysiological processes and systems in humans. These systems provide a path toward advancing knowledge in various areas of neuroscience, particularly related to neuro-regeneration and -degeneration as well as the creation of new therapeutics for the treatment of neurological diseases and disorders, such as Alzheimer’s disease, Parkinson’s disease, brain cancer, and nerve injuries.

229

230

CHAPTER 6 Brain-on-a-chip systems

References American Society of Clinical Oncology, 2017. Cancer.net Banerjee, A., Arha, M., Choudhary, S., Ashton, R.S., Bhatia, S.R., Schaffer, D.V., et al., 2009. Biomaterials 30, 4695. Bang, S., Na, S., Jang, J.M., Kim, J., Jeon, N.L., 2016. Adv. Healthc. Mater. 5, 159. Bavli, D., Prill, S., Ezra, E., Levy, G., Cohen, M., Vinken, M., et al., 2016. Proc. Natl. Acad. Sci. U.S.A. 113, E2231. Bittner, S.M., Guo, J.L., Mikos, A.G., 2018. Bioprinting 12, e00032. Bodas, D., Khan-Malek, C., 2006. Microelectron. Eng. 83, 1277. Booth, R., Kim, H., 2012. Lab Chip 12, 1784. Campisi, M., Shin, Y., Osaki, T., Hajal, C., Chiono, V., Kamm, R.D., 2018. Biomaterials 180, 117. Cesewski, E., Haring, A.P., Tong, Y., Singh, M., Thakur, R., Laheri, S., et al., 2018. Lab Chip 18, 2087. Charkhkar, H., Meyyappan, S., Matveeva, E., Moll, J.R., McHail, D.G., Peixoto, N., et al., 2015. Brain Res. 1629, 1. Chwalek, K., Tang-Schomer, M.D., Omenetto, F.G., Kaplan, D.L., 2015. Nat. Protoc. 10, 1362. Coquinco, A., Cynader, M., 2015. Microfluidic and Compartmentalized Platforms for Neurobiological Research. Springer, p. 139. Coquinco, A., Kojic, L., Wen, W., Wang, Y.T., Jeon, N.L., Milnerwood, A.J., et al., 2014. Mol. Cell. Neurosci. 60, 43. Coronado-Vela´zquez, D., Betanzos, A., Serrano-Luna, J., Shibayama, M., 2018. J. Eukaryotic Microbiol. Fan, Y., Nguyen, D.T., Akay, Y., Xu, F., Akay, M., 2016. Sci. Rep. 6, 25062. Fantuzzo, J.A., Hart, R.P., Zahn, J.D., Pang, Z.P., 2018. Dev. Dyn. 248, 65. Fennema, E., Rivron, N., Rouwkema, J., van Blitterswijk, C., de Boer, J., 2013. Trends Biotechnol. 31, 108. Folch, J., Petrov, D., Ettcheto, M., Abad, S., Sa´nchez-Lo´pez, E., Garcı´a, M.L., et al., 2016. Neural Plast. 2016, 8501693. Francardo, V., 2018. Behav. Brain Res. 352, 142. Griep, L., Wolbers, F., De Wagenaar, B., Ter Braak, P., Weksler, B., Romero, I.A., et al., 2013. Biomed. Microdevices 15, 145. Grosberg, A., Alford, P.W., McCain, M.L., Parker, K.K., 2011. Lab Chip 11, 4165. Gu, Q., Tomaskovic-Crook, E., Lozano, R., Chen, Y., Kapsa, R.M., Zhou, Q., et al., 2016. Adv. Healthc. Mater. 5, 1429. Gupta, M.K., Meng, F., Johnson, B.N., Kong, Y.L., Tian, L., Yeh, Y.-W., et al., 2015. Nano Lett. 15, 5321. Hackam, D.G., Redelmeier, D.A., 2006. JAMA 296, 1727. Haring, A.P., Sontheimer, H., Johnson, B.N., 2017a. Stem Cell Rev. Rep. 13, 381. Haring, A.P., Khan, A.U., Liu, G., Johnson, B.N., 2017b. Adv. Opt. Mater. 5, 1. Haring, A.P., Tong, Y., Halper, J., Johnson, B.N., 2018. Adv. Healthc. Mater. 7, 1800213. Haring, A.P., Thompson, E.G., Tong, Y., Laheri, S., Cesewski, E., Sontheimer, H., et al., 2019. Biofabrication . Helms, H.C., Abbott, N.J., Burek, M., Cecchelli, R., Couraud, P.-O., Deli, M.A., et al., 2016. J. Cereb. Blood Flow Metab. 36, 862.

References

Houschyar, K., Momeni, A., Pyles, M., Cha, J., Maan, Z., Duscher, D., et al., 2016. Plast. Surg. Int. 2016, 4175293. Huang, Y., Williams, J.C., Johnson, S.M., 2012. Lab Chip 12, 2103. Huh, D., Matthews, B.D., Mammoto, A., Montoya-Zavala, M., Hsin, H.Y., Ingber, D.E., 2010. Science 328, 1662. Huh, D., Hamilton, G.A., Ingber, D.E., 2011. Trends Cell Biol. 21, 745. Hynd, M.R., Frampton, J.P., Dowell-Mesfin, N., Turner, J.N., Shain, W., 2007. J. Neurosci. Methods 162, 255. Ionescu, A., Zahavi, E.E., Gradus, T., Ben-Yaakov, K., Perlson, E., 2016. Eur. J. Cell Biol. 95, 69. Jang, K.-J., Mehr, A.P., Hamilton, G.A., McPartlin, L.A., Chung, S., Suh, K.-Y., et al., 2013. Integr. Biol. 5, 1119. Jie, M., Mao, S., Liu, H., He, Z., Li, H.-F., Lin, J.-M., 2017. Analyst 142, 3629. Johnson, B.N., Lancaster, K.Z., Hogue, I.B., Meng, F., Kong, Y.L., Enquist, L.W., et al., 2016. Lab Chip 16, 1393. Jonsson, A., Inal, S., Uguz, I., Williamson, A.J., Kergoat, L., Rivnay, J., et al., 2016. Proc. Natl. Acad. Sci. U.S.A. 113, 9440. Jorfi, M., D’Avanzo, C., Tanzi, R.E., Kim, D.Y., Irimia, D., 2018. Sci. Rep. 8, 2450. Kam, N.W.S., Jan, E., Kotov, N.A., 2008. Nano Lett. 9, 273. Kang, H.-W., Lee, S.J., Ko, I.K., Kengla, C., Yoo, J.J., Atala, A., 2016. Nat. Biotechnol. 34, 312. Kehoe, S., Zhang, X., Boyd, D., 2012. Injury 43, 553. Kim, T.-H., Lee, K.-B., Choi, J.-W., 2013. Biomaterials 34, 8660. Kochanek, K.D., Murphy, S.L., Xu, J., Tejada-Vera, B., 2016. National vital statistics reports: from the Centers for Disease Control and Prevention, National Center for Health Statistics. Nat. Vital Stat. Syst. 65, 1. Kolesky, D.B., Truby, R.L., Gladman, A.S., Busbee, T.A., Homan, K.A., Lewis, J.A., 2014. Adv. Mater. 26, 3124. Kolesky, D.B., Homan, K.A., Skylar-Scott, M.A., Lewis, J.A., 2016. Proc. Natl. Acad. Sci. U.S.A. 113, 3179. Kunze, A., Valero, A., Zosso, D., Renaud, P., 2011. PLoS One 6, e26187. Lee, C.-Y., Chang, C.-L., Wang, Y.-N., Fu, L.-M., 2011. Int. J. Mol. Sci. 12, 3263. Lorenz, H., Despont, M., Fahrni, N., LaBianca, N., Renaud, P., Vettiger, P., 1997. J. Micromech. Microeng. 7, 121. Lozano, R., Stevens, L., Thompson, B.C., Gilmore, K.J., Gorkin Iii, R., Stewart, E.M., et al., 2015. Biomaterials 67, 264. Ma, V.Y., Chan, L., Carruthers, K.J., 2014. Arch. Phys. Med. Rehabil. 95, 986. Maclean, F.L., Ims, G.M., Horne, M.K., Williams, R.J., Nisbet, D.R., 2018. Adv. Mater. 30, 1805209. Madou, M.J., 2011. Manufacturing Techniques for Microfabrication and Nanotechnology , vol. 2. CRC Press. Mak, I.W., Evaniew, N., Ghert, M., 2014. Am. J. Transl. Res. 6, 114. Menken, M., Munsat, T.L., Toole, J.F., 2000. Arch. Neurol. 57, 418. Murbach, J.M., Currlin, S., Widener, A., Tong, Y., Chhatre, S., Subramanian, V., et al., 2018. MRS Commun. 8, 1043. Murphy, S.V., Atala, A., 2014. Nat. Biotechnol. 32, 773.

231

232

CHAPTER 6 Brain-on-a-chip systems

Kochanek K.D., Murry S.L., Xu J., Tejada-Vera B., National Vital Statistics Report 64, 2016, 1. Obien, M.E.J., Deligkaris, K., Bullmann, T., Bakkum, D.J., Frey, U., 2015. Front. Neurosci. 8, 423. Park, S.Y., Park, J., Sim, S.H., Sung, M.G., Kim, K.S., Hong, B.H., et al., 2011. Adv. Mater. 23, H263. Park, J., Lee, B.K., Jeong, G.S., Hyun, J.K., Lee, C.J., Lee, S.-H., 2015. Lab Chip 15, 141. Prabhakarpandian, B., Shen, M.-C., Nichols, J.B., Mills, I.R., Sidoryk-Wegrzynowicz, M., Aschner, M., et al., 2013. Lab Chip 13, 1093. Preusser, M., De Ribaupierre, S., Wo¨hrer, A., Erridge, S.C., Hegi, M., Weller, M., et al., 2011. Ann. Neurol. 70, 9. Ryan, J.R., Chen, T., Nakaji, P., Frakes, D.H., Gonzalez, L.F., 2015. World Neurosurg. 84, 1333. Sackmann, E.K., Fulton, A.L., Beebe, D.J., 2014. Nature 507, 181. Saha, K., Keung, A.J., Irwin, E.F., Li, Y., Little, L., Schaffer, D.V., et al., 2008. Biophys. J. 95, 4426. Sutherland, R.M., 1988. Science 240, 177. Szeto, J.Y.Y., Lewis, S.J.G., 2016. Curr. Neuropharmacol. 14, 326. Tibbitt, M.W., Anseth, K.S., 2009. Biotechnol. Bioeng. 103, 655. Tonda-Turo, C., Origlia, N., Mattu, C., Accorroni, A., Chiono, V., 2018. Curr. Med. Chem. 25, 1. Tong, Y., Murbach, J.M., Subramanian, V., Chhatre, S., Delgado, F., Martin, D.C., et al., 2018. MRS Adv. 3, 2365. Watkins, S., Robel, S., Kimbrough, I.F., Robert, S.M., Ellis-Davies, G., Sontheimer, H., 2014. Nat. Commun. 5, 4196. Wevers, N.R., Kasi, D.G., Gray, T., Wilschut, K.J., Smith, B., van Vught, R., et al., 2018. Fluids Barriers CNS 15, 23. Wong, K.V., Hernandez, A., 2012. ISRN Mech. Eng. 2012. Yi, Y., Park, J., Lim, J., Lee, C.J., Lee, S.-H., 2015. Trends Biotechnol. 33, 762.

CHAPTER

Kidney-on-a-chip

7 Filippo Zanetti

PMI R&D, Philip Morris Products S.A., Neuchaˆtel, Switzerland

Introduction Structure and function of the kidney The kidneys control the body’s balance of fluids and eliminated waste products by regulating the mechanisms of filtration, reabsorption, and secretion through the nephron, the functional unit of the kidney. The initial filtering unit of the nephron is the renal corpuscle, which includes the Bowman’s space, encapsulating the capillaries (glomerulus), and the renal tubules (proximal tubule and distal convoluted tubule), which reabsorb the remaining molecules into the bloodstream (Zhuo and Li, 2013; Lote and Lote, 1994) (Fig. 7.1). The proximal tubule reabsorbs most of the sodium chloride and sodium bicarbonate, completes the reabsorption of glucose, and is the sole site of transport for amino acids and important anions, such as citrate and phosphate. The kidney is also a metabolically active organ, an important site of gluconeogenesis, involved in vitamin D metabolism (Meyer et al., 1999; Jones et al., 2012), and the primary site of drug clearance (Yeung and Himmelfarb, 2019). The distal convoluted tubule plays a critical role in sodium, potassium, and divalent cation homeostasis, reabsorbing 5% 10% of filtered sodium and chloride, participating in potassium cation secretion, and maintaining systemic calcium and magnesium homeostasis (Mccormick and Ellison, 2015). Each nephron separates water, ions, and small molecules from the blood through ultrafiltration. Every day, approximately 180 L of filtrates flow through the renal tubule lumen, exposing the renal apical cells to fluid shear stress (Raghavan et al., 2014). Fluid shear stress is responsible for renal cell polarization, which occurs through differential gene expression and consequent cytoskeletal reorganization (Essig et al., 2001; Jang et al., 2010; Kaysen et al., 1999). In humans, fluid shear stress has been estimated to range from 0.7 to 1.2 dyn/cm2 (Essig et al., 2001). Proximal tubule epithelial cells (PTECs) are constantly subjected to transepithelial osmotic gradient and fluid shear stress; these cells can sense the flow with a single sensory organelle, the primary cilium (Praetorius, 2015), and transmit it into the cells to reorganize the junctional Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00007-3 © 2020 Elsevier Inc. All rights reserved.

233

FIGURE 7.1 Anatomical view of the kidney and schema of kidney transport mechanisms. (A) Sagittal section of the kidney, showing the cortex, medulla, and renal pelvis. (B) The proximal tubule extends from the Bowman’s capsule (cortex) and leads to the loop of Henle, which is connected to the distal tubule, ending in the connecting duct system. (C) Proximal tubule filtration. Solutes in the plasma (blue and orange) are freely filtered in the glomerulus, while proteins/drugs bound to proteins (red) are actively secreted by transporters expressed on the membrane of tubular epithelial cells. Important solutes are reabsorbed in the bloodstream (blue), while the other solutes (orange) and metabolites/drugs (red) are eliminated with urine. (D) Transport systems in the membrane of the proximal tubular epithelial cells facing the lumen of the proximal tubule (apical side) and the interstitium/blood (basolateral side). ATP-binding cassette transporters use ATP as source of energy to mediate the efflux; solute carriers-mediated transport is driven by electrochemical gradients of the compounds. Megalin and cubilin are endocytic receptors that reabsorb low-molecular-weight proteins (e.g., vitamins and hormones). ATP, Adenosine triphosphate; BCRP, breast cancer resistance protein; KG, ketoglutarate; MATE, multidrug and toxin extrusion transporter; MRP, multiresistance protein; NaPi, sodium phosphate transporter; OA, organic anion; OAT, organic anion transporter; OC, organic cation; OCT, organic cation transporter; OCTN, organic cation/carnitine transporter; P-gp, p-glycoprotein (multidrug resistance protein 1); SLCO, solute carrier organic anion; URAT, urate transporter.

Introduction

complex and the cytoskeleton and enhance apical endocytosis (Essig et al., 2001; Raghavan et al., 2014). The processes occurring in the tubular system (influx, efflux, and intracellular metabolism) are factors in the development of drug-induced kidney diseases (Nigam et al., 2015; Fisel et al., 2014). Influx of solutes occurs at the basal (absorption) and apical (reabsorption) membranes of PTECs and is mediated by solute carrier transporters, which regulate the uptake of both xenobiotics and endogenous substrates circulating in the blood from the basolateral membrane of the proximal tubule cells. These include the organic anion transporters (OATs, members of the solute carrier organic anion (SLCO) gene subfamily) and organic cation transporters (OCTs). Intracellular accumulation of internalized solutes is prevented by an efficient efflux mechanism from the apical membrane of PTECs to the lumen. Efflux occurs by solute carrier transporters, including OAT4, multidrug and toxin extrusion protein (MATE1/2-K), urate anion exchanger 1 (URAT1), and organic cation/carnitine transporters 1 and 2. Adenosine triphosphate binding cassette transporters are also present on the apical membrane of PTECs to mediate the efflux; these include multidrug resistance protein 1 (MDR1 or P-gp), breast cancer resistance protein, and multidrug resistance associated proteins 2 and 4.

Kidney pathologies Kidney pathologies can be categorized as acute kidney injury and chronic kidney disease, and the incidence of both types has increased over time (Coresh et al., 2007; Hsu et al., 2007; Xue et al., 2006). Acute kidney injury can result in permanent kidney damage, leading to the development or worsening of chronic kidney diseases (Basile, 2004; Venkatachalam et al., 2010). Acute kidney injury manifests when the blood supply to the kidney is interrupted (prerenal acute kidney injury) or when the urinary flow in the ureters is blocked (postrenal acute kidney injury) (Makris and Spanou, 2016). These events can occur, for example, in patients with renal hypoperfusion (Blantz, 1998) or urinary stone disease and can result in impaired renal blood flow and inflammatory processes that also reduce the glomerular filtration rate (Mahmud and Mahmud, 2017). In addition, acute kidney injury can originate from injuries in the different parts of the kidney (intrinsic acute kidney injury), such as the tubules (e.g., caused by acute tubular necrosis), the glomerulus (caused by glomerulonephritis), and the vascular system (Makris and Spanou, 2016). As the kidney receives 25% of cardiac output (Schnellmann, 2001), and renal tubules connect to blood vessels, the kidney is particularly susceptible to ischemic damage. Any systemic or intrarenal circulation failure may exert a notable effect on renal perfusion, resulting in vasoconstriction, cell death, and inflammation (Bonventre, 2010; Bonventre and Yang, 2011; Sharfuddin and Molitoris, 2011). The kidney is the second major target of drugs and chemicals, after the liver. Heavy metals, chemicals, fungal toxins, and pharmacotherapy can cause

235

236

CHAPTER 7 Kidney-on-a-chip

nephrotoxicity, leading to acute kidney injury that is often associated with increased morbidity and mortality (Hoste et al., 2015; Choudhury and Ahmed, 2006). Twenty percent of acute kidney injury is believed to be associated with drug-induced nephrotoxicity (Hoste et al., 2015; Awdishu and Mehta, 2017). These drugs include nonsteroidal antiinflammatory compounds, antibiotics, chemotherapeutics, and radiocontrast agents (Fanos and Cataldi, 2001; Murray and Brater, 1993; Perazella and Moeckel, 2010; Weisbord and Du Cheryon, 2018). Acute kidney injury accounts for 5% 7% of hospitalizations in the United States (Tolwani, 2012; Lameire et al., 2013). The global burden of kidney injury increased between 1988 and 2003, apparently because of the increase in drug prescription associated with nephrotoxicity (Tomlinson et al., 2013; Lameire et al., 2013). Acute kidney injury prolongs hospitalization and hospital expenses, for an annual cost of up to $8 billion in the United States and d434 d620 million in the United Kingdom (Fischer et al., 2005). Developing drugs with low nephrotoxic potential would minimize kidney injury; however, with the current preclinical research tools available, the toxic potential of candidate compounds is often underestimated (Szeto and Chow, 2005; Redfern et al., 2010). Drug attrition estimated from preclinical studies is only 2%, but this level increases to 9% in clinical trials, and toxic effects of new medicines are reported in 20% of new medicines during postmarket surveillance (Laverty et al., 2011). Kidney dysfunction increases the risk of adverse drug events by affecting intestinal, hepatic, and renal drug metabolism (Yeung et al., 2014; Dreisbach and Lertora, 2008). Understanding the mechanism of nephrotoxicity is therefore crucial to developing safer drugs. Current models used in preclinical kidney research do not fully replicate the structure and function of the kidney and are poor predictors of drug-induced kidney diseases. These limitations must be overcome by more reliable, predictive in vitro models.

Challenges associated with developing in vitro kidney models The kidney is composed of more than 10 cell types that are organized in a threedimensional (3D) network with a complex vascular system and surrounded by extracellular matrix (ECM) (Tiong et al., 2014). All of these features must be taken into account when developing a 3D in vitro model that replicates renal structure and function. Given the physiological and pathological importance of blood and secretory circulation in the kidney, a kidney model should incorporate a microfluidic platform mimicking luminal and tissue circulation. Culturing cells under flow rather than static conditions mimics the in vivo situation and enables the study of cell signal transduction (Essig et al., 2001; Raghavan et al., 2014; Jang et al., 2013), gene and protein expression (Choucha Snouber et al., 2012; Kaysen et al., 1999; Snouber et al., 2012), and responses to drugs (Jang et al., 2013; Snouber et al., 2012).

Introduction

To fully replicate the cellular complexity of the kidney, various cell types should be cocultured to enable cell cell interactions, specific signaling pathways, and immune cell recruitment (Linas and Repine, 1999; Tasnim and Zink, 2012). Moreover, it is critical that an in vitro renal system expresses functional transporters for appropriate method validation and data interpretation as well as a biological response to drugs that resembles the in vivo response. A 3D conformation resembling the architectural structure of the kidney should be achieved in in vitro models (Fitzgerald et al., 2015). In the tubular structures, for example, cell behavior is influenced by the curved shape of the tubules (Shen et al., 2013; Fitzgerald et al., 2015). Finally, the cell types selected to populate the chip should retain their human characteristics and enable the measurement of biomarkers relevant to the in vivo situation.

Microphysiological kidney models Two-dimensional versus three-dimensional systems Animal models are used to study the whole-kidney response, but interspecies differences in blood circulation, speed of metabolizing drugs, and transporter expression preclude extrapolation to clinical parameters (Heinonen et al., 2017; Kim and Takayama, 2015; Chu et al., 2013). Moreover, animal testing raises ethical concerns (Hartung, 2009). Isolated ex vivo human kidney slices and tubules are the most representative of the in vivo environment, but their limited viability hinders their use in drug research (Yeung and Himmelfarb, 2019). Primary cells isolated from human donors can survive up to 2 weeks and can be transfected with transporters, but their expression and activity levels cannot be controlled well (Yeung and Himmelfarb, 2019). Two-dimensional (2D) cultures have long been used in in vitro research and are inexpensive, easy to handle, and useful in studying the molecular mechanisms and signal transduction underlying biological processes. 2D cultures have been used extensively to study renal physiology, pharmacology, and pathology, but they cannot mimic the renal microenvironment because, for example, they lack apical basal polarization and cannot replicate transepithelial transport or experience fluid shear stress, as they are grown under static conditions. A model that includes ECM, compartmentalization, fluid shear stress, and 3D spatial distribution would better resemble kidney structures and be more suitable for preclinical studies. 3D models are superior to 2D models because they mimic the scaffolding and microenvironment of the native tissues more closely. The lack of polarization in 2D models can be achieved in 3D models via porous membranes, although conventional porous membrane-based protocols do not succeed in replicating the renal tubular epithelial cell phenotype completely (Vinaiphat et al., 2018). A recent development in Transwell insert protocols, based on the use of artificial

237

238

CHAPTER 7 Kidney-on-a-chip

urine on the apical side of the Transwell insert to mimic renal tubular fluid, has improved cell polarization (Vinaiphat et al., 2018). Renal tubule cells cultured on hollow fibers with a curved surface, although showing confluent morphology as on flat surfaces, are better models of renal functions, possibly because of increased mechanical stress owing to membrane curvature (Shen et al., 2013). In 3D cell cultures, nephrotoxicity biomarkers appear more reliable than in 2D cultures; a 3D model using kidney cortical epithelial cells was superior to a 2D model in predicting drug-induced chronic toxicity by detecting kidney injury molecule 1 and neutrophil gelatinase-associated lipocalin (Desrochers et al., 2013). The levels of nephrotoxicity-associated markers, such as inflammatory cytokines, nephrotoxicity-associated genes, and cytochrome enzymes, were closer to in vivo levels in a 3D organoid than in a 2D kidney culture model (Astashkina et al., 2012). The effects of two drugs known to induce nephrotoxicity, ifosfamide and cisplatin, have been shown to be limited or null in 2D models but to mimic in vivo effects in 3D models (Jang et al., 2013, Snouber et al., 2012). Diekjurgen and Grainger (2018) demonstrated that gelembedded 3D organoids exhibited greater expression of proximal tubule transporters than 2D cell cultures.

Cellular models Several cellular models are available for kidney research. PTECs are particularly susceptible to drug toxicity; therefore the development of in vitro models of proximal tubules has been a focus for the pharmaceutical industry (Perazella, 2009; Tiong et al., 2014). Both primary and immortalized cells from proximal tubules have been used in research. The most widely used immortalized cell lines are the Madin Darby canine kidney tubular epithelial cells (Ferrell et al., 2010; Kim et al., 2016; Mu et al., 2013), the Lilly Laboratories porcine tubular epithelial cells (Ozgen et al., 2004), and opossum proximal tubular epithelial cells (Ferrell et al., 2012; Shen et al., 2015). Animal-derived cells exhibit species-specific differences in the expression of SLCOs (Hagenbuch and Stieger, 2013), the P-gp transporter (Gottesman and Pastan, 1993; Schinkel et al., 1997), and flavincontaining monooxygenases (Lohr et al., 1998). For these reasons, the use of animal-derived cell cultures for preclinical testing is not recommended. Human renal cortex immortalized cultures are also available (e.g., human kidney 2 proximal tubular epithelial cells) (Adler et al., 2016). These cells retain the proximal tubule cell phenotypes and their functional transporters and are sensitive to toxicants (Jenkinson et al., 2012; Ryan et al., 1994). However, immortalized cell lines, although inexpensive and easy to maintain, do not replicate the primary cell phenotype, lack functional differentiation (Jenkinson et al., 2012), and exhibit poor expression of human-specific transporters (Kuteykin-Teplyakov et al., 2010). Moreover, their continued growth may result in multilayered structures that obstruct the hollow fiber lumen (Ozgen et al., 2004). Primary human cells are a better model than immortalized cells, and human PTECs are frequently used in in vitro research (Sanechika et al., 2011).

Kidney-on-a-chip models

These cells have been shown to retain the expression and functionality of several transporters (Brown et al., 2008). Although they offer the most functionality (Huang et al., 2015), they can be contaminated by other renal cell types (Van Der Hauwaert et al., 2013) and retain a low number of replication passages (maximum of 12 doublings) (Narayanan et al., 2013). Primary cells isolated from human tissues can retain donor-to-donor variability; this can be a strength or a limitation, as multidonor studies are required to elucidate responses to stimuli. A promising cell model is the embryonic stem cell-derived human PTEC-like cell, although it has not been sufficiently characterized in a microfluidic system (Narayanan et al., 2013). Induced pluripotent stem cells have also been tested for developing terminally differentiated podocytes for glomerulus-on-a-chip models (Allison, 2017; Ashammakhi et al., 2017; Musah et al., 2017).

Kidney-on-a-chip models A kidney-on-a-chip model is the combination of a kidney model (ideally a 3D model) with a platform recreating the microenvironment or structure of the kidney in vivo. The kidney-on-a-chip model should allow coculturing of various renal cell types and retain functional cell cell interactions (such as those occurring between glomerular vascular endothelial cells and podocytes), functional transporter expression, and metabolic and endocrine functions (Wilmer et al., 2016). A kidney-on-a-chip model can be designed to ensure high-throughput screening for drug toxicity testing (Fig. 7.2) to study the glomerular filtration processes, determine the pharmacokinetics, and increase the precision of drug dose evaluation. A model that closely resembles the human situation would enable a more accurate prediction of drug-induced nephrotoxicity and reduce the number of animals used in preclinical testing. Models of the glomerulus and the proximal and distal tubules have been developed, but the creation of a full nephron-on-a-chip platform integrating these components has not yet been achieved.

Glomerulus-on-a-chip The glomerulus is the filtering unit of the kidney and is composed of a network of capillaries and highly differentiated epithelial cells, the podocytes, which regulate selective filtration of blood into an ultrafiltrate that will become ultimately urine (Greka and Mundel, 2012). Developing an in vitro system that mimics glomerular function is of great interest to researchers. Zhou et al. (2016) developed a chip with compartmentalized channels resembling the structure of the glomerulus. They used immortalized human glomerular endothelial cells and murine podocyte precursor cells, arranged in layers lining two channels to model hypertensive nephropathy, and successfully replicated cytoskeletal rearrangements, cellular damage, and glomerular leakage by fluid flow. Wang et al. (2017) developed a

239

240

CHAPTER 7 Kidney-on-a-chip

Human clinical trials Animal models Pros: Established models Physiological relevance Complete organism

Cons: Species specifity Ethical concerns High costs

3D static culture models Pros: Established model More similar to the in vivo situation than 2D cultures Compatible with high-throughput setups

Next-generation models Pros: Improved complexity and functionality (e.g., vascularization and renal functionality similar to the in vivo situation) Fully validated User-friendly Cost-effective Highly predictive Compatible with “omics” analysis More versatile Replacement of animal models

High-throughput models Pros: Improved physiological relevance (perfusion and 3D) Possible to connect multiple organs High-throughput High content screening compatible

Cons: Cons: No mechanical stimuli/microfluidics

Lack of validation

Poorly predictive

2D models Pros: Established model Compatible with high-throughput setups Easy to handle Cost-effective

Cons: Lack of physiological relevance (e.g., no spatial arrangement) Poorly predictive

Microfluidic kidney-on-a-chip Pros: Improved physiological relevance (perfusion and 3D) Possible to connect multiple organs Standardized fabrication techniques Tissue-specific environment at microscale Cons: Low throughput Lack of validation Highly specialized No standard design

Early 3D perfusion Pros: Improved physiological relevance (perfusion and 3D) Microscale Cons: Highly specialized fabrication techniques Low-throughput Lack of validation

FIGURE 7.2 In vitro models: past, present, and future.

Kidney-on-a-chip models

chip with rat glomerular endothelial cells, basement membrane, and podocytes; the model replicated diabetic nephropathy and reproduced high-glucose pathological responses. The major limitation to the development of glomerulus-on-a-chip models has been the lack of functional human kidney podocytes. Musah et al. (2017) recently developed a glomerulus-on-a-chip device with terminally differentiated podocytes from human-induced pluripotent stem cells. Their system mimicked the tissue tissue interface and molecular filtration properties of the glomerular capillary wall; doxorubicin exposure caused podocyte death and proteinuria, resembling the nephrotoxicity observed in vivo.

Proximal tubule-on-a-chip The proximal renal tubule is the primary site of drug clearance and a target for drug-induced toxicity in the nephron and is therefore of primary importance in preclinical assessment of candidate compounds. The ability to control reabsorption and secretion of molecules is a crucial aspect of accurately reproducing the proximal tubule function in vitro. Hollow fibers provide a tubular model that resembles the structure of the proximal tubule and are used as a scaffold to culture renal proximal tubule cells. Ng et al. (2013) seeded human PTECs on the inner surface of hollow fibers coated with hydrogel, showing the formation of a monolayer with functional transport capabilities. A model incorporating hollow fibers exhibited secretory clearance of albumin-bound uremic toxins and albumin reabsorption (Jansen et al., 2016). Jang et al. (2013) developed a cost-effective proximal tubule-on-a-chip device from primary human PTECs seeded on the upper surface of an ECM-coated polyester membrane that splits the main channel of the device into an apical “luminal” channel and a basal “interstitial” space. They showed that cells exposed to fluid shear stress in this system gained polarity, columnar shape, primary cilia, and functional transporters, while cells cultured under static conditions did not. Sciancalepore et al. (2014) cultured human renal progenitor cells in a microfluidic system and showed that under fluid shear stress, the cells polarized and decreased the permeability of urea and creatinine, unlike cells cultured under static conditions. Weber et al. (2016) cultured human PTECs in a microfluidic device consisting of tubular structures surrounded by a chamber coated with an ECM, demonstrating the formation of cellular tubular structures, cell polarization, cellular morphology, basolateral and apical transport, and intracellular enzymatic function; cell viability was maintained for up to 4 weeks. Bioprinting is an alternative technique for shaping the structures of a proximal tubule-on-a-chip device (Sochol et al., 2016). Homan et al. (2016) bioprinted a complex tubular architecture circumscribed by PTECs and actively perfused through the open lumen with Pluronic ink on a fibrinogen ECM. Human PTECs coated the internal tubules and exhibited improved morphology in comparison with their 2D counterparts; moreover, the cells responded in a concentration-dependent

241

242

CHAPTER 7 Kidney-on-a-chip

manner after exposure to cyclosporin A. This system was stable for 2 months, indicating potentially improved capabilities for bioprinted devices. King et al. used bioprinting to design a proximal tubule incorporating renal fibroblasts, endothelial cells, and epithelial cells. This model was stable for at least 30 days and replicated features of renal fibrosis (King et al., 2017). The increased complexity required to replicate a complete nephron-on-a-chip device could be achieved in the future by 3D printing (Sochol et al., 2016).

Distal tubule-/Collecting duct-on-a-chip Few groups have attempted to develop a distal tubule-/collecting duct-on-a-chip model. Baudoin et al. (2007) cultured Madin Darby canine kidney cells on a chip with fibronectin-coated polydimethylsiloxane (PDMS) microchannels and microchambers. A second layer connected the culture chambers to a fluidic network, but flow rates resembling the in vivo situation (50 µL/min) impaired cell proliferation and survival. Jang and Suh (2010) achieved a functional tubule with PDMS hollow fibers coated with collagen. The diameter (50 µm) and fluid shear stress (1 dyn/cm2) resembled those of native renal tubules (Essig et al., 2001; Jameson et al., 2013). Cellular functionality was improved with smaller fiber diameters, suggesting that the surface curvature may affect cell function.

Kidney-on-a-chip: future perspectives Assembling the nephron components is not sufficient in developing a kidney-ona-chip platform. Structural parts, such as the thick ascending limb, the interstitium, and the cortical collecting duct, have not been well characterized yet; the structural functional rearrangements and nutrient, waste, hormone, and drug transport (Huh et al., 2012) must also be taken into consideration when designing a complete kidney-on-a-chip device. A first attempt to enable interaction between tubular cells and blood vessels was performed by Mu et al. (2013) by culturing Madin Darby canine kidney cells and primary umbilical vein endothelial cells. More recently, Homan et al. (2019) developed a millifluidic culture system using human pluripotent stem cells and endothelial cells to form organoids that are apically perfused. The authors observed that exposure to high fluid shear stress (1 4.27 mL/min) during development led to increased vascularization and enhanced maturity of the tubular and glomerular compartments, although this system does not yet ensure perfusion of the microvascular network.

High-throughput technologies Drug screening requires a high-throughput, standardized format to provide robust and reproducible results for preclinical studies. Developments in microfabrication

Kidney-on-a-chip: future perspectives

techniques (e.g., photolithography, etching, micromolding) have enabled a more affordable and widespread use of the microfluidic systems. For example, the OrganoPlate system (Mimetas, Leiden, The Netherlands) can process 96 independent cultures in a microfluidic chip on a standard microtiter plate. OrganoPlate is membrane-free, instead using a gel-media boundary that allows transepithelial transport. The platform is versatile and was previously tested for continuous perfusion, coculturing, tumor cell invasion and aggregation, and toxicity assays (Trietsch et al., 2013). Recent studies tested this platform for nephrotoxicity (Suter-Dick et al., 2018; Vormann et al., 2018; Vriend et al., 2018), modeling the functions of the proximal tubule, and confirmed the expression and function of renal transporters and cell polarization (Vormann et al., 2018; Vriend et al., 2018), the response to cisplatin (Vormann et al., 2018), and target miRNAs indicative of proximal tubule damage (Suter-Dick et al., 2018). An interesting high-throughput application of organs-on-a-chip technology is the combination of the device with mass spectrometry (Oedit et al., 2015) for screening candidate compounds by monitoring the genetic, proteomic, and metabolomic changes in response to exposure (Wilmer et al., 2016). More recently, Peel et al. (2019) proposed a framework for automated high-content confocal imaging of organs-on-a-chip that could apply to kidney tubule chips.

Applications of kidney-on-a-chip platforms Organ-on-a-chip technology replicates the human physiological environment better than single 2D and 3D cultures. 2D models have been used to study cisplatininduced nephrotoxicity in proximal tubular cell cultures (Schetz et al., 2005; Yao et al., 2007) but cannot replicate the physiological structure of an organ. With the development of 3D organoids, a more physiological organization of the tissue has been achieved. However, a static model lacks the natural flow dynamics necessary to polarize the cells and render the functions of the organ model physiologically relevant (Jang and Suh, 2010; Webb, 2017). The integration of microfluidic technology into 3D models enables better replication of the organ physiology (Kimura et al., 2018). One vasculature-on-a-chip device was shown to replicate vasoconstriction, a property that would be of interest in the context of testing drugs that alter renal blood flow (Yasotharan et al., 2015). Tubular structures replicated in kidney-on-a-chip models may be suitable for studying epithelial repair mechanisms, which involve the replacement of damaged or lost tubule cells by dedifferentiation, proliferation, migration, and redifferentiation of surviving epithelial cells (Bonventre, 2003; Humphreys et al., 2008, 2011). Fluid shear stress and primary cilia are likely to contribute to dedifferentiation and proliferation in the epithelial repair (Low et al., 2006; Patel et al., 2008; Shillingford et al., 2006) and can be reproduced in proximal tubule-on-a-chip devices, enabling the assessment of epithelial repair. Applications of the kidney-on-a-chip technology are expected to involve the study of various disease conditions, such as cisplatin-induced nephrotoxicity

243

244

CHAPTER 7 Kidney-on-a-chip

(Moll et al., 2013) or fibrosis (King et al., 2017). Urinary stone disease is another condition that can be modeled on-chip; Wei et al. (2012) cultured submandibular epithelial cells in channels and found that addition of calcium dichloride and trisodium phosphate to the channel facilitated the formation of calcium phosphate stones from calcium present in the tubules. Similar replication could be achieved in kidney models. Integration of the kidney-on-a-chip device into a multiorgan system could enable the study of systemic and secondary toxicity of drugs and the inflammatory response (Oleaga et al., 2016; Webb, 2017). Multiorgan-on-a-chip applications are discussed in Chapter 8, Heart-on-a-chip, of this book. Apart from pharmacological research, kidney-on-a-chip technology could be used in organ transplantation. Improving current dialysis techniques is of great clinical importance, as therapies employing hemodialysis only replicate glomerular filtration, leading to the accumulation of compounds such as uremic toxins (Fleming, 2011). Unlike the nephron, which filters and secretes toxins while reabsorbing electrolytes and small molecules, hemodialysis involves one-way diffusion/convection across a semipermeable membrane. A kidney-on-a-chip device could culture tubular progenitor cells or immortalized PTECs on semipermeable hollow fiber membranes coated with ECM proteins. Jansen et al. (2015) cultured PTECs in hollow fiber membranes, demonstrating the activity of OCT2. Humes et al. (2004) developed a kidney-on-a-chip device composed of a synthetic hemofilter lined with PTECs; when tested in phase I/II clinical trials, the device, capable of metabolic and endocrinologic activity, improved survival in patients suffering from acute kidney injury. Attempts to miniaturize artificial kidneys for implantation have also been made (Humes et al., 2014).

Opportunities and challenges Kidney-on-a-chip platforms seek to provide solutions to modeling the kidney in vitro. Using microfluidic systems and coculturing various renal cell types enable the reproduction of biological processes (Wilmer et al., 2016), while the integration of sensors permits continuous monitoring of cellular behavior (Henry et al., 2017). High-throughput devices coupled with mass spectrometry and highcontent imaging offer new opportunities for drug screening (Oedit et al., 2015; Peel et al., 2019), and the use of human-induced pluripotent stem cells overcomes the limitations of animal-derived cultures (Allison, 2017; Ashammakhi et al., 2017; Musah et al., 2017). The improvement in the organ-on-a-chip field is estimated to create a market of $6.13 billion by 2025, opening new avenues in translational research (Ashammakhi et al., 2018b). Several challenges remain, however, in optimizing a complete kidney-on-a-chip platform. First, the lifespan of primary human cells is limited, and human cell lines cannot fully represent the in vivo biology (Wikswo, 2017). The choice of chip biomaterials is also of major importance. Design standardization and materials validation would be necessary to achieve large-scale

References

application. PDMS, for example, has been shown to absorb and leach test compounds, an undesirable property in biomaterials (Heath et al., 2015; Regehr et al., 2009; Toepke and Beebe, 2006). To become a standard in drug screening, kidney-on-a-chip technology should be validated for sensitivity and specificity, meaning a replicated response to known nephrotoxicants and no response to compounds that do not target the kidney (Wilmer et al., 2016). Moreover, these devices should retain reproducible biological and standardized characteristics between chips or chambers in a high-throughput array (Wilmer et al., 2016). Although the extant devices are suitable to assess drug resorption, distribution, metabolism, and excretion in vitro, there is still no consensus regarding a single-chip system for drug testing and no standards of method validation and integration with laboratory tools (Ashammakhi et al., 2018b). Advances in cell biology and engineering are already providing some solutions. 3D bioprinting—and more recently, four-dimensional bioprinting, which adds conformational changes potential by using stimuli-responsive biomaterials and/or cells—enables the design of complex structures and their reproducibility in manufacturing (Gladman et al., 2016; Li et al., 2016; Liu et al., 2016; Ma et al., 2016; Tibbits, 2014; Tumbleston et al., 2015; Ashammakhi et al., 2018a). Developments in nanotechnologies are also promising in improving the precision of cell diagnosis and therapy. Huh et al. (2012) developed an alveolus-on-a-chip device coupled with a vasculature system that simulated chemotherapy-induced lung edema and could be used to study renal basal membrane permeability, occurring in diabetic nephropathy or nephrotic syndrome. Ribas et al. (2017) studied the effects of physiological and pathological strain on vascular aging using induced pluripotent stem cells from a patient with progeria syndrome in a microfluidic system. Cells derived from patients with genetic conditions such as cystic kidney disease could be used to replicate features of the disease in vitro. A bioprinted vascular epithelium has been used to study thrombosis, and tunable antibodies facilitated the study of interactions with circulating cancer cells (Xu et al., 2016; Yoon et al., 2016; Zhang et al., 2016). In the future, kidney-on-a-chip platforms could be harnessed to evaluate vascular damage- and immune-mediated conditions, such as hemolytic uremic syndrome and thrombotic microangiopathy, as well as the mechanisms of tolerance and rejection in transplanted kidneys.

References Adler, M., Ramm, S., Hafner, M., Muhlich, J.L., Gottwald, E.M., Weber, E., et al., 2016. A quantitative approach to screen for nephrotoxic compounds in vitro. J. Am. Soc. Nephrol. 27, 1015 1028. Allison, S.J., 2017. Bioengineering: kidney glomerulus-on-a-chip. Nat. Rev. Nephrol. 13, 382.

245

246

CHAPTER 7 Kidney-on-a-chip

Ashammakhi, N., Elkhammas, E.A., Hasan, A., 2017. Glomerulus-on-a-chip. Life up. Transplantation 101, e343 e344. Ashammakhi, N., Ahadian, S., Zengjie, F., Suthiwanich, K., Lorestani, F., Orive, G., et al., 2018a. Advances and future rerspectives in 4D bioprinting. Biotechnol. J. 13, e1800148. Ashammakhi, N., Wesseling-Perry, K., Hasan, A., Elkhammas, E., Zhang, Y.S., 2018b. Kidney-on-a-chip: untapped opportunities. Kidney Int. 94, 1073 1086. Astashkina, A.I., Mann, B.K., Prestwich, G.D., Grainger, D.W., 2012. Comparing predictive drug nephrotoxicity biomarkers in kidney 3-D primary organoid culture and immortalized cell lines. Biomaterials 33, 4712 4721. Awdishu, L., Mehta, R.L., 2017. The 6R’s of drug induced nephrotoxicity. BMC Nephrol. 18, 124. Basile, D.P., 2004. Rarefaction of peritubular capillaries following ischemic acute renal failure: a potential factor predisposing to progressive nephropathy. Curr. Opin. Nephrol. Hypertens. 13, 1 7. Baudoin, R., Griscom, L., Monge, M., Legallais, C., Leclerc, E., 2007. Development of a renal microchip for in vitro distal tubule models. Biotechnol. Prog. 23, 1245 1253. Blantz, R.C., 1998. Pathophysiology of pre-renal azotemia. Kidney Int. 53, 512 523. Bonventre, J.V., 2003. Dedifferentiation and proliferation of surviving epithelial cells in acute renal failure. J. Am. Soc. Nephrol. 14 (Suppl 1), S55 S61. Bonventre, J.V., 2010. Pathophysiology of AKI: injury and normal and abnormal repair. Contrib. Nephrol. 165, 9 17. Bonventre, J.V., Yang, L., 2011. Cellular pathophysiology of ischemic acute kidney injury. J. Clin. Invest. 121, 4210 4221. Brown, C.D., Sayer, R., Windass, A.S., Haslam, I.S., Broe, D.E., D’haese, M.E., et al., 2008. Characterisation of human tubular cell monolayers as a model of proximal tubular xenobiotic handling. Toxicol. Appl. Pharmacol. 233, 428 438. Choucha Snouber, L., Jacques, S., Monge, M., Legallais, C., Leclerc, E., 2012. Transcriptomic analysis of the effect of ifosfamide on MDCK cells cultivated in microfluidic biochips. Genomics 100, 27 34. Choudhury, D., Ahmed, Z., 2006. Drug-associated renal dysfunction and injury. Nat. Clin. Pract. Nephrol. 2, 80 91. Chu, X., Bleasby, K., Evers, R., 2013. Species differences in drug transporters and implications for translating preclinical findings to humans. Expert Opin. Drug Metab. Toxicol. 9, 237 252. Coresh, J., Selvin, E., Stevens, L.A., Manzi, J., Kusek, J.W., Eggers, P., et al., 2007. Prevalence of chronic kidney disease in the United States. JAMA 298, 2038 2047. Desrochers, T.M., Suter, L., Roth, A., Kaplan, D.L., 2013. Bioengineered 3D human kidney tissue, a platform for the determination of nephrotoxicity. PLoS One 8, e59219. Diekjurgen, D., Grainger, D.W., 2018. Drug transporter expression profiling in a threedimensional kidney proximal tubule in vitro nephrotoxicity model. Pflugers Arch. 470, 1311 1323. Dreisbach, A.W., Lertora, J.J., 2008. The effect of chronic renal failure on drug metabolism and transport. Expert Opin. Drug Metab. Toxicol. 4, 1065 1074. Essig, M., Terzi, F., Burtin, M., Friedlander, G., 2001. Mechanical strains induced by tubular flow affect the phenotype of proximal tubular cells. Am. J. Physiol. Renal Physiol. 281, F751 F762.

References

Fanos, V., Cataldi, L., 2001. Renal transport of antibiotics and nephrotoxicity: a review. J. Chemother. 13, 461 472. Ferrell, N., Desai, R.R., Fleischman, A.J., Roy, S., Humes, H.D., Fissell, W.H., 2010. A microfluidic bioreactor with integrated transepithelial electrical resistance (TEER) measurement electrodes for evaluation of renal epithelial cells. Biotechnol. Bioeng. 107, 707 716. Ferrell, N., Ricci, K.B., Groszek, J., Marmerstein, J.T., Fissell, W.H., 2012. Albumin handling by renal tubular epithelial cells in a microfluidic bioreactor. Biotechnol. Bioeng. 109, 797 803. Fischer, M.J., Brimhall, B.B., Lezotte, D.C., Glazner, J.E., Parikh, C.R., 2005. Uncomplicated acute renal failure and hospital resource utilization: a retrospective multicenter analysis. Am. J. Kidney Dis. 46, 1049 1057. Fisel, P., Renner, O., Nies, A.T., Schwab, M., Schaeffeler, E., 2014. Solute carrier transporter and drug-related nephrotoxicity: the impact of proximal tubule cell models for preclinical research. Expert Opin. Drug Metab. Toxicol. 10, 395 408. Fitzgerald, K.A., Malhotra, M., Curtin, C.M., O’brien, F.J., O’driscoll, C.M., 2015. Life in 3D is never flat: 3D models to optimise drug delivery. J. Controlled Release 215, 39 54. Fleming, G.M., 2011. Renal replacement therapy review: past, present and future. Organogenesis 7, 2 12. Gladman, A.S., Matsumoto, E.A., Nuzzo, R.G., Mahadevan, L., Lewis, J.A., 2016. Biomimetic 4D printing. Nat. Mater. 15, 413 418. Gottesman, M.M., Pastan, I., 1993. Biochemistry of multidrug resistance mediated by the multidrug transporter. Annu. Rev. Biochem. 62, 385 427. Greka, A., Mundel, P., 2012. Cell biology and pathology of podocytes. Annu. Rev. Physiol. 74, 299 323. Hagenbuch, B., Stieger, B., 2013. The SLCO (former SLC21) superfamily of transporters. Mol. Aspects Med. 34, 396 412. Hartung, T., 2009. Toxicology for the twenty-first century. Nature 460, 208 212. Heath, D.E., Sharif, A.R., Ng, C.P., Rhoads, M.G., Griffith, L.G., Hammond, P.T., et al., 2015. Regenerating the cell resistance of micromolded PEG hydrogels. Lab Chip 15, 2073 2089. Heinonen, I.H., Sorop, O., Merkus, D., Duncker, D.J., 2017. Comparative physiology and pathophysiology of the coronary circulation. Physiological Assessment of Coronary Stenoses and the Microcirculation. Springer. Henry, O.Y.F., Villenave, R., Cronce, M.J., Leineweber, W.D., Benz, M.A., Ingber, D.E., 2017. Organs-on-chips with integrated electrodes for trans-epithelial electrical resistance (TEER) measurements of human epithelial barrier function. Lab Chip 17, 2264 2271. Homan, K.A., Gupta, N., Kroll, K.T., Kolesky, D.B., Skylar-Scott, M., Miyoshi, T., et al., 2019. Flow-enhanced vascularization and maturation of kidney organoids in vitro. Nat. Methods 16, 255 262. Homan, K.A., Kolesky, D.B., Skylar-Scott, M.A., Herrmann, J., Obuobi, H., Moisan, A., et al., 2016. Bioprinting of 3D convoluted renal proximal tubules on perfusable chips. Sci. Rep. 6, 34845. Hoste, E.A., Bagshaw, S.M., Bellomo, R., Cely, C.M., Colman, R., Cruz, D.N., et al., 2015. Epidemiology of acute kidney injury in critically ill patients: the multinational AKI-EPI study. Intensive Care Med. 41, 1411 1423.

247

248

CHAPTER 7 Kidney-on-a-chip

Hsu, C.Y., Mcculloch, C.E., Fan, D., Ordonez, J.D., Chertow, G.M., Go, A.S., 2007. Community-based incidence of acute renal failure. Kidney Int. 72, 208 212. Huang, J.X., Kaeslin, G., Ranall, M.V., Blaskovich, M.A., Becker, B., Butler, M.S., et al., 2015. Evaluation of biomarkers for in vitro prediction of drug-induced nephrotoxicity: comparison of HK-2, immortalized human proximal tubule epithelial, and primary cultures of human proximal tubular cells. Pharmacol. Res. Perspect. 3, e00148. Huh, D., Torisawa, Y.S., Hamilton, G.A., Kim, H.J., Ingber, D.E., 2012. Microengineered physiological biomimicry: organs-on-chips. Lab Chip 12, 2156 2164. Humes, H.D., Buffington, D., Westover, A.J., Roy, S., Fissell, W.H., 2014. The bioartificial kidney: current status and future promise. Pediatr. Nephrol. 29, 343 351. Humes, H.D., Weitzel, W.F., Bartlett, R.H., Swaniker, F.C., Paganini, E.P., Luderer, J.R., et al., 2004. Initial clinical results of the bioartificial kidney containing human cells in ICU patients with acute renal failure. Kidney Int. 66, 1578 1588. Humphreys, B.D., Czerniak, S., Dirocco, D.P., Hasnain, W., Cheema, R., Bonventre, J.V., 2011. Repair of injured proximal tubule does not involve specialized progenitors. Proc. Natl. Acad. Sci. U.S.A. 108, 9226 9231. Humphreys, B.D., Valerius, M.T., Kobayashi, A., Mugford, J.W., Soeung, S., Duffield, J. S., et al., 2008. Intrinsic epithelial cells repair the kidney after injury. Cell Stem Cell 2, 284 291. Jameson, J.L., Loscalzo, J., Loscalzo, C.D.O.M.J., 2013. Harrison’s Nephrology and AcidBase Disorders, Second ed. McGraw-Hill Publishing. Jang, K.J., Suh, K.Y., 2010. A multi-layer microfluidic device for efficient culture and analysis of renal tubular cells. Lab Chip 10, 36 42. Jang, K.-J., Cho, H.S., Kang, D.H., Bae, W.G., Kwon, T.-H., Suh, K.-Y., 2010. Fluidshear-stress-induced translocation of aquaporin-2 and reorganization of actin cytoskeleton in renal tubular epithelial cells. Integr. Biol. 3, 134 141. Jang, K.J., Mehr, A.P., Hamilton, G.A., Mcpartlin, L.A., Chung, S., Suh, K.Y., et al., 2013. Human kidney proximal tubule-on-a-chip for drug transport and nephrotoxicity assessment. Integr. Biol. (Camb.) 5, 1119 1129. Jansen, J., De Napoli, I.E., Fedecostante, M., Schophuizen, C.M., Chevtchik, N.V., Wilmer, M.J., et al., 2015. Human proximal tubule epithelial cells cultured on hollow fibers: living membranes that actively transport organic cations. Sci. Rep. 5, 16702. Jansen, J., Fedecostante, M., Wilmer, M.J., Peters, J.G., Kreuser, U.M., Van Den Broek, P. H., et al., 2016. Bioengineered kidney tubules efficiently excrete uremic toxins. Sci. Rep. 6, 26715. Jenkinson, S.E., Chung, G.W., Van Loon, E., Bakar, N.S., Dalzell, A.M., Brown, C.D., 2012. The limitations of renal epithelial cell line HK-2 as a model of drug transporter expression and function in the proximal tubule. Pflugers Arch. 464, 601 611. Jones, G., Prosser, D.E., Kaufmann, M., 2012. 25-Hydroxyvitamin D-24-hydroxylase (CYP24A1): its important role in the degradation of vitamin D. Arch. Biochem. Biophys. 523, 9 18. Kaysen, J.H., Campbell, W.C., Majewski, R.R., Goda, F.O., Navar, G.L., Lewis, F.C., et al., 1999. Select de novo gene and protein expression during renal epithelial cell culture in rotating wall vessels is shear stress dependent. J. Membr. Biol. 168, 77 89. Kim, S., Lesherperez, S.C., Kim, B.C., Yamanishi, C., Labuz, J.M., Leung, B., et al., 2016. Pharmacokinetic profile that reduces nephrotoxicity of gentamicin in a perfused kidney-on-a-chip. Biofabrication 8, 015021.

References

Kim, S., Takayama, S., 2015. Organ-on-a-chip and the kidney. Kidney Res. Clin. Pract. 34, 165 169. Kimura, H., Sakai, Y., Fujii, T., 2018. Organ/Body-on-a-chip based on microfluidic technology for drug discovery. Drug Metab. Pharmacokinet. 33, 43 48. King, S.M., Higgins, J.W., Nino, C.R., Smith, T.R., Paffenroth, E.H., Fairbairn, C.E., et al., 2017. 3D proximal tubule tissues recapitulate key aspects of renal physiology to enable nephrotoxicity testing. Front. Physiol. 8, 123. Kuteykin-Teplyakov, K., Luna-Tortos, C., Ambroziak, K., Loscher, W., 2010. Differences in the expression of endogenous efflux transporters in MDR1-transfected versus wildtype cell lines affect P-glycoprotein mediated drug transport. Br. J. Pharmacol. 160, 1453 1463. Lameire, N.H., Bagga, A., Cruz, D., De Maeseneer, J., Endre, Z., Kellum, J.A., et al., 2013. Acute kidney injury: an increasing global concern. Lancet 382, 170 179. Laverty, H., Benson, C., Cartwright, E., Cross, M., Garland, C., Hammond, T., et al., 2011. How can we improve our understanding of cardiovascular safety liabilities to develop safer medicines? Br. J. Pharmacol. 163, 675 693. Li, Y.C., Zhang, Y.S., Akpek, A., Shin, S.R., Khademhosseini, A., 2016. 4D bioprinting: the next-generation technology for biofabrication enabled by stimuli-responsive materials. Biofabrication 9, 012001. Linas, S.L., Repine, J.E., 1999. Endothelial cells regulate proximal tubule epithelial cell sodium transport. Kidney Int. 55, 1251 1258. Liu, J., Hwang, H.H., Wang, P., Whang, G., Chen, S., 2016. Direct 3D-printing of cellladen constructs in microfluidic architectures. Lab Chip 16, 1430 1438. Lohr, J.W., Willsky, G.R., Acara, M.A., 1998. Renal drug metabolism. Pharmacol. Rev. 50, 107 141. Lote, C.J., Lote, C.J., 1994. Principles of Renal Physiology. Springer. Low, S.H., Vasanth, S., Larson, C.H., Mukherjee, S., Sharma, N., Kinter, M.T., et al., 2006. Polycystin-1, STAT6, and P100 function in a pathway that transduces ciliary mechanosensation and is activated in polycystic kidney disease. Dev. Cell 10, 57 69. Ma, X., Qu, X., Zhu, W., Li, Y.S., Yuan, S., Zhang, H., et al., 2016. Deterministically patterned biomimetic human iPSC-derived hepatic model via rapid 3D bioprinting. Proc. Natl. Acad. Sci. U.S.A. 113, 2206 2211. Mahmud, H.M., Mahmud, S.M., 2017. Acute kidney injury in Ureteric Stones: Single centre short term analysis. Pak. J. Med. Sci. 33, 808 812. Makris, K., Spanou, L., 2016. Acute kidney injury: definition, pathophysiology and clinical phenotypes. Clin. Biochem. Rev. 37, 85 98. Mccormick, J.A., Ellison, D.H., 2015. Distal convoluted tubule. Compr. Physiol. 5, 45 98. Meyer, C., Dostou, J.M., Gerich, J.E., 1999. Role of the human kidney in glucose counterregulation. Diabetes 48, 943 948. Moll, S., Ebeling, M., Weibel, F., Farina, A., Rosario, D.E.L., Hoflack, A.A., et al., 2013. Epithelial cells as active player in fibrosis: findings from an in vitro model. PLoS One 8, e56575. Mu, X., Zheng, W., Xiao, L., Zhang, W., Jiang, X., 2013. Engineering a 3D vascular network in hydrogel for mimicking a nephron. Lab Chip 13, 1612 1618. Murray, M., Brater, D.C., 1993. Renal toxicity of the nonsteroidal anti-inflammatory drugs. Annu. Rev. Pharmacol. Toxicol. 33, 435 465.

249

250

CHAPTER 7 Kidney-on-a-chip

Musah, S., Mammoto, A., Ferrante, T.C., Jeanty, S.S.F., Hirano-Kobayashi, M., Mammoto, T., et al., 2017. Mature induced-pluripotent-stem-cell-derived human podocytes reconstitute kidney glomerular-capillary-wall function on a chip. Nat. Biomed. Eng. 1. Narayanan, K., Schumacher, K.M., Tasnim, F., Kandasamy, K., Schumacher, A., Ni, M., et al., 2013. Human embryonic stem cells differentiate into functional renal proximal tubular-like cells. Kidney Int. 83, 593 603. Ng, C.P., Zhuang, Y., Lin, A.W.H., Teo, J.C.M., 2013. A fibrin-based tissue-engineered renal proximal tubule for bioartificial kidney devices: development, characterization and in vitro transport study. Int. J. Tissue Eng. 2013, 1 10. Nigam, S.K., Wu, W., Bush, K.T., Hoenig, M.P., Blantz, R.C., Bhatnagar, V., 2015. Handling of drugs, metabolites, and uremic toxins by kidney proximal tubule drug transporters. Clin. J. Am. Soc. Nephrol. 10, 2039 2049. Oedit, A., Vulto, P., Ramautar, R., Lindenburg, P.W., Hankemeier, T., 2015. Lab-on-aChip hyphenation with mass spectrometry: strategies for bioanalytical applications. Curr. Opin. Biotechnol. 31, 79 85. Oleaga, C., Bernabini, C., Smith, A.S., Srinivasan, B., Jackson, M., Mclamb, W., et al., 2016. Multi-Organ toxicity demonstration in a functional human in vitro system composed of four organs. Sci. Rep. 6, 20030. Ozgen, N., Terashima, M., Aung, T., Sato, Y., Isoe, C., Kakuta, T., et al., 2004. Evaluation of long-term transport ability of a bioartificial renal tubule device using LLC-PK1 cells. Nephrol. Dial. Transplant. 19, 2198 2207. Patel, V., Li, L., Cobo-Stark, P., Shao, X., Somlo, S., Lin, F., et al., 2008. Acute kidney injury and aberrant planar cell polarity induce cyst formation in mice lacking renal cilia. Hum. Mol. Genet. 17, 1578 1590. Peel, S., Corrigan, A.M., Ehrhardt, B., Jang, K.J., Caetano-Pinto, P., Boeckeler, M., et al., 2019. Introducing an automated high content confocal imaging approach for organs-onchips. Lab Chip 19, 410 421. Perazella, M.A., 2009. Renal vulnerability to drug toxicity. Clin. J. Am. Soc. Nephrol. 4, 1275 1283. Perazella, M.A., Moeckel, G.W., 2010. Nephrotoxicity from chemotherapeutic agents: clinical manifestations, pathobiology, and prevention/therapy. Semin. Nephrol. 30, 570 581. Praetorius, H., 2015. The primary cilium as sensor of fluid flow: new building blocks to the model. A review in the theme: cell signaling: proteins, pathways and mechanisms. Am. J. Physiol. Cell Physiol. 308, C198 C208. Raghavan, V., Rbaibi, Y., Pastor-Soler, N.M., Carattino, M.D., Weisz, O.A., 2014. Shear stress-dependent regulation of apical endocytosis in renal proximal tubule cells mediated by primary cilia. Proc. Natl. Acad. Sci. U.S.A. 111, 8506 8511. Redfern, W., Ewart, L., Hammond, T., Bialecki, R., Kinter, L., Lindgren, S., et al., 2010. Impact and frequency of different toxicities throughout the pharmaceutical life cycle. Toxicologist 114, 1081. Regehr, K.J., Domenech, M., Koepsel, J.T., Carver, K.C., Ellison-Zelski, S.J., Murphy, W. L., et al., 2009. Biological implications of polydimethylsiloxane-based microfluidic cell culture. Lab Chip 9, 2132 2139. Ribas, J., Zhang, Y.S., Pitrez, P.R., Leijten, J., Miscuglio, M., Rouwkema, J., et al., 2017. Biomechanical strain exacerbates inflammation on a progeria-on-a-chip model. Small 13, 1603737.

References

Ryan, M.J., Johnson, G., Kirk, J., Fuerstenberg, S.M., Zager, R.A., Torok-Storb, B., 1994. HK-2: an immortalized proximal tubule epithelial cell line from normal adult human kidney. Kidney Int. 45, 48 57. Sanechika, N., Sawada, K., Usui, Y., Hanai, K., Kakuta, T., Suzuki, H., et al., 2011. Development of bioartificial renal tubule devices with lifespan-extended human renal proximal tubular epithelial cells. Nephrol. Dial. Transplant. 26, 2761 2769. Schetz, M., Dasta, J., Goldstein, S., Golper, T., 2005. Drug-induced acute kidney injury. Curr. Opin. Crit. Care 11, 555 565. Schinkel, A.H., Mayer, U., Wagenaar, E., Mol, C.A., Van Deemter, L., Smit, J.J., et al., 1997. Normal viability and altered pharmacokinetics in mice lacking mdr1-type (drugtransporting) P-glycoproteins. Proc. Natl. Acad. Sci. U.S.A. 94, 4028 4033. Schnellmann, R.G., 2001. Toxic responses of the kidney, Casarett & Doull’s Toxicology: The Basic Science of Poisons, Sixth ed. McGraw Hill, New York, pp. 491 514. Sciancalepore, A.G., Sallustio, F., Girardo, S., Gioia Passione, L., Camposeo, A., Mele, E., et al., 2014. A bioartificial renal tubule device embedding human renal stem/progenitor cells. PLoS One 9, e87496. Sharfuddin, A.A., Molitoris, B.A., 2011. Pathophysiology of ischemic acute kidney injury. Nat. Rev. Nephrol. 7, 189 200. Shen, C., Meng, Q., Zhang, G., 2013. Increased curvature of hollow fiber membranes could up-regulate differential functions of renal tubular cell layers. Biotechnol. Bioeng. 110, 2173 2183. Shen, C., Zhang, G., Wang, Q., Meng, Q., 2015. Fabrication of collagen gel hollow fibers by covalent cross-linking for construction of bioengineering renal tubules. ACS Appl. Mater. Interfaces 7, 19789 19797. Shillingford, J.M., Murcia, N.S., Larson, C.H., Low, S.H., Hedgepeth, R., Brown, N., et al., 2006. The mTOR pathway is regulated by polycystin-1, and its inhibition reverses renal cystogenesis in polycystic kidney disease. Proc. Natl. Acad. Sci. U.S.A. 103, 5466 5471. Snouber, L.C., Letourneur, F., Chafey, P., Broussard, C., Monge, M., Legallais, C., et al., 2012. Analysis of transcriptomic and proteomic profiles demonstrates improved MadinDarby canine kidney cell function in a renal microfluidic biochip. Biotechnol. Prog. 28, 474 484. Sochol, R.D., Gupta, N.R., Bonventre, J.V., 2016. A role for 3D printing in kidney-on-achip platforms. Curr. Transplant. Rep. 3, 82 92. Suter-Dick, L., Mauch, L., Ramp, D., Caj, M., Vormann, M.K., Hutter, S., et al., 2018. Combining extracellular miRNA determination with microfluidic 3D cell cultures for the assessment of nephrotoxicity: a proof of concept study. AAPS J. 20, 86. Szeto, C.C., Chow, K.M., 2005. Nephrotoxicity related to new therapeutic compounds. Renal Failure 27, 329 333. Tasnim, F., Zink, D., 2012. Cross talk between primary human renal tubular cells and endothelial cells in cocultures. Am. J. Physiol. Renal Physiol. 302, F1055 F1062. Tibbits, S., 2014. 4D printing: multi-material shape change. Archit. Des. 84, 116 121. Tiong, H.Y., Huang, P., Xiong, S., Li, Y., Vathsala, A., Zink, D., 2014. Drug-induced nephrotoxicity: clinical impact and preclinical in vitro models. Mol. Pharm. 11, 1933 1948. Toepke, M.W., Beebe, D.J., 2006. PDMS absorption of small molecules and consequences in microfluidic applications. Lab Chip 6, 1484 1486.

251

252

CHAPTER 7 Kidney-on-a-chip

Tolwani, A., 2012. Continuous renal-replacement therapy for acute kidney injury. N. Engl. J. Med. 367, 2505 2514. Tomlinson, L.A., Abel, G.A., Chaudhry, A.N., Tomson, C.R., Wilkinson, I.B., Roland, M. O., et al., 2013. ACE inhibitor and angiotensin receptor-II antagonist prescribing and hospital admissions with acute kidney injury: a longitudinal ecological study. PLoS One 8, e78465. Trietsch, S.J., Israels, G.D., Joore, J., Hankemeier, T., Vulto, P., 2013. Microfluidic titer plate for stratified 3D cell culture. Lab Chip 13, 3548 3554. Tumbleston, J.R., Shirvanyants, D., Ermoshkin, N., Janusziewicz, R., Johnson, A.R., Kelly, D., et al., 2015. Additive manufacturing. Continuous liquid interface production of 3D objects. Science 347, 1349 1352. Van Der Hauwaert, C., Savary, G., Gnemmi, V., Glowacki, F., Pottier, N., Bouillez, A., et al., 2013. Isolation and characterization of a primary proximal tubular epithelial cell model from human kidney by CD10/CD13 double labeling. PLoS One 8, e66750. Venkatachalam, M.A., Griffin, K.A., Lan, R., Geng, H., Saikumar, P., Bidani, A.K., 2010. Acute kidney injury: a springboard for progression in chronic kidney disease. Am. J. Physiol. Renal Physiol. 298, F1078 F1094. Vinaiphat, A., Charngkaew, K., Thongboonkerd, V., 2018. More complete polarization of renal tubular epithelial cells by artificial urine. Cell Death Discov. 4, 47. Vormann, M.K., Gijzen, L., Hutter, S., Boot, L., Nicolas, A., Van Den Heuvel, A., et al., 2018. Nephrotoxicity and kidney transport assessment on 3D perfused proximal tubules. AAPS J. 20, 90. Vriend, J., Nieskens, T.T.G., Vormann, M.K., Van Den Berge, B.T., Van Den Heuvel, A., Russel, F.G.M., et al., 2018. Screening of drug-transporter interactions in a 3D microfluidic renal proximal tubule on a chip. AAPS J. 20, 87. Wang, L., Tao, T., Su, W., Yu, H., Yu, Y., Qin, J., 2017. A disease model of diabetic nephropathy in a glomerulus-on-a-chip microdevice. Lab Chip 17, 1749 1760. Webb, S.P.D., 2017. Cells in the third dimension. Biotechniques 62, 93 98. Weber, E.J., Chapron, A., Chapron, B.D., Voellinger, J.L., Lidberg, K.A., Yeung, C.K., et al., 2016. Development of a microphysiological model of human kidney proximal tubule function. Kidney Int. 90, 627 637. Wei, Z., Amponsah, P.K., Al-Shatti, M., Nie, Z., Bandyopadhyay, B.C., 2012. Engineering of polarized tubular structures in a microfluidic device to study calcium phosphate stone formation. Lab Chip 12, 4037 4040. Weisbord, S.D., Du Cheryon, D., 2018. Contrast-Associated Acute Kidney Injury is a Myth: No. Springer. Wikswo, J.P., 2017. Looking to the future of organs-on-chips: interview with Professor John Wikswo. Future Sci. OA 3, FSO163. Wilmer, M.J., Ng, C.P., Lanz, H.L., Vulto, P., Suter-Dick, L., Masereeuw, R., 2016. Kidney-on-a-chip technology for drug-induced nephrotoxicity screening. Trends Biotechnol. 34, 156 170. Xu, Z., Li, E., Guo, Z., Yu, R., Hao, H., Xu, Y., et al., 2016. Design and construction of a multi-organ microfluidic chip mimicking the in vivo microenvironment of lung cancer metastasis. ACS Appl. Mater. Interfaces 8, 25840 25847. Xue, J.L., Daniels, F., Star, R.A., Kimmel, P.L., Eggers, P.W., Molitoris, B.A., et al., 2006. Incidence and mortality of acute renal failure in Medicare beneficiaries, 1992 to 2001. J. Am. Soc. Nephrol. 17, 1135 1142.

References

Yao, X., Panichpisal, K., Kurtzman, N., Nugent, K., 2007. Cisplatin nephrotoxicity: a review. Am. J. Med. Sci. 334, 115 124. Yasotharan, S., Pinto, S., Sled, J.G., Bolz, S.S., Gunther, A., 2015. Artery-on-a-chip platform for automated, multimodal assessment of cerebral blood vessel structure and function. Lab Chip 15, 2660 2669. Yeung, C.K., Himmelfarb, J., 2019. Kidneys on chips: emerging technology for preclinical drug development. Clin. J. Am. Soc. Nephrol. 14, 144 146. Yeung, C.K., Shen, D.D., Thummel, K.E., Himmelfarb, J., 2014. Effects of chronic kidney disease and uremia on hepatic drug metabolism and transport. Kidney Int. 85, 522 528. Yoon, H.J., Shanker, A., Wang, Y., Kozminsky, M., Jin, Q., Palanisamy, N., et al., 2016. Tunable thermal-sensitive polymer graphene oxide composite for efficient capture and release of viable circulating tumor cells. Adv. Mater. 28, 4891 4897. Zhang, Y.S., Davoudi, F., Walch, P., Manbachi, A., Luo, X., Dell’erba, V., et al., 2016. Bioprinted thrombosis-on-a-chip. Lab Chip 16, 4097 4105. Zhuo, J.L., Li, X.C., 2013. Proximal nephron. Compr. Physiol. 3, 1079. Zhou, M., Zhang, X., Wen, X., Wu, T., Wang, W., Yang, M., et al., 2016. Development of a functional glomerulus at the organ level on a chip to mimic hypertensive nephropathy. Sci. Rep. 6, 31771.

253

CHAPTER

Heart-on-a-chip

8

Pierre Gaudriault, Dario Fassini and Antoni Homs-Corbera Cherry Biotech SAS, Rennes, France

Introduction The heart is a biological pump with electrical and mechanical characteristics that move blood through the vascular system and, consequently, through all other organs of the body. It is a vital organ that maintains homeostasis by effectively delivering nutrients and removing waste. Heart diseases and adverse events such as myocardial infarction, hypertrophy, and atherosclerosis are the leading cause of death worldwide (PhRMA, 2012) and a major focus of treatment development. Half of all drugs withdrawn from the market in the last 30 years caused cardiac electrophysiological dysfunction and muscle damage. This highlights the limitations of current drug-testing methods in evaluating cardiac effects. Cardiotherapies and cardiotoxicity are currently studied using two-dimensional (2D) static cultures of cardiomyocytes (Ralphe and de Lange, 2013) or, alternatively, animal models. Both methods have limitations; 2D cell cultures are oversimplified and fail to reproduce cell orientation and cardiac tissue physiology, while animal models frequently do not predict human response. Cardiovascular in vitro models that use human cells are desirable because of their relatively lower costs, their potential to better mimic human physiology, and societal concerns regarding animal testing. However, the complex dynamics are a strong limiting factor when aiming to mimic the main physiological functions of the heart in vitro. The conventional 2D cardiovascular in vitro models cannot replicate physiological conditions regulated by physical stimulations, such as electrical signaling, mechanical stress, or shear stress; these are important factors in setting the alignment, structure, and phenotype of cardiac cells. The versatility of microfluidic systems provides more physiologically relevant in vitro models of heart tissue. Microfluidic platforms, and by extension organon-a-chip devices, can provide continuous media perfusion to cells, allowing control of shear stresses and specific spatial distributions of various cell types. Furthermore, they can also be integrated with other technologies to stimulate cells mechanically or electrically. Recreating a heart-on-a-chip (HoC) device poses several challenges, including how to overcome the limited proliferation capacity of mature cardiomyocytes, Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00008-5 © 2020 Elsevier Inc. All rights reserved.

255

256

CHAPTER 8 Heart-on-a-chip

how to mimic the alignment and orientation of these cells when present in native heart tissue, and how to achieve synchronous beating of these cardiomyocytes in vitro. Nevertheless, recent advances in combining microfluidics, tissue engineering, and stem cell technologies have laid the foundation for building these physiologically relevant in vitro HoC models that could be further personalized with cells derived from specific patients. These more humanized models are good candidates to mimic the human in vivo system than currently used animal models, potentially reducing the burden of clinical trial failure during the development of new therapies. Departing from the classical heart models of 2D planar cell cultures, HoC technologies have achieved high degrees of complexity by reproducing physiological properties and incorporating mechanical, chemical, and electrical stimulations such as those present in the native organ. Using human cell arrangements and microfluidics to mimic the heart allows for a wide field of applications, such as drug discovery, toxicity and therapeutic screenings, and tissue regeneration. Existing HoC models have been used to study cardiopathologies, cardiotoxicity, and stem-cell differentiation. This chapter reviews the heart and current technologies that mimic the heart in vitro.

Anatomy and physiology of the heart The heart is the central element of the cardiovascular system, and its first role is to pump blood by mechanically contracting itself at a given rhythm. In this section, we review cardiac anatomy (Fig. 8.1) and physiology. More extensive indepth reviews are available elsewhere in the medical literature (Katz, 2010; Betts, 2013).

Dimension, location, and envelope The human heart, weighing approximately 250 350 g and measuring approximately 14 16 cm, is located inside the mediastinum, a compartment between the two lungs inside the thorax. More specifically, it is placed above the diaphragm at the front of the spinal column and behind the sternum and the ribs; the conically shaped organ stretches from the second rib toward the fifth intercostal space (Katz, 2010; Betts, 2013). Because of its particular location, the heart is wrapped inside a double-walled “envelope” termed the pericardium. The role of the superficial wall, or fibrous pericardium, is to protect the heart and attach it to the diaphragm. The interior wall is a complex structure composed of two layers: the parietal pericardium, which is bound to the fibrous pericardium, and the visceral pericardium, which is also termed the epicardium. The compartment delimited by these two layers is

Heart wall

FIGURE 8.1 Illustration showing the anatomy of the heart from Eric Pierce, “Diagram of the human heart” (2006, CC BY-SA 3.0). This file is licensed under the Creative Commons Attribution-Share Alike 3.0 Unported license. https:// commons.wikimedia.org/wiki/File:Diagram_of_the_human_heart_(cropped).svg.

named the pericardial cavity and contains the pericardial fluid, which is serous and lubricates the heart to prevent friction from cardiac beating. The contracting period of the heart is called systole and lasts 0.25 0.30 seconds. The resting period of the heart cycle is called diastole and lasts approximately 0.5 seconds.

Heart wall The heart wall is divided into three layers—the epicardium, myocardium, and endocardium from the external to the internal layer—and contains many blood vessels (Katz, 2010; Betts, 2013). The epicardium (or visceral pericardium) is mainly composed of connective tissue containing elastic fibers and adipose tissue (especially in elderly persons). The myocardium is the heart muscle and contains cardiac muscle cells, which are known as cardiomyocytes. This muscle provides the capacity of contraction and accounts for the major part of the heart’s mass. Cardiomyocytes are attached by connective tissues to form fiber bundles in a spiral or roundish shape; these bundles connect all the parts of the

257

258

CHAPTER 8 Heart-on-a-chip

heart. The collagen and elastic fibers of the connective tissues form a network that reinforces the structure of the heart. Where a large blood vessel emerges from the heart, these cardiac bundles form roundish shapes to maintain the vessel and prevent its dilation from continuous internal pressure. This network of cardiac bundle and connective tissues controls the propagation of the electrical influx, which is mandatory for a synchronized contraction of the heart. The thickness of the myocardium varies inside the heart. The endocardium is the internal layer of the heart; an endothelium that covers all the cavities and valves, and that it is a continuation of the endothelium of blood vessels that reach the heart. The endocardium is smooth and functions to decrease the friction of blood against the heart walls.

Heart cavities To pump blood, the heart is equipped with four cavities: two atria in the upper part and two ventricles in the lower part (Katz, 2010; Betts, 2013). The atria are separated by the interatrial septum and the ventricles by the interventricular septum. The right and left atria are morphologically similar. The front walls of both atria are covered by muscular bundles that form the pectinate muscle. The interatrial septum contains the fossa ovalis, a remnant of the foramen ovale, one of the two embryonic cardiac shunts. The atria are the entry point of blood into the heart and receive blood from general and pulmonary circulation. The atria are small, and they do not need to contract strongly to send blood to the ventricles. Their walls are thin with a small amount of muscle. During the heart cycle, atrial volume increases only slightly. The superior vena cava, inferior vena cava, and coronary sinus enter the right atrium, bringing blood from organs located above (superior vena cava) and below (inferior vena cava) the diaphragm, and from the heart itself (coronary sinus). While the right atrium collects blood from the general circulation, the left atrium receives blood from four pulmonary veins connected to the lungs. Ventricles account for much of the cardiac mass and eject blood from the heart to the circulation by contraction. The right ventricle ejects the blood into the left and right pulmonary arteries that carry blood to the lung to be oxygenated. The left ventricle pumps blood into the aorta, the largest artery, which provides blood to all organs by successive ramification. Because ventricles pump blood, their walls are thicker than the atrial walls. Furthermore, the myocardium is thicker in the left ventricle than in the right one. At the bottom of each ventricle are irregular muscle structures named trabeculae carneae. The papillary muscles are attached to the trabeculae carneae on one side and to the cardiac valves on the other side.

Mechanism of cardiac contraction

The cardiac valves Blood flow is unidirectional in the heart, passing from the atria to the ventricles to enter the arteries emerging from the top part of the heart. Four valves regulate this circuit: two between the atria and ventricles (atrioventricular valves), one in the aorta (aortic valve), and one at the basis of the pulmonary artery (pulmonary valve, also termed the pulmonic valve) (Katz, 2010; Betts, 2013). Atrioventricular valves separate the atria from the underlying ventricles, preventing the blood from returning to the atria when the ventricles contract. The main difference between the right and left atrioventricular valves is the number of cusps (endocardium structures reinforced by connective tissue) composing each valve. The right atrioventricular valve comprises three cusps termed the tricuspid valve, and the left comprises two cusps termed the bicuspid or mitral valve. Thin white collagen bundles, the chordae tendineae, attach each cusp of the atrioventricular valves to the papillary muscles of the ventricles. During the systole, the pressure increases in the ventricles, driving the closing of the valve. Aortic and pulmonary valves are located at the junction with the aorta and pulmonary arteries and prevent the blood from returning into the ventricle at ventricle contraction. Both valves, also termed sigmoid valves, are formed of three semilunar valves. Unlike the mechanism controlling atrioventricular valves, blood pressure increases inside the ventricles and open the sigmoid valves. As valvulopathy and valve toxicity are major concerns in drug surveys and clinical trials, creating human-relevant cardiac models is important and should account for the anatomical complexity of these critical structures.

Microscopic anatomy of the heart muscle The heart muscle is striated muscle, similar to skeletal muscle, and its contractions are generated by the slipping of myofilaments. Cardiomyocytes are short, thick, branched, and anastomosed; each contains one or two large nuclei (Fig. 8.2). The intercellular space is composed of connective tissue (endomysium) containing multiple capillary vessels. Unlike skeletal muscle fiber, cardiomyocytes are bound by their plasma membrane to intercalated discs. These structures contain gap junctions, desmosomes, and fascia adherens, and allow the cardiomyocytes to work in a single functional syncytium while connecting them both physically and electrically (Katz, 2010; Betts, 2013).

Mechanism of cardiac contraction The mechanism of cardiac contraction follows three principles (Katz, 2010; Betts, 2013):

259

260

CHAPTER 8 Heart-on-a-chip

FIGURE 8.2 Schematic picture of cardiomyocytes organization and arrangement in the heart. Image from Servier Medical Art by Servier (CC BY 3.0). This file is licensed under the Creative Commons Attribution 3.0 Unported license. https://smart.servier.com/ wp-content/uploads/2016/10/cellules_coeur.png.

1. As the heart muscle works as a functional syncytium, owing to the passage of ions through open junctions, the whole heart contracts as one unit, unlike skeletal striated muscle. 2. The cardiac refractory period is equal to 250 ms—roughly the same duration as the contraction. This refractory period corresponds to the nonexcitable period following a contraction in which sodium channels remain open. In the heart, this long refractory period serves to prevent extended contractions that would disrupt the pumping function of the organ. 3. Heart muscle cells are self-excitable and can spontaneously and rhythmically depolarize to produce an action potential that then propagates within the heart.

Physiology of the heart muscle The inner capacity of the heart to beat is spontaneous. Without nerves, the heart will continue to beat (a phenomenon observed in heart transplant procedures), but the heart is physiologically coupled with the nervous system to regulate its rhythm.

Spontaneous production of action potential The coordinated and spontaneous capacity of the heart to contract is due to two factors: (1) the presence of open junctions, as previously described, and (2) the electrical conduction system of the heart, comprising noncontractile cells that create action potentials and propagate them from atria to ventricles.

Excitation and action potential propagation

FIGURE 8.3 Action potential cycle.

Cells from a particular area of the electrical conduction system, termed pacemaker cells, can slowly depolarize toward their excitation threshold, triggering the creation of the action potential (Katz, 2010; Betts, 2013). The mechanism is thought to start with the slow reduction in permeability of the pacemaker cell membrane to K1 ions and concurrent regulation of Na1 inside the cells (hyperpolarization-activated cyclic nucleotide-gated channels). Thus the balance between K1 loss and Na1 influx is disrupted, and the inner side of the membrane becomes less and less negative. Once a membrane potential of 240 mV is reached, both the Ca21 (T-type) and Na1 channels open to allow these two ion types to enter the cells, inverting membrane potential (ascending phase of the action potential). The falling phase of the action potential is created by the increase of K1 membrane permeability (K1 ions move toward the interstitial space). The membrane becomes more and more negative, and once the resting membrane potential is reached, K1 channels close and the cells are ready for a new spontaneous slow depolarization. A typical action potential cycle is represented in Fig. 8.3.

Excitation and action potential propagation The electrical conduction system of the heart is activated sequentially as follows: 1. The sinoatrial node, located in the right atria, is the pacemaker area of the heart and depolarizes spontaneously 75 100 times/min, depending on the action of inhibitory nerves and hormones. The sinoatrial node rhythm is responsible for heartbeat frequency.

261

262

CHAPTER 8 Heart-on-a-chip

2. From the sinoatrial node, the action potential travels for approximately 0.04 seconds in the atria to reach the atrioventricular node, where it is delayed for about 0.1 second, allowing the atria to complete their contraction cycle before the ventricles begin their cycle. A reduction in fiber diameters reduces the speed of the action potential. 3. The atria and ventricles are not electrically connected by open junctions, and the bundle of His (1893), located at the bottom of the interatrial septum, is the only electrical link between the atria and ventricles. The bundle branches to the right and left, and the action potential travels through both branches inside the interventricular septum to the heart apex. 4. Purkinje fibers end the action potential’s journey inside the interventricular septum inside the walls of each ventricle. As the myocardium is thicker in the left ventricle, the Purkinje fiber network is denser and more elaborate in this part of the heart. The Purkinje fibers also innervate papillary muscles that can contract to close the atrioventricular valves. In a healthy heart, this process takes approximately 220 ms from the sinoatrial node to the last depolarization of the cells in the ventricular myocardium. Ventricle contraction happens almost immediately after the action potential is generated. In the ventricle, the contraction starts at the apex and follows the Purkinje fiber network, opening the aortic and pulmonary valves. The atrioventricular node and the Purkinje fibers can spontaneously depolarize at a lower frequency than that of the sinoatrial node (50 and 30 times/min, respectively), and can thus become the pacemaker of the heart when the sinoatrial node fails.

Heart nerve connections The heart is innervated by both the sympathetic and parasympathetic systems. The sympathetic system accelerates heartbeat frequency by reducing the excitation threshold of the sinoatrial node via the release of noradrenaline (also called norepinephrine). The parasympathetic system reduces heartbeat frequency by hyperpolarizing the plasma membrane after the release of acetylcholine. This hyperpolarization is due to the opening of K1 channels in the myocardium cells. In the resting situation, both the sympathetic and parasympathetic systems act on the heart, but the parasympathetic stimuli predominate, meaning that heartbeat frequency is constantly reduced by the parasympathetic system.

Electrocardiography The electrical currents that are generated and spread in the heart also transmit through other body tissues and fluids and are detectable at the surface of the skin.

In vitro models of the heart

This is the basis of the electrocardiogram, a readout graph of voltage versus time derived from noninvasive recording of potentials at the skin surface. In conventional electrocardiography, 10 electrodes are placed on the patient’s chest and limbs and the magnitudes of the cardiac electrical potentials are measured over time from 12 angles or leads. The electrocardiogram of a healthy heart displays five sequential waves with typical shapes that indicate the magnitude and direction of depolarization during the cardiac cycle. The first or P wave has a low amplitude and a duration of about 0.08 seconds. It corresponds to atrial depolarization by the sinoatrial node. Approximately 0.1 second after the beginning of the P wave, the atria contract. The following waves—Q, R, and S waves—are grouped as the QRS complex, which is the result of the ventricles depolarizing. The asymmetric shape of the QRS complex reflects the differences between ventricle sizes and the time to depolarization of each ventricle. The QRS complex has a mean duration of 0.08 seconds. Finally, the T wave is the observation of ventricle repolarization and lasts 0.16 seconds. The repolarization of the atria occurs during the QRS complex, and, therefore it is not typically visible on the electrocardiogram. In a healthy heart, the sequence or electric wave is rather constant, making it possible to detect abnormalities in the electrical conduction system. For example, a tall QRS complex indicates ventricular hypertrophy, a flat or inverted T wave may indicate myocardial ischemia or left ventricular hypertrophy, a peaked T wave can be an early sign of myocardial infarction or hyperkalemia, and an elongated QT interval reflects an abnormality in cardiac repolarization, increasing the risk of arrhythmia.

In vitro models of the heart Considering the vital role of the heart and the mortality burden of cardiovascular diseases, it is not surprising that scientists have long attempted to develop relevant and ethically sustainable heart models. Models to advance research on heart diseases or to better predict the cardiac effects of drugs must be reproducible and robust, meaning that experiments performed in different labs must generate similar outcomes (validation), and that the data generated must be directly comparable. The ideal models must also be robust enough to be translated into clinical practice; this implies that bias related to the phenotypic heterogeneity of the population should be considered, or in an ideal situation, the model can be patient specific. Traditionally, efforts to develop scalable and reliable heart models, in particular in vitro models, have centered on three main lines of investigation: elucidating and manipulating the events that follow myocardial ischemia, studying the physiology of rare heart diseases, and assessing drug safety.

263

264

CHAPTER 8 Heart-on-a-chip

Traditional two-dimensional cell culture heart models Traditional 2D cell cultures are generally obtained using flat and rigid polystyrene Petri dishes or flasks. Cells are seeded and, once attached to the plastic substrate, begin to divide and grow in a monolayer. This approach has been used for decades and is still considered a valuable approach to investigating specific aspects of cardiomyocyte biology. This model has gained popularity for elucidating the factors underlying the hypoxia resulting from an ischemic event. Flat cardiomyocyte monolayers are rendered hypoxic by changing the gas concentrations (i.e., 1% O2, 5% CO2, and 94% N2) and replacing the medium (e.g., with a nutrientdepleted medium) for a fixed period before returning to normal gas concentrations (normoxia) and culture conditions. This simple model allows mechanistic study of the contribution of each factor without increasing the complexity by inclusion of other cells types (e.g., fibroblasts, endothelial cells, inflammatory/immune cells, and platelets) or circulating factors (e.g., hormones, neurotransmitters, and cytokines) (Heusch, 2017).

Two-dimensional cell culture heart models for ischemia Ischemia occurs when the blood supply to a tissue is reduced. The tissue receives less oxygen and fewer nutrients, potentially leading to necrosis downstream of the occlusion. When ischemia involves a significant portion of the heart or large coronary arteries, the muscle tissue suffering from lack of oxygen and nutrients dies leading to myocardial infarction which is still the most frequent cause of death in many countries. Experimental models of myocardial ischemia are used to investigate and determine the mechanisms and genes involved in the physiological responses to ischemic events. While the model can accept some approximations, it must still approach the natural complexity of human physiology as much as possible (Heusch, 2017; Lindsey et al., 2018). Typical effects of simulating cardiac ischemia in 2D models include apoptosis and necrosis (Long et al., 1997; Tanaka et al., 1994), mitochondrial damage (Brodarac et al., 2015; Canfield et al., 2012), and the generation of reactive oxygen species. Standard cytotoxicity assays based on membrane permeability are generally the preferred method to evaluate cellular viability (Brodarac et al., 2015; Canfield et al., 2012; Date et al., 2002), while programmed cell death (apoptosis) is assessed by specific staining protocols (Brodarac et al., 2015; Date et al., 2002; Kang et al., 2000). Levels of reactive oxygen species can be measured easily (Date et al., 2002), while assessments of cell morphology, beating activity, or action potential can be used to evaluate the effects of a temporary ischemia (Heinzel et al., 2006; Kang et al., 2000). Table 8.1 presents a summary of 2D cell culture models previously used for studying ischemia.

Traditional two-dimensional cell culture heart models

Table 8.1 Summary of cells used in in vitro two-dimensional heart models. Cell type/source

Advantages

Disadvantages

References

Adult primary cardiomyocytes

Cellular physiology closer to in vivo situations Easy to culture and expand

Must be collected from patients

Kang et al. (2000), Maddaford et al. (1999), and Portal et al. (2013)

Bias deriving from a single donor Different physiology from adult cells

Liu et al. (2018)

Immortalized adult cardiomyocytes Neonatal cardiomyocytes

Cardiac progenitor cells Induced pluripotent stem cell-derived cardiomyocytes

Can be expanded and reimplanted Can be patientspecific and/or disease-specific

Must be collected from patients Difficulties in obtaining a fully differentiated phenotype

Date et al. (2002), Negoro et al. (2001), and Tanaka et al. (1994) Bauer et al. (2012)

Brodarac et al. (2015) and Canfield et al. (2012)

Two-dimensional cell culture heart models for drug discovery Before reaching the market, a drug must undergo a long and expensive evaluation of its efficacy and safety. The pharmaceutical industry assesses both the benefits and risks to human health following a well-established set of preclinical tests before providing all the data to regulatory authorities and requesting permission to test the drugs on patients (Eichler et al., 2008; Pignatti et al., 2011; Senderowicz, 2010; Silbergeld et al., 2015). Traditionally, preclinical trials aim to assess the doses and the risk of a compound using in vitro models and in vivo animal studies. In vitro tests are generally performed with established 2D cell cultures following the rigorous protocols in Test Guideline 10993 of the International Organization for Standardization. Tests performed on 2D cultures are still not sufficient to simulate the complexity and possible reactions on a whole-organism level (e.g., immune system reactions, production of secondary metabolites during drug catabolism, and drug pharmacokinetics). For that reason, the current gold standard and required step before entering the clinical trial phase is animal testing, which is time-consuming, expensive, ethically questionable, and often fails to predict the effects on humans (Akhtar, 2015; Hartung, 2010; Kessel and Frank, 2007). The assumption that animals will respond to drugs in a manner similar to human reactions has led to deaths (McKenzie et al., 1995; Moore, 2016; Suntharalingam et al., 2006). Moreover, drugs that may exert positive effects in humans might be blocked by adverse outcomes in animal testing. A comparative review of preclinical and human data for coffee consumption (Bonita et al., 2007) reported that consumption of a caffeine dose equivalent to that of a

265

266

CHAPTER 8 Heart-on-a-chip

normal daily uptake was detrimental to renal structure and functions in rats but not in humans (Tofovic et al., 2002). These discrepancies have put the use of animal testing under question, and the ethical issues around animal testing have promoted efforts to replace, refine, and reduce the use of animals in research. As mentioned, cardiotoxicity is a major cause of clinical trial failures and drug withdrawals (Arrowsmith and Miller, 2013; Cook et al., 2014; Laverty et al., 2011; Suntharalingam et al., 2006), indicating that the preclinical models currently in use are not sufficient to predict adverse cardiac effects. While these models are relatively efficient in evaluating acute and severe secondary effects, they fail to assess long-term or chronic effects (Arrowsmith and Miller, 2013; Cook et al., 2014; Cross et al., 2015; Onakpoya et al., 2016). Acute myocardial toxicity is generally assessed using a relatively simple and standardized test with engineered cell lines expressing the human ether-a-go-go-related gene (hERG) (Bridgland-Taylor et al., 2006; Colatsky et al., 2016). hERG codes for a protein known as Kv11.1, the alpha subunit of a potassium ion channel involved in regulating the electrical activity of the heart (Lamothe et al., 2016; Thomas et al., 2006). Molecules with apparently low structural affinity to the ligand of the receptor have been shown to modify its activity, inducing long QT syndrome and sudden death (Lamothe et al., 2016).

3D cell culture models of the heart Traditional 2D cell cultures are neither physiological nor systemic and animal models are not fully predictive of human toxicity; these approaches are not sufficient to reliably predict the efficacy and safety of new drugs. A prominent cause of drug withdrawal from the market between 1990 and 2006 was unpredicted cardiotoxicity (Beauchamp et al., 2015). During the same period, a new technology was developed and is now widely used by both research laboratories and pharmaceutical companies. Organoid models (Simian and Bissell, 2017) are a form of 3D cell culture. Although there are several ways to obtain organoids, which mimic some aspects of organs, they all rely on the cellular capacity to self-organize. The importance of cell architecture, polarization, cell cell interactions, and the extracellular matrix (ECM) in regulating and promoting cellular differentiation and physiology has been shown (Edmondson et al., 2014; Kenny et al., 2007; Rimann and GrafHausner, 2012; Yamada and Cukierman, 2007). Recent reviews on spheroids are available (Simian and Bissell, 2017; Takahashi, 2019). These 3D heart microtissues offer several advantages in studying the biology, physiology, and pharmacology of the human heart. Unlike 2D cardiomyocyte cultures, models organizing cells in the third dimension have prolonged viability and better retain their contractile activity (Benam et al., 2015; Eschenhagen et al., 2015; Mitcheson et al., 1998; Polonchuk et al., 2017). Table 8.2 lists examples of

3D cell culture models of the heart

Table 8.2 3D heart models based on organoids. Cell type(s)

Extracellular matrix

Cardiomyocytes

Alginate hydrogel

Neonatal rat cardiomyocytes/ cardiac fibroblasts Neonatal rat cardiomyocytes/ cardiac fibroblasts/ endothelial cells Embryonic-derived cardiomyocytes/ cardiac fibroblasts/ endothelial cells iPSC-derived human cardiomyocytes/ induced cardiac fibroblasts/endotelial cells iPSC-derived human cardiomyocytes

Fibrin and Matrigel hydrogel

Cardiac progenitor cells

None

iPSC-derived human cardiomyocytes

Fibronectin/Gelatin coating

None (cell sheet technique)

Poly-L-lactic acid and polylactic glycolic acid None

None

Results

References

Functional sarcomeric structure maintained up to 14 days Functional muscle strips

Decker et al. (1991)

Improved engraftment and functionality in a postmyocardial infarction in vivo model Vascularized cardiac tissue

Sekine et al. (2008)

Reproducible and relevant model based on iPSCs

Polonchuk et al. (2017)

Reproducible and relevant model for drug toxicity in iPSCs Higher NOTCH expression than twodimensional cultures Responses to doxorubicin, E-4031, and isoproterenol comparable to the effects on humans

Beauchamp et al. (2015)

Chan et al. (2015)

Caspi et al. (2007)

Mauretti et al. (2017) Takeda et al. (2018)

iPSC, Induced pluripotent stem cell.

heart 3D models/organoids of varying complexity. The amount of scientific literature produced regarding organoid models has increased exponentially since the year 2000 (Fig. 8.4). Decker et al. cultured adult cardiac myocytes for 14 days in a matrix of alginate hydrogels, obtaining cells similar to the ones found in vivo both in terms of spindle-like shape and myofibrillar volume density (Decker et al., 1991). Chan et al. (2015) obtained functional multistrip cardiac muscle fibers using a sacrificial outer mold. Sodium alginate was mixed with 20:1 primary neonatal rat cardiac myocytes/transfected human embryonic kidney cells (Plasmid AAV(pAAV)-CAG-ChR2-GFP to express a light-gated ion channel) and injected into

267

268

CHAPTER 8 Heart-on-a-chip

FIGURE 8.4 Number of publications related to organoid cultures found in Google Scholar.

the hollow channels to mimic the architecture of muscle fibers. The resulting syncytium generated contractions upon blue-light stimulation. Sekine et al. (2008) demonstrated the promotion of neovascularization by culturing neonatal rat cardiomyocytes and endothelial cells in three cell-sheet layers. A model based on human embryonic stem cells (hESCs) seeded into a porous poly-L-lactic acid/ polylactic glycolic acid ECM allowed the formation of synchronously contracting human cardiac tissues containing endothelial vessel networks (Caspi et al., 2007). In the same work, the authors demonstrated the positive effects of mural cells in decreasing endothelial cell death and increasing endothelial cell proliferation, increasing the viability and proliferation of new cardiomyocytes. The obtained 3D heart models exhibited cardiac-specific molecular, ultrastructural, and functional properties (Caspi et al., 2007). A spheroid model mimicking features of human heart morphology, biochemistry, and pharmacology was developed from induced pluripotent stem cell (iPSC) derived cardiomyocytes, endothelial cells, and fibroblasts (Beauchamp et al., 2015). Similar results in terms of drug responses have been obtained using only human iPSC-derived cardiomyocytes coated with nanolayers of gelatin/fibronectin (Takeda et al., 2018).

The future of 3D cell culture heart models: bioprinting Additive manufacturing, more commonly termed 3D printing, constructs complex 3D objects according to digital design specifications. Although based on different principles, additive manufacturing technologies rely on a layer-by-layer approach and in general have a direct relationship between accuracy of the construct, dimensions, and time to completion. These constraints have been defined

3D cell culture models of the heart

as the biofabrication window (Malda et al., 2013). Unlike 3D printing, techniques that use cell-laden biopolymers (bioinks) to construct objects are called bioprinting or 3D bioprinting. Although additive manufacturing is not new, only in the past decade has it been harnessed for tissue engineering. In this section, we briefly review the state of bioprinted heart models; more in-depth information on 3D bioprinting of heart models is available in a recent review (Ong et al., 2018). The main limitation of current 3D bioprinting technologies is their dependence on the rheological and bioactive properties of the bioinks. A sheer thin material makes for an ideal bioink; this material should be easily cross-linked to calibrate its mechanical properties in the presence of cells. Given the importance of the extracellular microenvironment in regulating cell physiology, the ideal bioink should also exhibit some bioactivity. The polymer should promote cell adhesion, and if iPSCs are used, it should promote full differentiation into the desired phenotype. Bioinks used to produce 3D heart models are based on four main biopolymer formulations: alginate (Gaetani et al., 2012), hyaluronic acid/gelatin (Gaetani et al., 2015), collagen (Jakab et al., 2010, 2008), and native ECM (Pati et al., 2014). Sodium alginate is a polymer obtained from brown seaweed and forms a hydrogel spontaneously and rapidly when mixed with divalent cations, typically Ca21. This property has been exploited in tissue engineering, regeneration medicine, and drug-delivery systems (Lee and Mooney, 2012). Gaetani et al. (2012) mixed sodium alginate with human cardiomyocyte progenitor cells and obtained a patch that retained the functional properties of cardiomyocytes (Ong et al., 2018). In another study, Gaetani et al. (2015) used a formulation containing hyaluronic acid, gelatin, and human cardiomyocyte progenitor cells that were shown to have good grafting properties and to promote better tissue remodeling in infarcted rats. The advantage of the latter approach is that both hyaluronic acid and gelatin, which are products of partial hydrolysis of collagen, naturally occur in the ECM and can be easily remodeled using fibroblast and endothelial cell enzymes. Models based on collagen (Jakab et al., 2008) and native ECM (Jang et al., 2016; Pati et al., 2014; Shim et al., 2012) rely on natural ECM components (collagen, glycosaminoglycans, and proteoglycans), so that cells are embedded in their natural scaffolding. This advantage has been shown in the work of Jang et al. (2016), where cardiac progenitor cells actively proliferated and expressed more differentiated features owing to the biomimetic bioinks used and matching the stiffness of the hydrogel to that of the target tissue using vitamin B2 and ultraviolet light irradiation. The role of the ECM in regulating cellular and tissue physiology is well known. Collagen, the main component of the ECM, is an organized structure that supports cells and confers specific mechanical properties on tissues and organs. Apart from transmitting, dissipating, storing, and releasing energy, collagen is involved in multiple physiological processes. Dedicated molecule-binding motifs allow the extracellular storage and release of cellular mediators during embryogenesis and tissue regeneration (Cziro´k et al., 2004; Di Lullo et al., 2002; Gelse et al., 2003; Schuppan et al., 1998). The mechanical properties of tissues are

269

270

CHAPTER 8 Heart-on-a-chip

determined by collagen orientation and the various levels of interfibrillar crosslinks and associated extracellular components. Yamada and Cukierman (2007) demonstrated that ECM stiffness can be sensed by cells, affecting their adhesion, cytoskeletal structure, and proliferation, as well as the fate of stem cells (Engler et al., 2006; Lv et al., 2015). In the heart, type I and III collagen fibers are organized in discrete bundles; type I collagen fibers exhibit relatively higher stiffness, while type III collagen fibers have poorer mechanical properties and are susceptible to plastic deformation that is not recovered after the straining stress is released (Collier et al., 2012). Alterations in collagen composition and cross-linking have been associated with various pathologies (Collier et al., 2012). Collagens evolve in time and undergo a natural aging process. Collagen is relatively poorly cross-linked in newborns, but the levels of covalent cross-linking and glycation increase with age (Bailey et al., 1998). The properties of the ECM are therefore important considerations when designing hydrogel-based models for realistic phenotypes, especially when designing models that simulate the heart in elderly individuals. Ischemic in vitro models would not benefit from collagen formulations such as Matrigel or rat tail collagen, because they are poorly cross-linked and glycated. In the elderly, fibrotic pathways promote the development of some pathologies (Horn and Trafford, 2016).

Limitations of traditional two-dimensional model and 3D heart models Models based on single/multiple cell phenotype monolayers as well as 3D spheroids and organoids are still largely used in academic labs and R&D departments of pharma companies. Although such models cannot fully mimic the heart physiology or fully recapitulate pathological conditions the scalability of such models is still a major advantage. 3D organoids/spheroids models have slowly replaced 2D models because they are more physiologically relevant and can better mimic the physiology of heart cells and thus predict the effects of a drug in vivo (see Table 8.2). Nevertheless, the heart is a complex organ where the cells are subjected to different mechanical and electrical stimuli that can be difficult to recreate in 3D in vitro models. The effects of shear stress and mechanical stimulation are essential factors needed to obtain and maintain the phenotype of endothelial cells as well as heart myocytes. Nevertheless, the methods used to produce and keep in culture such models is limiting the possibilities of providing such relevant stimuli. Spheroids and organoids are relatively small and should not adhere to the culture support to keep their spherical shape; applying shear stress and rhythmic mechanical stress is thus a significant engineering challenge.

Organ-on-a-chip models of the heart

Organ-on-a-chip models of the heart: a new opportunity to mimic cardiac physiology, pathology, and toxicity Accurate mimicking of cardiac tissue physiology is challenged by the complexity of the cardiac cell microenvironment. In the native heart, cardiomyocytes are subjected to continuous deformations from the contraction and relaxation phases of the heartbeat. This produces a cyclic stress along the axis of the locally oriented fibers of the ECM, activating signaling pathways that are sensitive to these mechanical stresses and ultimately influencing the characteristics of the cardiomyocytes (Salameh et al., 2010; Trepat et al., 2007; Zhuang et al., 2000). Apart from a well-organized 3D design of cells and ECM structures, a reliable HoC model should include the biochemical and electromechanical stimuli of the myocardial environment. The geometry of microsystems and microenvironmental conditions have been shown to affect the morphology, alignment, and proliferation of H9C2 rat embryonic cardiomyocytes (Kobuszewska et al., 2017). Microfluidic devices composed of polydimethylsiloxane (PDMS) and glass with varying microchannel geometries have been developed to study the heart. The microstructures typically comprised three parallel microchannels separated by two rows of micropillars, following a design similar to that of Zervantonakis et al. (Mathur et al., 2015; Zervantonakis et al., 2012), a circular chamber, and a longitudinal channel. The structures were tested under static and perfused conditions and showed that elongated morphology, parallel orientation, and a large number of cardiac cells were obtained under perfusion in the longitudinal microchannel. A 2D microfluidic construction, also based on the work of Zervantonakis et al., was used by Mathur et al. (2015) to generate a HoC model incorporating and culturing cardiomyocytes derived from stem cells. This cardiac microphysiological system maintained cell viability and functionality over multiple weeks and was used in pharmacological studies, although the model failed to integrate mechanical stimulation. Another HoC model made of PDMS structured as three microchannels separated by micropillar arrays was used to reconstruct the interface between blood vessels and myocardial tissue (Ren et al., 2013). The system was seeded through the central channel with H9C2 cells after coating the microchannels with type I collagen, and the adjacent channels were used to supply the culture with Dulbecco’s modified Eagle medium (DMEM). Hypoxia was successfully simulated by perfusing the culture with a mitochondrial oxidative phosphorylation uncoupler in one of the adjacent channels. Models based on integrated flexible membranes that generate mechanical forces on superficially 2D-grown cardiac tissues also fail to incorporate the complex 3D cell-to-cell and ECM-to-cell interactions of the native myocardium (Ghafar-Zadeh et al., 2011; Nguyen et al., 2015, 2013). Nonetheless, this type of mechanical stimulation has been shown to contribute to calcium regulation in

271

272

CHAPTER 8 Heart-on-a-chip

cardiomyocytes and to enhance cell proliferation and protein synthesis during embryogenesis (Nguyen et al., 2015). One HoC platform was designed to construct aligned and functional 3D human pluripotent stem cell derived cardiac microtissues (Thavandiran et al., 2013). The system was based on cardiac microwires and was shown to be useful in generating a tachycardic model of arrhythmogenesis. The platform leveraged mechanical stress to improve the assembly, maturation, and functionality of the cardiac tissue while avoiding the need for tissue vascularization by miniaturizing the model. Collagen was used to control the self-assembly of the 3D tissues, and point electrodes were used for electromechanical stimulation of the cells, resulting in the expression of genes associated with cardiac maturation and electrical propagation similar to that seen in the in vivo situation. This platform could not regulate tissue properties and lacked a microfluidic environment to reproduce the generated tissues effectively or to test the concentration-dependent effect of drugs and other compounds. However, it demonstrated a bottom-up approach to architecting cardiac microtissues with functional properties that mimic native tissue in vitro through a combination of technologies. In vivo, cardiac cells experience 3D and multidirectional stress fields that vary by cardiac region. Calibration and control of the applied stress fields in vitro are therefore important when mimicking cardiac physiology. Under specific physiological and pathological conditions, cardiac fibroblasts activate, propagate, differentiate, and critically alter the amount of myocardial ECM, with notable consequences for myocardial functioning. Uniaxial and uniform cyclic stress has been shown to improve the functionality of cardiac cell monolayers grown in vitro in microdevices fabricated using the flexible and elastic PDMS (Ugolini et al., 2016). Cells were cultured on a 5-mm2 membrane supporting up to 8% stress, and actuation was achieved by the application of negative fluidic pressures to connected hollowed and microchanneled structures placed strategically around the membranes. The authors concluded that cellular morphology (shape, area, alignment, and nuclear form) was affected by the cyclic stresses and by the activation of mechanotransduction-associated transcription factors. However, this device could not control the mechanical stress on 3D tissues precisely, limiting its usefulness. An alternative model was presented later by Marsano et al. (2016) and aimed to better mimic the mechanical stimulation experienced in vivo by cardiac cells in portions of the myocardium. This 3D HoC platform was also constructed by assembling layers of PDMS and leveraging the material’s elastic mechanical properties. Again, fluidic pressure, in this case positively applied into microstructures placed under the cell culture region, was used to create membrane deformation and ultimately cell stress. This model incorporated microfluidics to deliver drugs and chemicals efficiently into the in vitro heart structure and could generate geometrically well-defined 3D cell constructs with the aid of cell-laden hydrogel prepolymers. A relatively simple mechanism provided a controlled uniform uniaxial cyclic stress to 3D heart cell cultures for HoC platforms, and microengineered

Organ-on-a-chip models of the heart

cardiac tissues of well-defined dimensions were successfully reproduced in vitro from both neonatal rat and human iPSC-derived cardiomyocytes. A sustained and synchronous beating activity of cultured cells was observed, as was a response to electric pacing signals provided by the insertion of electrodes. The system responded to low concentrations of isoprenaline, replicating the chronotropic effect previously observed in hESC-derived heart tissue (Schaaf et al., 2011; Turnbull et al., 2014) and in iPSC-derived cardiomyocyte monolayers (Jonsson et al., 2011) at higher doses. Another, less versatile HoC system was developed to use 3D chip structures in heart models with physiological relevance (Aung et al., 2016). The device defined an inner circular chamber made of PDMS and glass, with multiple branched microchannels leading in and out of the chamber and converging into a single inlet and outlet. The circular chamber contained a three-layer sandwich structure made of two polyacrylamide hydrogel layers containing embedded 200-nm green fluorescent particles and surrounding a central part with selectively photopolymerized spots of gelatin methacrylate containing isolated primary neonatal cardiomyocytes. The structure was fixed to the bottom and top of the circular chamber and allowed real-time contractile stress assessment by monitoring the fluorescent beads during cell beating. These engineered cardiac tissues increased their beating frequency and stress magnitude when exposed to 0.1 μg/mL of exogenous epinephrine. A similar system with embedded nanomaterials has been used to test cardiac contractility (Ahn et al., 2018). This HoC was inspired by mussels and uses adhesive and aligned polydopamine/polycaprolactone nanofibers to emulate the 3D native ECM environment of the myocardium (Ahn et al., 2018). Previously developed thin-film technology sensors and PDMS microdevices provided continuous and noninvasive readouts of the contractile stress and beating frequency of an engineered human iPSC-derived cardiomyocyte tissue (Lind et al., 2017). The original sensors were coated with nanofibers that supported the formation of anisotropic and contractile muscle tissues, and the effect of titanium dioxide and silver nanoparticles on the contractile function was assessed and related to cardiotoxicity. At different concentrations, a decrease in contractile function was observed that was consistent with structural damage recorded using other methods. To solve the problem of large-scale fabrication reproducibility, flexibility, and accuracy of microscale gelatin hydrogel structures, Nawroth et al. (2018) proposed a biocompatible laser-etching approach. The method allows the construction of groove and pillar structures with feature dimensions on the order of 10 30 μm and a standard deviation of about 0.3 μm. Patterned heart chip substrates were seeded with neonatal rat cardiomyocytes and with human iPSC-derived cardiomyocytes to generate cardiac tissues that achieved structural organization, contractile function, and long-term viability. However, the study was a proof-of-concept effort, and no associated functional HoC device incorporating microfluidics was presented.

273

274

CHAPTER 8 Heart-on-a-chip

Recent advances in HoC models include the combination of microfluidic technologies with tissue engineering and nanotechnologies. The most relevant work in this sense is the engineered scale model of the heart ventricle developed by MacQueen et al. (2018); the investigators guided tissue assembly using polycaprolactone/gelatin nanofibrous 3D substrates that promote anisotropic myocardial tissue genesis. The heart ventricle ECM-inspired scaffold was placed inside a microfluidic bioreactor that included both intra- and extraventricular flow loops and access ports for catheters. The scaffold was seeded with both neonatal rat ventricular myocytes and human iPSC-derived cardiomyocytes. After 3 5 days of culture, synchronous ventricle contraction developed spontaneously and displayed stable differentiated excitation patterns when simulating healthy and injured ventricles. These patterns repeated at 4 5 Hz for healthy (plane waves) and injured (pinned spiral waves) tissues. The results demonstrated the potential of this microfluidic bioreactor model to mimic arrhythmogenic heart diseases, and followed previous work by this research group (Grosberg et al., 2011) where an anisotropic ventricular myocardium was engineered by seeding myocytes into a combination of poly(N-isopropylacrylamide), fibronectin, and PDMS microconstructs and patterns on a glass substrate. Even though the myocytes were grown in a 2D conformation and microfluidics were not used, contractibility, cytoskeletal architecture, and action potential propagation were assessed, and a drug response curve to epinephrine showed the expected chronotropic effects. The same technique was used later by this research group with a polycarbonate (PC) and aluminum microfluidic device incorporating temperature control and electrical stimulation to demonstrate a potential high-throughput HoC platform for pharmacological studies (Agarwal et al., 2013). In this work, the device was used to test the positive inotropic effect of isoproterenol on cardiac contractility. Other alternatives to 3D HoC structures with seeded cells have included the use of biopsy specimens or previously grown spheroids. Cheah et al. (2010) constructed a HoC model with microfluidic perfusion, electrostimulation, and realtime electrochemical monitoring of reactive oxygen species from preserved biopsy specimens. The device had a PDMS microfluidic chamber with five embedded electrodes bonded to a Petri dish; samples of rat right ventricular tissue and human right atrial tissue maintained viability up to 5 and 3.5 hours, respectively. Cellular damage was monitored by quantifying the release of lactate dehydrogenase, the levels of which were low during basal conditions but peaked 10 15 minutes after the damage was induced. Reactive oxygen species were monitored in situ using amperometry and correlated with the release of lactate dehydrogenase, providing a method for real-time monitoring of specimen viability while experimenting in vitro. Christoffersson et al. presented a HoC model based on human iPSC-derived cardiac spheroids (Christoffersson et al., 2018), with a microfluidic system coupled to confocal high-content microscopy that provided quantitative data on the effects of pharmacological compounds on cell outgrowth. Commercially available polymeric microscope slides with six parallel microfluidic channels and coated with laminin were used for the HoC device, and cardiac

Effects of heart-on-a-chip models on cell differentiation

spheroids were seeded inside the channels, incubated in cell culture medium, and fixed to a rocking platform to generate perfusion by gravity. The system was demonstrated to run for 48 hours and an assay on cardiac spheroids with six relevant compounds (doxorubicin, endothelin 1, acetylsalicylic acid, isoproterenol, phenylephrine, and amiodarone) at three concentrations was performed. The results were not conclusive for all six compounds because of a lack of comparative data, but some expected results were successfully replicated even at low concentrations. No significant differences between static and dynamic culture conditions were observed in the experiments, perhaps owing to the rocking platform and a limited medium volume. A recent attempt to leverage organ-on-a-chip technologies to replicate the interactions of multiple organs has also involved HoC models. An integrated heart/cancer-on-a-chip model using human healthy cardiac cells and HepG2 hepatocellular carcinoma cells incorporated closed blood circulation loops with integrated microvalves and a micropump for accurate fluidic control (Kamei et al., 2017). Cells were cultivated on the chip after coating the culture chambers with Matrigel as ECM. The known cardiotoxic effect of the anticancer drug doxorubicin was evaluated in vitro by reproducing the generation of toxic metabolites by HepG2 cells and their delivery to the cardiac cells through blood circulation. The results suggested that doxorubicin is not toxic to cardiomyocytes, but its metabolite doxorubicinol is the underlying cause of the observed cardiotoxicity. Recreating these types of complex environments on a chip was anticipated a few years ago (Moya et al., 2013).

Effects of heart-on-a-chip models on cell differentiation An important feature for HoC application in therapeutics testing and personalized or precise medicine is its use in combination with stem cells. We have already reviewed some models that used iPSCs, but differentiation of stem cells into cardiac cells on a chip is a subject of interest by itself (Jastrzebska et al., 2016). A PDMS and glass bioreactor with embedded 1.3-mm electrodes was used to evaluate cardiogenesis and the generation of reactive oxygen species in embryoid bodies derived from hESCs (Serena et al., 2009). The experiments showed that electrical field stimulation generated intracellular endogenous reactive oxygen species associated with spontaneous contractions, expression of troponin T, and sarcomeric organization on the embryoid bodies. Thus cardiac differentiation of hESCs was linked to the generation of reactive oxygen species following the application of electrical fields. Advances toward achieving cardiomyocyte differentiation on a chip continued with the work of Tandon et al. (2010), who proposed a microfluidic system with an array of integrated interdigitated electrodes to culture heart ventricle cells from neonatal rats and human adipose tissue-derived

275

276

CHAPTER 8 Heart-on-a-chip

stem cells. After 6 days of culturing, electrical stimulation increased the proliferation, elongation, and alignment of these two cell types. Attempts have also been made to generate 3D functional cardiac tissue onchip for preclinical testing in vitro in devices with integrated electrical stimulation (Xiao et al., 2014). These wire-like tissues were created from both primary neonatal rat cardiomyocytes and hESC-derived cardiomyocytes and could be perfused and electrically stimulated. The device was made of PDMS and consisted of a 3D structure defining a microchannel that contained a biowire template and was sandwiched between two cell culture chambers. A polytetrafluoroethylene tubing template was anchored between the two ends of the PDMS structure suspended in the central microchannel before introducing ECM-emulating hydrogels and the cells. A system to perfuse media and drugs through the biowire was incorporated also in PDMS, using a glass substrate as a base, and transversal, as well as longitudinal, to-the-biowire electrodes were integrated in the device to allow electrical stimulation. Both cell types beat spontaneously after formation of these cardiac cylindrical tissues. The phenotype of the cardiomyocytes was improved when electrical stimulus was applied through the integrated carbon rod electrodes. The orientation of the cells followed the electrical field and the parallel field provided a tissue stiffness for the biowires similar to the in vivo tissue. Furthermore, the release of nitric oxide (NO) from sodium nitroprusside perfused through the biowires tubes slowed down the spontaneous beating of the cardiac biowires based on hESC-derived cardiomyocytes. A different approach was demonstrated by introducing hemodynamic-induced mechanical stress to H9c2 embryonic rat myoblasts on a chip (Giridharan et al., 2010). This microfluidic cardiac cell culture model was composed of PC plates and an embedded PDMS membrane intended to replicate the influence of mechanical stress on cardiac cell structure, contractile function, and gene expression. A phenotype similar to the in vivo situation was achieved for the differentiated cardiomyocytes at conditions mimicking normal mechanical stresses. Distinctively to static controls, the cells cultured in the HoC model showed alignment of F-actin stress fibers. The model was used to emulate pathological states such as heart failure, hypertension, hypotension, bradycardia, and tachycardia. Similarly, another microfluidic platform was developed allowing simultaneously versatile cell seeding arrangements, long-term cell viability, programmable uniaxial cyclic stretch, and optical observation (Wan et al., 2011); the microdevice was constructed of PDMS and leveraged the stretchability and transparency of this material. The chip layout included three microfluidic channels separated by two gel regions that allowed culturing embryoid bodies. Uniaxial cyclic stretch was applied and controlled by a precision linear motor. Cardiomyogenic differentiation in murine embryonic stem cells was enhanced on the chip compared with differentiation in conventional well plates, but the application of uniaxial stretch appeared to interrupt differentiation, potentially reducing long-term cardiogenesis. More recently, a microfluidic device providing electrical, mechanical, and biochemical stimulation was presented and validated by controlling human

Effects of heart-on-a-chip models on cell differentiation

bone-marrow mesenchymal stem cells (Pavesi et al., 2015) cultured in DMEM. The microdevice was constructed in PDMS and electrodes were generated by selectively embedding carbon nanotubes (CNTs) in its structure. The device consisted of a sandwiched 3D structure allowing microfluidic perfusion and mechanical strain through controlled pneumatic channels and an embedded PDMS membrane. Electrical stimulation was provided by conductive regions of a CNT and PDMS mixture. This microbioreactor enabled morphological and geneexpression analyses on the cellular response to stimuli, as well as fluorescence immunostaining, confocal imaging, and harvesting previously stimulated cells for standard molecular biology assays. When mechanical stimulation at 7% strain was applied simultaneously with 5 V/cm electrical stimulation, more than 75% of the cytoskeletal fibers aligned within 30 degrees of the perpendicular direction to the stresses. Fig. 8.5 details the device and displays the direction of the strain and electrical stimulation. Both mechanical and electrical stimulations enhanced activation of cardiac myocyte markers. There was a stronger effect of mechanical strain than of electrical stimulation, and a strong synergy between both stimuli when applied simultaneously.

FIGURE 8.5 Organ-on-a-chip device providing full electrical, mechanical, and biochemical stimulation. Adapted from Pavesi, A., Adriani, G., Rasponi, M., Zervantonakis, I.K., Fiore, G.B., Kamm, R.D., 2015. Controlled electromechanical cell stimulation on-a-chip. Sci. Rep. 5, 11800. https://doi.org/10.1038/ srep11800. This work is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0).

277

Table 8.3 Heart-on-a-chip models and their evolution in recent years. materials device

Perfusion time

flow rate

External stimulation

Author

Year

Zhuang et al.

2000

Silicone, Teflon

None

None

Mechanical

Neonatal rat ventricular CM

Collagen

Physiological/ Arrhythmia model

Serena et al.

2009

None

None

Electrical

hESC (line H13)

Hydrogel

Cardiogenesis

Salameh et al.

2010

PDMS, glass, electrodes (stainless steel, titanium nitride, or titanium) Silicone membrane

None

None

Mechanical

Neonatal rat CM

Gelatin

Physiological model

Cheah et al.

2010

5/3.5 h

100 μL/min

Electrical

Biopsies of rat right ventricular/ Human right atrial tissue

None

Physiological model

PDMS, PS, Teflon, electrodes (platinum)

Cell types

Coating/ Extracellular matrix

Study

Tested biomarkers Gap junction protein Cx43 and the fascia adherens junction protein N-cadherin Intracellular ROS, TnT expression, and spontaneous contractions

Morphology, distribution and expression of connexin Cx43 and N-cadherin, and expression of transcription factors activator protein 1 and CREB (cAMP response element binding protein) in the nucleus LDH and ROS

Tandon et al.

2010

Giridharan et al.

2010

Wan et al.

2011

Jonsson et al.

Grosberg et al.

PDMS, glass, electrodes (indium tin oxide) PDMS, PC, latex

None

None

Electrical

Neonatal rat ventricular CM/ hASC

Collagen

Physiological model

Morphology and Cx43

24 h

8 mL/min

Mechanical

Rat embryonic cardiomyoblast line (H9c2)

Fibronectin

Morphology and intracellular F-actin

PDMS

None

None

Mechanical

2011

PS, glass, electrodes (gold)

None

None

None

mESC line (CGR8)/hMVEC hiPSC-derived CM line and mESC-derived CM line

Gelatin/ Collagen Gelatin

Physiological/ Heart failure/ Hypertension/ Hypotension/ Tachycardia/ Bradycardia model Cardiogenesis

2011

PDMS, glass, PNIPAAm, electrodes (platinum)

None

None

Electromechanical

Neonatal rat ventricular CM

Fibronectin

Physiological model/ Arrhythmia model/Drug effects (isoproterenol, carbachol, E4031, amlodipine, DMSO, zatebradine, and T-type calcium channel blockers) Epinephrine on physiological model

Morphology and differentiation Impedance measurements

Morphology, voltage membrane dye RH237, and distribution of actin, nuclei, and sarcomeric α-actinin.

(Continued)

Table 8.3 Heart-on-a-chip models and their evolution in recent years. Continued Perfusion time

Author

Year

Ren et al.

2013

PDMS, glass

15 min

0.2 20 μL/ min

None

Rat embryonic cardiomyoblast line (H9c2)

Collagen

Hypoxiainduced myocardial injury model

Nguyen et al.

2013

PDMS; electrodes (platinum)

4 days

5.2 mL/min (0.5 s pulses)

Electromechanical

Embryonic chick CM

Collagen/ Glutaraldehyde

Cardiogenesis

Thavandiran et al.

2013

PDMS, PS, electrodes (platinum)

None

None

Electromechanical

Neonatal rat CM/ hPSC HES2 line

Collagen/ Pluronic acid

Cardiogenesis/ Physiological/ Drug effects (epinephrine, lidocaine, verapamil, and 21 L-type Ca channel blocker) model

Agarwal et al.

2013

PDMS, glass, PNIPAAm, PC, aluminum, electrodes (platinum), medical-grade urethane adhesive

10 min

1 mL/min

Electrical

Neonatal rat ventricular CM

Fibronectin

Isoproterenol effects on physiological model

flow rate

External stimulation

Coating/ Extracellular matrix

materials device

Cell types

Study

Tested biomarkers Morphology, distribution of actin, nuclei, mitochondrial membrane potential, and caspase-3 activity Morphology, differentiation, contractility, and beat rate Morphology, α-actinin, cTnT, NKX2 5, DDR2, Cx43, SIRPA, cTnT, ANF, BNP, MYL2, MYL7, MYH6, MYH7, transmembrane action potentials, and intracellular calcium transients Morphology, contractility, distribution of nuclei, actin, and sarcomeric α-actinin

Turnbull et al.

2014

PDMS

None

None

Electrical

hESC-CM

Collagen/ Matrigel

Physiological/ Drug effects (calcium chloride and verapami) model

Xiao et al.

2014

PDMS, glass, PTFE, electrodes (carbon)

4 days

2 μL/min

Electrical

Neonatal rat CM and hESCCM

Collagen

Physiological/ Drug effects (NO) model

Pavesi et al.

2015

PDMS, CNT

None

None

Electromechanical

hMSC

Fibronectin

Cardiogenesis/ Physiological model

Mathur et al.

2015

PFMS, glass

5 days

20 μL/h

None

hiPSC

Fibronectin

Physiological model/Drug effects [isoproterenol (β-adrenergic agonist), E4031 (hERG blocker), verapamil (multi-ion channel blocker), and metoprolol (β-adrenergic antagonist)]

Morphology, contractility, twitch stress, beat rate, action potential, expression of cardiac TNNT2, α-sarcomeric actinin (ACTN2), Cx-43, α-SMA, α-cardiac muscle actin (ACTC1), α-MHC, β-MHC, ANF, and SERCA2a Morphology, distribution of cTnT, Cx-43, α-actinin, F-actin fibers, and nuclei Morphology, distribution of actin fibers, nuclei, and expression of GATA4, MEF2C, MYH7, NKX2.5, TUBB, CX43, TNNT2, and OCT4 Morphology, contractility, beat rate, electrophisiology, distribution of sarcomeric alpha actinin, and activity of Ca21 channels

(Continued)

Table 8.3 Heart-on-a-chip models and their evolution in recent years. Continued

Cell types

Coating/ Extracellular matrix

Mechanical

Embryonic chick CM

Collagen/ Glutaraldehyde

None

Mechanical

Human adult CF

Fibronectin

None

None

Electromechanical

Neonatal rat CM and fibroblasts/ hiPSC-CM

Fibrin gel

Cardiogenesis/ Physiological model and effect of isoprenaline

14 days

n.a.

None

Neonatal mouse CM and fibroblasts

Methacrylated gelatin

Physiological model and effect of epinephrine

72 h

1 μL/min

None

H9c2

None

Cardiogenesis/ Physiological model and effect of verapamil

Author

Year

materials device

Perfusion time

Nguyen et al.

2015

PDMS

4 days

5.2 mL/min (0.5 s pulses)

Ugolini et al.

2016

PDMS

None

Marsano et al.

2016

PDMS, electrodes (carbon)

Aung et al.

2016

Kobuszewska et al.

2017

PDMS, glass, GelMA, PAm, 200 nm green fluorescent nanoparticles PDMS, glass

flow rate

External stimulation

Study Cardiogenesis/ Physiological model and effect of isoproterenol Physiological/ Cardiacfibrotic disease model

Tested biomarkers SERCA2a, α-actinin, PLB, TnT, and α-, β-MHC Morphology, proliferation markers Ki67 and PHH3, EdU DNA synthesis labeling, and YAP expression and distribution Morphology, contractility, beat rate, expression of α-sarcomeric actinin, MYH7, MYH6, MLC2v, MLC2a, troponin I, Cx43, and Ncadherin Morphology, contractility, and expression or distribution of Cx43 Morphology and proliferation

Lind et al.

2017

PDMS, PC, titanium gold thin film

None

None

None

Neonatal rat ventricular CM/ hiPSC-CM

Fibronectin

Kamei et al.

2017

PDMS, glass

24 h

26 nL/min

None

HepG2 hepatocellular carcinoma cell line and human primary CM

Fibronectin/ Gelatin/ Matrigel

Ahn et al.

2018

PDMS, glass, PNIPAAm, titanium gold thin film, PC, PDA/ PCL nanofiber

None

None

None

Neonatal rat ventricular CM

Fibronectin

Physiological model/Drug effects (isradipine, nicardipine, clofilium tosylate, PD118057, flecainide, salmeterol xinafoate, isoproterenol, desipramine, astemizole, domperidone, FK-506, mefloquine, and isradipine, across a vascular endothelial barrier with and without coexposure of TNF-α) Physiological model/ Drug effects (doxorubicin, doxorubicinol, and staurosporine) Cardiogenesis/ Physiological model/Drug effects (silver nanoparticles and titania nanoparticles)

Morphology, beat rate, contractibility, twitch stress, and IC50

Morphology, proliferation, and EdU DNA synthesis labeling

Morphology, contractility, expression of α-sarcomeric actinin, and LDH toxicity test

(Continued)

Table 8.3 Heart-on-a-chip models and their evolution in recent years. Continued materials device

Perfusion time

flow rate

External stimulation

Cell types

Coating/ Extracellular matrix

Author

Year

Study

MacQueen et al.

2018

PC, silicone, stainless steel, PV catheter

6 months

N.a. (pressure pulses)

Mechanical

Neonatal rat ventricular CM/ hiPSC-CM

PCL gelatin nanofibers

Cardiogenesis/ Physiological model/ Arrhythmia model/Drug effects (isoproterenol)

Christoffersson et al.

2018

Ibidi polymer

48 h

N.a. (rocking table)

None

hiPSC-CM spheroids

Laminin

Physiological model/Drug effects (doxorubicin, endothelin-1, acetylsalicylic acid, isoproterenol, phenylephrine, amiodarone)

Tested biomarkers Morphology, contractility, PV measurement, echocardiography, expression and distribution of Factin fibers, calcium and sarcomeric α-actinin Morphology, nuclei, cTnT, α-actinin, MHC, and NKX2.5

Notice the complexity of the models and the fact that fluidic flow, which is intrinsic of the heart function, has only been introduced recently and in a limited number of studies. Furthermore, only one model combines electromechanics and fluidics in order to improve the heart physiology model. α-MHC, α-Myosin heavy chain; β-MHC, β-myosin heavy chain; α-SMA, α-smooth muscle actin; ANF, atrial natriuretic factor; CF, cardiac fibroblasts; CM, cardiomyocytes; CNT, carbon nanotubes; cTnT, cardiac troponin T; Cx43, connexin43; DMSO, dimethyl sulfoxide; EdU, 5-ethynyl-2bdeoxyuridine; GelMA, methacrylated gelatin; hASC, human adipose tissue-derived stem cells; hESC, human embryonic stem cells; hESC-CM, human embryonic stem cell-derived cardiomyocytes; hiPSC, human induced pluripotent stem cell; hMSC, human mesenchymal stem cells; hMVEC, human microvascular endothelial cells; hPSC, human pluripotent stem cell; IC50, half-maximal inhibitory concentration; LDH, lactate dehydrogenase; mESC, murine embryonic stem cell; NO, nitric oxide; PAm, polyacrylamide; PC, polycarbonate; PCL, polycaprolactone; PDA, polydopamine; PDMS, polydimethylsiloxane; PHH3, phospho-histone-H3; PLB, phospholamban; PNIPAAm, poly(N-isopropylacrylamide); PS, polystyrene; PTFE, polytetrafluoroethylene; PV, pressure volume; ROS, reactive oxygen species; SERCA2a, sarcoplasmic reticulum calcium ATPase2a; SIRPA, signal-regulatory protein α; TNNT2, troponin-T type 2; TnT, troponin T; YAP, YES-associated protein.

References

Final remarks and future directions Current advances in tissue engineering, materials, 3D printing, microtechnologies, and microfluidics are addressing the requirements for in vitro models of the heart. However, despite the large numbers of publications and research studies on the topic (Table 8.3), and their developments and designs, current HoC models are still far from large-scale fabrication and distribution. Limitations arise from fabrication complexity, material costs, and lack of standards in drug-discovery or cardiopathology studies. Standardization of techniques and materials, advances in manufacturing processes, and increased private sector investment will help these technologies become feasible realities. The efficacy and efficiency of these technologies have thus far been demonstrated only partially and in laboratory settings. Reproducible and massproducible HoC models should be established and systematically tested and validated against previous methods used for characterizing the therapeutic and toxicological effects of compounds. The continuous advancement of complementary technologies and their integration with currently available models, as well as a better understanding of complex biological systems, should provide, together with increased societal and industrial interest, a strong basis for the effective application of these technologies in improving human health while reducing overall costs and the use of animals.

Acknowledgments Dr. Dario Fassini was funded through European Unions’s Horizon 2020 Marie Sklodowska-Curie Individual Fellowships Programme (H2020-MSCA-IF-2017) under grant agreement no 752277; the book chapter content benefited also of the exchanges and inter% through the project CISTEM of the European Unions’s Horizon 2020 Marie actions funded Sklodowska-Curie Research and Innovation Staff Exchange Programme (H2020-MSCARISE-2017) under grant agreement no 778354. The content of this book chapter reflects, % the authors’ view, and the European Union is not however, only Cherry Biotech SAS and liable for any use that may be made of the information contained therein.

References Akhtar, A., 2015. The flaws and human harms of animal experimentation. Camb. Q. Healthc. Ethics. 24, 407 419. Available from: https://doi.org/10.1017/ S0963180115000079. Agarwal, A., Goss, J.A., Cho, A., McCain, M.L., Parker, K.K., 2013. Microfluidic heart on a chip for higher throughput pharmacological studies. Lab Chip 13, 3599. Available from: https://doi.org/10.1039/c3lc50350j.

285

286

CHAPTER 8 Heart-on-a-chip

Ahn, S., Ardon˜a, H.A.M., Lind, J.U., Eweje, F., Kim, S.L., Gonzalez, G.M., et al., 2018. Mussel-inspired 3D fiber scaffolds for heart-on-a-chip toxicity studies of engineered nanomaterials. Anal. Bioanal. Chem. 410, 6141 6154. Available from: https://doi.org/ 10.1007/s00216-018-1106-7. Arrowsmith, J., Miller, P., 2013. Phase II and Phase III attrition rates 2011 2012. Nat. Rev. Drug Discov. 12, . Available from: https://doi.org/10.1038/nrd4090569 569. Aung, A., Bhullar, I.S., Theprungsirikul, J., Davey, S.K., Lim, H.L., Chiu, Y.-J., et al., 2016. 3D cardiac μtissues within a microfluidic device with real-time contractile stress readout. Lab Chip 16, 153 162. Available from: https://doi.org/10.1039/C5LC00820D. Bailey, A.J., Paul, R.G., Knott, L., 1998. Mechanisms of maturation and ageing of collagen. Mech. Ageing Dev. 106, 1 56. Bauer, M., Kang, L., Qiu, Y., Wu, J., Peng, M., Chen, H.H., et al., 2012. Adult cardiac progenitor cell aggregates exhibit survival benefit both in vitro and in vivo. PLoS One 7, e50491. Available from: https://doi.org/10.1371/journal.pone.0050491. Beauchamp, P., Moritz, W., Kelm, J.M., Ullrich, N.D., Agarkova, I., Anson, B.D., et al., 2015. Development and characterization of a scaffold-free 3D spheroid model of induced pluripotent stem cell-derived human cardiomyocytes. Tissue Eng., C: Methods 21, 852 861. Available from: https://doi.org/10.1089/ten.TEC.2014.0376. Benam, K.H., Dauth, S., Hassell, B., Herland, A., Jain, A., Jang, K.-J., et al., 2015. Engineered in vitro disease models. Annu. Rev. Pathol. Mech. Dis. 10, 195 262. Available from: https://doi.org/10.1146/annurev-pathol-012414-040418. Betts, J.G., 2013. Anatomy & Physiology, Openstax College. pp. 787 846. Bonita, J.S., Mandarano, M., Shuta, D., Vinson, J., 2007. Coffee and cardiovascular disease: in vitro, cellular, animal, and human studies. Pharmacol. Res. 55, 187 198. Available from: https://doi.org/10.1016/j.phrs.2007.01.006. Bridgland-Taylor, M.H., Hargreaves, A.C., Easter, A., Orme, A., Henthorn, D.C., Ding, M., et al., 2006. Optimisation and validation of a medium-throughput electrophysiology-based hERG assay using IonWorkst HT. J. Pharmacol. Toxicol. Methods 54, 189 199. Available from: https://doi.org/10.1016/j.vascn.2006.02.003. ˇ c, T., Oberwallner, B., Mahmoodzadeh, S., Neef, K., Albrecht, J., et al., Brodarac, A., Sari´ 2015. Susceptibility of murine induced pluripotent stem cell-derived cardiomyocytes to hypoxia and nutrient deprivation. Stem Cell Res. Ther. 6, 83. Available from: https:// doi.org/10.1186/s13287-015-0057-6. Canfield, S.G., Sepac, A., Sedlic, F., Muravyeva, M.Y., Bai, X., Bosnjak, Z.J., 2012. Marked hyperglycemia attenuates anesthetic preconditioning in human-induced pluripotent stem cell-derived cardiomyocytes. Anesthesiology 117, 735 744. Available from: https://doi.org/10.1097/ALN.0b013e3182655e96. Caspi, O., Lesman, A., Basevitch, Y., Gepstein, A., Arbel, G., Habib, I.H.M., et al., 2007. Tissue engineering of vascularized cardiac muscle from human embryonic stem cells. Circ. Res. 100, 263 272. Available from: https://doi.org/10.1161/01. RES.0000257776.05673.ff. Chan, V., Neal, D.M., Uzel, S.G.M., Kim, H., Bashir, R., Asada, H.H., 2015. Fabrication and characterization of optogenetic, multi-strip cardiac muscles. Lab Chip 15, 2258 2268. Available from: https://doi.org/10.1039/c5lc00222b. Cheah, L.-T., Dou, Y.-H., Seymour, A.-M.L., Dyer, C.E., Haswell, S.J., Wadhawan, J.D., et al., 2010. Microfluidic perfusion system for maintaining viable heart tissue with real-time electrochemical monitoring of reactive oxygen species. Lab Chip 10, 2720 2726. Available from: https://doi.org/10.1039/c004910g.

References

Christoffersson, J., Meier, F., Kempf, H., Schwanke, K., Coffee, M., Beilmann, M., et al., 2018. A cardiac cell outgrowth assay for evaluating drug compounds using a cardiac spheroid-on-a-chip device. Bioengineering (Basel, Switzerland) 5, 36. Available from: https://doi.org/10.3390/bioengineering5020036. Colatsky, T., Fermini, B., Gintant, G., Pierson, J.B., Sager, P., Sekino, Y., et al., 2016. The comprehensive in vitro proarrhythmia assay (CiPA) initiative—update on progress. J. Pharmacol. Toxicol. Methods 81, 15 20. Available from: https://doi.org/10.1016/j. vascn.2016.06.002. Collier, P., Watson, C.J., van Es, M.H., Phelan, D., McGorrian, C., Tolan, M., et al., 2012. Getting to the heart of cardiac remodeling; how collagen subtypes may contribute to phenotype. J. Mol. Cell. Cardiol. 52, 148 153. Available from: https://doi.org/10.1016/ j.yjmcc.2011.10.002. Cook, D., Brown, D., Alexander, R., March, R., Morgan, P., Satterthwaite, G., et al., 2014. Lessons learned from the fate of AstraZeneca’s drug pipeline: a five-dimensional framework. Nat. Rev. Drug Discov. 13, 419 431. Available from: https://doi.org/ 10.1038/nrd4309. Cross, M.J., Berridge, B.R., Clements, P.J.M., Cove-Smith, L., Force, T.L., Hoffmann, P., et al., 2015. Physiological, pharmacological and toxicological considerations of druginduced structural cardiac injury. Br. J. Pharmacol. 172, 957 974. Available from: https://doi.org/10.1111/bph.12979. Cziro´k, A., Rongish, B.J., Little, C.D., 2004. Extracellular matrix dynamics during vertebrate axis formation. Dev. Biol. 268, 111 122. Available from: https://doi.org/ 10.1016/J.YDBIO.2003.09.040. Date, T., Belanger, A.J., Mochizuki, S., Sullivan, J.A., Liu, L.X., Scaria, A., et al., 2002. Adenovirus-mediated expression of p35 prevents hypoxia/reoxygenation injury by reducing reactive oxygen species and caspase activity. Cardiovasc. Res. 55 (309 319). Available from: https://doi.org/10.1016/S0008-6363(02)00412-1. Decker, M.L., Behnke-Barclay, M., Cook, M.G., La Pres, J.J., Clark, W.A., Decker, R.S., 1991. Cell shape and organization of the contractile apparatus in cultured adult cardiac myocytes. J. Mol. Cell. Cardiol. 23, 817 832. Di Lullo, G.A., Sweeney, S.M., Ko¨rkko¨, J., Ala-Kokko, L., San Antonio, J.D., 2002. Mapping the ligand-binding sites and disease-associated mutations on the most abundant protein in the human, type I collagen. J. Biol. Chem. 277, 4223 4231. Available from: https://doi.org/10.1074/jbc.M110709200. Edmondson, R., Broglie, J.J., Adcock, A.F., Yang, L., 2014. Three-dimensional cell culture systems and their applications in drug discovery and cell-based biosensors. Assay Drug Dev. Technol. 12, 207 218. Available from: https://doi.org/10.1089/adt.2014.573. Eichler, H.-G., Pignatti, F., Flamion, B., Leufkens, H., Breckenridge, A., 2008. Balancing early market access to new drugs with the need for benefit/risk data: a mounting dilemma. Nat. Rev. Drug Discov. 7, 818 826. Available from: https://doi.org/10.1038/ nrd2664. Engler, A.J., Sen, S., Sweeney, H.L., Discher, D.E., 2006. Matrix elasticity directs stem cell lineage specification. Cell 126, 677 689. Available from: https://doi.org/10.1016/j. cell.2006.06.044. Eschenhagen, T., Mummery, C., Knollmann, B.C., 2015. Modelling sarcomeric cardiomyopathies in the dish: from human heart samples to iPSC cardiomyocytes. Cardiovasc. Res. 105, 424 438. Available from: https://doi.org/10.1093/cvr/cvv017. Gaetani, R., Doevendans, P.A., Metz, C.H.G., Alblas, J., Messina, E., Giacomello, A., et al., 2012. Cardiac tissue engineering using tissue printing technology and human

287

288

CHAPTER 8 Heart-on-a-chip

cardiac progenitor cells. Biomaterials 33, 1782 1790. Available from: https://doi.org/ 10.1016/j.biomaterials.2011.11.003. Gaetani, R., Feyen, D.A.M., Verhage, V., Slaats, R., Messina, E., Christman, K.L., et al., 2015. Epicardial application of cardiac progenitor cells in a 3D-printed gelatin/hyaluronic acid patch preserves cardiac function after myocardial infarction. Biomaterials 61, 339 348. Available from: https://doi.org/10.1016/j.biomaterials.2015.05.005. Gelse, K., Po¨schl, E., Aigner, T., 2003. Collagens--structure, function, and biosynthesis. Adv. Drug Deliv. Rev. 55, 1531 1546. Ghafar-Zadeh, E., Waldeisen, J.R., Lee, L.P., 2011. Engineered approaches to the stem cell microenvironment for cardiac tissue regeneration. Lab Chip 11, 3031. Available from: https://doi.org/10.1039/c1lc20284g. Giridharan, G.A., Nguyen, M.-D., Estrada, R., Parichehreh, V., Hamid, T., Ismahil, M.A., et al., 2010. Microfluidic cardiac cell culture model (μCCCM). Anal. Chem. 82, 7581 7587. Available from: https://doi.org/10.1021/ac1012893. Grosberg, A., Alford, P.W., McCain, M.L., Parker, K.K., 2011. Ensembles of engineered cardiac tissues for physiological and pharmacological study: heart on a chip. Lab Chip 11, 4165 4173. Available from: https://doi.org/10.1039/c1lc20557a. Heinzel, F., Luo, Y., Dodoni, G., Boengler, K., Petrat, F., Dilisa, F., et al., 2006. Formation of reactive oxygen species at increased contraction frequency in rat cardiomyocytes. Cardiovasc. Res. 71, 374 382. Available from: https://doi.org/10.1016/j. cardiores.2006.05.014. Hartung, T., 2010. Food for thought. . . on alternative methods for chemical safety testing. ALTEX 27, 3 14. Heusch, G., 2017. Critical issues for the translation of cardioprotection. Circ. Res. 120, 1477 1486. Available from: https://doi.org/10.1161/CIRCRESAHA.117.310820. His, W.J., 1893. Die Ta¨tigkeit des embryonalen Herzens und deren Bedeutung fu¨r die Lehre von der Herzbewegung beim Erwachsenen. Arb. aus der medizinischen Klin. zu Leipzig 1, 14 50. Horn, M.A., Trafford, A.W., 2016. Aging and the cardiac collagen matrix: novel mediators of fibrotic remodelling. J. Mol. Cell. Cardiol. 93, 175 185. Available from: https://doi. org/10.1016/j.yjmcc.2015.11.005. Jakab, K., Norotte, C., Damon, B., Marga, F., Neagu, A., Besch-Williford, C.L., et al., 2008. Tissue engineering by self-assembly of cells printed into topologically defined structures. Tissue Eng., A 14, 413 421. Available from: https://doi.org/10.1089/ tea.2007.0173. Jakab, K., Norotte, C., Marga, F., Murphy, K., Vunjak-Novakovic, G., Forgacs, G., 2010. Tissue engineering by self-assembly and bio-printing of living cells. Biofabrication 2, 022001. Available from: https://doi.org/10.1088/1758-5082/2/2/022001. Jang, J., Kim, T.G., Kim, B.S., Kim, S.-W., Kwon, S.-M., Cho, D.-W., 2016. Tailoring mechanical properties of decellularized extracellular matrix bioink by vitamin B2induced photo-crosslinking. Acta Biomater. 33, 88 95. Available from: https://doi.org/ 10.1016/j.actbio.2016.01.013. Jastrzebska, E., Tomecka, E., Jesion, I., 2016. Heart-on-a-chip based on stem cell biology. Biosens. Bioelectron. 75, 67 81. Available from: https://doi.org/10.1016/j. bios.2015.08.012. Jonsson, M.K.B., Wang, Q.-D., Becker, B., 2011. Impedance-based detection of beating rhythm and proarrhythmic effects of compounds on stem cell-derived cardiomyocytes.

References

Assay Drug Dev. Technol. 9, 589 599. Available from: https://doi.org/10.1089/ adt.2011.0396. Kamei, K., Kato, Y., Hirai, Y., Ito, S., Satoh, J., Oka, A., et al., 2017. Integrated heart/cancer on a chip to reproduce the side effects of anti-cancer drugs in vitro. RSC Adv. 7, 36777 36786. Available from: https://doi.org/10.1039/C7RA07716E. Kang, P.M., Haunstetter, A., Aoki, H., Usheva, A., Izumo, S., 2000. Morphological and molecular characterization of adult cardiomyocyte apoptosis during hypoxia and reoxygenation. Circ. Res. 87, 118 125. Katz, A.M., 2010. Physiology of the Heart, fifth ed. Lippincott Williams & Wilkins, Philadelphia, PA. Kenny, P.A., Lee, G.Y., Myers, C.A., Neve, R.M., Semeiks, J.R., Spellman, P.T., et al., 2007. The morphologies of breast cancer cell lines in three-dimensional assays correlate with their profiles of gene expression. Mol. Oncol. 1, 84 96. Available from: https://doi.org/10.1016/j.molonc.2007.02.004. Kessel, M., Frank, F., 2007. A better prescription for drug-development financing. Nat. Biotechnol. 25, 859 866. Available from: https://doi.org/10.1038/nbt0807-859. Kobuszewska, A., Tomecka, E., Zukowski, K., Jastrzebska, E., Chudy, M., Dybko, A., et al., 2017. Heart-on-a-chip: an investigation of the influence of static and perfusion conditions on cardiac (H9C2) cell proliferation, morphology, and alignment. SLAS Technol. 22, 536 546. Available from: https://doi.org/10.1177/2472630317705610. Lamothe, S.M., Guo, J., Li, W., Yang, T., Zhang, S., 2016. The human ether-a-go-gorelated gene (hERG) potassium channel represents an unusual target for proteasemediated damage. J. Biol. Chem. 291, 20387 20401. Available from: https://doi.org/ 10.1074/jbc.M116.743138. Laverty, H., Benson, C., Cartwright, E., Cross, M., Garland, C., Hammond, T., et al., 2011. How can we improve our understanding of cardiovascular safety liabilities to develop safer medicines? Br. J. Pharmacol. 163, 675 693. Available from: https://doi.org/ 10.1111/j.1476-5381.2011.01255.x. Lee, K.Y., Mooney, D.J., 2012. Alginate: properties and biomedical applications. Prog. Polym. Sci. 37, 106 126. Available from: https://doi.org/10.1016/j.progpolymsci. 2011.06.003. Lind, J.U., Yadid, M., Perkins, I., O’Connor, B.B., Eweje, F., Chantre, C.O., et al., 2017. Cardiac microphysiological devices with flexible thin-film sensors for higherthroughput drug screening. Lab Chip 17, 3692 3703. Available from: https://doi.org/ 10.1039/C7LC00740J. Lindsey, M.L., Bolli, R., Canty, J.M., Du, X.-J., Frangogiannis, N.G., Frantz, S., et al., 2018. Guidelines for experimental models of myocardial ischemia and infarction. Am. J. Physiol. Heart Circ. Physiol. 314, H812 H838. Available from: https://doi.org/ 10.1152/ajpheart.00335.2017. Liu, J., Volkers, L., Jangsangthong, W., Bart, C.I., Engels, M.C., Zhou, G., et al., 2018. Generation and primary characterization of iAM-1, a versatile new line of conditionally immortalized atrial myocytes with preserved cardiomyogenic differentiation capacity. Cardiovasc. Res. 114 (14), 1848 1859. Available from: https://doi.org/10.1093/cvr/ cvy134. Long, X., Boluyt, M.O., Hipolito, M.L., Lundberg, M.S., Zheng, J.S., O’Neill, L., et al., 1997. p53 and the hypoxia-induced apoptosis of cultured neonatal rat cardiac myocytes. J. Clin. Invest. 99, 2635 2643. Available from: https://doi.org/10.1172/JCI119452.

289

290

CHAPTER 8 Heart-on-a-chip

Lv, H., Li, L., Sun, M., Zhang, Y., Chen, L., Rong, Y., et al., 2015. Mechanism of regulation of stem cell differentiation by matrix stiffness. Stem Cell Res. Ther. 6, 103. Available from: https://doi.org/10.1186/s13287-015-0083-4. MacQueen, L.A., Sheehy, S.P., Chantre, C.O., Zimmerman, J.F., Pasqualini, F.S., Liu, X., et al., 2018. A tissue-engineered scale model of the heart ventricle. Nat. Biomed. Eng. 2, 930 941. Available from: https://doi.org/10.1038/s41551-018-0271-5. Maddaford, T.G., Hurtado, C., Sobrattee, S., Czubryt, M.P., Pierce, G.N., 1999. A model of low-flow ischemia and reperfusion in single, beating adult cardiomyocytes. Am. J. Physiol. Circ. Physiol. 277, H788 H798. Available from: https://doi.org/10.1152/ ajpheart.1999.277.2.H788. Malda, J., Visser, J., Melchels, F.P., Ju¨ngst, T., Hennink, W.E., Dhert, W.J.A., et al., 2013. 25th Anniversary article: engineering hydrogels for biofabrication. Adv. Mater. 25, 5011 5028. Available from: https://doi.org/10.1002/adma.201302042. Marsano, A., Conficconi, C., Lemme, M., Occhetta, P., Gaudiello, E., Votta, E., et al., 2016. Beating heart on a chip: a novel microfluidic platform to generate functional 3D cardiac microtissues. Lab Chip 16, 599 610. Available from: https://doi.org/10.1039/ C5LC01356A. Mathur, A., Loskill, P., Shao, K., Huebsch, N., Hong, S., Marcus, S.G., et al., 2015. Human iPSC-based cardiac microphysiological system for drug screening applications. Sci. Rep. 5, 8883. Available from: https://doi.org/10.1038/srep08883. Mauretti, A., Rossi, F., Bax, N.A.M., Miano, C., Miraldi, F., Goumans, M.J., et al., 2017. Spheroid three-dimensional culture enhances Notch signaling in cardiac progenitor cells. MRS Commun. 7, 496 501. Available from: https://doi.org/10.1557/ mrc.2017.82. McKenzie, R., Fried, M.W., Sallie, R., Conjeevaram, H., Di Bisceglie, A.M., Park, Y., et al., 1995. Hepatic failure and lactic acidosis due to fialuridine (FIAU), an investigational nucleoside analogue for chronic hepatitis B. N. Engl. J. Med. 333, 1099 1105. Available from: https://doi.org/10.1056/NEJM199510263331702. Mitcheson, J.S., Hancox, J.C., Levi, A.J., 1998. Cultured adult cardiac myocytes: future applications, culture methods, morphological and electrophysiological properties. Cardiovasc. Res. 39, 280 300. Moore, N., 2016. Lessons from the fatal French study BIA-10-2474. BMJ 353, i2727. Available from: https://doi.org/10.1136/bmj.i2727. Moya, M., Tran, D., George, S.C., 2013. An integrated in vitro model of perfused tumor and cardiac tissue. Stem Cell Res. Ther. 4, S15. Available from: https://doi.org/ 10.1186/scrt376. Nawroth, J.C., Scudder, L.L., Halvorson, R.T., Tresback, J., Ferrier, J.P., Sheehy, S.P., et al., 2018. Automated fabrication of photopatterned gelatin hydrogels for organ-onchips applications. Biofabrication 10, 025004. Available from: https://doi.org/10.1088/ 1758-5090/aa96de. Negoro, S., Kunisada, K., Fujio, Y., Funamoto, M., Darville, M.I., Eizirik, D.L., et al., 2001. Activation of signal transducer and activator of transcription 3 protects cardiomyocytes from hypoxia/reoxygenation-induced oxidative stress through the upregulation of manganese superoxide dismutase. Circulation 104, 979 981. Nguyen, M.-D., Tinney, J.P., Yuan, F., Roussel, T.J., El-Baz, A., Giridharan, G., et al., 2013. Cardiac cell culture model as a left ventricle mimic for cardiac tissue generation. Anal. Chem. 85, 8773 8779. Available from: https://doi.org/10.1021/ac401910d.

References

Nguyen, M.-D., Tinney, J.P., Ye, F., Elnakib, A.A., Yuan, F., El-Baz, A., et al., 2015. Effects of physiologic mechanical stimulation on embryonic chick cardiomyocytes using a microfluidic cardiac cell culture model. Anal. Chem. 87, 2107 2113. Available from: https://doi.org/10.1021/ac503716z. Onakpoya, I.J., Heneghan, C.J., Aronson, J.K., 2016. Post-marketing withdrawal of 462 medicinal products because of adverse drug reactions: a systematic review of the world literature. BMC Med. 14, 10. Available from: https://doi.org/10.1186/s12916016-0553-2. Ong, C.S., Nam, L., Ong, K., Krishnan, A., Huang, C.Y., Fukunishi, T., et al., 2018. 3D and 4D bioprinting of the myocardium: current approaches, challenges, and future prospects. Biomed. Res. Int. 2018, 1 11. Available from: https://doi.org/10.1155/2018/ 6497242. Pati, F., Jang, J., Ha, D.-H., Won Kim, S., Rhie, J.-W., Shim, J.-H., et al., 2014. Printing three-dimensional tissue analogues with decellularized extracellular matrix bioink. Nat. Commun. 5, 3935. Available from: https://doi.org/10.1038/ncomms4935. Pavesi, A., Adriani, G., Rasponi, M., Zervantonakis, I.K., Fiore, G.B., Kamm, R.D., 2015. Controlled electromechanical cell stimulation on-a-chip. Sci. Rep. 5, 11800. Available from: https://doi.org/10.1038/srep11800. PhRMA, 2012. Pharmaceutical Industry Profile 2012. Pharmaceutical Research and Manufacturers of America Pignatti, F., Gravanis, I., Herold, R., Vamvakas, S., Jonsson, B., Marty, M., 2011. The European Medicines Agency: an overview of its mission, responsibilities, and recent initiatives in cancer drug regulation. Clin. Cancer Res. 17, 5220 5225. Available from: https://doi.org/10.1158/1078-0432.CCR-11-0623. Polonchuk, L., Chabria, M., Badi, L., Hoflack, J.-C.C., Figtree, G., Davies, M.J., et al., 2017. Cardiac spheroids as promising in vitro models to study the human heart microenvironment. Sci. Rep. 7, 7005. Available from: https://doi.org/10.1038/s41598-01706385-8. Portal, L., Martin, V., Assaly, R., D’anglemont de Tassigny, A., Michineau, S., Berdeaux, A., et al., 2013. A model of hypoxia-reoxygenation on isolated adult mouse cardiomyocytes. J. Cardiovasc. Pharmacol. Ther. 18, 367 375. Available from: https://doi.org/ 10.1177/1074248412475158. Ralphe, J.C., de Lange, W.J., 2013. 3D engineered cardiac tissue models of human heart disease: learning more from our mice. Trends Cardiovasc. Med. 23, 27 32. Available from: https://doi.org/10.1016/j.tcm.2012.08.007. Ren, L., Liu, W., Wang, Y., Wang, J.-C., Tu, Q., Xu, J., et al., 2013. Investigation of hypoxia-induced myocardial injury dynamics in a tissue interface mimicking microfluidic device. Anal. Chem. 85, 235 244. Available from: https://doi.org/10.1021/ ac3025812. Rimann, M., Graf-Hausner, U., 2012. Synthetic 3D multicellular systems for drug development. Curr. Opin. Biotechnol. 23, 803 809. Available from: https://doi.org/10.1016/j. copbio.2012.01.011. Salameh, A., Wustmann, A., Karl, S., Blanke, K., Apel, D., Rojas-Gomez, D., et al., 2010. Cyclic mechanical stretch induces cardiomyocyte orientation and polarization of the gap junction protein connexin43. Circ. Res. 106 (1592 602). Available from: https:// doi.org/10.1161/CIRCRESAHA.109.214429.

291

292

CHAPTER 8 Heart-on-a-chip

Schaaf, S., Shibamiya, A., Mewe, M., Eder, A., Sto¨hr, A., Hirt, M.N., et al., 2011. Human engineered heart tissue as a versatile tool in basic research and preclinical toxicology. PLoS One 6, e26397. Available from: https://doi.org/10.1371/journal.pone.0026397. Schuppan, D., Schmid, M., Somasundaram, R., Ackermann, R., Ruehl, M., Nakamura, T., et al., 1998. Collagens in the liver extracellular matrix bind hepatocyte growth factor. Gastroenterology 114, 139 152. Sekine, H., Shimizu, T., Hobo, K., Sekiya, S., Yang, J., Yamato, M., et al., 2008. Endothelial cell coculture within tissue-engineered cardiomyocyte sheets enhances neovascularization and improves cardiac function of ischemic hearts. Circulation 118, S145 S152. Available from: https://doi.org/10.1161/CIRCULATIONAHA.107.757286. Senderowicz, A.M., 2010. Information needed to conduct first-in-human oncology trials in the United States: a view from a former FDA medical reviewer. Clin. Cancer Res. 16, 1719 1725. Available from: https://doi.org/10.1158/1078-0432.CCR-09-2766. Serena, E., Figallo, E., Tandon, N., Cannizzaro, C., Gerecht, S., Elvassore, N., et al., 2009. Electrical stimulation of human embryonic stem cells: cardiac differentiation and the generation of reactive oxygen species. Exp. Cell Res. 315, 3611 3619. Available from: https://doi.org/10.1016/j.yexcr.2009.08.015. Shim, J.-H., Lee, J.-S., Kim, J.Y., Cho, D.-W., 2012. Bioprinting of a mechanically enhanced three-dimensional dual cell-laden construct for osteochondral tissue engineering using a multi-head tissue/organ building system. J. Micromech. Microeng. 22, 085014. Available from: https://doi.org/10.1088/0960-1317/22/8/085014. Silbergeld, E.K., Mandrioli, D., Cranor, C.F., 2015. Regulating chemicals: law, science, and the unbearable burdens of regulation. Annu. Rev. Public Health 36, 175 191. Available from: https://doi.org/10.1146/annurev-publhealth-031914-122654. Simian, M., Bissell, M.J., 2017. Organoids: a historical perspective of thinking in three dimensions. J. Cell Biol. 216, 31 40. Available from: https://doi.org/10.1083/jcb.201610056. Suntharalingam, G., Perry, M.R., Ward, S., Brett, S.J., Castello-Cortes, A., Brunner, M.D., et al., 2006. Cytokine storm in a phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N. Engl. J. Med. 355, 1018 1028. Available from: https://doi.org/10.1056/ NEJMoa063842. Takahashi, T., 2019. Organoids for drug discovery and personalized medicine. Annu. Rev. Pharmacol. Toxicol. 59, 447 462. Available from: https://doi.org/10.1146/annurevpharmtox-010818-021108. Takeda, M., Miyagawa, S., Fukushima, S., Saito, A., Ito, E., Harada, A., et al., 2018. Development of in vitro drug-induced cardiotoxicity assay by using three-dimensional cardiac tissues derived from human induced pluripotent stem cells. Tissue Eng., C: Methods 24, 56 67. Available from: https://doi.org/10.1089/ten.TEC.2017.0247. Tanaka, M., Ito, H., Adachi, S., Akimoto, H., Nishikawa, T., Kasajima, T., et al., 1994. Hypoxia induces apoptosis with enhanced expression of Fas antigen messenger RNA in cultured neonatal rat cardiomyocytes. Circ. Res. 75, 426 433. Tandon, N., Marsano, A., Maidhof, R., Numata, K., Montouri-Sorrentino, C., Cannizzaro, C., et al., 2010. Surface-patterned electrode bioreactor for electrical stimulation. Lab Chip 10, 692 700. Available from: https://doi.org/10.1039/b917743d. Thavandiran, N., Dubois, N., Mikryukov, A., Masse, S., Beca, B., Simmons, C.A., et al., 2013. Design and formulation of functional pluripotent stem cell-derived cardiac microtissues. Proc. Natl. Acad. Sci. U.S.A. 110, E4698 E4707. Available from: https://doi. org/10.1073/pnas.1311120110.

References

Thomas, D., Karle, C.A., Kiehn, J., 2006. The cardiac hERG/IKr potassium channel as pharmacological target: structure, function, regulation, and clinical applications. Curr. Pharm. Des. 12, 2271 2283. Tofovic, S.P., Kost, C.K., Jackson, E.K., Bastacky, S.I., 2002. Long-term caffeine consumption exacerbates renal failure in obese, diabetic, ZSF1 (fa-fa(cp)) rats. Kidney Int. 61, 1433 1444. Available from: https://doi.org/10.1046/j.1523-1755.2002.00278.x. Trepat, X., Deng, L., An, S.S., Navajas, D., Tschumperlin, D.J., Gerthoffer, W.T., et al., 2007. Universal physical responses to stretch in the living cell. Nature 447, 592 595. Available from: https://doi.org/10.1038/nature05824. Turnbull, I.C., Karakikes, I., Serrao, G.W., Backeris, P., Lee, J.-J., Xie, C., et al., 2014. Advancing functional engineered cardiac tissues toward a preclinical model of human myocardium. FASEB J. 28, 644 654. Available from: https://doi.org/10.1096/fj.13228007. Ugolini, G.S., Rasponi, M., Pavesi, A., Santoro, R., Kamm, R., Fiore, G.B., et al., 2016. On-chip assessment of human primary cardiac fibroblasts proliferative responses to uniaxial cyclic mechanical strain. Biotechnol. Bioeng. 113, 859 869. Available from: https://doi.org/10.1002/bit.25847. Wan, C., Chung, S., Kamm, R.D., 2011. Differentiation of embryonic stem cells into cardiomyocytes in a compliant microfluidic system. Ann. Biomed. Eng. 39, 1840 1847. Available from: https://doi.org/10.1007/s10439-011-0275-8. Xiao, Y., Zhang, B., Liu, H., Miklas, J.W., Gagliardi, M., Pahnke, A., et al., 2014. Microfabricated perfusable cardiac biowire: a platform that mimics native cardiac bundle. Lab Chip 14, 869 882. Available from: https://doi.org/10.1039/C3LC51123E. Yamada, K.M., Cukierman, E., 2007. Modeling tissue morphogenesis and cancer in 3D. Cell 130, 601 610. Available from: https://doi.org/10.1016/j.cell.2007.08.006. Zervantonakis, I.K., Hughes-Alford, S.K., Charest, J.L., Condeelis, J.S., Gertler, F.B., Kamm, R.D., 2012. Three-dimensional microfluidic model for tumor cell intravasation and endothelial barrier function. Proc. Natl. Acad. Sci. U.S.A. 109, 13515 13520. Available from: https://doi.org/10.1073/pnas.1210182109. Zhuang, J., Yamada, K.A., Saffitz, J.E., Kle´ber, A.G., 2000. Pulsatile stretch remodels cellto-cell communication in cultured myocytes. Circ. Res. 87, 316 322.

293

CHAPTER

Caenorhabditis elegans-ona-chip: microfluidic platforms for high-resolution imaging and phenotyping

11

Sudip Mondal and Adela Ben-Yakar Department of Mechanical Engineering, The University of Texas at Austin, Austin, TX, United States

Introduction Major market segments, including the pharmaceutical, manufacturing, cosmetics, agrochemistry, petroleum, food, and apparel industries, seek methods to assess the toxicity of their products and raw materials (NC3R, 2007; Crawford et al., 2017; Maertens et al., 2014; Maertens and Hartung, 2018; Collins et al., 2008; Hartung, 2010). Exposure to toxicants in the environment and consumer products can be life-threatening and lead to adverse health effects, and many industries are moving to characterize their raw materials with complete toxicity profiles (e.g., developmental and reproductive toxicity, genotoxicity, irritation, inflammatory response, neuronal toxicity) (Crawford et al., 2017). The new mandates and regulations on animal use have led government agencies and industry partners to implement initiatives such as ToxCast, Tox21, Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH), Institute for Health and Consumer Protection guidelines, and green chemistry (Crawford et al., 2017; Boyd et al., 2016; Dix et al., 2007). One of the main goals of these initiatives is to fully characterize all chemicals on the market. Thousands of chemicals have been listed and await toxicity profiling as relevant models and new technologies are becoming available. US government agencies have recommended the testing of all compounds in as many as 15 concentrations, generally ranging from approximately 5 nM 100 μM, to generate a dose response curve that would significantly reduce the false-positive and false-negative rates observed in traditional screening methods (Collins et al., 2008; Inglese et al., 2006). To protect individuals and the environment, many industries follow strict standards in profiling chemicals for informed decisionmaking regarding their continued use. In drug discovery, abandonment of a candidate compound at the preclinical stage is mainly attributed to specific organ

Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00009-7 © 2020 Elsevier Inc. All rights reserved.

363

364

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

toxicities, with the most common causes being cardiovascular toxicity and adverse effects on the central nervous system (Cook et al., 2014; Guengerich, 2011).

Limitations of current toxicity screening models Compound attrition rate in the pharmaceutical industry during the early clinical phase has been estimated at 40% and is commonly attributed to organ-specific toxicity (Guengerich, 2011). This is likely because the reliable identification of organ-specific toxicity in nonclinical models during the earlier stages of the drug-development process has been challenging. Cellular models are inexpensive, efficient, and ethically compatible preliminary screening alternatives to using large-animal models (Pridgeon et al., 2018; Almeida et al., 2017). However, there is a notable translational gap between these cellular models and animal models (Clark et al., 1999; Roberts et al., 2015). To bridge this gap, researchers are turning to advanced three-dimensional (3D) in vitro models such as organon-a-chip devices, spheroids, and organoids, as well as humanized small-animal models, such as the nematode Caenorhabditis elegans, Drosophila, and zebrafish, to elucidate toxicity mechanisms more effectively than in vitro twodimensional (2D) cultures (Moser, 2007; Chen et al., 2015; Meyer and Williams, 2014; De Simone et al., 2018; Hartley and Brennand, 2017; Noyes et al., 2016). Incorporation of humanized models and multiorgan platforms (Cho and Yoon, 2017; Kimura et al., 2018; Alepee et al., 2014; Ishida, 2018; Selimovic et al., 2013; Bovard et al., 2018; Forsythe et al., 2018) into the early drug discovery pipeline can improve the attrition rate and provide safer drug candidates for clinical trials. Neurotoxicology is one of the most challenging organ-specific toxicity forms because it requires system-level studies to capture complex interneuronal interactions in an age-dependent manner (Harry et al., 1998; Dorea, 2011; Coecke et al., 2006). Current screening methods for neurotoxicity involve in vitro culture models of neurons, neuro-spheres, and brain organoids (Miranda et al., 2018; Jensen and Parmar, 2006; Dekkers et al., 2019; Pasca et al., 2015; Birey et al., 2017). While using human-induced pluripotent stem cells for organoids and organ-on-achip models provides distinct advantages, these cellular platforms cannot replicate the full complexity of the nervous system, and no current in vitro model provides a system-level neurotoxicity profile similar to those measured in animal models (Boyd et al., 2016; Noyes et al., 2016; Boyd et al., 2010).

Small animal models: Caenorhabditis elegans To study toxicity in vivo without ethical concerns while enabling system-level testing, researchers have turned to small-animal models such as C. elegans, Drosophila, and zebrafish. C. elegans is the only small-animal model that can be grown in large quantities over a few days for large-scale screens and uniquely suited for high-throughput studies thanks to its short lifespan, transparent body,

Introduction

small size, and facile genetic manipulations. It is a simple yet complex model with differentiated cells and organs, well-defined lifespan, and a completely mapped nervous system. The conserved cellular pathways, 60% gene homology with humans, and established gene-editing tools make C. elegans an ideal model for system-level toxicity studies. With C. elegans, researchers can characterize toxicity mechanisms across multiple levels of biological organization, from gene receptor interactions to cellular processes to phenotypic alterations such as developmental, reproductive, and neuronal defects (Meyer and Williams, 2014; Kaletta and Hengartner, 2006; Avila et al., 2012; Ruszkiewicz et al., 2018; Leung et al., 2008; Hunt, 2017; Giacomotto and Segalat, 2010). Various mutant and transgenic C. elegans strains have been developed by toxicologists to model the effects of toxicity and delineate the underlying mechanistic pathways (Boyd et al., 2016; Li et al., 2019; Ju et al., 2014; Koch et al., 2014; Caito et al., 2013; McVey et al., 2016; Harlow et al., 2016, 2018; Meyer et al., 2016; Negga et al., 2011, 2012). While C. elegans has been used for in vivo toxicity profiling, the lack of highcontent screening technologies and automated phenotyping precluded the use of C. elegans in high-throughput screening. One of the major obstacles to adapting this model into high-throughput screening is the requirement to immobilize a large number of the animals in a rapid manner and image them at high resolutions for high-content phenotyping. In current high-throughput studies, C. elegans are commonly immobilized using anesthetics, and their gross phenotypes are analyzed in conventional multiwell plates using low-resolution imaging methods (Xiong et al., 2017; Wahlby et al., 2012; Leung et al., 2011, 2013; Gosai et al., 2010; Allard et al., 2013; Dupuy et al., 2007). While these early large-scale screens are encouraging, there is an urgent need for new methods with higher resolution imaging capabilities to capture the cellular and subcellular phenotypes necessary for multiparametric toxicology profiling (Alexander et al., 2014; Mondal et al., 2018; Caceres Ide et al., 2012; Keil et al., 2017; Ben-Yakar, 2019).

Conception of Caenorhabditis elegans-on-a-chip Microfluidics, the science of fluid flow to manipulate objects in micron-sized devices using the forces generated by the flow, is well-positioned to enable anesthetic-free immobilization of C. elegans for high-resolution, high-throughput imaging. C. elegans (approximately 1 mm in length and 50 μm in diameter) fit within the microfluidic devices and can be manipulated by the fluid principles. A suitable microenvironment for the nematode can be created on the chips with standard animal culture conditions (Ben-Yakar, 2019; Ben-Yakar et al., 2009; Cornaglia et al., 2017; Bakhtina and Korvink, 2014). Moreover, microfluidics can provide optical access for high-resolution imaging of the immobilized animals and is compatible with automation for rapid high-content phenotyping. Implementation of microfluidic technologies in C. elegans studies has produced novel devices with precise immobilization capabilities for high-resolution

365

366

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

imaging (Mondal et al., 2011, 2012, 2016, 2018; Caceres Ide et al., 2012; Keil et al., 2017; Bakhtina and Korvink, 2014; Chokshi et al., 2009; Chronis et al., 2007; Guo et al., 2008; Hulme et al., 2007; Krajniak et al., 2013; Krajniak and Lu, 2010), laser axotomy (Guo et al., 2008; Chung and Lu, 2009; Go¨kc¸e et al., 2014; Gokce et al., 2017), phenotyping (Mondal et al., 2016, 2018; Chung et al., 2008; Letizia et al., 2018; San-Miguel et al., 2016), sorting (Chung et al., 2008), mutant screening (Lee et al., 2013; Crane et al., 2012), electrophysiology recordings (Lockery et al., 2012), and behavioral studies (Churgin et al., 2017; Hulme et al., 2010). Microfluidics has enabled several key discoveries and revolutionized the field of C. elegans genetics and disease modeling through high-throughput and high-content screening (Mondal et al., 2018; Cornaglia et al., 2017; Mondal et al., 2016), large-scale mutant screening (Crane et al., 2012), automated aging studies (Cornaglia et al., 2015; Xian et al., 2013), and long-term developmental studies (Keil et al., 2017). This chapter describes microfluidics-based platforms for screening C. elegans to address questions in fundamental biology and to study the efficacy of chemical compounds in human diseases models.

Caenorhabditis elegans as a whole-animal model in scientific research This soil-dwelling nematode was first described in 1900 and was proposed as a model organism by Sydney Brenner in 1965 (Brenner, 1974). It has since been widely adopted as a model system in genetics, reproductive, and neurobiology. In laboratories, the nematode is fed Escherichia coli bacteria on solid nematode growth medium or in liquid (Stiernagle, 2006). The Caenorhabditis Genetics Center at the University of Minnesota (https://cgc.umn.edu/) curates, maintains, and distributes C. elegans strains for research and development. Strains are grown following standardized culture conditions to achieve consistent phenotypes at 20  C. The short lifecycle (approximately 21 days in total) includes four larval stages, and the nematodes become young adults in 3 days, at which time they begin to lay eggs. A nematode can produce approximately 300 eggs within its gravid period (approximately three further days), then continue to grow and, from around day 8 onward, exhibit signs of aging such as reduced motility, neuronal degeneration, and increased rates of bacterial infection (Pincus and Slack, 2010; Son et al., 2019; Chow et al., 2006). C. elegans was the first animal to have its genome sequenced in full (C. elegans Sequencing Consortium, 1998). Fluorescent probes enable the monitoring of cells and proteins of interest in vivo without the need to dissect the animals. C. elegans has six chromosomes and approximately 20,000 protein-coding genes that are expressed over 959 cells in an adult hermaphrodite. The nematode has been used in biology to study development, growth, maturation, regeneration,

Caenorhabditis elegans as a whole-animal model in scientific research

degeneration, aging, and disease pathologies. While many human disease related genes are present in the nematode’s genome, disease-associated human genes of interest can easily be expressed in the nematode by the insertion of the wild-type copy or mutant genes. In animals with larger genomes and multiple copies, elucidating gene function can be challenging, but a single copy in the nematode is simple to modify in loss-of-function or gain-of-function mutants. Nematodes have been instrumental in identifying genetic mechanisms of diseases (Baskoylu et al., 2018; Sahn et al., 2017).

Advantages of Caenorhabditis elegans as a model for neuroscience and neurotoxicity An adult hermaphrodite C. elegans has 302 neurons and 56 glial cells (Hobert, 2010). All 118 morphologically distinct neuron classes and 7,000 connections have been identified and mapped (Richmond, 2005). C. elegans shares neurotransmitters such as acetylcholine, dopamine, serotonin, γ-aminobutyric acid, and glutamate with humans. The nervous system possesses sensory neurons, interneurons, polymodal neurons, and motor neurons (Bargmann, 2012). Researchers can study neuronal circuits using chemical, thermal, food, and olfactory stimuli and can visualize individual cells or subcellular components in intact animals. High-resolution imaging tools enable the tracking of morphological changes of the nervous system at various developmental stages, advancing our understanding of maturation, aging, and age-related conditions (Keil et al., 2017; Melentijevic et al., 2017). C. elegans uses sensillal and nonsensillal receptors to sense, process, and navigate thermal, chemical, and oxygen gradients. Microfluidic platforms have been developed that can control surrounding environmental cues and stimuli to study nematode neuromuscular mechanics. The approaches to studying neuronal circuit behavior have included chemical olfactory exposure, optogenetic stimulation, chemical or thermal gradient changes, mechanical stimulation, and pressure delivery (Sengupta and Samuel, 2009; Fang-Yen et al., 2015). The forces delivered by individual animals can be quantified using a piezoelectric pillar or deflected cantilever measurements from nematode locomotion (Doll et al., 2009). Deviations from the measured values owing to a mutation or disease protein can thus be attributed to motor defects (Rahman et al., 2018). Neuropathologies similar to those in humans can develop in C. elegans when the nematodes are exposed to environmental toxicants and agrochemical products (Meyer and Williams, 2014; Avila et al., 2012; Leung et al., 2008; Harlow et al., 2016). Significant overlap in the activity of chemicals has also been found between C. elegans and zebrafish (Boyd et al., 2010, 2016). C. elegans disease models can be developed by exposure to environmental neurotoxicants or by overexpressing toxic proteins (Braungart et al., 2004; Nass et al., 2002; Zhou et al., 2013). Dopaminergic and serotonergic neurons are of specific interest because they are involved in a broad range of physiological

367

368

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

functions, including egg-laying, pharyngeal pumping, defecation, and locomotion. When exposed to 1-methyl-4-phenylpyridinium, for example, dopaminergic neurons degenerate in a concentration-dependent manner with age (Pu and Le, 2008). This degeneration and the subsequent induced mobility defects are dependent on the dopamine transporter gene and occur through mitochondrial transport chain I pathways, similar to the mechanisms reported for human Parkinson’s disease like symptoms (Pu and Le, 2008; Wang et al., 2007). Neurotoxicants such as 6hydroxy dopamine and rotenone have been shown to cause a significant reduction in speed (Nass et al., 2002; Zhou et al., 2013; Offenburger et al., 2018; Chikka et al., 2016). However, neurotoxic effects may be suppressed by pretreating nematodes with acetaminophen, which is known to be neuroprotective (Dexter et al., 2012; Salam et al., 2013). Proteins such as α-synuclein or multiple repeats of glutamate are toxic to C. elegans cells and cause cellular degeneration over time (Bodhicharla et al., 2012; Morley et al., 2002; Faber et al., 1999; Brignull et al., 2006). The protein aggregation and cellular degeneration phenotypes (Fig. 11.1) have been used to identify genetic interactors in the search for therapeutic targets and disease mechanisms in Huntington’s disease and Parkinson’s disease (van Ham et al., 2008; Nollen et al., 2004).

Challenges in using Caenorhabditis elegans for large-scale studies Large-scale culture and maintenance protocols have been developed for growing the animals in both solid media and liquid culture. The liquid culture is preferable

FIGURE 11.1 Caenorhabditis elegans neurodegeneration models. (A) Morphological changes in C. elegans dopaminergic neurons expressing green fluorescent protein in healthy (top) and degenerated (bottom) states. The degeneration was induced by the chemical treatment with a neurotoxin. (B) Length-dependent polyglutamine aggregation models in C. elegans. Scale bar: 25 μm (A) and 100 μm (B). Adapted from (A) Nass, R., et al., 2002. Neurotoxin-induced degeneration of dopamine neurons in Caenorhabditis elegans. Proc. Natl. Acad. Sci. U.S.A. 99 (5), 3264 3269; (B) Morley, J.F., et al., 2002. The threshold for polyglutamine-expansion protein aggregation and cellular toxicity is dynamic and influenced by aging in Caenorhabditis elegans. Proc. Natl. Acad. Sci. U.S.A. 99 (16), 10417 10422.

Caenorhabditis elegans as a whole-animal model in scientific research

for easy exposure, transfer, distribution, and sorting of the nematodes. Chemical treatment of C. elegans on solid media is uneven, and the stability of the chemical cannot be controlled over time. In liquid culture on multiwell plates the nematodes remain in a 3D environment and are exposed to a constant chemical concentration. Liquid culture methodologies have been leveraged to screen C. elegans using fluorometric measurements (Leung et al., 2011; Rodriguez-Palero et al., 2018) and fluorescence imaging (Wahlby et al., 2012; Gosai et al., 2010), as well as in food-clearing (Gomez-Amaro et al., 2015), mortality (Lucanic et al., 2018; Solis and Petrascheck, 2011), and swimming (Hunt, 2017) assays. Advancements in automation and integration with robotic liquid handling systems have enabled large-scale C. elegans screens in standard multiwell plates (Gosai et al., 2010). Biological processes such as growth, tissue development, cell viability, and autophagy can be assessed using transgenic C. elegans in multiwell plates with fluorescence microscopy and image acquisition and data analysis modules (Gosai et al., 2010). Although current technologies provide insights into disease mechanisms and chemical modulators, their optical interrogation platforms have low resolution and can identify only gross phenotypes and from a few animals per population. High-resolution imaging and fully automated image analysis would be required to score cellular and subcellular phenotypes and enable ultrafast screening of C. elegans strains with subtle phenotypes. As such, the development of new tools to unlock the full potential of nematode-based models, such as C. elegans-on-a-chip, is the focus of extensive current investigation.

Recent advances in the development of Caenorhabditis elegans-ona-chip Microfluidics is the science of using the forces generated by fluid flow to manipulate objects in micron-sized devices; it has been applied in the areas of life sciences, physics, electronics, sensors, aerospace, and mechanics. New microfluidics-based devices may achieve higher speeds and improved functionality compared with current technology, and combining these platforms with improved imaging modalities will enable greater specificity and could revolutionize high-throughput and high-content screening of C. elegans. There is considerable interest in this field of research, and the number of publications in the field of C. elegans biology and microfluidic technologies has greatly increased over the past two decades (Fig. 11.2). C. elegans-based research outcomes have also been reported in major international journals and conferences. Microfluidic platforms are commercially available for single-cell sequencing (Gong et al., 2018; Xin et al., 2016), cell counting (Dekker et al., 2018), electroporation (Geng and Lu, 2013), and electric measurements (Ionescu-Zanetti et al., 2005; Lau et al., 2006). With advancements in more complex cellular and 3D models, device architecture must be upgraded with compatible electrical, optical, or genetic tools to enable the use of microfluidic platforms with higher order

369

370

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

FIGURE 11.2 PubMed Central (PMC) publication list with “C. elegans,” “microfluidics,” and “C. elegans AND microfluidics” keywords. The search was conducted for publications indexed between January 2000 and December 2018. The number of publications displayed is per calendar year.

model systems. Research groups in academia and industry have already adopted micron-sized platforms that provide a suitable microenvironment for cellular and tissue samples (Kim et al., 2012, 2015; Sontheimer-Phelps et al., 2019; Gupta et al., 2016), but there are limited resources for platforms that are suitable for whole organisms. Electrical signals from cellular samples or C. elegans neurons can provide information on resistance, capacitance, and inductance, which can then be used to identify the health of the samples. Although various large-scale chips have been fabricated for large 3D cellular models, such as spheroids, no such technology exists to date for fabricating chips with only a few tens of micron size channels that can house small model organisms. Microfluidic systems have been shown to have unprecedented precision and sensitivity for a diverse array of biological applications, including single-cell analysis (Reece et al., 2016), enzymatic assays (Hadd et al., 1997; Lagally et al., 2001), polymerase chain reaction (Lagally et al., 2001; Khandurina et al., 2000), and circulating tumor detection (Cho et al., 2018; Khamenehfar and Li, 2016). Microfluidic technologies also use more complex structures such as multilayer cellular architecture, multicellular networks, organoids, multiorgan platforms, and organism-on-a-chip systems. Miniaturization reduces sample volume while adding additional functionality; robotic fluidic workstations are integrated with biological workflows to address research needs while reducing labor and consumable expenses, optimizing space, and scaling up data-rich approaches to quantitative analysis. These automated platforms meet the current industry standards for in vitro research while providing greater value with in vivo biology.

The use of microfluidic platforms for Caenorhabditis elegans research One of the major challenges of high-throughput assays is to integrate a workflow that relies on multiwell formats. This can be accomplished with automated liquid

The use of microfluidic platforms for Caenorhabditis elegans research

handling systems; however, although microfluidic technology enables fast and automated control of chemical and biological samples, simple interfaces between the microfluidic platforms and standard laboratory equipment are lacking. Multiple interfaces are required to establish highly cumbersome connections of thin tubes to the chip for buffer flow and connections of pneumatic pumps for the operation of on-chip valves using individually addressable syringe or pressure reservoirs. Because of these limitations, research laboratories are directing their focus instead on constructing tools in multiwell formats (Mondal et al., 2016; Churgin et al., 2017; Ghorashian et al., 2013). C. elegans research in particular has benefited greatly by the development of microfluidic technologies that allow precise flow control in delivering chemical stimuli and simple animal manipulation, which are otherwise labor-intensive and time-consuming. Microfluidic devices also ameliorate the effects of anesthetics and glues in developmental and long-term studies that require the nematodes to be immobilized (Chokshi et al., 2009; Guo et al., 2008; Mondal et al., 2011). Anesthetic agents may exert adverse effects on in vivo neuronal transport by halting transport in the early larval stages and reducing the flux of vesicles in older animals (Mondal et al., 2011). Microfluidic immobilization is a more physiologically relevant assay for in vivo measurements; nematodes immobilized under flexible membranes are kept stationary without affecting cellular functionality. Similar techniques were used to demonstrate transport parameters in Drosophila and zebrafish larvae (Mondal et al., 2011, 2012; Mondal and Koushika, 2014). The next sections discuss microfluidic technologies that have been developed and used specifically for C. elegans studies.

Serial versus parallel platforms There are multiple microfluidic platforms for C. elegans research, which vary widely by features and capabilities. These can be classified into two major categories: serial and parallel platforms. In a serial microfluidic platform (Fig. 11.3) a single population of C. elegans is loaded at the entrance of the microfluidic chip with up to a few hundred animals at a time. After the initial loading, a single animal is pushed through the chip by buffer flow to be positioned inside the microfluidic chip. Positive pressure at the chip entrance drives the flow of the buffer toward the exit that guides the animals into the imaging region for high-resolution optical interrogation. Alternately, vacuum suction is applied at the chip exit while connecting a buffer reservoir at the entrance and pulling buffer, which in turn pulls the animals through the micron-sized channels. The flow of the animals through a narrow channel is monitored by a high-speed camera or geometrical structures where the animals are held for high-resolution imaging. The process is repeated serially to capture, image, analyze, and sort multiple animals for phenotyping. In parallel microfluidic devices (Fig. 11.4), C. elegans from multiple populations are loaded onto a single chip simultaneously, and multiple animals from

371

372

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

FIGURE 11.3 Schematic showing approaches for immobilizing Caenorhabditis elegans nematodes serially in microfluidic devices. Four major classes of serial microfluidic chips have been used for studying regeneration (Guo et al., 2008), sorting (Chung et al., 2008), phenotyping (Crane et al., 2012), and stimulation and calcium imaging (Chalasani et al., 2007). The microfluidic chips are shown on the left; results are shown on the right.

each population are analyzed simultaneously or serially. However, few platforms exist to handle multiple nematodes in parallel. The majority of microfluidic devices implement complex on-chip valves, multiple input ports, complex microfluidic channel geometries, and large chip dimensions that are impossible to scale up and develop into a multiwell device. To develop a high-throughput platform and meet industry standards, the side-by-side format must be achieved with a defined well-to-well spacing, specific plate footprint, simple operation, and the option to integrate with existing automated liquid handling systems. Our laboratory previously developed a multiwell platform designed in a 96-well format and demonstrated the platform’s applicability for high-throughput C. elegans screening (Mondal et al., 2016).

The use of microfluidic platforms for Caenorhabditis elegans research

FIGURE 11.4 Schematic showing approaches for immobilizing Caenorhabditis elegans nematodes in parallel in microfluidic devices. Four major classes of the serial chips are presented. Parallel chips have been used for studying aging (Hulme et al., 2010), developmental biology (Keil et al., 2017), regeneration (Gokce et al., 2017), and phenotypic screening (Mondal et al., 2016). The microfluidic chips are shown on the left; results are shown on the right.

First-generation serial microfluidic platforms Most of the first-generation microfluidic devices developed for C. elegans studies were serial platforms. In these devices, one or more nematodes were flowed serially into the interrogation area for optical, electrical, mechanical, and chemical analysis. These devices have contributed to our knowledge in areas such as longterm disease progression, in vivo nerve regeneration, and aging research in live animals. The assays typically require manipulation of the animals without affecting their physiology. In such platforms, animals are analyzed one at a time and oriented specifically at one location for optical interrogation in a repeatable

373

374

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

manner. One of the advantages of microfluidic platforms is the use of a thin glass substrate at the bottom of the chip, enabling high-numerical-aperture objectives for good optical access and high-efficiency data collection. One of the early microfluidic devices used tapered channels to immobilize C. elegans and squeeze the animals inside a narrow channel (Chronis et al., 2007). This device included a tapered channel to hold a single animal with the tip of its head protruding out of the channel into a flow channel to sense the chemical flowing through the flow channel. Multiple inputs were connected to the flow channel, and the flow was switched in a synchronized manner to expose the nematode nose to various chemicals. The chemicals were switched rapidly, and the nematode’s response could be monitored optically as output from fluorescencetagged sensory neurons (Fig. 11.3). The platform, termed “olfactory chip,” delivered and altered specific stimuli and recorded the sensory neuron, interneuron, and behavioral response using an engineered microenvironment.

Second-generation serial microfluidic platforms The second generation of microfluidic devices used methods such as a flexible membrane to mechanically trap and manipulate single nematodes (Guo et al., 2008; Wang and Jin, 2011; Rohde et al., 2007; Zeng et al., 2008), small suction channels to capture single animals from a population (Zeng et al., 2008), the application of carbon dioxide (Chokshi et al., 2009), electrical fields to induce electrotaxis (Chuang et al., 2011; Rezai et al., 2012; Tong et al., 2013), cold fluid to induce temporary paralysis (Chung et al., 2008; Crane et al., 2012), and surface acoustic wave perturbation (Ding et al., 2012). C. elegans have been immobilized by a flexible membrane for axotomy using femtosecond laser ablation of individual mechanosensory neurons (Guo et al., 2008) (Fig. 11.3). The work demonstrated a fast recovery of injured axons in the absence of anesthetics; axonal regrowth occurred within 60 90 minutes, unlike the 6 9 hours required by conventional anesthetic assays. Regeneration was confirmed through gain of behavioral responses postaxotomy. Another research group developed a similar microfluidic device for laser axotomy in C. elegans and imaged axonal regeneration using two-photon microscopy (Rohde et al., 2007; Zeng et al., 2008; Samara et al., 2010). They later used this platform to screen chemicals as potential modulators of regeneration (Samara et al., 2010), incorporating micromanipulators to deliver animals from a standard multiwell plate into the microfluidic chip. The serial microfluidic chips require only one imaging location, into which the animals are brought serially. After imaging, the animals can then be discarded or transported to another chip region for sorting or storage. These platforms enable the serial processing of hundreds of C. elegans nematodes for phenotyping. Chung et al. (2008) developed an automated platform to load, image, analyze, and sort individual C. elegans using fluorescence-based morphology and intensity readouts. The platform used a low-temperature stage to immobilize the animals during imaging and identify cellular and subcellular features, screening

The use of microfluidic platforms for Caenorhabditis elegans research

approximately 400 nematodes per hour (Fig. 11.3). Unlike manual phenotyping and screening experiments that require several months to complete, these microfluidic platforms achieve the screens within a few days and with minimal human intervention. However, challenges associated with delivering multiple populations serially into the chip have precluded the use of serial devices in large-scale, highthroughput studies.

Parallel microfluidic platforms Parallelization of trapping channels is another route to high-throughput studies of C. elegans nematodes on-chip. Tens to hundreds of channels for immobilization can be arranged in a parallel fluidic circuit, to load and house multiple nematodes from a single population in individual trapping channels/chambers in a single device. This approach allows for simultaneous imaging of multiple animals with lower magnification objectives. Being primarily single-layer devices, which are fabricated using a single mold, they are simple to fabricate and operate (Hulme et al., 2007; Chronis, 2010). The addition of on-chip valves can further improve chip performance by enabling more precise sample manipulation and positioning for automated imaging (Guo et al., 2008; Chung and Lu, 2009; Go¨kc¸e et al., 2014; Chung et al., 2008; Samara et al., 2010). A few devices with a parallel configuration have been developed to image a large number of animals and detect small changes in cellular processes. Hulme et al. (2007) pioneered a device consisting of straight channels with tapered geometry to immobilize C. elegans delivered from a common input port. The single-layer device was designed as a branched network with up to seven branchings to create 128 traps and hold C. elegans using channel widths from 100 μm down to 10 μm (Fig. 11.5). A similar design was adopted to immobilize and house several nematodes for developmental analysis during the entire lifespan (Hulme et al., 2010). As an individual animal is immobilized in the trap, it blocks the flow of buffer through that particular channel, pushing other nematodes into other channels. However, this platform required 15 20 minutes of loading time, and the tapering channel design resulted in a large variation in the aspect ratio (width to height) along the length of the channel. In such configurations, we found that the animals tend to rotate themselves along their sides as the aspect ratio becomes smaller than one (Ghorashian, 2010). Further innovations included the development of a similar tapered chip to immobilize nematodes for long-term imaging and single-synapse ablation (Allen et al., 2008) and incorporation of electrodes to the underside of the device to measure pharyngeal pumping rates in several nematodes simultaneously following exposure to antiparasitic compounds (Lockery et al., 2012; Weeks et al., 2016). Gokce et al. (2017) subsequently developed a multifunction device for nerve regeneration studies, with functionalities such as axotomy, postsurgery housing for recovery, and postrecovery imaging on one microfluidic chip. The chip features 20 immobilization channels with gradually reduced heights in addition to

375

376

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

FIGURE 11.5 Parallel microfluidic chips for parallel or serial processing of Caenorhabditis elegans nematodes. (A) A single-layer device with tapered channel geometry. The traps are arranged in parallel to immobilize several nematodes for imaging studies. (B) Image of a dye-filled multifunction chip with 20 immobilization channels. The inset on the right shows a closeup image of all immobilization channels filled with nematodes at an applied pressure of approximately 65 kPa. Adapted from (A) Hulme, S.E., et al., 2007. A microfabricated array of clamps for immobilizing and imaging C. elegans. Lab Chip 7 (11), 1515 1523; (B) Gokce, S.K., et al., 2017. A multi-trap microfluidic chip enabling longitudinal studies of nerve regeneration in Caenorhabditis elegans. Sci. Rep. 7 (1), 9837.

the tapering channel widths to preserve the aspect ratio around 1.0; this geometry maintains the animals in lateral orientation and enables precise focusing of a femtosecond laser beam for on-chip laser axotomy and optical measurement of neuronal regeneration of up to 20 animals simultaneously. Research into the biology of aging, for example, requires low-resolution timelapse imaging of multiple C. elegans nematodes from a single population to quantify lifespan. Platforms with isolated chambers have been fabricated to house and track multiple animals over time (Letizia et al., 2018; Churgin et al., 2017; Xian et al., 2013). Parallel devices are useful for rapid, high-volume imaging of hundreds of animals within individual channels.

Disadvantages of current platforms Although most serial and parallel devices are efficient in providing phenotypic information from a large number of animals from a single population, these devices cannot be scaled up for large-scale platforms. While single-layer devices are simple to fabricate, they lack the ability to control animal orientation and have relatively slow loading configurations. In contrast the fabrication of devices with two-layer geometry and pressurized on-chip valves is complex, preventing their use in large-scale applications. Currently, with single-copy nematode strains that exhibit subtle fluorescence signal changes, researchers require a large number of animals from each population for analysis. In the following section, we describe a novel microfluidic platform that enables high-throughput and

Large-scale microfluidics for phenotyping multiple populations

high-content imaging of several thousands of C. elegans nematodes from 96 populations at high resolution to identify alterations in the fluorescence signal.

Large-scale microfluidics for phenotyping multiple populations of Caenorhabditis elegans nematodes The need to phenotype nematodes following chemical exposures relative to control populations requires the development of large-scale microfluidic devices that can process multiple populations and large numbers of animals. Such large-scale devices must be easy to operate (without cumbersome interfaces such as multiple inputs and complex tubing manifolds), compatible with automated platforms such as liquid handling and robotics systems with a standard side-by-side footprint, provide a sufficient number of samples per test in a limited space, and compatible with high-resolution imaging (proximity to the imaging optics and optically favorable orientation).

Development of the vivoChip To develop the first large-scale microfluidic platform that overcomes the challenges mentioned before, our laboratory at the University of Texas at Austin designed a hybrid device whose exterior has a standard multiwell format for interfacing with commercial automated liquid handling platforms and whose interior consists of microfabricated channels that connect directly to the wells to handle a large number of animals (Fig. 11.6). The vivoChip was specifically designed with 96 on-chip wells and 40 parallel trapping channels per well (Fig. 11.6A and B) (Mondal et al., 2016). The microfluidic immobilization channels have a tapered geometry and multiple heights to keep an aspect ratio of approximately 1.0 and maintain the animals in their natural lateral orientation when pushed inside the channel by pressure. The animals’ lateral orientation is best suited for optical imaging of cellular processes along the ventral cord. The vivoChip has singleinput and single-output fluid interfaces for simple connectivity and flow control. A gasket system applies pressure through a single common input. By applying a specific pressure sequence, 40 immobilization channels across 96 wells can trap approximately 4,000 animals across the entire chip simultaneously within less than 3 minutes in an optically favorable orientation and at predetermined locations. After immobilization a constant pressure keeps the animals trapped for imaging. The chip with the gasket is mounted onto a microscope stage to image all channels using automated image-acquisition software that controls stage movement, focuses on the sample, and records 3D images within 15 minutes. The images can then be analyzed using custom-written analysis software to identify fluorescencetagged phenotypes from each animal. Degeneration and aggregation phenotypes are scored within 15 minutes from 5,760 images (96 wells 3 4 fields of view per

377

FIGURE 11.6 The first large-scale microfluidics platform (vivoChip) for high-throughput and high-resolution fluorescence imaging of Caenorhabditis elegans nematodes for high-content phenotypic screening. (A) Schematic of the automated vivoChip setup for high-content screening of C. elegans nematodes. (B) The gasket system holding a 96well vivoChip. (C) The vivoChip was used to screen a large number of SCAPP animals to identify norbenzomorphans with neuroprotective roles. The panel below shows the neurodegeneration percentages for all 10 compounds tested on a single chip at four concentrations (20, 2, 0.2, and 0.02 μM). The green and orange lines represent mean 6 standard deviation of degeneration in the healthy and vehicle-treated, unhealthy SCAPP populations, respectively. (D) Phenotypic screen of approximately 1000 US Food and Drug Administration approved drugs using a polyglutamine aggregation model. The panel below shows the aggregation number per unit length of the animal. The solid and dashed lines represent the average 6 three standard deviations of the aggregate numbers. SCAPP, Single-copy APP. Adapted from Ben-Yakar, A., 2019. High-content and high-throughput in vivo drug screening platforms using microfluidics. Assay Drug Dev. Technol. 17 (1), 8 13.

Large-scale microfluidics for phenotyping multiple populations

well 3 15 stacks per field of view). Data extracted from the images are saved into multidimensional arrays for statistical analysis.

Practical applications of the vivoChip We used the 96-well vivoChip platform to investigate structure activity relationships of various compounds in a C. elegans disease model of the human amyloid precursor protein (Fig. 11.6C) (Mondal et al., 2018). Another high-content screening of approximately 1,000 FDA-approved compounds identified four that reduced aggregation of polyglutamine (Fig. 11.6D) in a nematode strain relevant for studying Huntington’s disease (Mondal et al., 2016). Phenotypic screening images the entire body in 3D to capture the off-target effects of the exposure of interest. To image the entire animal, which is 1 mm in length, a large field of view camera and a 0.3 numerical aperture were used at 10 3 magnification. One-micron resolutions obtained using these settings are sufficient for definitive readouts. At 10 3 magnification, the camera captures 10 parallel immobilization channels (1.5 3 1.5 mm2). On a single 96-well chip, approximately 3,650 animals (at 90% trapping efficiency) can be imaged within 15 minutes or 14,600 animals per hour from multiple populations. A fully automated software can capture 15 z-stacks fluorescence images of 10 channels simultaneously. A graphical user interface includes an image analysis algorithm to phenotype each animal and display multiparameter scores such as the size of the fluorescence blobs, the length of the animals, the intensity of the fluorescence, and the number of fluorescence blobs per unit body length of each animal. One of the most important features of vivoChip is its unique design that can orient animals in an optically favorable position during immobilization. The challenge in imaging subtle changes in C. elegans is related to the need for immobilization in a lateral orientation (Mondal et al., 2018). Most animals commonly rotate themselves to a nonlateral position when immobilized in tapering channels, as the width of the channels decreases relative to their height. To overcome this challenge, chip designs incorporating valves, suction, sieves, and curved channel geometries have been suggested (Caceres Ide et al., 2012; Go¨kc¸e et al., 2014; Chung et al., 2008; Lee et al., 2013; Zeng et al., 2008). While providing the required orientation conditions for high-resolution imaging, these sophisticated chip geometries cannot be scaled up to achieve high throughput. The unique trap geometry of the vivoChip design includes reducing both height and width concurrently to maintain their aspect ratio, preventing the animals from rotating. Under controlled pressure the animals orient themselves rapidly in a lateral position with a 90% 95% success rate during trapping, depending on the strain. The vivoChip platform is well positioned to fulfill the need for alternative animal testing at high throughput and high content during the early phases of drug discovery. A 384well vivoChip has also been developed, and it can immobilize 30 animals per population and image all 11,520 animals (30 3 384) within 35 minutes in 3D for fluorescence-based phenotypic analysis.

379

380

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

Future directions Microfluidic devices enable high-resolution imaging of fluorescently tagged C. elegans nematodes in systemic biological studies. New automation tools, precise optomechanics, small-volume liquid manipulators, and automated analysis facilitate the adoption of microfluidic technologies in various life-science markets that require large-scale automated screens. An alternative to the multitrap approach would be to perform imaging without immobilization. However, there are two major challenges to immobilization-free screening, which requires serial analysis. First, the animals must be delivered to the imaging location. Second, as camera-based, high-resolution imaging is conventionally used for high-content phenotyping of C. elegans, immobilization is required; it may take up to a few seconds to image z-stacks of a single animal even under full automation (BenYakar et al., 2009; Mondal et al., 2016; Crane et al., 2012). To address these challenges, our laboratory has designed a population-delivery chip with on-chip wells and multiplexing on-chip valves system that can deliver multiple populations and up to 200 nematodes per population serially to another platform for optical interrogation (Ghorashian et al., 2013). The multiplexer chip is designed in a multiwell format and has been validated for a 16-well plate (Ghorashian et al., 2013) and a 64-well plate (Ben-Yakar et al., 2013; Ghorashian, 2013) (Fig. 11.7A and C). The 64-well delivery chip incorporates automated control of the on-chip pneumatic valves to operate delivery within 2.5 seconds per well (Ghorashian, 2013). A buffer plug between two successive populations eliminates cross-mixing and prevents air bubbles, two factors in the stability of fluid flow profiles and error-free chip operations. Our laboratory also developed an ultrafast 3D flow-cytometry method for 3D imaging of C. elegans nematodes. As the population delivery time is 2.5 seconds, the 3D analysis of 50 200 animals should be performed within 2 seconds. The samples should thus flow at a speed of 1 m/s, similar to the speeds achieved by the COPAS Biosort large-particle flow cytometer from Union Biometrica (Pulak, 2006). The COPAS instrument, however, has poor resolution and cannot distinguish subtle phenotypic changes. Current literature for large-particle flow cytometers has only demonstrated speeds of up to 1 mm/s because of the low frame rates of current imaging methods (McGorty et al., 2015; Wu et al., 2013). To our knowledge, imaging at 1 m/s has only been achieved so far in 2D bright-field cytometry (Goda et al., 2009). Capturing blur-free images of C. elegans nematodes moving at 1 m/s through a flow cytometer requires an imaging method of 1 million frames per second (fps). Our laboratory developed line excitation array detection (LEAD) fluorescence microscopy (Fig. 11.7B), an imaging technique capable of 0.8 million fps (Martin et al., 2018). This technology incorporates ultrafast line scanning and sensitive array detection, enabling frame rates that are higher by two orders of magnitude than other fluorescence imaging systems while maintaining high sensitivity and pixel rate. The LEAD flow cytometry platform does not require animal

Future directions

FIGURE 11.7 Increasing throughput and content in screening Caenorhabditis elegans phenotypes. (A) Schematic of an integrated multiplexer chip for pushing nematodes through the channel at approximately 1 m/s for fast serial imaging. (B) The image of a 64-well multiplexer chip. The on-chip valves control the population delivery from a specific well through the exit. (C) Line excitation array detection fluorescence microscope, implemented as a three-dimensional whole-animal flow cytometer capturing 0.8 million frames per second. A longitudinal tellurium dioxide AOD driven by a chirped frequency scans the laser excitation beam across an angled plane on the sample. The excited plane is imaged onto 14 channels of a 16-channel PMT array, capturing a full frame each scan cycle. The microscope is combined with a microfluidic device that delivers hundreds of C. elegans nematodes at 1 m/s through the excitation region (inset). The image represents a three-dimensional reconstruction of a C. elegans nematode (with polyglutamine tagged with yellow fluorescent protein) imaged over 0.79 ms using line excitation array detection fluorescence microscopy. Scale bar: 50 μm. AOD, acousto-optic deflector; PMT, photomultiplier tube.

immobilization and can image hundreds of animals within 1 second. A 3.5-μm average resolution in all three dimensions and speeds over 1 m/s did not produce motion blur, and imaging was 1000 times faster than the speed reported for 3D cytometers (Fig. 11.7D and E) (McGorty et al., 2015; Wu et al., 2013). The LEAD microscope successfully identified a concentration-dependent reduction in the number of polyglutamine aggregates in the nematodes following exposure to the antiarrhythmic drug dronedarone. The LEAD 3D flow cytometer for wholeanimal screening makes it possible to screen up to 10,000 compounds in under a day.

381

382

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

Integrating this system with the population-delivery chip (Ghorashian et al., 2013, 2016) will yield the fastest C. elegans screening modality for ultrahigh throughput at the same cost and speed of in vitro screening. Screening technologies rely heavily on rapid, efficient, and automated data acquisition and analysis software packages. Machine-learning and artificial intelligence based algorithms are being incorporated into screening platforms to fully automate the assays at increasing speeds. The vivoChip platform could be modified in the future to house and feed C. elegans nematodes, expose them to chemical compounds, and image their physical traits (e.g., size, number of eggs) and behavioral output on-chip before performing phenotypic analyses using highresolution imaging that requires immobilization. The development of new genetic models and imaging platforms will facilitate the adoption of C. elegans into mainstream phenotypic screening in drug discovery, mode-of-action studies, in vivo toxicology assays, and investigations of structure activity relationships. The benefits for the pharmaceutical industry, of high-throughput, rapid and accurate screening tools to eliminate potentially toxic compounds at the preclinical stage and reduce the number of expensive drug failures during clinical development, are clear. The benefits for consumers, of safer, and potentially cheaper drug therapies are also appealing. While significant progress has been made to date toward development of the large-scale platforms of C. elegans-on-a-chip, their wide-adoption requires additional research to enable rapid analysis of a large number of images, and development and validation of C. elegans models to ensure assays can recapitulate existing in vivo data before this novel technology and C. elegans models can become a keystone of toxicologic analysis.

Acknowledgments We thank the National Institutes of Health for financial support (NIH Director’s Transformative Award from the National Institute on Aging, R01 AG041135). We thank the Targeted Therapeutic Drug Discovery and Development Program (TTDDDP) at The University of Texas at Austin for assistance with compound libraries. The TTDDDP core facility is supported by Cancer Prevention Research Institute of Texas grant RP110532.

References Alepee, N., et al., 2014. State-of-the-art of 3D cultures (organs-on-a-chip) in safety testing and pathophysiology. ALTEX 31 (4), 441 477. Alexander, A.G., Marfil, V., Li, C., 2014. Use of Caenorhabditis elegans as a model to study Alzheimer’s disease and other neurodegenerative diseases. Front. Genet. 5, 279. Allard, P., et al., 2013. A C. elegans screening platform for the rapid assessment of chemical disruption of germline function. Environ. Health Perspect. 121 (6), 717 724.

References

Allen, P.B., et al., 2008. Single-synapse ablation and long-term imaging in live C. elegans. J. Neurosci. Methods 173 (1), 20 26. Almeida, A., Sarmento, B., Rodrigues, F., 2017. Insights on in vitro models for safety and toxicity assessment of cosmetic ingredients. Int. J. Pharm. 519 (1-2), 178 185. Avila, D., Helmcke, K., Aschner, M., 2012. The Caenorhabiditis elegans model as a reliable tool in neurotoxicology. Hum. Exp. Toxicol. 31 (3), 236 243. Bakhtina, N.A., Korvink, J.G., 2014. Microfluidic laboratories for C. elegans enhance fundamental studies in biology. RSC Adv. 4 (9), 4691 4709. Bargmann, C.I., 2012. Beyond the connectome: how neuromodulators shape neural circuits. Bioessays 34 (6), 458 465. Baskoylu, S.N., et al., 2018. Single copy/knock-in models of ALS SOD1 in C. elegans suggest loss and gain of function have different contributions to cholinergic and glutamatergic neurodegeneration. PLoS Genet. 14 (10), e1007682. Ben-Yakar, A., 2019. High-content and high-throughput in vivo drug screening platforms using microfluidics. Assay Drug Dev. Technol. 17 (1), 8 13. Ben-Yakar, A., Ghorashian, N., Go¨kc¸e, S.K., Guo, S.X., Everett, N., Bourgeois, F., 2013. Microfluidic Devices for the Rapid and Automated Processing of Sample Populations. US Patent 10,052,631. Ben-Yakar, A., Chronis, N., Lu, H., 2009. Microfluidics for the analysis of behavior, nerve regeneration, and neural cell biology in C. elegans. Curr. Opin. Neurobiol. 19 (5), 561 567. Birey, F., et al., 2017. Assembly of functionally integrated human forebrain spheroids. Nature 545 (7652), 54 59. Bodhicharla, R., et al., 2012. Effects of alpha-synuclein overexpression in transgenic Caenorhabditis elegans strains. CNS Neurol. Disord. Drug Targets 11 (8), 965 975. Bovard, D., et al., 2018. A lung/liver-on-a-chip platform for acute and chronic toxicity studies. Lab Chip 18 (24), 3814 3829. Boyd, W.A., et al., 2010. Medium- and high-throughput screening of neurotoxicants using C. elegans. Neurotoxicol. Teratol. 32 (1), 68 73. Boyd, W.A., et al., 2016. Developmental effects of the ToxCast Phase I and Phase II chemicals in Caenorhabditis elegans and corresponding responses in zebrafish, rats, and rabbits. Environ. Health Perspect. 124 (5), 586 593. Braungart, E., et al., 2004. Caenorhabditis elegans MPP 1 model of Parkinson’s disease for high-throughput drug screenings. Neurodegener. Dis. 1 (4-5), 175 183. Brenner, S., 1974. The genetics of Caenorhabditis elegans. Genetics 77 (1), 71 94. Brignull, H.R., et al., 2006. Modeling polyglutamine pathogenesis in C. elegans. Methods Enzymol. 412, 256 282. C. elegans Sequencing Consortium, 1998. Genome sequence of the nematode C. elegans: a platform for investigating biology. Science 282 (5396), 2012 2018. Caceres Ide, C., et al., 2012. Laterally orienting C. elegans using geometry at microscale for high-throughput visual screens in neurodegeneration and neuronal development studies. PLoS One 7 (4), e35037. Caito, S.W., Valentine, W.M., Aschner, M., 2013. Dopaminergic neurotoxicity of S-ethyl N,N-dipropylthiocarbamate (EPTC), molinate, and S-methyl-N,N-diethylthiocarbamate (MeDETC) in Caenorhabditis elegans. J. Neurochem. 127 (6), 837 851. Chalasani, S.H., et al., 2007. Dissecting a circuit for olfactory behaviour in Caenorhabditis elegans. Nature 450 (7166), 63 70.

383

384

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

Chen, P., et al., 2015. Manganese-induced neurotoxicity: from C. elegans to humans. Toxicol. Res. (Camb) 4 (2), 191 202. Chikka, M.R., et al., 2016. The mitochondria-regulated immune pathway activated in the C. elegans intestine is neuroprotective. Cell Rep. 16 (9), 2399 2414. Cho, H., et al., 2018. Microfluidic technologies for circulating tumor cell isolation. Analyst 143 (13), 2936 2970. Cho, S., Yoon, J.Y., 2017. Organ-on-a-chip for assessing environmental toxicants. Curr. Opin. Biotechnol. 45, 34 42. Chokshi, T.V., Ben-Yakar, A., Chronis, N., 2009. CO2 and compressive immobilization of C. elegans on-chip. Lab Chip 9 (1), 151 157. Chow, D.K., et al., 2006. Sarcopenia in the Caenorhabditis elegans pharynx correlates with muscle contraction rate over lifespan. Exp. Gerontol. 41 (3), 252 260. Chronis, N., 2010. Worm chips: microtools for C. elegans biology. Lab Chip 10 (4), 432 437. Chronis, N., Zimmer, M., Bargmann, C.I., 2007. Microfluidics for in vivo imaging of neuronal and behavioral activity in Caenorhabditis elegans. Nat. Methods 4 (9), 727 731. Chuang, H.S., et al., 2011. Dielectrophoresis of Caenorhabditis elegans. Lab Chip 11 (4), 599 604. Chung, K., Lu, H., 2009. Automated high-throughput cell microsurgery on-chip. Lab. Chip 9 (19), 2764 2766. Chung, K., Crane, M.M., Lu, H., 2008. Automated on-chip rapid microscopy, phenotyping and sorting of C. elegans. Nat. Methods 5 (7), 637 643. Churgin, M.A., et al., 2017. Longitudinal imaging of Caenorhabditis elegans in a microfabricated device reveals variation in behavioral decline during aging. Elife 6, e26652. Clark, D.L., et al., 1999. Predictive value of preclinical toxicology studies for platinum anticancer drugs. Clin. Cancer Res. 5 (5), 1161 1167. Coecke, S., et al., 2006. The value of alternative testing for neurotoxicity in the context of regulatory needs. Environ. Toxicol. Pharmacol. 21 (2), 153 167. Collins, F.S., Gray, G.M., Bucher, J.R., 2008. Toxicology. Transforming environmental health protection. Science 319 (5865), 906 907. Cook, D., et al., 2014. Lessons learned from the fate of AstraZeneca’s drug pipeline: a five-dimensional framework. Nat. Rev. Drug Discov. 13 (6), 419 431. Cornaglia, M., et al., 2015. An automated microfluidic platform for C. elegans embryo arraying, phenotyping, and long-term live imaging. Sci. Rep. 5, 10192. Cornaglia, M., Lehnert, T., Gijs, M.A.M., 2017. Microfluidic systems for high-throughput and high-content screening using the nematode Caenorhabditis elegans. Lab Chip 17 (22), 3736 3759. Crane, M.M., et al., 2012. Autonomous screening of C. elegans identifies genes implicated in synaptogenesis. Nat. Methods 9 (10), 977 980. Crawford, S.E., et al., 2017. Green toxicology: a strategy for sustainable chemical and material development. Environ. Sci. Eur. 29 (1), 16. De Simone, U., et al., 2018. Human 3D cultures as models for evaluating magnetic nanoparticle CNS cytotoxicity after short- and repeated long-term exposure. Int. J. Mol. Sci. 19, 7. Dekker, S., et al., 2018. From chip-in-a-lab to lab-on-a-chip: a portable Coulter counter using a modular platform. Microsyst. Nanoeng. 4, 34. Dekkers, J.F., et al., 2019. High-resolution 3D imaging of fixed and cleared organoids. Nat. Protoc. 14, 1756 1771.

References

Dexter, P.M., Caldwell, K.A., Caldwell, G.A., 2012. A predictable worm: application of Caenorhabditis elegans for mechanistic investigation of movement disorders. Neurotherapeutics 9 (2), 393 404. Ding, X., et al., 2012. On-chip manipulation of single microparticles, cells, and organisms using surface acoustic waves. Proc. Natl. Acad. Sci. U.S.A. 109 (28), 11105 11109. Dix, D.J., et al., 2007. The ToxCast program for prioritizing toxicity testing of environmental chemicals. Toxicol. Sci. 95 (1), 5 12. Doll, J.C., et al., 2009. SU-8 force sensing pillar arrays for biological measurements. Lab Chip 9 (10), 1449 1454. Dorea, J.G., 2011. Integrating experimental (in vitro and in vivo) neurotoxicity studies of low-dose thimerosal relevant to vaccines. Neurochem. Res. 36 (6), 927 938. Dupuy, D., et al., 2007. Genome-scale analysis of in vivo spatiotemporal promoter activity in Caenorhabditis elegans. Nat. Biotechnol. 25 (6), 663 668. Faber, P.W., et al., 1999. Polyglutamine-mediated dysfunction and apoptotic death of a Caenorhabditis elegans sensory neuron. Proc. Natl. Acad. Sci. U.S.A. 96 (1), 179 184. Fang-Yen, C., Alkema, M.J., Samuel, A.D., 2015. Illuminating neural circuits and behaviour in Caenorhabditis elegans with optogenetics. Philos. Trans. R. Soc. Lond. B: Biol. Sci. 370 (1677), 20140212. Forsythe, S.D., et al., 2018. Environmental toxin screening using human-derived 3D bioengineered liver and cardiac organoids. Front. Public Health 6, 103. Geng, T., Lu, C., 2013. Microfluidic electroporation for cellular analysis and delivery. Lab Chip 13 (19), 3803 3821. Ghorashian, N., 2010. Parallelized microfluidic devices for high-throughput nerve regeneration studies in Caenorhabditis elegans. Biomedical Engineering. The University of Texas at Austin. Ghorashian, N., 2013. Automated, microfluidic platforms to facilitate nerve degeneration studies with C. elegans. Biomedical Engineering. The University of Texas at Austin. Ghorashian, N., et al., 2013. An automated microfluidic multiplexer for fast delivery of C. elegans populations from multiwells. PLoS One 8 (9), e74480. Ghorashian, N., Go¨kc¸e, S.K., Ben-Yakar, A., 2016. In: Love, C. (Ed.), Microfluidic Systems for Whole-Animal Screening with C. elegans. Wiley Online Library. Giacomotto, J., Segalat, L., 2010. High-throughput screening and small animal models, where are we? Br. J. Pharmacol. 160 (2), 204 216. Goda, K., Tsia, K.K., Jalali, B., 2009. Serial time-encoded amplified imaging for real-time observation of fast dynamic phenomena. Nature 458 (7242), 1145 1149. Go¨kc¸e, S.K., et al., 2014. A fully automated microfluidic femtosecond laser axotomy platform for nerve regeneration studies in C. elegans. PLoS One 9 (12), e113917. Gokce, S.K., et al., 2017. A multi-trap microfluidic chip enabling longitudinal studies of nerve regeneration in Caenorhabditis elegans. Sci. Rep. 7 (1), 9837. Gomez-Amaro, R.L., et al., 2015. Measuring food intake and nutrient absorption in Caenorhabditis elegans. Genetics 200 (2), 443 454. Gong, H., Do, D., Ramakrishnan, R., 2018. Single-cell mRNA-seq using the Fluidigm C1 system and integrated fluidics circuits. Methods Mol. Biol. 1783, 193 207. Gosai, S.J., et al., 2010. Automated high-content live animal drug screening using C. elegans expressing the aggregation prone serpin alpha1-antitrypsin Z. PLoS One 5 (11), e15460.

385

386

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

Guengerich, F.P., 2011. Mechanisms of drug toxicity and relevance to pharmaceutical development. Drug Metab. Pharmacokinet. 26 (1), 3 14. Guo, S.X., et al., 2008. Femtosecond laser nanoaxotomy lab-on-a-chip for in vivo nerve regeneration studies. Nat. Methods 5 (6), 531 533. Gupta, N., et al., 2016. Microfluidics-based 3D cell culture models: utility in novel drug discovery and delivery research. Bioeng. Transl. Med. 1 (1), 63 81. Hadd, A.G., et al., 1997. Microchip device for performing enzyme assays. Anal. Chem. 69 (17), 3407 3412. van Ham, T.J., et al., 2008. C. elegans model identifies genetic modifiers of alphasynuclein inclusion formation during aging. PLoS Genet. 4 (3), e1000027. Harlow, P.H., et al., 2016. The nematode Caenorhabditis elegans as a tool to predict chemical activity on mammalian development and identify mechanisms influencing toxicological outcome. Sci. Rep. 6, 22965. Harlow, P.H., et al., 2018. Comparative metabolism of xenobiotic chemicals by cytochrome P450s in the nematode Caenorhabditis elegans. Sci. Rep. 8 (1), 13333. Harry, G.J., et al., 1998. In vitro techniques for the assessment of neurotoxicity. Environ. Health Perspect. 106 (Suppl 1), 131 158. Hartley, B.J., Brennand, K.J., 2017. Neural organoids for disease phenotyping, drug screening and developmental biology studies. Neurochem. Int. 106, 85 93. Hartung, T., 2010. Food for thought. . . on alternative methods for chemical safety testing. ALTEX 27 (1), 3-14. Hobert, O., 2010. Neurogenesis in the nematode Caenorhabditis elegans. WormBook. pp. 1 24. Hulme, S.E., et al., 2007. A microfabricated array of clamps for immobilizing and imaging C. elegans. Lab Chip 7 (11), 1515 1523. Hulme, S.E., et al., 2010. Lifespan-on-a-chip: microfluidic chambers for performing lifelong observation of C. elegans. Lab Chip 10 (5), 589 597. Hunt, P.R., 2017. The C. elegans model in toxicity testing. J. Appl. Toxicol. 37 (1), 50 59. Inglese, J., et al., 2006. Quantitative high-throughput screening: a titration-based approach that efficiently identifies biological activities in large chemical libraries. Proc. Natl. Acad. Sci. U.S.A. 103 (31), 11473 11478. Ionescu-Zanetti, C., et al., 2005. Mammalian electrophysiology on a microfluidic platform. Proc. Natl. Acad. Sci. U.S.A. 102 (26), 9112 9117. Ishida, S., 2018. Organs-on-a-chip: current applications and consideration points for in vitro ADME-Tox studies. Drug Metab. Pharmacokinet. 33 (1), 49 54. Jensen, J.B., Parmar, M., 2006. Strengths and limitations of the neurosphere culture system. Mol. Neurobiol. 34 (3), 153 161. Ju, J., et al., 2014. Neurotoxic evaluation of two organobromine model compounds and natural AOBr-containing surface water samples by a Caenorhabditis elegans test. Ecotoxicol. Environ. Saf. 104, 194 201. Kaletta, T., Hengartner, M.O., 2006. Finding function in novel targets: C. elegans as a model organism. Nat. Rev. Drug Discov. 5 (5), 387 398. Keil, W., et al., 2017. Long-term high-resolution imaging of developing C. elegans larvae with microfluidics. Dev. Cell 40 (2), 202 214. Khamenehfar, A., Li, P.C., 2016. Microfluidic devices for circulating tumor cells isolation and subsequent analysis. Curr. Pharm. Biotechnol. 17 (9), 810 821.

References

Khandurina, J., et al., 2000. Integrated system for rapid PCR-based DNA analysis in microfluidic devices. Anal. Chem. 72 (13), 2995 3000. Kim, H.J., et al., 2012. Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow. Lab Chip 12 (12), 2165 2174. Kim, J., et al., 2015. Organ-on-a-chip: development and clinical prospects toward toxicity assessment with an emphasis on bone marrow. Drug Saf. 38 (5), 409 418. Kimura, H., Sakai, Y., Fujii, T., 2018. Organ/Body-on-a-chip based on microfluidic technology for drug discovery. Drug Metab. Pharmacokinet. 33 (1), 43 48. Koch, K., et al., 2014. Caenorhabditis elegans as model system in pharmacology and toxicology: effects of flavonoids on redox-sensitive signalling pathways and ageing. ScientificWorldJournal 2014, 920398. Krajniak, J., Lu, H., 2010. Long-term high-resolution imaging and culture of C. elegans in chip-gel hybrid microfluidic device for developmental studies. Lab Chip 10 (14), 1862 1868. Krajniak, J., et al., 2013. C.L.I.P.—continuous live imaging platform for direct observation of C. elegans physiological processes. Lab Chip 13 (15), 2963 2971. Lagally, E.T., Medintz, I., Mathies, R.A., 2001. Single-molecule DNA amplification and analysis in an integrated microfluidic device. Anal. Chem. 73 (3), 565 570. Lau, A.Y., et al., 2006. Open-access microfluidic patch-clamp array with raised lateral cell trapping sites. Lab Chip 6 (12), 1510 1515. Lee, H., et al., 2013. Quantitative screening of genes regulating tryptophan hydroxylase transcription in Caenorhabditis elegans using microfluidics and an adaptive algorithm. Integr. Biol. 5 (2), 372 380. Letizia, M.C., et al., 2018. Microfluidics-enabled phenotyping of a whole population of C. elegans worms over their embryonic and post-embryonic development at singleorganism resolution. Microsyst. Nanoeng. 4, 6. Leung, C.K., Deonarine, A., Strange, K., Choe, K.P., 2011. High-throughput screening and biosensing with fluorescent C. elegans strains. J. Vis. Exp. (51), e2745. Available from: https://doi.org/10.3791/2745. Leung, C.K., et al., 2013. An ultra high-throughput, whole-animal screen for small molecule modulators of a specific genetic pathway in Caenorhabditis elegans. PLoS One 8 (4), e62166. Leung, M.C., et al., 2008. Caenorhabditis elegans: an emerging model in biomedical and environmental toxicology. Toxicol. Sci. 106 (1), 5 28. Li, J., et al., 2019. An industry perspective: a streamlined screening strategy using alternative models for chemical assessment of developmental neurotoxicity. Neurotoxicology 73, 17 30. Lockery, S.R., et al., 2012. A microfluidic device for whole-animal drug screening using electrophysiological measures in the nematode C. elegans. Lab Chip 12 (12), 2211 2220. Lucanic, M., Garrett, T., Gill, M.S., Lithgow, G.J., 2018. A simple method for high throughput chemical screening in Caenorhabditis Elegans. J. Vis. Exp. (133), e56892. Available from: https://doi.org/10.3791/56892. Maertens, A., Hartung, T., 2018. Green toxicology-know early about and avoid toxic product liabilities. Toxicol. Sci. 161 (2), 285 289. Maertens, A., et al., 2014. Green toxicology. ALTEX 31 (3), 243 249. Martin, C., et al., 2018. Line excitation array detection fluorescence microscopy at 0.8 million frames per second. Nat. Commun. 9 (1), 4499.

387

388

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

McGorty, R., et al., 2015. Open-top selective plane illumination microscope for conventionally mounted specimens. Opt. Express 23 (12), 16142 16153. McVey, K.A., et al., 2016. Exposure of C. elegans eggs to a glyphosate-containing herbicide leads to abnormal neuronal morphology. Neurotoxicol. Teratol. 55, 23 31. Melentijevic, I., et al., 2017. C. elegans neurons jettison protein aggregates and mitochondria under neurotoxic stress. Nature 542 (7641), 367 371. Meyer, D., Williams, P.L., 2014. Toxicity testing of neurotoxic pesticides in Caenorhabditis elegans. J. Toxicol. Environ. Health B: Crit. Rev. 17 (5), 284 306. Meyer, D., et al., 2016. Differential toxicities of nickel salts to the nematode Caenorhabditis elegans. Bull. Environ. Contam. Toxicol. 97 (2), 166 170. Miranda, C.C., et al., 2018. Towards multi-organoid systems for drug screening applications. Bioengineering (Basel) 5 (3), e49. Mondal, S., et al., 2011. Imaging in vivo neuronal transport in genetic model organisms using microfluidic devices. Traffic 12 (4), 372 385. Mondal, S., Ahlawat, S., Koushika, S.P., 2012. Simple microfluidic devices for in vivo imaging of C. elegans, Drosophila and zebrafish. J. Vis. Exp. (67), e3780. Available from: https://doi.org/10.3791/3780. Mondal, S., Koushika, S.P., 2014. Microfluidic devices for imaging trafficking events in vivo using genetic model organisms. Methods Mol. Biol. 1174, 375 396. Mondal, S., et al., 2016. Large-scale microfluidics providing high-resolution and highthroughput screening of Caenorhabditis elegans poly-glutamine aggregation model. Nat. Commun. 7, 13023. Mondal, S., et al., 2018. High-content microfluidic screening platform used to identify sigma2R/Tmem97 binding ligands that reduce age-dependent neurodegeneration in C. elegans SC_APP model. ACS Chem. Neurosci. 9, 1014 1026. Morley, J.F., et al., 2002. The threshold for polyglutamine-expansion protein aggregation and cellular toxicity is dynamic and influenced by aging in Caenorhabditis elegans. Proc. Natl. Acad. Sci. U.S.A. 99 (16), 10417 10422. Moser, V.C., 2007. Animal models of chronic pesticide neurotoxicity. Hum. Exp. Toxicol. 26 (4), 321 331. NC3R, 2007. Toxicity Testing in the Twenty-First Century: A Vision and a Strategy. The National Academy Press. Nass, R., et al., 2002. Neurotoxin-induced degeneration of dopamine neurons in Caenorhabditis elegans. Proc. Natl. Acad. Sci. U.S.A. 99 (5), 3264 3269. Negga, R., et al., 2011. Exposure to Mn/Zn ethylene-bis-dithiocarbamate and glyphosate pesticides leads to neurodegeneration in Caenorhabditis elegans. Neurotoxicology 32 (3), 331 341. Negga, R., et al., 2012. Exposure to glyphosate- and/or Mn/Zn-ethylene-bis-dithiocarbamate-containing pesticides leads to degeneration of gamma-aminobutyric acid and dopamine neurons in Caenorhabditis elegans. Neurotox. Res. 21 (3), 281 290. Nollen, E.A., et al., 2004. Genome-wide RNA interference screen identifies previously undescribed regulators of polyglutamine aggregation. Proc. Natl. Acad. Sci. U.S.A. 101 (17), 6403 6408. Noyes, P.D., Garcia, G.R., Tanguay, R.L., 2016. Zebrafish as an in vivo model for sustainable chemical design. Green Chem. 18 (24), 6410 6430. Offenburger, S.L., et al., 2018. 6-OHDA-induced dopaminergic neurodegeneration in Caenorhabditis elegans is promoted by the engulfment pathway and inhibited by the transthyretin-related protein TTR-33. PLoS Genet. 14 (1), e1007125.

References

Pasca, A.M., et al., 2015. Functional cortical neurons and astrocytes from human pluripotent stem cells in 3D culture. Nat. Methods 12 (7), 671 678. Pincus, Z., Slack, F.J., 2010. Developmental biomarkers of aging in Caenorhabditis elegans. Dev. Dyn. 239 (5), 1306 1314. Pridgeon, C.S., et al., 2018. Innovative organotypic in vitro models for safety assessment: aligning with regulatory requirements and understanding models of the heart, skin, and liver as paradigms. Arch. Toxicol. 92 (2), 557 569. Pu, P., Le, W., 2008. Dopamine neuron degeneration induced by MPP 1 is independent of CED-4 pathway in Caenorhabditis elegans. Cell Res. 18 (9), 978 981. Pulak, R., 2006. Techniques for analysis, sorting, and dispensing of C. elegans on the COPAS flow-sorting system. Methods Mol. Biol. 351, 275 286. Rahman, M., et al., 2018. NemaFlex: a microfluidics-based technology for standardized measurement of muscular strength of C. elegans. Lab Chip 18 (15), 2187 2201. Reece, A., et al., 2016. Microfluidic techniques for high throughput single cell analysis. Curr. Opin. Biotechnol. 40, 90 96. Rezai, P., et al., 2012. Electrical sorting of Caenorhabditis elegans. Lab Chip 12 (10), 1831 1840. Richmond, J., Synaptic function. The C. elegans Research Community, WormBook. Erik M. Jorgensen and Joshua M. Kaplan, 2005: p. 1 14, Available from: https://doi/ 10.1895/wormbook.1.69.1. Roberts, R.A., et al., 2015. Translational biomarkers of neurotoxicity: a health and environmental sciences institute perspective on the way forward. Toxicol. Sci. 148 (2), 332 340. Rodriguez-Palero, M.J., et al., 2018. An automated method for the analysis of food intake behaviour in Caenorhabditis elegans. Sci. Rep. 8 (1), 3633. Rohde, C.B., et al., 2007. Microfluidic system for on-chip high-throughput whole-animal sorting and screening at subcellular resolution. Proc. Natl. Acad. Sci. U.S.A. 104 (35), 13891 13895. Ruszkiewicz, J.A., et al., 2018. C. elegans as a model in developmental neurotoxicology. Toxicol. Appl. Pharmacol. 354, 126 135. Sahn, J.J., et al., 2017. Sigma 2 receptor/Tmem97 agonists produce long lasting antineuropathic pain effects in mice. ACS Chem. Neurosci. 8 (8), 1801 1811. Salam, S., et al., 2013. A microfluidic phenotype analysis system reveals function of sensory and dopaminergic neuron signaling in C. elegans electrotactic swimming behavior. Worm 2 (2), e24558. Samara, C., et al., 2010. Large-scale in vivo femtosecond laser neurosurgery screen reveals small-molecule enhancer of regeneration. Proc. Natl. Acad. Sci. U.S.A. 107 (43), 18342 18347. San-Miguel, A., et al., 2016. Deep phenotyping unveils hidden traits and genetic relations in subtle mutants. Nat. Commun. 7, 12990. Selimovic, S., Dokmeci, M.R., Khademhosseini, A., 2013. Organs-on-a-chip for drug discovery. Curr. Opin. Pharmacol. 13 (5), 829 833. Sengupta, P., Samuel, A.D., 2009. Caenorhabditis elegans: a model system for systems neuroscience. Curr. Opin. Neurobiol. 19 (6), 637 643. Solis, G.M., Petrascheck, M., 2011. Measuring Caenorhabditis elegans life span in 96 well microtiter plates. J. Vis. Exp. (49), e2496. Available from: https://doi.org/10.3791/ 2496.

389

390

CHAPTER 11 Caenorhabditis elegans-on-a-chip: microfluidic

Son, H.G., et al., 2019. Age-dependent changes and biomarkers of aging in Caenorhabditis elegans. Aging Cell 18 (2), e12853. Sontheimer-Phelps, A., Hassell, B.A., Ingber, D.E., 2019. Modelling cancer in microfluidic human organs-on-chips. Nat. Rev. Cancer 19 (2), 65 81. Stiernagle, T., 2006. Maintenance of C. elegans. The C. elegans Research Community, WormBook. pp. 1 11, doi/10.1895/wormbook.1.101.1. Tong, J., et al., 2013. Microfluidic-based electrotaxis for on-demand quantitative analysis of Caenorhabditis elegans’ locomotion. J. Vis. Exp. (75), e50226. Wahlby, C., et al., 2012. An image analysis toolbox for high-throughput C. elegans assays. Nat. Methods 9 (7), 714 716. Wang, Y.M., Pu, P., Le, W.D., 2007. ATP depletion is the major cause of MPP 1 induced dopamine neuronal death and worm lethality in alpha-synuclein transgenic C. elegans. Neurosci. Bull. 23 (6), 329 335. Wang, Z., Jin, Y., 2011. Genetic dissection of axon regeneration. Curr. Opin. Neurobiol. 21 (1), 189 196. Weeks, J.C., et al., 2016. Microfluidic platform for electrophysiological recordings from host-stage hookworm and Ascaris suum larvae: a new tool for anthelmintic research. Int. J. Parasitol. Drugs Drug Resist 6 (3), 314 328. Wu, J., Li, J., Chan, R.K., 2013. A light sheet based high throughput 3D-imaging flow cytometer for phytoplankton analysis. Opt. Express 21 (12), 14474 14480. Xian, B., et al., 2013. WormFarm: a quantitative control and measurement device toward automated Caenorhabditis elegans aging analysis. Aging Cell 12 (3), 398 409. Xin, Y., et al., 2016. Use of the Fluidigm C1 platform for RNA sequencing of single mouse pancreatic islet cells. Proc. Natl. Acad. Sci. U.S.A. 113 (12), 3293 3298. Xiong, H., Pears, C., Woollard, A., 2017. An enhanced C. elegans based platform for toxicity assessment. Sci. Rep. 7 (1), 9839. Zeng, F., Rohde, C.B., Yanik, M.F., 2008. Sub-cellular precision on-chip small-animal immobilization, multi-photon imaging and femtosecond-laser manipulation. Lab Chip 8 (5), 653 656. Zhou, S., Wang, Z., Klaunig, J.E., 2013. Caenorhabditis elegans neuron degeneration and mitochondrial suppression caused by selected environmental chemicals. Int. J. Biochem. Mol. Biol. 4 (4), 191 200.

CHAPTER

Gut-on-a-chip microphysiological systems for the recapitulation of the gut microenvironment

9

Seung Hwan Lee1,2,3 and Jong Hwan Sung4 1

Department of Bionano Engineering, Hanyang University, Ansan, 2 Nanosensor Research Institute, Hanyang University, Ansan, 3 Department of Bionanotechnology, Hanyang University, Ansan, 4 Department of Chemical Engineering, Hongik University, Seoul,

Republic Republic Republic Republic

of Korea of Korea of Korea of Korea

Introduction The major functions of the small intestine are digestion and absorption. The small intestinal wall is composed of four layers—the mucosa, submucosa, muscularis externa, and serosa—and is connected by the vascular and nervous systems to perform gastrointestinal functions with mechanical movement, including peristalsis and mixing processes (Lee and Sung, 2018). The mucosa consists of an epithelial layer, lamina propria, and muscularis mucosae (Li et al., 2013). The epithelial layer has barrier functions for effective absorption or prevention of orally administered food and drugs (Lee et al., 2016). The barrier function of the small intestine affects the physiological response to drugs as well as drug bioavailability and efficacy (Olson et al., 2000). Thus intestinal barrier functions should be taken into account when developing drugs. In addition, the intestine interacts with other organs, commensal microbes, and the immune system (Garrett et al., 2010; Round and Mazmanian, 2009; Bein et al., 2018). Therefore abnormalities in its function can cause several diseases such as inflammatory bowel disease, viral, bacterial as well as parasitic infections, enteropathies, and cancer (Rahmani et al., 2019; Schirmer et al., 2018; Blutt et al., 2018; Shin and Kim, 2018). Several in vitro models that simulate the barrier function of the intestinal epithelial layer have been developed and applied to assess drug absorption, permeability, and metabolism. Early approaches involved the extraction of gut tissues from animals, incorporating intestinal membrane segments, cannulated everted sacs, and the Ussing chamber, which mimic the in vivo environment as they contain cells and enzyme that are present in vivo (Kaplan and Cotler, 1972; Lampen et al., 1998; Lasker and Rickert, 1978). However, because the supply of extracted tissue is limited and the long-term maintenance of intestinal function is difficult, Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00010-3 © 2020 Elsevier Inc. All rights reserved.

295

296

CHAPTER 9 Gut-on-a-chip microphysiological systems

high-throughput drug screening applications are limited (Brandon et al., 2003). To replace these in vitro models, a Caco-2 cell model and parallel artificial membrane permeability assay have been developed (Hubatsch et al., 2007; Reis et al., 2010). Although these models are widely accepted for drug screening, the absorption kinetics could vary depending on drug types or transport pathways, resulting in inaccuracies in drug identification (Sun et al., 2008). Furthermore, it is not possible to coculture the Caco-2 cells with the microbiome using conventional in vitro models (Bein et al., 2018). This limitation makes it difficult to study intestinal diseases and their pathophysiology. Therefore it is necessary to develop improved in vitro models that can mimic the function and microenvironment of in vivo intestine more closely.

Gut models with 3D structures mimicking intestinal epithelial layer topology To evaluate drug absorption and permeability across the intestinal membrane, 2D monolayer culture systems have been developed. A representative example is the Transwell system, which is composed of a well plate and well inserts. The bottom of the insert is a semipermeable membrane on which cells can be cultured. In most studies, Caco-2 cells, an intestinal epithelial cell line derived from colon carcinoma, are cultured on the membrane to form cell layers, and then drugs are added to the apical side of the Caco-2 layer. Drug absorption and permeability can be assessed by analyzing the amount of drug molecules in the basolateral side. However, this model often yields results that are different from known in vivo values in terms of drug absorption and permeability (Artursson et al., 2001). One of the reasons for this discrepancy is the difference between the structure of in vivo gut epithelium and the in vitro model (Abbott, 2003; Anselme and Bigerelle, 2011). The gut epithelium has distinctive 3D structures, consisting of villi and crypts (Bein et al., 2018; Nicholson et al., 2005), that not only increase the surface area, but also affect the physiology of gut epithelial cells, such as enzymatic activity and the expression of transporter proteins in the cells (Li et al., 2013). Therefore the development of 3D structures in vitro is important for the reproduction of the physiological functions and characteristics of the in vivo gut system. For this purpose, researchers have developed structures that are topographically similar to the intestinal villi using novel microfabrication technologies. For example, porous polymer membranes covering silicon pillar structures were constructed, on which Caco-2 cells were cultured (Esch et al., 2012). Villi structures have also been constructed using hydrogels, which are more compatible with cells compared with other polymers or silicon-based materials such as polydimethylsiloxane (PDMS) (Cushing and Anseth, 2007). In addition, hydrogels can provide a physiological environment for cells and facilitate cell extracellular matrix

Gut models with 3D structures mimicking intestinal epithelial

interactions. Furthermore, using laser ablation and sacrificial molding, collagen scaffolds mimicking villus-like structures were fabricated (Sung et al., 2011) (Fig. 9.1A). Caco-2 cells were cultured for 3 weeks on 3D structures and different polarizations were observed at the top and bottom of the villi structures. At the top, the cells differentiated into a more columnar and polarized form than did those at the base. These changes affected the transepithelial electrical resistance (TEER) and permeability coefficients of the drugs. Compared with the results using the traditional 2D Caco-2 model, the lower TEER value and higher permeability coefficients of the model drug, atenolol, indicated that the 3D culture of Caco-2 cells mimicked the human intestinal system more closely than the 2D monolayer culture. In addition, the 3D villi model induced increased protein expression of mucin even in the absence of goblet cells, which inhibited bacterial invasion and improved barrier function (Yu et al., 2012). Instead of developing approaches that imitate the villi structure, several studies have focused on replicating gut epithelium crypts in vitro. Caco-2 cells were cultured on crypt-like structures constructed using PDMS or hydrogel materials via microfabrication, and physiological factors were investigated and compared with those in 2D monolayer culture (Wang et al., 2009, 2010) (Fig. 9.1B). Consequently, mitochondrial activity was higher, whereas alkaline phosphatase activity and TEER were lower compared with 2D monolayer culture. Similarly, parylene C was chemically deposited onto decellularized porcine intestine to mimic intestinal topology (Koppes et al., 2016). In another approach, primary crypts derived from murine colon were cultured in 2D or 3D to generate a millimeter-scale epithelium in vitro (Wang et al., 2014). These approaches have demonstrated the importance of 3D structures in mimicking in vivo intestinal function and provide insights into why the absorption and permeability coefficient of drugs may differ between in vitro and in vivo conditions. The combination of 3D villi and crypt structures will help mimic human gut functions more closely and enable a more accurate evaluation of drug permeability coefficients. These researches suggest that mimicking the 3D topology of the intestinal epithelium is one of the key factors to improve the physiological relevance of in vitro gut models. Although these 3D topology-based studies have been successful in mimicking the in vivo intestinal environment, most of these systems have incorporated the Caco-2 cancer cell lines. Since this cancer cell line contains genetic mutations, it has limited use in creating a physiologically realistic model (Mochel et al., 2018). To overcome this limitation, the 3D villi and crypt structures are combined with the intestinal organoids. The organoids have submillimeter size clusters that retain the structure and function of the original organs (Yin et al., 2016). Intestinal organoids, such as minigut, can be derived from adult stem cells or pluripotent stem cells (Dotti and Salas, 2018). Since Sato et al. (2009) succeeded in deriving organoids using mouse-derived intestinal cell lines, several studies have attempted to create human intestinal cell-derived organoids. However, these intestinal organoids can only be produced in the form of an enclosed lumen in a hydrogel; this

297

FIGURE 9.1 Gut models for mimicking 3D topology of the intestinal layers. (A) Schematic illustration displaying the fabrication of the 3D villi model using hydrogel. (B) Schematic illustration showing the hydrogel membrane-based cell culture system mimicking 3D crypt-like topology. The image was reproduced with permission (A) Sung, J.H., et al., 2011. Microscale 3-D hydrogel scaffold for biomimetic gastrointestinal (GI) tract model. Lab Chip 11 (3), 389 392; (B) Wang, L., et al., 2010. Synergic effects of crypt-like topography and ECM proteins on intestinal cell behavior in collagen based membranes. Biomaterials 31 (29), 7586 7598.

Microfluidics-based gut models for mimicking the dynamic environment

limits the access of drugs, chemicals, and microbiota when compared to the in vivo environment (Rahmani et al., 2019). Wang et al. (2017, 2018) designed micropatterned collagen scaffolds using PDMS-based stamps. These scaffolds have 3D topography including pillars and can mimic the intestinal villi and crypt structures. The organoids were dissociated and cultured on the 3D collagen scaffolds as a monolayer. This approach enabled various reagents to easily access the luminal surface and hence induced a biochemical gradient along the crypt villus axis. In addition, the authors demonstrated that a gradient of growth factors induces migration of cells and formation of cell-lineage where stem cell and progenitor cell zones are formed in the crypt well and differentiated cell zones are formed in the villi structure. These approaches show that the 3D structure can reproduce the in vivo environment more closely through a combination of microtechnology and organoids cell culture.

Microfluidics-based gut models for mimicking the dynamic environment The in vivo intestinal wall represents a dynamic microenvironment in terms of food digestion, nutrient absorption, and mechanical movement, including peristalsis and segmentation. These movements facilitate chyme and blood flow postprandially, and result in physiological changes in intestinal cells (Sanderson, 1999). However, conventional culture systems, such as flasks, well plates, and Transwell systems, are generally static. In these static systems, cells adhere to the bottom of a flask or membrane and grow in culture medium. Therefore conventional 2D static culture fails to faithfully recreate the in vivo physiological microenvironment and the cellular characteristics of the intestine. To overcome these limitations, physiological parameters, such as mechanical movement and flow of medium, have been integrated into culture systems using microtechnology. In several studies, microfluidics-based flow environments have been applied to cell culture, using top and bottom microchannels separated by an integrated porous membrane. Caco-2 cells were cultured on the porous membrane and formed a layer resembling the intestinal wall. The absorption and transportation of molecules were analyzed to measure the absorption permeability of the Caco-2 cell layer (Imura et al., 2009; Kimura et al., 2008; Yeon and Park, 2009). Further, analysis of barrier functions of the gut epithelium under the presence of a Caco-2 cell layer was performed. Chi et al. (2015) developed a microfluidic cell culture device and showed that Caco-2 cells grew into a 3D structure on the porous membrane between the top and bottom PDMS layers, forming villus-like structures under fluidic conditions. In addition, the secretion of mucin protein, muc-2, increased, and formation of tight junctions with expression of occluding was observed, with inhibition of paracellular transport of large molecules.

299

300

CHAPTER 9 Gut-on-a-chip microphysiological systems

More recently, Ingber’s group developed a microfluidics-based model with an integrated porous membrane between the top and bottom microchannels (Kim et al., 2012; Kim and Ingber, 2013). A cyclic suction module was added to induce stretching and relaxation of the porous membrane in the system. An interesting novelty of their work was that a mechanical movement was introduced into the system. Caco-2 cells were cultured on the porous membrane, and culture medium was perfused through the cell layer. The perfusion of the medium introduced shear force and peristalsis, which were caused by cyclic strain in the same manner as that in the in vivo microenvironment (Fig. 9.2A). The system demonstrated the formation of microvilli structures using Caco-2 cells and the possibility of coculture with intestinal microbes. Intestinal barrier function, catalytic activity of enzymes, mucus generation, and drug metabolism were improved compared with those of conventional static models. Shim et al. (2017) studied the effect of combining 3D villi structure with fluidics technology on the physiology of Caco-2 cells. Caco-2 cells were cultured

FIGURE 9.2 Microfluidic system-based gut models. (A) Schematic illustration of the microfluidic system-based gut-on-a-chip for mimicking dynamic flow and cyclic strain. (B) Schematic illustration of the microfluidics-based guton-a-chip system that combines the 3D villi structure model with a microfluidic system. The image was reproduced with permission (A) Kim, H.J., et al., 2012. Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow. Lab on a Chip 12 (12), 2165 2174; (B) Shim, K.-Y., et al., 2017. Microfluidic gut-on-a-chip with three-dimensional villi structure. Biomed. Microdevices 19 (2), 37.

Microfluidics-based gut models for mimicking the dynamic environment

under three different conditions: 2D monolayer culture on Transwells, 2D monolayer culture on microfluidic chips, and 3D culture on microfluidic chips (Fig. 9.2B). Various physiological functions were then compared, with a focus on absorptive and metabolic functions of the gut epithelium. Aminopeptidase activity was most enhanced in the monolayer culture under fluidic conditions. When the Caco-2 cells were cultured on 3D villi structures under fluidic conditions, the activity and absorptive permeability of P450 3A4 increased whereas the efflux transport decreased, suggesting that the combination of various environmental factors can contribute to significant changes in the gut cell functions. Although these studies have been successful in closely mimicking in vivo intestinal environment, most of them have used the Caco-2 cell line. As previously mentioned, Caco-2 cells originate from tumors and consequently harbor numerous mutations, making them unsuitable for genome-related research (Mochel et al., 2018). Organoids are another potential model system for simulating intestinal epithelial biology in a 3D matrix. However, the enclosed lumens that form in organoids are difficult to access, which can impede coculture with other cells, making it challenging to accurately mimic the intestinal environment (Bein et al., 2018; Rahmani et al., 2019). Recently, several studies have developed a human primary intestine model using dissociation of intestinal organoid in combination with microfluidic systems. Kasendra et al. (2018) isolated primary epithelial cells from intestinal biopsies, cultured them as organoids and then dissociated them by using enzymes. Next, intestinal stem cells were obtained and cultured on integrated porous membranes inside a microfluidic chip. In addition, primary human intestinal microvascular endothelial cells were cultured on the opposite side of the membrane, which facilitated faster confluency of the epithelial cells as well as increased the efficiency of monolayer formation and barrier function. Under cyclic movement and flow conditions, villi-like protrusions, multi-lineage differentiation, brush border enzyme activity, and secretion of mucin were demonstrated. Particularly, the results of transcriptome comparison indicated that the expression pattern of genes related to digestion, proliferation, and host defense mechanism is closer to the duodenum of the in vivo system as compared to that of static culture, intestine organoid, or Caco-2 cell line-based microfluidic systems. Workman et al. used human intestinal organoids that were derived from induced pluripotent stem cells (iPSCs) (Workman et al., 2018). The organoids derived from iPSCs contain epithelial and mesenchymal cells. Since the presence of mesenchymal cells disrupts the expansion of epithelial cells, the epithelial cells were sorted and seeded into the microfluidic chip. Under flow conditions, the villous-like projections, polarization of cells with brush borders, presence of Paneth cells, goblet cells, enterocytes, and entero-endocrine cells were demonstrated. The authors also demonstrated the response of the epithelial layer to interferon-gamma (IFN-γ), which is a cytokine associated with inflammatory bowel disease, through the upregulation of indoleamine 2,3-dioxygenase 1 and guanylate binding protein 1. Furthermore, the presence of IFN-γ and tumor necrosis factor-alpha caused an increase in permeability of the epithelial layer.

301

302

CHAPTER 9 Gut-on-a-chip microphysiological systems

This combination of organoid and microfluidic system provided a more physiologically and pathophysiologically realistic model than the Caco-2-based in vitro models. Through the addition of various components used in in vivo systems, the system will be able to mimic the in vivo system more closely and can be applied for drug screening and personalized medicine (Ronaldson-Bouchard and VunjakNovakovic, 2018; Ma et al., 2012). In this section, we summarized microfluidics-based models and the importance of a dynamic microenvironment. The combination of dynamic conditions with 3D structures provides synergetic effects on mimicking in vivo gut functions. The gut microenvironment is composed of a complex network of multiple factors. Elucidating how these environmental factors contribute to the physiology of the gut epithelium will be a critical task going forward.

First-pass metabolism models After food and drugs are administered orally, they undergo a process known as first-pass metabolism whereby they are absorbed and metabolically transformed in the gut and liver before entering systemic circulation (Nakanishi and Tamai, 2015). Depending on the outcome of first-pass metabolism, the effect of the administered drugs on target organs or tissues can vary. Therefore it is important to replicate the metabolic function of the gut and liver for drug screening applications. In one approach, the Transwell system was used to coculture Caco-2 and HepG2 cells, representing the gut and liver, respectively (Lau et al., 2004). Choi et al. (2004) compared Transwell-based coculture with monoculture and demonstrated the importance of coculture in reproducing first-pass metabolism. Their study showed the superiority of continuous perfusion-based coculture for HepG2 cell growth and metabolic enzyme activity. However, the Transwell system has limitations in replicating the time-dependent dynamics between the gut and liver. Because the gut, liver, and other organs are connected via blood vessels, absorbed food and drugs cross the intestinal epithelium and are then sequentially transferred to the liver for metabolization under dynamic conditions. Based on these limitations, researchers have proposed various models of microfluidics-based coculture for gut and liver cells to simulate first-pass metabolism. In one model, the microfluidic system consisted of chambers that represented the gut and liver, connected by microfluidic channels, whereby the flow of cell culture medium represented blood flow. The drugs were first injected into the gut chamber, where absorption and metabolism of the intestine were simulated. The drug molecules that passed through the gut cells entered the liver chamber, in which metabolic reactions were mimicked (Ghaemmaghami et al., 2012; Inamdar and Borenstein, 2011). Leclerc’s group combined cell culture inserts with a microfluidic system, in which Caco-2 cells were cultured in inserts and HepG2

First-pass metabolism models

cells were cultured in the microfluidic system (Bricks et al., 2014). To simulate first-pass metabolism, phenacetin was applied to and passed through the Caco-2 cell layer. It was then directed toward the HepG2 cells, where it was metabolized into acetaminophen. The transformation of phenacetin in the coculture system was greater than that in the monoculture system. Based on these observations, mathematical and pharmacokinetic models were developed to simulate first-pass metabolism and modeling-based drug development. However, immortalized cell lines have limitations in the replication of in vivo reactions. To address these limitations, Groothuis’ group integrated precision-cut slices, which were extracted from rats, into a microfluidic system (van Midwoud et al., 2010). Because cell culture medium continuously flowed from the first compartment containing intestinal slices to the second compartment containing liver slices, metabolites formed by the intestinal slices were delivered to the liver slices. Under these conditions, intestinal and liver metabolism were maintained for several hours. When the intestinal and liver slices were exposed to bile acid to simulate bile acid homeostasis, fibroblast growth factor 15 was expressed in the intestinal slices. In this case, the expression of CYP7A1 was lower than that when the liver slices were only exposed to bile acid, and the system was able to replicate in vivo regulation. One of the disadvantages of microfluidic systems is that an external pump is needed to control the flow. The presence of a pump causes unexpected problems for the operation of the culture system, such as bubble formation and bacterial contamination, making the culture system difficult to apply (Sung and Shuler, 2009). To overcome these issues, researchers have integrated external pumps within the culture system or developed alternatives without a pump altogether. Marx et al. integrated a peristaltic micropump into a microchip. Transwell inserts were used to culture human primary intestinal epithelial cells and the inserts were integrated with a microfluidic system to connect with liver spheroids (Maschmeyer et al., 2015). Shuler’s group proposed a pumpless solution and applied it to demonstrate a plug-and-play system (Esch et al., 2016). Gut epithelial cells and liver cells were cultured in separate chips, and the cells were combined after they were allowed to mature. Through an integrated passive valve system, unidirectional flow from the gut to the liver was demonstrated. Because these systems reflect the correct in vivo fluid-to-tissue ratio, they can be used to predict the effect of drugs using mathematical models. Moreover, the minimization of external systems enables the establishment of simple, user-friendly models and allows the implementation of organ-on-a-chip in diverse applications. The purpose of mimicking first-pass metabolism is to study the effect of drugs on cells. Researchers have proposed the addition of another chamber, connected to the gut and liver chambers via microfluidic channels, that could provide information on the effect of metabolites during first-pass metabolism on other tissues. Sato et al. focused on the effect of anticancer agents on tumor cells by culturing Caco-2, HepG2, and MCF-7 cells in a microsystem (Imura et al., 2010). The anticancer agent was transported to the MCF-7 cells after crossing the Caco-2 layer and being exposed to HepG2 cells. The weaker effect of the anticancer agent in

303

304

CHAPTER 9 Gut-on-a-chip microphysiological systems

the presence of liver metabolism, based on the viability of MCF-7 cells, was demonstrated. Furthermore, drugs were mixed with synthetic digestive juices to simulate gastrointestinal digestion. Gastric juices affected alkaline drugs such as Tegafur through neutralization and caused degradation (Imura et al., 2012). In a similar study, Fujii et al. integrated lung cells (L-2 or A549) or adipocytes (3T3-L1) into the system and demonstrated the effect of liver metabolism on the efficacy of drugs on cancer cells (Kimura et al., 2015). Shuler’s group developed an in vitro microscale cell culture analog where Caco-2, HepG2, and L2 cells were cultured and connected to investigate the effect of acetaminophen and 50nm carboxylated polystyrene nanoparticles on the body after experiencing firstpass metabolism (Esch et al., 2014; McAuliffe et al., 2008). To obtain accurate information on the effect of oral drug uptake, HT29-MTX cells were cocultured with the addition of a mucus layer or chyme mimic solution (Mahler et al., 2009). In this section, we reviewed the recent development of first-pass metabolism models. Microsystem-based in vitro models to predict drug efficacy and bioavailability have the potential to overcome the limitations of conventional drug screening applications. Although there is still room for improvement, these approaches will guide future development of in vitro models.

Gut microbe coculture models The human gut epithelium is populated with a complex community of gut microbiota. Recent studies have suggested that the gut microbiota play a significant role in maintaining health. Further, the alteration of the normal microbiota, also known as dysbiosis, has been associated with the development of several diseases (Wells et al., 2011). Although it is important to understand how these gut microbes interact with gut epithelium and affect human health, in vitro models for the study of these interactions have been lacking. Most previous research had to depend on the analysis of human fecal samples or the use of animal models. Recent progress in the development of novel in vitro models that incorporate both the gut epithelium and gut microbes holds great promise for future research in this area. Development of gut microbiota coculture systems requires a combination of microfluidic technology, cell culture techniques, microbiology, as well as various analytical techniques. More conventional approaches use a series of macroscale (hundreds of milliliters) reactors to culture gut cells and gut microbes to simulate the human gut (Marzorati et al., 2014; Payne et al., 2012). These macroscale systems offer the advantage of having a similar length scale to the human intestine; however, one disadvantage is that it is difficult to fine-tune the gut microenvironment. Recent progress in microfabrication technology includes the development of microscale systems that enable the coculture of gut cells and microbes. For

Gut microbe coculture models

example, Chi et al. (2015) introduced Salmonella bacteria to a microfluidic system with gut cells, and examined the adherence of the bacteria on the gut epithelium. A more detailed study on the interaction of the gut with microbes includes the use of a microfluidic chip that enables the independent culture of eukaryotic cells and bacteria by utilizing a pneumatically actuated system to segregate the two (Kim et al., 2010). Commensal Escherichia biofilm with eukaryotic HeLa cells was first formed, followed by the introduction of enterohemorrhagic E. coli into the system, basically mimicking the sequence of a gastrointestinal tract infection. More recently, gut bacteria and gut epithelium were cocultured for a prolonged period of time. A microfluidic gut chip exhibiting peristaltic movement was used to simulate the mechanical aspect of the gut microenvironment. By coculturing Caco-2 cells with Lactobacillus rhamnosus GG for more than 1 week, the interaction of gut microbe with gut epithelium in the context of disease was studied (Kim et al., 2012). This system is useful in relating the contribution of intestinal bacterial overgrowth and inflammation to mechanical deformation (Kim et al., 2016). A different approach focused more on the 3D tissue microenvironment of the gut. By mimicking the 3D villi structure of the human gut and coculturing Caco-2 cells with bacteria, including Salmonella, Pseudomonas, Staphylococcus, and Lactobacillus, investigators observed the establishment of microbial niches along the crypt villi axis, as well as the effect of probiotics on preventing invasion by pathogenic bacteria (Costello et al., 2014). This illustrates the diverse and complex nature of the gut microenvironment, and combinations of these microenvironment factors contribute to the interaction of gut microbes with gut epithelium. One important aspect of the gut microenvironment is the oxygen concentration, relating to the fact that a significant portion of the gut microbiota consists of anaerobic bacteria. Thus mimicking the gut environment inevitably involves the generation of a hypoxic environment for anaerobic bacteria, while maintaining a healthy oxygen environment for gut cells. Microfluidics can be an ideal solution for this, since it allows accurate control of flow and mass transfer. Shah et al. (2016), reported a modular, microfluidic-based model, termed HuMiX or humanmicrobial cross-talk, which enables the recapitulation of the human microbe interface in the gut. Authors used their system to coculture gut cells and commensal L. rhamnosus GG and obligate anaerobe Bacteroides caccae under anaerobic conditions. A simpler and clever approach was reported, in which 50-mL conical culture tubes were used to develop a coculture system under anaerobic conditions (Sadabad et al., 2015). In this system, liquid medium was placed on top of solid agar containing bacteria in a conical tube. During an 18 36 hour coculture experiment, it was observed that Caco-2 cells promoted growth and metabolism of Faecalibacterium prausnitzii, while F. prausnitzii suppressed inflammation and oxidative stress in Caco-2 cells. Although this approach is perhaps too simple to apply to a wide range of microbes and experimental conditions, it successfully shows the implications of gut microbe gut epithelium interaction in terms of maintaining health and homeostasis. More refined methods for controlling

305

306

CHAPTER 9 Gut-on-a-chip microphysiological systems

microenvironmental factors, including 3D topography, fluidic shear, and oxygen concentration, will provide more robust in vitro platforms to establish successful gut microbe cocultures. One of the major challenges that remain is to coculture diverse populations of gut microbiota rather than small sub-populations.

Conclusion In the present chapter, we reviewed the gut-on-a-chip systems for reproducing in vivo gut microenvironment. The gut microenvironment is extremely complex and involves mechanical stimuli, such as fluidic shear and peristaltic movements, a wide range of cell types, 3D structures, and gut microbiota. Recreating all these elements would be a daunting task, but previous studies have shown that even partially mimicking these environmental factors can have significant effects on the physiology of gut cells. Thus in the present chapter, we reviewed various in vitro systems developed for mimicking the in vivo environment more closely such as 3D structure-based culture systems, microfluidics-based systems, connections with liver organ models, and coculture with microbiome. Future studies should focus on overcoming the challenges of combining all microenvironmental factors such as 3D topology with flow and cyclic strains as well as coculture with other cells, organs, and microbiome. Furthermore, since the intestine is a large organ and some spatial variations exist along the entire organ, simulating the different segments of the intestine in the in vitro models is another challenge that needs to be overcome. In addition, the intestinal organoids were combined with previously developed gut-on-a-chip systems to take a step forward in reproducing the in vivo systems better. Nevertheless, there is still more potential for improvements such as the integration of other stem cells, immune cells, and microbiome and connection with other organ-derived organoid systems (Liu et al., 2016). These approaches will enable reproduction of the in vivo environment and will provide valuable information regarding the gut physiology and pathophysiology. Therefore the use of microphysiological system-based body-on-a chip models can facilitate the development of personalized medicine, therapies, and novel drugs.

Acknowledgments This work was supported by Ministry of Trade, Industry and Energy (MOTIE), Republic of Korea (10050154, Establishment of Infrastructure for industrialization of Korean Useful Microbes, R0004073) and Hongik University Research Fund. This work was supported by the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (MSIT) (2018R1C1B5085757), the Ministry of Education (2018R1A6A1A03024231), and the Research Institute of Engineering & Technology of Hanyang University.

References

References Abbott, A., 2003. Cell Culture: Biology’s New Dimension. Nature Publishing Group. Anselme, K., Bigerelle, M., 2011. Role of materials surface topography on mammalian cell response. Int. Mater. Rev. 56 (4), 243 266. Artursson, P., Palm, K., Luthman, K., 2001. Caco-2 monolayers in experimental and theoretical predictions of drug transport1. Adv. Drug Delivery Rev. 46 (1-3), 27 43. Bein, A., et al., 2018. Microfluidic organ-on-a-chip models of human intestine. Cell. Mol. Gastroenterol. Hepatol. 5 (4), 659 668. Blutt, S.E., et al., 2018. Engineered human gastrointestinal cultures to study the microbiome and infectious diseases. Cell. Mol. Gastroenterol. Hepatol. 5 (3), 241 251. Brandon, E.F., et al., 2003. An update on in vitro test methods in human hepatic drug biotransformation research: pros and cons. Toxicol. Appl. Pharmacol. 189 (3), 233 246. Bricks, T., et al., 2014. Development of a new microfluidic platform integrating cocultures of intestinal and liver cell lines. Toxicol. In Vitro 28 (5), 885 895. Chi, M., et al., 2015. A microfluidic cell culture device (μFCCD) to culture epithelial cells with physiological and morphological properties that mimic those of the human intestine. Biomed. Microdevices 17 (3), 58. Choi, S.H., et al., 2004. Enhanced cytochrome P450 capacities of Caco-2 and Hep G2 cells in new coculture system under the static and perfused conditions: evidence for possible organ-to-organ interactions against exogenous stimuli. Mater. Sci. Eng. C 24 (3), 333 339. Costello, C.M., et al., 2014. 3-D intestinal scaffolds for evaluating the therapeutic potential of probiotics. Mol. Pharmaceutics 11 (7), 2030 2039. Cushing, M.C., Anseth, K.S., 2007. Hydrogel cell cultures. Science 316 (5828), 1133 1134. Dotti, I., Salas, A., 2018. Potential use of human stem cell derived intestinal organoids to study inflammatory bowel diseases. Inflamm. Bowel Dis. 24 (12), 2501 2509. Esch, M.B., et al., 2012. On chip porous polymer membranes for integration of gastrointestinal tract epithelium with microfluidic ‘body-on-a-chip’devices. Biomed. Microdevices 14 (5), 895 906. Esch, M.B., et al., 2014. Body-on-a-chip simulation with gastrointestinal tract and liver tissues suggests that ingested nanoparticles have the potential to cause liver injury. Lab Chip 14 (16), 3081 3092. Esch, M.B., et al., 2016. Modular, pumpless body-on-a-chip platform for the co-culture of GI tract epithelium and 3D primary liver tissue. Lab Chip 16 (14), 2719 2729. Garrett, W.S., Gordon, J.I., Glimcher, L.H., 2010. Homeostasis and inflammation in the intestine. Cell 140 (6), 859 870. Ghaemmaghami, A.M., et al., 2012. Biomimetic tissues on a chip for drug discovery. Drug Discov. Today 17 (3 4), 173 181. Hubatsch, I., Ragnarsson, E.G., Artursson, P., 2007. Determination of drug permeability and prediction of drug absorption in Caco-2 monolayers. Nat. Protoc. 2 (9), 2111. Imura, Y., et al., 2009. A microfluidic system to evaluate intestinal absorption. Anal. Sci. 25 (12), 1403 1407. Imura, Y., Sato, K., Yoshimura, E., 2010. Micro total bioassay system for ingested substances: assessment of intestinal absorption, hepatic metabolism, and bioactivity. Anal. Chem. 82 (24), 9983 9988.

307

308

CHAPTER 9 Gut-on-a-chip microphysiological systems

Imura, Y., Yoshimura, E., Sato, K., 2012. Micro total bioassay system for oral drugs: evaluation of gastrointestinal degradation, intestinal absorption, hepatic metabolism, and bioactivity. Anal. Sci. 28 (3), 197 199. Inamdar, N.K., Borenstein, J.T., 2011. Microfluidic cell culture models for tissue engineering. Curr. Opin. Biotechnol. 22 (5), 681 689. Kaplan, S.A., Cotler, S., 1972. Use of cannulated everted intestinal sac for serial sampling as a drug absorbability (permeability) screen. J. Pharm. Sci. 61 (9), 1361 1365. Kasendra, M., et al., 2018. Development of a primary human small Intestine-on-a-Chip using biopsy-derived organoids. Sci. Rep. 8 (1), 2871. Kim, H.J., Ingber, D.E., 2013. Gut-on-a-chip microenvironment induces human intestinal cells to undergo villus differentiation. Integr. Biol. 5 (9), 1130 1140. Kim, H.J., et al., 2012. Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow. Lab Chip 12 (12), 2165 2174. Kim, H.J., et al., 2016. Contributions of microbiome and mechanical deformation to intestinal bacterial overgrowth and inflammation in a human gut-on-a-chip. Proc. Natl. Acad. Sci. U.S.A. 113 (1), E7 E15. Kim, J., Hegde, M., Jayaraman, A., 2010. Co-culture of epithelial cells and bacteria for investigating host pathogen interactions. Lab Chip 10 (1), 43 50. Kimura, H., et al., 2008. An integrated microfluidic system for long-term perfusion culture and on-line monitoring of intestinal tissue models. Lab Chip 8 (5), 741 746. Kimura, H., et al., 2015. An on-chip small intestine liver model for pharmacokinetic studies. J. Lab. Autom. 20 (3), 265 273. Koppes, A.N., et al., 2016. Complex, multi-scale small intestinal topography replicated in cellular growth substrates fabricated via chemical vapor deposition of Parylene C. Biofabrication 8 (3), 035011. Lampen, A., et al., 1998. Metabolism and transport of the macrolide immunosuppressant sirolimus in the small intestine. J. Pharmacol. Exp. Ther. 285 (3), 1104 1112. Lasker, J., Rickert, D.E., 1978. Absorption and glucuronylation of diethylstilbestrol by the rat small intestine. Xenobiotica 8 (11), 665 672. Lau, Y.Y., et al., 2004. Evaluation of a novel in vitro Caco-2 hepatocyte hybrid system for predicting in vivo oral bioavailability. Drug Metab. Dispos. 32 (9), 937 942. Lee, S.H., Sung, J.H., 2018. Organ-on-a-chip technology for reproducing multiorgan physiology. Adv. Healthcare Mater. 7 (2), 1700419. Lee, S.H., et al., 2016. Microtechnology-based organ systems and whole-body models for drug screening. Biotechnol. J. 11 (6), 746 756. Li, N., et al., 2013. Development of an improved three-dimensional in vitro intestinal mucosa model for drug absorption evaluation. Tissue Eng., C: Methods 19 (9), 708 719. Liu, F., et al., 2016. Drug discovery via human-derived stem cell organoids. Front. Pharmacol. 7, 334. Ma, L., et al., 2012. Towards personalized medicine with a three-dimensional microscale perfusion-based two-chamber tissue model system. Biomaterials 33 (17), 4353 4361. Mahler, G.J., et al., 2009. Characterization of a gastrointestinal tract microscale cell culture analog used to predict drug toxicity. Biotechnol. Bioeng. 104 (1), 193 205. Marzorati, M., et al., 2014. The HMIt module: a new tool to study the host-microbiota interaction in the human gastrointestinal tract in vitro. BMC Microbiol. 14 (1), 133.

References

Maschmeyer, I., et al., 2015. Chip-based human liver intestine and liver skin cocultures A first step toward systemic repeated dose substance testing in vitro. Eur. J. Pharm. Biopharm. 95, 77 87. McAuliffe, G.J., et al., 2008. Development of a gastrointestinal tract microscale cell culture analog to predict drug transport. Mol. Cell. Biomech. 5 (2), 119. van Midwoud, P.M., et al., 2010. A microfluidic approach for in vitro assessment of interorgan interactions in drug metabolism using intestinal and liver slices. Lab Chip 10 (20), 2778 2786. Mochel, J.P., et al., 2018. Intestinal stem cells to advance drug development, precision, and regenerative medicine: a paradigm shift in translational research. AAPS J. 20 (1), 17. Nakanishi, T., Tamai, I., 2015. Interaction of drug or food with drug transporters in intestine and liver. Curr. Drug Metab. 16 (9), 753 764. Nicholson, J.K., Holmes, E., Wilson, I.D., 2005. Gut microorganisms, mammalian metabolism and personalized health care. Nat. Rev. Microbiol. 3 (5), 431. Olson, H., et al., 2000. Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul. Toxicol. Pharm. 32 (1), 56 67. Payne, A.N., et al., 2012. Advances and perspectives in in vitro human gut fermentation modeling. Trends Biotechnol. 30 (1), 17 25. Rahmani, S., et al., 2019. Intestinal organoids: a new paradigm for engineering intestinal epithelium in vitro. Biomaterials 194, 195 214. Reis, J.M., Sinko, B., Serra, C.H.R., 2010. Parallel artificial membrane permeability assay (PAMPA)—is it better than Caco-2 for human passive permeability prediction? Mini Rev. Med. Chem. 10 (11), 1071 1076. Ronaldson-Bouchard, K., Vunjak-Novakovic, G., 2018. Organs-on-a-chip: a fast track for engineered human tissues in drug development. Cell Stem Cell 22 (3), 310 324. Round, J.L., Mazmanian, S.K., 2009. The gut microbiota shapes intestinal immune responses during health and disease. Nat. Rev. Immunol. 9 (5), 313. Sadabad, M.S., et al., 2015. A simple coculture system shows mutualism between anaerobic faecalibacteria and epithelial Caco-2 cells. Sci. Rep. 5, 17906. Sanderson, I.R., 1999. The physicochemical environment of the neonatal intestine. Am. J. Clin. Nutr. 69 (5), 1028s 1034s. Sato, T., et al., 2009. Single Lgr5 stem cells build crypt villus structures in vitro without a mesenchymal niche. Nature 459 (7244), 262. Schirmer, M., et al., 2018. Dynamics of metatranscription in the inflammatory bowel disease gut microbiome. Nat. Microbiol. 3 (3), 337. Shah, P., et al., 2016. A microfluidics-based in vitro model of the gastrointestinal human microbe interface. Nat. Commun. 7, 11535. Shim, K.-Y., et al., 2017. Microfluidic gut-on-a-chip with three-dimensional villi structure. Biomed. Microdevices 19 (2), 37. Shin, W., Kim, H.J., 2018. Pathomimetic modeling of human intestinal diseases and underlying host-gut microbiome interactions in a gut-on-a-chip. Methods in Cell Biology. Elsevier, pp. 135 148. Sun, H., et al., 2008. The Caco-2 cell monolayer: usefulness and limitations. Expert Opin. Drug Metabol. Toxicol. 4 (4), 395 411. Sung, J.H., Shuler, M.L., 2009. Prevention of air bubble formation in a microfluidic perfusion cell culture system using a microscale bubble trap. Biomed. Microdevices 11 (4), 731 738.

309

310

CHAPTER 9 Gut-on-a-chip microphysiological systems

Sung, J.H., et al., 2011. Microscale 3-D hydrogel scaffold for biomimetic gastrointestinal (GI) tract model. Lab Chip 11 (3), 389 392. Wang, L., et al., 2009. Influence of micro-well biomimetic topography on intestinal epithelial Caco-2 cell phenotype. Biomaterials 30 (36), 6825 6834. Wang, L., et al., 2010. Synergic effects of crypt-like topography and ECM proteins on intestinal cell behavior in collagen based membranes. Biomaterials 31 (29), 7586 7598. Wang, Y., et al., 2014. In vitro generation of colonic epithelium from primary cells guided by microstructures. Lab Chip 14 (9), 1622 1631. Wang, Y., et al., 2017. A microengineered collagen scaffold for generating a polarized crypt-villus architecture of human small intestinal epithelium. Biomaterials 128, 44 55. Wang, Y., et al., 2018. Formation of human colonic crypt array by application of chemical gradients across a shaped epithelial monolayer. Cell. Mol. Gastroenterol. Hepatol. 5 (2), 113 130. Wells, J.M., et al., 2011. Epithelial crosstalk at the microbiota mucosal interface. Proc. Natl. Acad. Sci. U.S.A. 108 (Supplement 1), 4607 4614. Workman, M.J., et al., 2018. Enhanced utilization of induced pluripotent stem cell derived human intestinal organoids using microengineered chips. Cell. Mol. Gastroenterol. Hepatol. 5 (4), 669 677. e2. Yeon, J.H., Park, J.-K., 2009. Drug permeability assay using microhole-trapped cells in a microfluidic device. Anal. Chem. 81 (5), 1944 1951. Yin, X., et al., 2016. Engineering stem cell organoids. Cell Stem Cell 18 (1), 25 38. Yu, J., et al., 2012. In vitro 3D human small intestinal villous model for drug permeability determination. Biotechnol. Bioeng. 109 (9), 2173 2178.

CHAPTER

Computational pharmacokinetic modeling of organ-on-chip devices and microphysiological systems

10

Andrzej Przekwas and Mahadevabharath R. Somayaji CFD Research Corporation, Huntsville, AL, United States

Introduction Drug discovery and development protocols The current drug discovery and development process, conceived more than 50 years ago, remains a costly, slow, inefficient, and risk-laden process that delivers products with problematic safety and efficacy (Paul et al., 2010; Morgan et al., 2018; Somayaji, 2016). Approximately 95% of drugs that enter clinical trials postanimal testing do not make it to the market because of the significant phylogenetic distance between laboratory animals and humans. Furthermore, the total number of new drugs commercialized in recent years has not matched the rise in pharmaceutical research and development spending, from $108B in 2006 to $141B in 2015 (Schuhmacher et al., 2016). Reducing the high product attrition rates remains a key challenge for the pharmaceutical industry, and most attrition occurs in full clinical trials (Kola and Landis, 2004). Over the past two decades, the number of failures of smallmolecule drug candidates owing to poor pharmacokinetics (PK) or bioavailability has diminished significantly, while the lack of efficacy and low margins of safety became the major causes of Phase II and III attrition (Kola and Landis, 2004; Paul et al., 2010). Too often, the toxicity of a candidate drug is identified during clinical trials and not in early drug screening and preclinical studies (Maxmen, 2011). Current preclinical drug testing relies on conventional in vitro static cell culture assays of immortalized or primary cell lines because of their robustness, simplicity, and high-throughput operation. Two-dimensional (2D) in vitro cell cultures and barriers have been used successfully to evaluate drug metabolism, intrinsic clearance, and transport, but are inadequate for testing the efficacy or toxicity of new drug candidates. Three-dimensional (3D) in vitro cell cultures Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00011-5 © 2020 Elsevier Inc. All rights reserved.

311

312

CHAPTER 10 Computational pharmacokinetic modeling

have been developed to investigate spatial morphology and concentration gradients, and examples include cancer cell spheroids (Hirschhaeuser et al., 2010; Kondo et al., 2011; Fennema et al., 2013; Kwapiszewska et al., 2014) and hepatocyte spheroids (Tostoes et al., 2012; Wang et al., 2015; Bell et al., 2016; Tong et al., 2016). Compared with 2D cell cultures, 3D tumor spheroid models have shown reduced efficacy for several antineoplastic agents, similar to observations in clinical trials (Edmondson et al., 2014). However, these cell cultures are not accurate models of the organ microenvironments, where multiple cell types arranged in a complex 3D architecture receive and transmit a host of biochemical, mechanical, and physiological cues. Prior to clinical trials, drug candidates are routinely evaluated in animals to determine the safe dose for first-in-human studies and assess the compound’s safety and efficacy profiles. Animal studies also are used to evaluate the mechanism of action, discover new drug targets and biomarkers, and elucidate wholebody PK parameters, such as absorption, distribution, metabolism, and excretion, as well as pharmacodynamic (PD) and toxicity data (McGonigle and Ruggeri, 2014). Several preclinical systems—ranging from zebrafish to mice to primates— are used to identify drug safety and efficacy in animal models that exhibit different aspects of human diseases such as cancer, diabetes, and autoimmune conditions. Because of considerable interspecies differences in anatomy and physiology, direct extrapolation of drug response from animals to humans remains controversial (Olson et al., 2000; Sharma and McNeill, 2009; Fine and VunjakNovakovic, 2017). Simplified allometric scaling based on body mass is also inadequate because of the inherent anatomical, physiological, and genetic differences (e.g., life span, heart and respiration rate, metabolism, immunity). Extrapolation of in vitro data to humans is even more challenging. In vitro dissolution and cell culture devices are used for in vitro-to-in vivo correlation of compound properties or extrapolation of PK parameters (e.g., hepatic clearance) to in vivo PK parameters (FDA, 1997; Rostami-Hodjegan, 2012), but they are not adequate during quantitative translation to human PK and PD responses (Danhof et al., 2008). New computational tools are needed for translation of data from physiologically interconnected organ-on-chip (OoC) devices to the human in vivo situation (Stokes et al., 2015).

Organ-on-chip technology and prospects for human-on-a-chip devices To address these drug discovery and development challenges, U.S. research agencies such as the Defense Advanced Research Projects Agency (DARPA), the National Institutes of Health (NIH), the Food and Drug Administration (FDA), and the Defense Threat Reduction Agency have initiated and sponsored multimillion-dollar research programs to aid in the development of novel preclinical models for generating human-relevant drug data. These initiatives have led to

Introduction

the development of preclinical microfluidic microphysiological systems (MPS) for predicting human-relevant responses to drugs (Hartung and Zurlo, 2012; Marx et al., 2012; Sutherland et al., 2013). An MPS comprises several in vitro OoC devices linked into a single human-on-a-chip (HoC) system (Fig. 10.1). While HoC systems are primarily intended for characterizing and testing new compounds, the ultimate goal of an HoC system is to demonstrate the predictive capability of in vitro-to-in vivo (human) extrapolation of drug efficacy or toxicity. Typically, FDA-approved drugs with well-characterized in vitro and in vivo behaviors are used in HoC systems for calibration before testing the PK properties of new compounds. While in vitro testing using multiwell plates is customary in the drug discovery pipeline, the use of in vitro microfluidic HoC technology for predicting clinical drug responses is a recent undertaking (Marx et al., 2012; Gao et al., 2013; Baudoin et al., 2013; Somayaji et al., 2016, 2018). Several projects have been awarded to develop various organ devices, and selected teams [Wyss Institute, Massachusetts Institute of Technology (MIT), Los Alamos National Laboratory, and US academia] were tasked with developing a microfluidic platform integrating multiple organs into an MPS (Huh et al., 2010; 2012; Bhatia, and Ingber, 2014; Capulli et al., 2014; Ebrahimkhani et al., 2014; Esch et al., 2015; Wikswo and Porter, 2015). The teams fabricate and test individual OoC devices and stem cell biology tools for functionalizing the OoC devices with human-derived cells (Fabre et al., 2014).

Computational models for in vitro and in vivo pharmacology Over the past few decades, mathematical models have gained popularity at various stages of the drug discovery and development process. Simulations of molecular dynamics are now routinely used in computer-aided drug design, target identification, and prediction of compound physicochemical properties (Durrant and McCammon, 2011; Borhani and Shaw, 2012; Zhao and Caflisch, 2019). Continuum-based models at various scales, ranging from compartmental to distributed high-fidelity (spatiotemporal- 1D/2D/3D), are being used during drug development to support aspects of product formulation, dissolution, absorption, distribution, metabolism, excretion, PK, PD, and toxicity (Jusko, 2013; Visser et al., 2014; Knight-Schrijver et al., 2016; Pinero et al., 2018). Most mathematical models used for simulating in vitro transport or metabolism (e.g., permeability and hepatic metabolic clearance), as well as in vivo PK, use simple onecompartment lumped formulations. These models are typically described with ordinary differential equations using lumped (or averaged) quantities to represent complex biophysical and biological processes. For example, well-established physiologically based PK (PBPK) models represent the whole body, with a small number of compartments representing individual organs connected to arterial and venous compartments. Organ models are typically represented as well-stirred reactors with drugs partitioned between blood and tissue subcompartments and assuming local equilibrium. For organs involving epithelial barriers with large

313

FIGURE 10.1 Example of various organ-on-a-chip devices and their linking into a microphysiological or human-on-a-chip system. Adapted from Huh, D., Hamilton, G.A., Ingber, D.E., 2011. From 3D cell culture to organs-on-chips. Trends Cell Biol. 21 (12), 745754 (Huh et al., 2011).

Introduction

FIGURE 10.2 Physical and mathematical representations of an organ-on-a-chip device and in vivo organ tissue barriers and the corresponding schematic of pharmacokinetic models.

surface area, gastrointestinal tract in particular, a number of spatially distributed compartments are used for more accurate simulation of drug transit and absorption (Yu and Amidon, 1999; Huang et al., 2009; Agoram et al., 2001; Jamei et al., 2009; Sjogren et al., 2016). Fig. 10.2 shows typical compartmental and mathematical representations of an in vitro organ tissue barrier and an in vivo PBPK model used for whole-body human PK studies. The main advantages of these mathematical models are their simplicity, computational speed, and small number of parameters to calibrate. Calibration of in vivo PK and PBPK models requires collection of drug, metabolite, and biomarker data from the blood, body fluids (e.g., urine), and in some cases, from relevant tissue/organs at several time points during the drug biodistribution and elimination phases. Some of the parameters in in vivo wholebody PBPK models can be estimated from in vitro experiments, complementary mathematical models, and in vitro-to-in vivo correlation protocols. For example, specialized in vitro permeability assays are used to estimate in vivo permeability (Hubatsch et al., 2007; Buckley et al., 2012; Larregieu and Benet, 2014; Bittermann and Goss, 2017), and in vitro hepatocyte cell cultures are used to estimate in vivo hepatic clearance (Zuegge et al., 2009; Chiba et al., 2009; Bowman and Benet, 2016; Wood et al., 2017). Over the past decade, a new approach called, quantitative systems pharmacology, was introduced to integrate computational disciplines such as systems biology, physiology, PK, PD, disease progression, genetics, biomarker kinetics, and in vitro-to-in vivo extrapolation

315

316

CHAPTER 10 Computational pharmacokinetic modeling

(Geerts et al., 2013; Visser et al., 2014; Nijsen et al., 2018; Hartmanshenn et al., 2018). Quantitative systems pharmacology also relies on advanced computational methods and algorithms such as multiscale modeling, machine learning, artificial intelligence, and cloud computing. Multiscale modeling is particularly attractive for combining lumped (compartmental) and distributed (spatiotemporal) first principlesbased mathematical models for capturing the intricate biophyiscal details of selected organs or tissues (e.g., lung airway barrier, hepatic sinusoid, cancer) (Sluka et al., 2016; Kannan et al., 2018; Norton et al., 2019; Kalra et al., 2019). First principlesbased models are especially suitable for modeling engineered in vitro devices such as cell cultures, Transwell inserts, and microfluidic and OoC devices for which the geometry, media, morphology, and cellular responses are well described and monitored in real time. This chapter outlines the current status and prospects for application of multiscale mathematical models in the development of OoC devices and multiorgan MPS, as well as their use for translation of in vitro results to humans.

Computational models for designing organ-on-a-chip devices Modeling transport in conventional in vitro cell cultures Conventional in vitro static cell cultures, including plated cultures and Transwell inserts, have been used to evaluate various aspects of drug PK and PD at the cellular level. The experimental data collected from these measurements are typically used to calibrate mathematical models, which in concert with “scaling” paradigms can then be used for extrapolating the results to humans. For example, in vitro hepatocyte cultures are typically used to estimate the in vitro enzymatic intrinsic clearance, Clint, and then to calculate the in vivo hepatic metabolic clearance, ClH, with well-stirred or plug flow (or tube) “reactor” models of the liver (Pelkonen and Turpeinen, 2007; Kratochwil et al., 2017). Due to the relative geometrical simplicity of the conventional in vitro cell culture designs, simple one-compartment mathematical models using ordinary differential equations can be used to represent relevant physicochemical and biological processes (Sun and Pang, 2008; Nagar et al., 2014). Fig. 10.3 presents a schematic of a conventional Transwell cell culture permeability barrier and an equivalent compartmental mathematical model to calculate drug transport from the apical (donor) compartment through the cellular barrier to the basolateral (receiver) compartment. A detailed description of the conventional lumped parameter and the first principlesbased models of drug transport across the barrier as follows. A compartmental mathematical model describing drug transport and metabolism can be derived using general mass balance equations with all flux terms integrated over the compartment boundaries. The resulting equations are expressed in

Computational models for designing organ-on-a-chip devices

FIGURE 10.3 Schematic of an in vitro Transwell cell culture barrier and a corresponding compartmental mathematical model of the barrier.

the form of a set of ordinary differential equations representing the species mass balance in each compartment. For the barrier model (Fig. 10.3B), the mass balance equations for the apical (a), cellular (c), and basolateral (b) compartments can be written as: dCa eff 5 2 Ja2c 1 Jc2a dt

(10.1)

dCc eff 5 Ja2c 2 Jc2b 2 Jc2a 2 Clc UCc dt

(10.2)

dCb 5 Jc2b dt

(10.3)

Va Vc

Vb

where Ci and Vi (i 5 a, c, b) represent the species concentration and the volume of individual compartments, and Ji2j describe mass fluxes from compartment i to j. These fluxes include contributions from both neutral and ionic parts of the efflux drug. Jc2a represents the active efflux from inside the cell to the apical compartment and Clc is the intracellular compound clearance rate, for example, hepatic metabolism, which is typically expressed using the MichaelisMenten kinetics model (Wagner et al., 1985): Clc 5

Vmax Km 1 Cc

(10.4)

where Vmax represents the maximum reaction rate and the Michaelis constant Km is the compound concentration at which the reaction rate is half of Vmax. An improved model accounting for the spatial resolution of the barrier involving compound partition into the cell membrane and binding to intracellular lipids and organelles, has been demonstrated (Kulkarni et al., 2016). The flux terms in the above compartmental mass balance equations can be expressed either in a lumped parameter form or using first principlesbased formulation. For example, the flux term between the apical (a) and cellular (c) compartments for both formulations can be expressed as:

317

318

CHAPTER 10 Computational pharmacokinetic modeling

Lumped: Ja2c 5 ka2c ðCa 2 Cc Þ

(10.5)

Ja2c 5 APeff :a2c ðCa 2 Cc Þ

(10.6)

Physics-based:

where ka2c is the lumped kinetic coefficient (termed diffusive clearance rate). The effective permeability in Eq. (10.6) can be calculated to account for compound permeation from the apical medium to the membrane, P1, and from the membrane to the cellular space, P2: Peff 5

P1 P2 P1 1 P2

(10.7)

and P1 5

δM kDM

1 1

δ1 D1

and

P2 5

δM kDM

1 1

δ2 D2

(10.8)

where δ1, δm, δ2 are effective thicknesses of the apical, membrane, and cellular layers (e.g., unstriped layers), and D1, Dm, D2 are the diffusion coefficients in each layer, respectively. Note that to calibrate the ka2c parameter in the lumped model, experimental data are needed for a specific barrier, while the physics flux model can be constructed using the barrier and semiempirical models of the barrier geometry/morphology and diffusion coefficient models (e.g., StokesEinstein model) (Peskir, 2003). In vitro cell-barrier experimental data are often used to estimate the effective (or apparent) permeability using the mass balance equation for the receiver (basolateral) compartment: Vb

dCb 5 2 APeff ðCb 2 Ca Þ dt

(10.9)

In these experiments the donor concentration is typically assumed to be constant and the receiver concentrationtime, Cb(t), is measured. If the initial Cb equals zero, we can approximate the time derivative and calculate the no-sink condition Peff: Vb

Cb 2 0 5 2 APeff ðCb 2 Ca Þ t

.

Peff 5

Vb Cb AUtðCa 2 Cb Þ

(10.10)

This expression can be further simplified if during the measurement period (0 2 t) a “sink” condition is assumed: CacCb. The above compartmental model can be extended to simulate pH-dependent drug ionization, accumulation in the barrier, effects of various transports, drug binding to Transwell walls, and other effects (Kra¨mer 2016). For compounds exhibiting rapid transport across the barrier, the model should account for concentration gradients in the apical and basolateral volumes. In such

Computational models for designing organ-on-a-chip devices

cases, a one-dimensional model is required, in which Eqs. (10.1)(10.3) are expressed as partial differential equations:   @C @C @C 5 D 2S @t @x @x

(10.11)

where the term S represents the generalized sink (e.g., metabolic clearance) term.

Modeling metabolic support in in vitro cell cultures Oxygen and glucose supply is an important factor for the growth and viability of cells in in vitro mammalian cultures, multilayered cell barriers, cell spheroids, and 3D tissue constructs (Sumaru et al., 2007). However, the chemical composition of typical cell culture medium deviates from physiological values. For example, Dulbecco’s modified Eagle’s medium contains 25 mM glucose, which is four times the physiological level (McKee and Komarova, 2017). Oxygen levels can affect cellular growth and differentiation, enzyme expression, and culture physiology such as liver zonation (Camp and Capitano, 2007; Thomas et al., 2011). Oxygen levels in cultured cells depend on the culture type: in shallow-medium Transwell inserts and plated cultures, oxygen levels can be controlled by adjusting the gas (O2, CO2) concentration levels in the incubator. In microfluidic devices, oxygen can be supplied through the package material and in dissolved form through the perfusing medium. The commonly used polydimethylsiloxane (PDMS)-based fabrication material in microfluidic devices offers high oxygen solubility and excellent oxygen supply via passive diffusion (Thomas et al., 2011). In static cell cultures, oxygen and glucose concentrations in the medium and inside the cells can be calculated using the mass balance equations (Eq. 10.11). The sink term is zero in the medium and finite in the cell layer, where it describes the oxygen and glucose metabolic rates. The oxygen consumption rate in a cell culture layer can be calculated as Sox 5

Vmax;ox ρcell UCox Km;ox 1 Cox

(10.12)

where Cox is the oxygen concentration in the cell layer, and ρcell 5 Ncells/Vol is the volumetric cell density, and Ncells is the number of cells in the culture volume. The maximum respiration rate per cell, Vmax,ox 5 qox 3 ρcell, can be estimated from oxygen consumption per cell, qox. For a typical coculture of hepatocytes and human umbilical vein endothelial cells, Vmax,ox can be estimated as 3.8 3 10216 mol/cell/s and 4 3 10217 mol/cell/s, respectively. The corresponding Michaelis constants (Km,ox 5 SlK0 m,ox) using the solubility are K0 m 5 5.6 mmHg for hepatocytes and K0 m 5 0.5 mmHg for human umbilical vein endothelial cells (Allen and Bhatia, 2003; Kim et al., 2013). Fig. 10.4 shows an example computational domain for one-dimensional transport of oxygen in an in vitro static cell culture and example simulation results of cell oxygen partial pressures for various

319

320

CHAPTER 10 Computational pharmacokinetic modeling

FIGURE 10.4 Schematic of a computational domain for one-dimensional simulations of an oxygen diffusion reaction in an in vitro static cell culture and the calculated cell layer pO2 as a function of medium height.

medium heights calculated using CoBi tools (http://medicalavatars.cfdrc.com). The parameters for the model were taken from previously reported simulations (Ng et al., 2014). Unlike oxygen, in static cell cultures, glucose must be supplied in the medium. For long-term cultures, the medium is periodically replaced to supply fresh nutrients, remove waste products, and collect samples for analysis. Typical glucose concentration (Cg) in fresh Dulbecco’s modified Eagle’s medium is approximately 4.5 g/l. Several mathematical models of glucose metabolism in cell cultures have been developed to account for glycolysis and oxidative phosphorylation at varying resolutions (Cortassa and Aon, 1994; Bertram et al., 2006; Chalhoub et al., 2007; Yang et al., 2015). The simplest approach is a global, one-step model for glucose and lactate mass balance equations (Eq. 10.12) with the sink terms as follows: Sg 5

Vmax;g ρcell UCg Km;g 1 Cg

(10.13)

A simple global anaerobic glucose reaction stoichiometry relationship can be used to calculate the lactate (l) production rate as Sl 5 2rl 3 Sg with rl 5 2 as the stoichiometric factor (the typical lactate-to-glucose ratio is 1.61.7), but a more frequently used approach combines the effects of both the anaerobic and aerobic metabolic pathways, and the lactate production rate (Sengers et al., 2005; Molavian et al., 2009; Burova et al., 2019) is calculated as 1 Sl 5 2 2Sg 1 Sox 3

(10.14)

The metabolic cell support models could be extended to simulate cell culture proliferation during the functionalization step (Lewis et al., 2005; Ma et al., 2007; Fadda et al., 2012). However, in most cases, drug efficacy testing is typically performed with fully functionalized (nonproliferating) cell cultures. The exceptions are tests of compound toxicity, infection, and cancer cell cultures where the number of viable cells changes over time.

Computational models for designing organ-on-a-chip devices

Modeling perfusion in organ-on-a-chip devices Microfluidic OoC devices should be able to reproduce the in vivo hydromechanical effects in vitro. Blood, interstitial, and mucosal fluid mechanics influence multiple physiological, biochemical, and pharmacological processes in living organisms. Furthermore, fluid flows directly affect the structure of the lymphatic vasculature and mucosal barriers, as well as cell morphology and cellular signaling (Huber et al., 2018). Contemporary computational fluid dynamics tools, such as ANSYS Fluent CFD-ACE 1 , COMSOL, and CoBi, provide the necessary capabilities for high-fidelity modeling of coupled fluid flow, mass transport, and biochemistry in various types of microfluidic devices (Przekwas et al., 2000a,b, 2006; Erickson, 2005; Kockmann et al., 2006; Barber and Emerson, 2008; Glatzel et al., 2008; Patrachari et al., 2012; Dereli-Korkut et al., 2014; Wu et al., 2015). Microfluidically perfused in vitro cell cultures and OoC devices can be modeled computationally with various levels of fidelity, ranging from a simple multi-compartmental approach to distributed high-fidelity 3D multiphysics models. While high-fidelity models may be more appropriate during OoC design for analyzing flow patterns, pressure drops, wall shear stress profiles, mechanical loads on membranes (Anderson and Knothe Tate, 2007), reduced-order models are more suitable for modeling long-term drug transport, and PK and PD effects. This section describes high-fidelity and multiscale models for designing and developing microfluidic OoC devices. High-fidelity simulations of microfluidic devices require a computational mesh generated from device geometry. Fig. 10.5 shows a few relevant examples of microfluidic OoC devices and representative 3D computational domains. As typical microfluidic devices have relatively simple geometries, it is not difficult to generate a quality hexahedral computational mesh for computational fluid dynamics simulations. The geometry/mesh models require specification of volume conditions (e.g., epi-channel, endo-channel, porous membrane, permeable sold, cell layer) and boundary conditions (e.g., inlets, outlets, walls, interfaces). OoC tissue-barrier devices typically involve membranes, which are functionalized by epithelial and endothelial cells. In barrier-type OoC devices, computational models should not only resolve the barrier thickness but also adjacent layers of epithelial and endothelial cell cultures, which are important for PK/PD studies. Initial physics-based design of fluidic devices can be conducted using simple algebraic relations and dimensionless numbers such as the Reynolds number, Re, which is the ratio of inertial forces to viscous forces: Re 5

ρudH μ

with

dH 5

4A Lw

(10.15)

where ρ is the fluid density, μ is the dynamic viscosity, u is the characteristic velocity, dH is the hydraulic diameter, A is the channel cross-sectional area, and Lw is the wetted perimeter of the channel. In OoC microchannels, the flow is

321

322

CHAPTER 10 Computational pharmacokinetic modeling

FIGURE 10.5 Examples of microfabricated organ-on-a-chip devices and computational domains for fluid flow and drug transport simulations.

typically laminar with dominant viscous forces (Re , 1) with a characteristic parabolic velocity profile in stream-wise direction, x. Flow in a tube of diameter d can then be expressed as: 8Q ux ðrÞ 5 2 πd

 2 ! 2r 12 d

(10.16)

where Q is the volumetric flow rate and r is the radial distance from the axis. The average velocity in the tube can be calculated as u¯ 5 Q/A. For a rectangular channel, the velocity profile in the stream-wise direction u(y,z) is expressed in the form of a Fourier series expansion (Tanyeri et al., 2011). Fig. 10.6 shows typical velocity profiles in rectangular microchannels with different channel heights. The analytical function for the velocity profile can be further used to calculate the wall shear stress, τ wall 5 2μdu(r)/dr |r5d/2, a quantity important for evaluating shear forces acting on biological cells attached to the channel walls. The pressure drop in the channel, Δp, for a given flow rate, Q, can be calculated as Q 5 Δp/R, where R is the hydraulic resistance. For a fully developed flow in a channel, it can be calculated using the HagenPoiseuille equation (Huber et al., 2018): R5

128μLx πdH4

(10.17)

Computational models for designing organ-on-a-chip devices

FIGURE 10.6 Schematic illustration of velocity profiles in a cross section of a rectangular microchannel where width is greater than height (W . H).

where Lx is the channel length. For a rectangular microchannel with a width of W and a height of H, the hydraulic resistance can be calculated using the following analytical relationship (Tanyeri et al., 2011): R5

12μLx  ;H ,W H WH 3 1 2 0:63 W

(10.18)

The above relationships yield an analytical breakdown of the pump requirements, pressure loads on membranes, and, using Kirchhoff network principles, flows in channel networks. Computational fluid dynamics tools are needed for more accurate and dynamic simulations. An in-depth discussion of fluid mechanics parameters in microfluidic devices has been given by Esch and Mahler (2019). To simulate fluid flow and mass transport in microfluidic devices, computational fluid dynamics tools solve the NavierStokes equations in the form of continuity and momentum equations. The mass conservation (continuity) equation:

-

  @ρ 1 rU ρ u 5 2 m_ f 2s @t

(10.19)

where u is the velocity vector and m_ f 2s is the mass transfer rate between the fluid and solid. The fluid medium momentum conservation equations:   @ρ u -1 rU ρ u u 5 2 rp 1 rUτ f 1 ρ g 1 F f 2s @t

-

(10.20)

where p is the fluid pressure, τ is the viscous-stress tensor, g is the gravity vector, and, for fluids carrying particles or cells, F f 2s is the fluid-particle momentum exchange source/sink term.

323

324

CHAPTER 10 Computational pharmacokinetic modeling

Species conservation equations:   @ρYi 1 rU ρ u Yi 5 rUðDi rYi Þ 2 SYi i 5 1; . . .; N @t

(10.21)

where Yi is the mass fraction of the ith constituent of the medium with N being the total number of constituents, Di is the diffusion coefficient, and SYi is the generalized sink term from chemical reactions and medium-particle mass exchange. In many applications, transport of very small (nanometer-size) particles can be simulated using species mass transport equations (Eq. 10.21). For larger particles such as biological cells, bacteria, or viruses, one can use the Lagrangian equations of motion for particle velocity along its trajectory (Marshall, 2009): -

mp -

d up 5 Fp dt

and -

up

5

d~ xp dt

(10.22)

where u p is the particle velocity vector, x p is the particle position vector, and mp is the particle mass. The forces acting on the particle F p include drag, gravity (buoyancy), Brownian diffusion, collision with walls, and electrostatic, magnetic, and other forces (Marshall, 2009). Because the incompressible flow fields in microfluidic OoC devices operating with constant pumping develop (stabilize) rapidly, a steady-state continuous flow can be assumed and one needs to solve only species transport, resulting in much faster simulations. A two-step simulation protocol is available in the CoBi framework: (1) simulate the steady-state flow (or a full time cycle for transient periodic flows) and store the solution, (2) simulate long-time drug delivery, transport, and PK/PD in the OoC device, solving species transport equations with ‘frozen’ flow fields. As mentioned, full 3D computational fluid dynamics models involve the complex task of mesh generation and require longer simulation times. Many fluid circuit and vascular network applications use reduced-order fluid flow models, solving two one-dimensional partial differential equations for the flow rate Q and pressure p (Olufsen, 1999; Sherwin et al., 2003). CoBi tools provide unique reduced-order modeling capability to represent complex tubing networks or microfluidic channels using quasi-3D (Q3D) geometries and solving full 3D NavierStokes equations for all three Cartesian velocity components. CoBi Q3D tools solve transport equations in the stream-wise direction numerically and in the cross-stream analytically (Kannan et al., 2017, 2018). The advantages of the Q3D method include improved representation of geometry, more accurate treatment of momentum terms, and natural 3D visualization of simulation results. Fig. 10.7 shows a comparison between full 3D and Q3D fluid flow simulations for selected examples including a straight tube, 90 degrees tube turn, and tube junction. The 3D 90 degrees tube turn model used approximately 20,000 cells, while the Q3D model used only 100 cells. Note that the predicted pressure distribution profiles in Q3D are very similar to those in the 3D model, but the simulation time is orders of magnitude faster.

Computational models for designing organ-on-a-chip devices

FIGURE 10.7 Comparison between CoBi 3D and Q3D laminar fluid flow simulations in example straight tube, sharp tube bend, and microchannel junction. 3D, Three-dimensional; Q3D, quasithree-dimensional.

FIGURE 10.8 Example organ-on-a-chip device connected to a microfluidic perfusion system.

Most of the contemporary computational fluid dynamics tools solve the above 3D equations using numerical finite-volume methods. Details of the numerical methods and solution algorithms in the CoBi 3D and Q3D CFD solver have been presented before (Jiang and Przekwas, 1994; Przekwas et al., 2006; Chen and Przekwas, 2010).

Modeling peripheral components of microfluidic organ-on-a-chip devices Microfluidic OoC devices require various peripheral components such as tubing, pumps, reservoirs, valves, sensors, and controllers (Fig. 10.8). To ensure

325

326

CHAPTER 10 Computational pharmacokinetic modeling

physiological relevance, the in vitro pumping system for a specific OoC device should replicate the in vivo fluid mechanics parameters (flow rate, pressure) for that organ. For example, tissue barriers require two separate pumps operating on apical and basolateral channels. Microreservoirs, used to either supply fresh medium or collect medium efflux, are typically open to the atmosphere, providing pressure ‘compliance’ to the system. The pressure, or volume, compliance of microreservoirs is essential to handle pumping of incompressible liquids in closed-loop microfluidic systems. Micropumps are particularly important peripheral components as they provide a constant supply of medium and, depending on their location, induce system overpressure if located on the inlet side pushing the fluid or underpressure if placed on the outlet side pulling the fluid (Byun et al., 2014). A steep fall in pressure within the microfluidic system or presence of fluid stagnation zones may lead to the undesired formation of bubbles, particularly if the OoC device is constructed of materials permeable to oxygen, such as PDMS (Sung et al., 2009). Mathematical modeling of fluid flow, and transport of drugs and metabolites in peripheral OoC components can be used to analyze drug delivery, in vitro PK, and experimental data and include calculations of volume and compound concentration in collection reservoirs. Peripheral microfluidic components can be modeled using reduced-order, ordinary differential equations or distributed Q3D models. Rather than modeling the pumps explicitly, it is easier to specify flow rate boundary conditions at the tube inlet (or outlet) at the pump location. Mathematical modeling of perfusing medium collected in microreservoirs can assist in calculating compound mass (concentration) eluted for the OoC during time-dependent compound delivery. A simple ordinary differential equation-based model (compartmental) of a collection reservoir should include both the volume and compound mass balance equations: dVR 5 Qin 2 V_ R;samp ðtÞ dt

(10.23)

dVR CR 5 Qin Cin 2 V_ R;samp ðtÞCR dt

(10.24)

where VR is the time-dependent medium volume in the reservoir, Qin is the medium volumetric flow rate entering the reservoir, CR is the concentration of the compound in the reservoir, Cin is the inflow concentration, and V_ R;samp ðtÞ is the intermittent medium volume sampled from the reservoir. If the volume of the interconnecting tubing is comparable to the other volumes and if the compound concentration is time-dependent, a spatially resolved model, for example, the Q3D, is more accurate. Dynamic simulation of a reservoir using a Q3D model involves the supply tube, the drain tube, and the time-dependent volume of the reservoir. Special treatments of pressure boundary conditions are needed: zero value at the drain tube exit and at the free surface of the fluid medium in the reservoir. For numerical stability, a moving mesh must be used in the reservoir to accommodate any supply-drain flux imbalance. Fig. 10.9 shows example CoBi

Models of drug pharmacokinetics in organ-on-a-chip devices

FIGURE 10.9 CoBi Q3D model used to simulate the dynamics of a reservoir with connected supply and drain tubes. Q3D, Quasi-three-dimensional.

Q3D simulations of dynamic supply-drain fluid flow simulations in a reservoir. The species mass transport equations in the reservoir are solved along with the fluid dynamics equations.

Models of drug pharmacokinetics in organ-on-a-chip devices Compartmental models of transport in organ-on-a-chip devices An OoC device is designed and functionalized to replicate in vivo human organ anatomy and physiology. Because exact replication is not practical, various simplifications and scaling protocols have been investigated (Wikswo et al., 2013; Somayaji et al., 2016, 2018). From a pharmacological perspective, assuming that the cellular physiology can be preserved in vitro, the OoC medium should expose the biological cells to a drug concentrationtime profile that is similar to the in vivo profile. For characterized drugs, PBPK models calibrated from clinical

327

328

CHAPTER 10 Computational pharmacokinetic modeling

data can predict the plasma-drug concentration profile, which can subsequently be used for the drug delivery protocol in the OoC device (e.g., setting an in vivolike inlet concentration profile for a liver OoC). For practical reasons, in both OoC experiments and computational models, a drug bolus administration protocol is typically applied at the OoC inlet. The ultimate computational challenge of OoC PK/PD simulations is to maximize the use of first principlesbased models with minimal calibration of in vitro data. Because the biophysics of drug transport in microfluidic channels can be accurately simulated in OoC devices using advanced physics-based models, the greater challenge is to reliably model the intracellular PK and PD effects. Ultimately, a complementary first principlesbased computational PBPK/PD model can then be used for in vivo simulations, and subsequently for model-based in vitro-to-in vivo translation (IVIVT). As mentioned in the previous section, PK/PD simulations in an OoC device can focus on modeling species (drug, metabolite) transport with precomputed and stored hydrodynamic results. The problem is then to solve spatiotemporal convection diffusionreaction systems of equations in the OoC for the duration of the drug exposure and PK/PD responses. To that end, in this section, we describe the reduced-order, ordinary differential equation-based modeling approach for tissue barrier OoC devices, followed by partial differential equation-based Q3D examples. The geometry of a Wyss Institute barrier-type OoC device (Fig. 10.10) (Huh et al., 2010, 2012; Kim et al., 2012; Prantil-Baun et al., 2018; Sontheimer-Phelps et al., 2019; Novak et al., 2019; Herland et al., 2019) can be ‘discretized’ in the stream-wise and cross-stream directions. In the following example, the OoC is divided into three stream-wise (axial) spatial compartments (proximal, central, distal) and seven transverse subcompartments in the cross-stream direction: top PDMS, top channel, epithelial layer, membrane, endothelial layer, bottom channel, and bottom PDMS.

FIGURE 10.10 Schematic of the geometric and morphological structure of the Wyss Institute organ-on-achip barrier model.

Models of drug pharmacokinetics in organ-on-a-chip devices

The reduced-order models can be derived from a general spatiotemporal transport equation: @C 1 rðvCÞ 5 rðDrCÞ 2 S @t

(10.25)

by integrating spatial terms for convection and diffusion into individual fluxes across control volume boundaries and treating them as generalized sink terms in the ordinary differential equations: V

@C 5 QðCin 2 CÞ 1 ðJ2 2J1 ÞTr 1 ðJ2 2J1 ÞAx 2 VUS @t

(10.26)

where C is the compound concentration inside the compartment, Cin is the inlet concentration, v is the fluid velocity, D is the diffusion coefficient of the compound, S is the generalized sink term, V is the compartment volume, Q is the volumetric flow rate, and J is the diffusive flux across the boundaries of the control volume in the axial and transverse directions. The axial flux and the transverse flux, for example, fluid (F) to membrane (M), can be expressed as D ðCF2 2 CF Þ δ2     CM CM 2 CF 5 SUPF2M 2 CF k k

J2;Ax 5 A J2;Tr 5 S

δM kDM

1 1

δF DF

(10.27) (10.28)

where A is the channel cross-sectional area, S is the channel-barrier area (nomenclature, not to be confused with the generalized sink term discussed above), CF is the drug concentration in the compartment, CF2 is the upstream compartment convention, CM is the concentration in the membrane, k is the partition coefficient, δM and δF are the half-distances to the FM interface, and PF2M is the effective permeability coefficient. Note that depending on the compound partition coefficient, k, there may be a concentration discontinuity at the fluidmembrane interface (Fig. 10.10). Detailed derivation of the reduced-order model for the barrier typical of Wyss Institute OoC devices can be found elsewhere (supplement of Herland et al., 2019). The model accounts for drug partitioning into the basal and apical PDMS layers of the OoC package material, for cellular drug metabolism, and for the efflux transporters (JPgp). For practical cases, two sets of species conservation equations must be solved for both the primary compound and for its metabolites. Both sets of equations are coupled via a drug metabolism reaction term in the epithelial cell layers (e.g., in liver hepatocytes), where a sink term of the primary compound (e.g., a prodrug) becomes a source term in the metabolite conservation equation. The drug conservation equation in the epithelial layer may include the compound metabolism term calculated using the organ-specific intrinsic clearance, CL,int. For example, the equation for the hepatocyte cell layer involves the intrinsic clearance. Some OoC devices are subject to oscillatory membrane/channel stretching (Huh et al., 2010), which is essential to achieve the physiological functions of the

329

330

CHAPTER 10 Computational pharmacokinetic modeling

FIGURE 10.11 Schematics of the Massachusetts Institute of Technology liver-on-a-chip device (Chen et al., 2017) (A) and a representative mathematical pharmacokinetics model of the device (B).

endothelial-epithelial cell barriers observed in vivo. Because of small strain and very short time constants of the membrane stretching (approximately 10% strain with 0.2 Hz frequency), the direct stretch effects on organ geometry can be neglected in the present reduced-order OoC model. However, the stretch effects are indirectly accounted for in the calibrated permeability term (Eq. 10.28). The same mathematical formulation can be used for compartmental modeling of OoC devices with side-by-side barriers (e.g., SynVivo, discussed next). Another well-established OoC platform, developed at MIT, uses an integrated microfluidic pump and a functionalized vertically stacked Transwell barrier (Fig. 10.11) (Domansky et al., 2010; Griffith et al., 2014; Sarkar et al., 2015; Chen et al., 2017; Tsamandouras et al., 2017; Bovard et al., 2018). The MIT Transwell insert-based liver-on-a-chip device (Fig. 10.11A) has a polystyrene scaffold with microchannels that enable medium flow form the basal compartment through the membrane and through the aggregate hepatic cell culture (hepatocytes and Kupffer cells) into the apical compartment. A recirculation pump maintains continuous flow between the two compartments. The advantages of a Transwell insert-based microfluidic OoC is direct visual access to the barrier, ability to sample medium directly from the organ interstitium, and good oxygen delivery through the airmedium interface. The geometrical complexity of the Transwell liver-on-a-chip device can be best represented by compartmental models with various levels of resolution. A simple “well-stirred” reactor model lumps the basal and apical fluid media and the cellular barrier into one compartment, as shown in Eq. (10.20) (Maass et al., 2017): VL

dCL 5 QLB;in CLB;in 1 QLA;in CLA;in 2 QLA;out CL 2 CLL CL dt

(10.29)

where CL is the average compound concentration in the liver, VL is the total medium volume in the liver device, QLB,in and QLA,in are the inlet flow rates into the basal and apical subcompartments, CLB,in and CLA,in are the inlet compound conventions, and CLL is the compound “hepatic” intrinsic clearance rate (by the hepatocytes). In this one-compartment model, all parameters are known and only CLL requires calibration. However, the clearance occurs in the much smaller

Models of drug pharmacokinetics in organ-on-a-chip devices

hepatocyte barrier volume. A better approximation would be to discretize the liver model into three compartments—basal, apical, and hepatic barrier—and to allow volume compliance. Simulation of incompressible flows in closed-volume systems can cause numerical problems (singularity), but incorporating a variable volume with a prescribed pressure boundary (a compliant volume) can resolve such problems. Below, it is assumed that the basal medium volume is constant (neglecting the basal membrane flexibility), but the apical medium volume, open to the atmosphere, is treated dynamically. The following model also allows for external medium inlets and outlets to the liver, replicating the physiological conditions. Neglecting medium sampling, the apical volume balance equation for this system is dVLA 5 QLB;in 1 QLA;in 2 QLA;out dt

(10.30)

The drug mass balance equations for the three compartments are   dCLB (10.31) 5 QLB;in CLB;in 1 QLR CLA 2 QLB;in 1 QLR CLB 2 JLB2H dt   CH dCH 5 (QLB;in 1 QLR )CLB 2 QLB;in 1 QLR VH 1 JLB2H 2 JH2LA 2 CLint  f u  CH dt KpH (10.32) VLB

VLA

 CH   dCLA  5 QLB;in 1 QLR 1 QLA;in CLA;in 2 QLB;in 1 QLR 1 QLA;in CLA 1 JH2LA dt KpH (10.33)

where CLB, CLA, and CH are the compound concentrations in the basal, apical, and barrier compartments, QLR is the intraliver volumetric recirculation rate, VH is the total volume of the hepatic cellular barrier (cells plus interstitium), KpH is the partition coefficient between the medium and hepatocytes, CLint is the intrinsic hepatic clearance term (due to the action of cytochrome P450 enyzmes), and fu is the unbound drug fraction. Other parameters are as above and in Fig. 10.11. The diffusive fluxes can be calculated as   CH JLB2H 5 P:A CLB 2 KpH   CH JH2LA 5 P:A 2CLA KpH

(10.34) (10.35)

A similar model can be formulated for the MIT Transwell insert-based gut-ona-chip device (Chen et al., 2017; Maass et al., 2017). Fig. 10.12 shows a schematic of the guton-a-chip platform and a representative mathematical PK model of the device. Similar to the liver device model, a three-compartment model can be developed for the gut-on-a-chip device that is composed of the basal volume (portal vein), apical volume (intestinal fluids), and the cellular barrier (enterocytes and other, e.g., stromal, cells).

331

332

CHAPTER 10 Computational pharmacokinetic modeling

FIGURE 10.12 Schematics of a gut-on-a-chip device (Chen et al., 2017) and a representative mathematical PK model of the device. PK, Pharmacokinetics.

A simplified two-compartment model for this Transwell insert device has been recently reported (Tsamandouras et al., 2017). The model consists of the vascular compartment (basolateral volume with a flow-through) and the apical compartment [apical medium (gut lumen) and the cellular barrier (enterocytes)]: VGB

dCGB 5 QGB;in ðCGB;in 2 CGB Þ1PG AG ðCGA 2 CGB Þ dt

(10.36)

dCGA 5 2PG AG ðCGA 2 CGB Þ 2 CLG;int CGA dt

(10.37)

VGA

where VGA, VGB, CGA, and CGB are the volumes and compound concentrations in the apical and basal compartments, QGB,in is the medium volumetric flow rate through the basal compartment, AGB,in is the barrier (Transwell insert) membrane area, PG is the effective permeability of the barrier, and CLG,int is the intrinsic intestinal clearance term (compound metabolic rate). Note that the clearance term is applied to the entire apical volume rather than to the barrier volume. A more complete model would also account for oral drug delivery by including a compound (and volume) addition to the gut apical compartment. The major disadvantage of compartmental models is their limited spatial resolution. As mentioned in the previous section, full 3D models can be computationally too expensive while Q3D type models may be the optimum approach.

Spatial models of transport in organ-on-a-chip devices The process of fabricating microfluidic devices prefers relatively simple geometric 2D horizontal layouts linearly extruded in the vertical direction, resulting in microchannels with rectangular cross sections. Mathematical modeling of species transport in such configurations can be simulated by solving a set of partial differential equations (Eq. 10.25) in 3D computational meshes or using CoBi Q3D modeling. To evaluate the accuracy of the Q3D model a simple microfluidic barrier has been set up with the tracer supplied through the basal channel, permeating

Models of drug pharmacokinetics in organ-on-a-chip devices

FIGURE 10.13 Comparison between 3D and quasi-3D simulations of tracer (C) transport through an idealized microfluidic barrier from the apical channel to the basal channel. 3D, Three-dimensional.

FIGURE 10.14 Comparison between 3D and quasi-3D simulations of tracer (C) transport through an idealized microfluidic organ-on-a-chip barrier. 3D, Three-dimensional.

via a thin membrane into the apical channel (Fig. 10.13). This 3D model has a 100 3 20 3 20 mesh while the Q3D model has a 100 3 1 3 1 mesh. A step function of the tracer concentration, C 5 1 (nondimensional), lasting 50 seconds is applied at the inlet of the basal channel. The concentration profiles at monitoring points (A-out, B-out) show excellent agreement between 3D and Q3D simulations. Fig. 10.14 shows a microfluidic OoC barrier configuration similar to that of the Wyss Institute OoC platform. A tracer with inlet concentration, C 5 1, supplied at the basal channel inlet, permeates the smaller cross-sectional apical channel via a thin membrane. The spatial distribution of the tracer in both channels is shown, with good agreement between 3D and Q3D model simulations.

333

334

CHAPTER 10 Computational pharmacokinetic modeling

FIGURE 10.15 Quasi-three-dimensional simulations of tracer transport across the SynVivo barrier and formation of the concentration gradient.

A significant number of OoC devices adopted an easier-to-fabricate horizontal barrier with side-by-side channels separated by a microfabricated permeable (slotted) barrier (Hung et al., 2005; Lee et al., 2007; Deosarkar et al., 2015; Adriani et al., 2017; Sun et al., 2018; Brown et al., 2019). In such side-by-side configuration, depending on the pressure gradient across the barrier, the medium and the drug can be transported across the slotted barrier by diffusion and convection. Fig. 10.15 shows a side-by-side SynVivo OoC barrier and CoBi Q3D model of medium flow and tracer transport. The model simulates transport of a tracer across the chip from the donor channel to the receiver channel. A steady-state concentration gradient is formed across the chip, with decreasing magnitude along the chip that could be used to study cellular polarization and chemotaxis. The computational models of the above OoC platforms, Transwell inserts, vertical barrier, and side-by-side barrier can be adapted to model organs such as the gut, liver, lung (bronchial and alveolar), renal proximal tubule, bloodbrain barrier, skin, and bone. One of parameters calculated from the OoC barrier experimental data is the barrier’s effective permeability, Peff. Unlike static cultures, the microfluidic Peff model should include convective terms. For the case of two microfluidic channels, donor (D) and receiver (R) separated by a membrane (cell barrier), the compound transport equation in the receiver can be written as VR

dCR 5 QR ðCRin 2 CR Þ 2 JMR dt

(10.38)

Assuming that CRin 5 0 and expressing the flux term using the effective permeability Peff we have

Models of drug pharmacokinetics in organ-on-a-chip devices

VR

dCR CR 2 0 5 2 QR CR 2 AUPeff ðCR 2 CD Þ  VR t dt

(10.39)

The effective permeability, Peff, can be calculated using the experimentally measured CR(t) for a known and constant donor drug concentration, CD: Peff 5

ðVR =t 1 QR ÞUCR AUðCD 2 CR Þ

(10.40)

Note that if QR 5 0 or QR{VR/t, the conventional Peff formula is recovered.

Multiscale models of organ-on-a-chip devices Whether simulating OoC devices with complex geometries or studying cell/tissue-scale structures embedded in microdevices, multiscale computational modeling is required. For example, detailed simulation of cellular spheroids immersed in a Transwell insert would require several spheroidal objects embedded in the medium pool. Simultaneous simulation of intraspheroid drug/metabolite transport and spheroid-medium exchange can be performed. Another OoC device with complex geometry for which a multiscale model is required is the liver bioreactor developed by the Charite´ university clinic (Fig. 10.16) (Zeilinger et al., 2011).

FIGURE 10.16 3D and quasi-3D multiscale model of drug transport and metabolism in the liver bioreactor of the Charite university clinic (Zeilinger et al., 2011). 3D, Three-dimensional.

335

336

CHAPTER 10 Computational pharmacokinetic modeling

The liver bioreactor tissue chamber contains a tissue coculture of primary human hepatocytes and nonparenchymal cells (Kupffer and endothelial cells) with an embedded network of two fiber types supplying the culture medium and gaseous oxygen (Fig. 10.16A). The liver bioreactor is integrated into a closed-loop microfluidic system with a pump, bubble trap reservoir, fresh medium supply, and waste removal (Fig. 10.16B). A multiscale CoBi model of the liver bioreactor couples a 3D mesh model of the tissue chamber and Q3D models for the embedded countercurrent fibers. The individual fibers (carrying medium and gas) are modeled using a Q3D tube geometry composed of the fluid lumen and a porous solid wall enabling two-way communications between fibers and tissue. The gas-carrying fibers exchange O2 and CO2 while the media fibers exchange substrates, drugs, metabolites, and biomarkers. The liver bioreactor model is coupled to the Q3D model of the peripheral system. The 3D multispecies transport equations (partial differential equations) are solved in the liver bioreactor interstitial space coupled to the cellular biochemistry equations (ordinary differential equations) solved in the cellular subcompartment. The cell model simulates metabolic pathways and fluxes, cellular metabolic control mechanisms, enzyme kinetic mechanisms, and hepatocyte-specific metabolic characteristics. The drug PK model accounts for the cellularity, drug intrinsic clearance (from cytochrome P450 enzymes), and drug transport across the cell layers. Readily measurable in vitro drug parameters such as log P/D, drug binding in plasma and hepatocytes, membrane fluxes, and metabolic clearance serve as model inputs. Fig. 10.16C shows simulation results for a dynamic distribution of a tracer, initially supplied into the bubble trap reservoir, into the system.

Models of drug pharmacokinetics in microphysiological systems The development of MPS has been spearheaded by DARPA- and NIH-funded projects (Esch et al., 2014; Marx et al., 2016; Vernetti et al., 2017; Zhang and Radisic, 2017; Shuler, 2019; Cirit and Stokes, 2018; Prantil-Baun et al., 2018; Novak et al., 2019). At present, there are very few published reports on computational modeling of multiorgan MPS, also termed body-on-a-chip or HoC platforms (Shuler and Esch, 2010; Somayaji et al., 2016; Sung et al., 2018). While in vivo human whole-body PBPK models have been used for decades and modeling standards have been established, a formal approach to modeling in vitro MPS does not exist, rendering the in vitro-to-in vivo translation task a challenging endeavor. This section briefly summarizes the current state of MPS modeling. The topological layout and connectivity between individual organs is constructed similar to that of the in vivo PBPK model in Fig. 10.2. In an MPS construct, the endothelial compartments of individual organs (vascular) are supplied

Models of drug pharmacokinetics in microphysiological systems

FIGURE 10.17 Schematics of example simplified physiological layouts of current microphysiological systems. (A) MIT gutliver system. (B) MIT gutliverkidney system. (C) Philip Morris International lungliver system. (D) Cornell University rotational frame multiorgan system. (E) Wyss Institute gutliverkidney system. MIT, Massachusetts Institute of Technology.

by the “arterial” fluids and are cleared into a common “venous” compartment. Some examples of exceptions include the gut OoC outlet draining into the liver OoC inlet (via the “portal vein”) and part of the medium from the kidney glomerulus draining into a “urinary bladder” collector. Medium flow control in epithelial compartments depends on specific tissues such as the gut lumen, nephron filtrate duct, lung epithelium, skin construct, and brain tissue. Current computational models of MPS comprise few organs, typically only gut and liver, interconnected in a simplified ‘physiological’ layout (Fig. 10.17) (Lee and Jun, 2019; Bovard et al., 2018; Maass et al., 2017; Tsamandouras et al., 2017; Prantil-Baun et al., 2018; Wang and Shuler, 2018; Lee and Jun, 2019). In such layouts, the arterial and venous compartments are typically combined into a single compartment, a mixer (Maass et al., 2017), or arteriovenous (AV) reservoir (Prantil-Baun et al., 2018).

337

338

CHAPTER 10 Computational pharmacokinetic modeling

FIGURE 10.18 Schematics of the MIT (A) gutliver mixer (MPS-1) and (B) gutliverkidney mixer (MPS-2) platforms with simplified physiological layouts. MIT, Massachusetts Institute of Technology; MPS, microphysiological system.

As in the single-OoC devices, multiorgan MPS can be represented using either the reduced-order (ordinary differential equation-based), distributed high-fidelity (spatiotemporal- 1D/2D/3D or Q3D) or multiscale models. Because of the system complexity, essentially all published mathematical models of MPS use the very simplified, ordinary differential equation-based, first-generation models. Here, we briefly describe first-generation mathematical models of MPS developed at MIT and at the Wyss Institute. We anticipate that Q3D-based models will provide a foundation for the second generation of MPS platforms. The MIT MPS is based on interconnected OoC devices linked to a “mixer” that potentially represents the vascular pool and blood in other (missing) organs (Maass et al., 2017; Tsamandouras et al., 2017). Fig. 10.18 shows the schematic of two versions of the system: gutliver mixer (MPS-1 and 2) gutliverkidney mixer (MPS-2). The mathematical models used to describe these systems are based on conventional compartmental PK models (Jones and Rowland-Yeo, 2013; Jusko, 2013). The gutliver-mixer model (Fig. 10.18A) solves species mass balance equations in each organ: Gut apical (lumen-enterocytes) VGa

dCGa 5 PG UAG ðCGb 2 CGa Þ 2 CLG UCGa dt

(10.41)

Gut basolateral (blood) dCGb 5 QGb ðCmix 2 CGb Þ 2 PG UAG ðCGb 2 CGa Þ dt

(10.42)

dCL 5 QGb ðCGb 2 CL Þ 1 QL ðCmix 2 CL Þ 2 CLL UCL dt

(10.43)

VGb

Liver VL

Models of drug pharmacokinetics in microphysiological systems

Mixer Vmix

dCmix 5 ðQGb 1 QL ÞðCL 2 Cmix Þ dt

(10.44)

where CGa, CGb, CL, and Cmix are drug concentrations in individual organs, and VGa, VGb, VL, and Vmix are the volumes of individual organs. AG is the gut barrier exchange area (Transwell insert membrane surface area), PG is the gut barrier effective permeability, and CLG and CLL are intrinsic drug clearance rates for the gut and liver, respectively. The flow rates, Q, are specified from the physiological proportions of the total outflow from the mixer (cardiac output) (Fig. 10.18A). A simplified form of the model, with one gut compartment (both apical and basolateral), was used to investigate the in vitro PK of diclofenac and hydrocortisone for different MPS layouts (gut only, liver only, and gut-liver) (Tsamandouras et al., 2017). The results were used to estimate the intrinsic parameters (e.g., permeability, metabolic clearance) in each simulated MPS. An improved version of the MIT MPS (MPS-2, shown in Fig. 10.18B) requires the inclusion of a kidney OoC device (Maass et al., 2017) and a “urinary bladder” reservoir to collect the filtrate from the kidney apical compartment. This version also assumes constant fresh medium supply into the mixer with the flow rate equal to the waste collection rate (glomerular filtration rate) into the “bladder.” The gutliver-mixer equations of the MPS-2 model are similar to those used in the MPS-1 model. The additional equations are Mixer Vmix

dCmix 5 QL CL 1 QKb CKb 1 QF CF 2 ðQL 1 QKb 1 QF ÞCmix dt

(10.45)

where QF and CF are the fresh medium constant resupply rate into the mixer and injected drug concentration (intravenous dose), respectively. Kidney apical (filtrate) VKa

dCKa 5 QKa ðCmix 2 CKa Þ 1 PK UAK ðCKb 2 CKa Þ dt

(10.46)

Kidney basal (vascular) VKb

dCKb 5 QKb ðCmix 2 CKa Þ 2 PK UAK ðCKb 2 CKa Þ dt

(10.47)

Bladder (waste) dVW 5 QKa dt VW

or

0 VW 5 VW QKa Ut

dCW 5 QKa ðCKa 2 CW Þ dt

(10.48) (10.49)

where CKa, CKb, and CW are drug concentrations in individual organs, and VKa, VKb, and VW are volumes of individual organs. Note that the waste compartment volume, VW, is a function of time but can be solved analytically by integrating

339

340

CHAPTER 10 Computational pharmacokinetic modeling

Eq. (10.48) for a constant flow rate, QKa. However, if any medium is sampled from the waste, Eq. (10.48) has to be solved numerically. A simplified form of the above model was used to investigate the in vitro PK of selected drugs and compare the results from the MPS-1 and MPS-2 model layouts (Maass et al., 2017). Integration of microfluidically linked organs into a closed-loop MPS poses several challenges with the peripheral devices (tubing, pumps, reservoirs, valves), such as potential leaks, bubble formation, flow rate controls, sampling, and medium resupply. The Wyss Institute MPS architecture uses OoC devices in which the endothelial and epithelial microchannels are connected to microreservoirs and peristaltic pumps on the suction side, transporting the medium from the inlet microreservoir through the OoC to the outlet microreservoir. Physiological communication between individual organs is accomplished using a robotic autosampling system to achieve periodic medium exchange between OoC microreservoirs and the AV reservoir. CoBi tools have been used to develop computational models of Wyss multiorgan MPS (Prantil-Baun et al., 2018; Novak et al., 2019; Herland et al., 2019). The geometry of each Wyss OoC device (Fig. 10.19) is represented with three spatial control volumes in the stream-wise direction and one control volume in the cross-stream direction. Each stream-wise section is subdivided into following subcompartments: PDMS package, top channel, epithelial layer, membrane, endothelial layer, bottom channel, and PDMS package. The transport equations for the basal (B) and apical (A) channels involve both the convective and diffusive terms which, for each zone i 5 1,2,3, can be expressed as VB;i

  dCB;i 5 QB;i CB;i21 2 CB;i 1 ΔJdiffB;i 1 JB2M;i dt

i 5 1; 2; 3

(10.50)

FIGURE 10.19 Geometrical model of the Wyss Institute organ-on-a-chip device and schematic of the barrier morphology used for model development (not to scale).

Models of drug pharmacokinetics in microphysiological systems

where ΔJdiff.i is the diffusive flux difference across the downstream and upstream control volume faces in the stream-wise direction (along the microchannel axis). ΔJdiffB;i 5

 AUD   AUD  CB;i11 2 CB;i 2 CB;i 2 CB;i21 Δx Δx

(10.51)

where A is the channel cross-sectional area and Δx is the channel subcompartment length. In many cases, species conservation equations for both the primary compound and its metabolites must be solved. Both sets of equations are coupled with a drug metabolism reaction term in the epithelial cell layers (e.g., liver hepatocytes). The Wyss OoC barrier model (Fig. 10.19) for each axial zone (i 5 1,2,3) involves compound conservation equations for the basal package (BP), basal medium (BM), endothelial barrier (E), membrane (M), epithelial barrier (H), apical medium (AM), and apical package (AP): VBP;i VBM;i

dCBP;i 5 2 JBP2BM 1 ΔJDiff ;BP;i dt

dCBM;i 5 QBM ðCBM;i21 2 CBM;i Þ 1 JBP2BM 2 JBM2E 1 ΔJDiff ;BM dt

(10.54)

dCM;i 5 JE2M 2 JM2H 1 ΔJDiff ;M;i dt

(10.55)

dCH;i 5 JM2H 2 JH2AM 2 fuH UClH;int UCH;i 2 kact CH;i dt

(10.56)

VM;i

VAM;i

(10.53)

dCE;i 5 JBM2E 2 JE2M dt

VE;i

VH;i

(10.52)

dCAM;i 5 QAM ðCAM;i21 2 CAM;i Þ 1 JH2AM 2 JAM2AP 1 ΔJDiff ;AM 1 kact CH;i dt VAP;i

dCAP;i 5 JAM2AP 1 ΔJDiff ;AP;i dt

(10.57) (10.58)

where QBM and QAM are the medium volumetric flow rates in the basal and apical channels, respectively. The epithelial cell layer (H) metabolic clearance term in Eq. (10.56) with ClH,int represents the intrinsic clearance. The last terms in Eqs. (10.56) and (10.57) represent the active transport between the epithelial cells and the apical medium, where kact is the efflux rate constant for each OoC device. The cross-stream passive diffusion fluxes between individual barrier layers are   CBP;i JBP;BM 5 SUPBP2BM fuPDMS 2 fuMedia CBM;i kP;PDMS   CE;i JBM;E 5 SUPBM2E fuMedia CBM;i 2 fuE kP   JE2M 5 SUPE2M fuE CE;i 2 fuM CM;i   JM2H 5 SUPM2H fuM CM;i 2 fuH CH;i

(10.59) (10.60) (10.61) (10.62)

341

342

CHAPTER 10 Computational pharmacokinetic modeling

  CH;i JH2AM 5 SUPH2AM fuH 2 fuMedia CAM;i kp   CAP;i JAM2AP 5 SUPAM2AP fuMedia CAM;i 2 fuPDMS kP;PDMS

(10.63) (10.64)

The stream-wise diffusive flux term, ΔJdiff, in BM and AM media, membrane (M), and PDMS package layers (basal package and apical package) are calculated using Eq. (10.51). In the preceding equations, kp is the partition coefficient for a specific compound, fu is the unbound fraction of the compound, S is the surface area normal to the cross-stream direction, and P is the permeability of each barrier interface estimated using Eqs. (10.56) and (10.57) and then calibrated computationally for each organ device (Herland et al., 2019). Fig. 10.20 illustrates the Wyss MPS layout of gut, liver, and kidney OoC platforms with their microreservoirs intermittently linked to the AV reservoir. Computational simulation of the Wyss MPS with intermittent medium exchange between the organs involves the above barrier model equations (Eqs. 10.5210.64) and additional equations for inlet/outlet tubing and inlet/outlet microreservoirs (Eqs. 10.6510.68). Because the organ microreservoirs are periodically emptied and replenished with fresh medium, dynamic fluid volume conservation equations must be solved for each microreservoir. The conservation equations for the basal (endothelial) inlet and outlet microreservoirs (IR, OR) and the inlet/outlet tubing (IT, OT) are dVBIR 5 V_ BIR;inj ðtÞ 2 QBM dt

(10.65)

dVBIR CBIR 5 V_ BIR;inj ðtÞCBIR;inj 2 QBM CBIR dt

(10.66)

dCBIt 5 QBM ðCBIR 2 CBIt Þ dt   dCBOt VBOt 5 QBM CBI;3 2 CBOt dt VBIt

(10.67) (10.68)

where V_ BIR;inj and V_ BOR;samp are the intermittent medium injections into the inlet microreservoir from the AV reservoir and sampling from the outlet microreservoir, respectively. CBIR,inj is the injected concentration into the inlet basal microreservoir (AV concentration for a linked system), and CBOt is the basal outlet tubing concentration. CBI,3 is the concentration at the basal medium outlet section (i 5 3 in Eq. 10.63). The conservation equations for the apical side have a similar form. The equations for apical microreservoirs are analogous. The physiological AV pool is represented by an AV reservoir with a dynamic medium volume, VAV(t), which can be periodically resupplied from the basal (venous) outlet microreservoirs of the liver and kidney, drained to supply basal inlet microreservoirs of all organ devices, replenished with fresh medium, and

FIGURE 10.20 Schematic layout of the Wyss Institute gutliverkidney microphysiological system used to set up CoBi simulations. Each organ chip has its own microreservoirs at the inlets and outlets from the basal (endothelial) and apical (epithelial) channels. —Modified from Herland, A., et al., 2019. Quantitative prediction of human drug pharmacokinetic responses enabled by fluidically coupled vascularized organ chips. Nature, in review.

344

CHAPTER 10 Computational pharmacokinetic modeling

sampled for analysis. The total and compound mass balance equations describe the AV reservoir dynamics:   dVAV 5 2 V_ BIR;g 1 V_ BIR;l 1 V_ BIR;k 1 V_ AIR;k 1 V_ sM 1 V_ BOR;l 1 V_ BOR;k 1 V_ fM dt   dVAV CAV 5 2 V_ BIR;g 1 V_ BIR;l 1 V_ BIR;k 1 V_ AIR;k 1 V_ sM CAV 1 V_ BOR;l CBOR;l dt 1 V_ BOR;k CBOR;k 1 V_ fM CfM

(10.69)

(10.70)

where the indices g, l, k, and fM represent gut, liver, kidney, and fresh medium, respectively. V_ sM and V_ fM are the volumetric rates of sampling and fresh medium supply to the AV reservoir. The above model has been used to simulate the PK of nicotine and cisplatin in the Wyss MPS (Herland et al., 2019). Transport simulation for highly lipophilic drugs in organ devices made with PDMS has to account for drug distribution into the package material. The above OoC model involves drug conservation equations in the PDMS package layers adjacent to the basal and apical channels (Eqs. 10.52 and 10.58). To calculate the medium-PDMS permeability and partition coefficients in flux terms, Eqs. (10.59) and (10.64) and data from a custom experiment are used. In such an experiment, a blank chip (with no cell layers) filled with pure flowing medium is subjected to a transient drug bolus injection to the apical channel for a specific time period, followed by a pure medium drug-washout period. The drug partitions to the apical PDMS, diffuses through the membrane into the basal channel, and partitions into the basal PDMS. Experimental drug concentrationtime data are collected from the apical and basolateral outlets during the loading and washout periods. The optimization method is used to calibrate the medium-PDMS partition coefficients and permeabilities.

Models based on in vitro-to-in vivo translation One of the key applications of in vitro HoC systems is to characterize the pharmacological behavior of new compounds under development and predict their expected clinical response using in vitro-to-in vivo extrapolation methods. Even the most advanced in vitro HoC systems would be an enormous simplification of the human body. Therefore there will always be a need for mathematical modelbased extrapolation of in vitro results to the in vivo situation. In vitro-to-in vivo correlation involves “a predictive mathematical model describing the relationship between an in-vitro property of a dosage form and an in-vivo response” (FDA, 1997). This doseresponse model establishes a statistical correlation between in vitro dissolution or permeation rate and in vivo PK properties such as peak serum concentration and the area under the curve (Amidon et al., 1995; O’Hara et al., 2001). In contrast, in vitro-to-in vivo extrapolation uses mechanistic but relatively simple models to extrapolate data from drug metabolism in cell cultures

Models based on in vitro-to-in vivo translation

or transport across barriers to in vivo organ clearance or absorption (Howell et al., 2012; Poulin and Haddad, 2013). In vitro organ data, along with drug physicochemical properties, are then used in PBPK models (Chen et al., 2012; Rostami-Hodjegan, 2012). Emerging physiologically interconnected multiorgan devices (HoC or body-on-a-chip) can generate large quantities of data that would require a new generation of computational tools to interpret and translate to the in vivo situation (Stokes et al., 2015; Cirit and Stokes, 2018). Thus far, extrapolation methods have largely relied on scaling the in vitro PK quantities obtained from static cell culture experiments to organ-level estimates using the cellularity (Barter et al., 2007; Jamei et al., 2014; Jones and RowlandYeo, 2013; Ito and Houston, 2005; Naritomi et al., 2001). Separate in vitro experiments are used to estimate drug dissolution, intestinal permeability (absorption), hepatic intrinsic clearance, or renal clearance rates. For example, in vivo clearance (CLint-scaled) can be scaled from the apparent in vitro clearance (CLintapp expressed in μL/min/mg for microsomes or μL/min/million cells for hepatocytes) from systems composed of recombinant enzymes, microsomes, or hepatocytes using physiological scaling factors such as human microsomal protein per gram of liver (MPPGL) and hepatocellularity per gram of liver (HPGL) and liver weight (LW), as shown in Eq. (10.71). In addition, intersystem extrapolation factors have been suggested to account for the differences in activity and expression between systems (Crewe et al., 2011). CLint2scaled 5 CLint2app =fuinc MPPGL: LW

(10.71)

CLint2scaled 5 CLint2app =fuinc HPGL: LW

(10.72)

Furthermore, in vivo hepatic clearance (CLh) and the hepatic availability (Fh 5 1 2 CLh/Qh) can be calculated using in vitro primary human hepatocyte culture data and the well-stirred liver model (Riley et al., 2005; Chiba et al., 2009). Scaled intrinsic clearance has been suggested as an input to the PBPK model (Jones and Rowland-Yeo, 2013). CLh 5

Qh Ufup UCLint;invivo Qh 1 fup UCLint;invivo

(10.73)

where Qh is the liver blood flow (20 mL/min/kg), fup is the unbound fraction in plasma, and the in vivo clearance is estimated using a physiology-based scaling factor, which converts the units of CLint,in vitro (μL/min/106 cells) to CLint,in vivo (mL/min/kg). While such methods are useful for characterizing the PK behavior of uncharacterized drugs, static in vitro cell cultures have some notable drawbacks. For example, loss of drug-metabolizing enzyme activity in suspended hepatocytes limits the accurate determination of the intrinsic clearance of slowly metabolized compounds (Chan et al., 2013). Furthermore, the absence of in vivo aspects of fluid flow and mechanical cues in these static in vitro systems affects cell differentiation and viability, enzyme and transporter expression, and ultimately the

345

346

CHAPTER 10 Computational pharmacokinetic modeling

transport metabolic activities of the cells, limiting the utility of these methods for studying diseases states. Although advances have resulted in the development of 2D and 3D cell culture models that incorporate multiple cell types, extracellular matrix molecules to study tissue function, cell signaling, and drug response in diseased states, these models cannot precisely quantify transcellular transport and have no multiscale architecture at the biological interfaces. These characteristics are essential to emulating certain organ-level functions in an in vitro system (Bhatia and Ingber, 2014; Kim et al., 2012). Microfluidic OoC devices can potentially address these limitations by replicating cellular functions and transporter expression. When different chips are linked, multiple PK quantities can be determined from the resulting HoC system. Although several morphologies exist, the two-channel designs provide flexibility in controlling drug infusion and therefore can be adapted for different organ configurations. For example, a drug can be introduced on the apical side of a gut-ona-chip device or on the endothelial side of a liver-on-a-chip device. In more sophisticated chip designs, the blood medium channel is lined with endothelial cells to better mimic in vivo physiology and functions and to augment the biological relevance of the data obtained from these chips (Osaki et al., 2018; Lee and Jun, 2019). HoC platforms that combine multiple in vitro microfluidic OoC devices in a physiologically consistent manner are being adopted for drug testing and characterization (Sutherland et al., 2013; Low and Tagle, 2017). PDMS is popular in fabricating OoC devices because of the availability of well-standardized casting methods for cost-effective fabrication and its intrinsic material characteristics: it is elastic, gas-permeable, transparent, and non-toxic (Low and Tagle, 2017; Mukhopadhyay, 2007; Van Meer et al., 2017). However, its intrinsic hydrophobic polymer structure can lead to nonspecific absorption of proteins, nonpolar organic solvents, small hydrophobic molecules, and therapeutic drugs (Gokaltun et al., 2017; Mukhopadhyay, 2007; Regehr et al., 2009). Therefore in microfluidic chips, in addition to coupled convective and diffusive transport, significant drug loss into the polymeric package can also distort the interpretation of on-chip data. Although the HoC concept appears promising, many of the current MPS platforms involve relatively few organs—typically two to five OoC devices linked in a physiologically plausible manner. While one may expect human-like responses from HoC devices with human-derived cells, at least for specific intensive quantities, such as normalized concentrations, rate constants, and barrier transport properties, innate in vitro platform-specific attributes can lead to variances with the in vivo environment and may also convolute the in vitro-to-in vivo correlations. Some of these attributes include design aspects, fabrication constraints, medium properties (e.g., protein content, additives), physiological linking (e.g., missing organs), cellular properties (e.g., source, expression), and process parameters (e.g., flow rates) (Somayaji et al., 2016). To address some of the inherent differences between OoC devices and their in vivo counterparts (i.e., the mutually exclusive system attributes), we propose

Models based on in vitro-to-in vivo translation

that advanced biologically and physiologically inspired computational modeling, discussed as follows, can support and accelerate device design and development, enable the creation of physiologically consistent organ-linking protocols and rational pharmacotherapy experiments, and contribute to robust extrapolation of in vitro on-chip outcomes to human responses, which is key for the rapid and successful development of new drugs (Somayaji et al., 2016, 2018). Furthermore, complementary experimental and computational models of physiologically connected OoC devices provide a unique opportunity to establish next-generation multiscale PBPK models coupling the whole body and spatially and morphologically resolved barrier models (Sluka et al., 2016; Clancy et al., 2016; Kannan et al., 2018). Computational models of morphologically resolved barriers in specific organs could be validated on HoC platforms and extended to multiscale PBPK models. Fig. 10.21 illustrates the overall conceptual approach to addressing the differences between in vitro and in vivo environments by computational modeling, which entails the application of two sequential steps: in vivo-to-in vitro replication (IVIVR) and in vitro-to-in vivo extrapolation (IVIVE). Philosophically, in vivo-to-in vitro replication aims to replicate the organismal microenvironments in vitro to mimic specific functions, whereas mathematically, in vivo-to-in vitro replication “can be viewed as a constrained optimization problem in which the objective is to minimize the discrepancy between the in vivo and in vitro microenvironments” (Somayaji et al., 2016, 2018). Chip designs identified via in vivo-to-in vitro replication are biologically consistent representations of the target system (in vivo) and can be fabricated in a laboratory. It is therefore good practice for HoC platform developers to envision the desired in vivo characteristics and functions of the OoC devices a priori while simultaneously addressing the microfluidic design and anomalies (Somayaji et al., 2018). The starting point for implementing these steps is the development of mathematical models of OoC devices that are based on in vivo physiology and pharmacology and described using first principles-based concepts of medium flow, drug transport across cellular barriers, and subcellular and metabolic losses. To this end, both reduced-order (multi-compartmental ordinary differential equationbased) models of individual OoC devices and high-fidelity (2D and 3D partial differential equation-based) models are used exclusively or in conjunction (multiscale linking) to accurately predict the drug transport within an OoC environment. Furthermore, to ensure that the models are effective for new drug testing, that is, that they can predict the in vivo behavior of a drug, the governing equations of the models must rely on physicochemical drug properties that can be readily measured using standard assays (Davies and Morris, 1993; Poulin and Theil, 2000; Rodgers et al., 2005; Rodgers and Rowland, 2006, 2007). Moreover, the parametric nature of first principles-based mathematical models should allow for the use of the same set of governing equations in both platforms. Therefore in vivo PBPK and in vitro HoC models represent mathematical equivalents of a living human, and the sequential application of the replication and extrapolation steps is

347

FIGURE 10.21 Illustration of in vivo-to-in vitro replication (IVIVR) and in vitro-to-in vivo extrapolation (IVIVE). Computational models can help emulate the in vivo microenvironment in the in vitro microfluidic chips by replication, and also help understand the clinical relevance of on-chip data by extrapolation.

Conclusions and future perspectives

thought to enable the intelligent exchange of data between the two platforms, guiding the testing and evaluation of new drug candidates (Somayaji et al., 2016, 2018). Following in vivo-to-in vitro replication-guided design, the next step is to extrapolate the in vitro outcomes to in vivo. To this end, given the inherent differences between the two systems, certain intensive PK/PD quantities such as intrinsic clearance rates, barrier transport parameters and mass transfer coefficients can be considered (Somayaji, et al., 2016). Fig. 10.22 illustrates the intensive quantities of an orally administered compound, which can be estimated from an in vitro three-organ HoC device linked with a common AV reservoir. Using the in vitro and in vivo estimates, ratios of intensive PK/PD quantities between in vivo and in vitro systems (termed scaling factors) can be estimated for compounds with well-characterized in vivo behavior and available clinical data. Scaling factor maps can then be established by plotting each scaling factor against the specific drug properties for a given HoC configuration and design. The nature of this type of scaling factor correlation will rely on the design characteristics of a given in vitro system, medium type, cell properties, cell culture conditions, and drug properties. A 1:1 quantitative match between the in vitro PK/PD/toxicity data and in vivo observations should not be used as a measure of the reliability and ‘functional closeness’ of a given in vitro system. Instead, the goal should be to obtain a consistent scaling factor correlation. We believe that model-guided replication can assist in ensuring that consistent and reliable scaling factor correlations are derived. After estimating the intensive PK/PD quantities associated with a given HoC system, the corresponding in vivo values can be deduced using scaling factor maps established from characterized compounds, and these scaled values can then be used as input data for a computational in vivo human PBPK model to predict PK/PD responses. Thus in the process of in vitro-to-in vivo extrapolation, a computational human PBPK model is necessary to derive the in vivo meaning of on-chip data and to make clinical predictions (Somayaji et al., 2016, 2018).

Conclusions and future perspectives Microfluidics-enabled cell culture technologies paved the way for OoC devices and their integrated systems and advanced the prospect of HoC systems to mimic in vivo physiological functions and pharmacological responses. Advanced physics- and biology-based computational models can assist in the design of these systems and in analyzing experimental data. This chapter reviewed recent achievements in the development of microfluidic OoC devices and MPS platforms and presented a detailed discussion of multiscale microphysiology-based computational models for optimizing design and data analysis. While the conventional in vitro static cell cultures could be represented by simple onecompartment well-stirred mathematical models, microfluidic OoC devices require

349

FIGURE 10.22 Conceptual illustration of a three-organ linked human-on-a-chip platform connected by a common AVR to mimic in vivo physiology. The system can be used to estimate the in vitro pharmacokinetic quantities of an orally administered compound for clinical extrapolation. AVR, Arteriovenous reservoir.

Acknowledgments

spatially distributed models that account for convection, diffusion, barrier transport, and reaction phenomena. To that end, in this chapter, we have discussed those approaches for modeling in vitro microfluidic OoC devices. A novel mathematical formulation using Q3D and multiscale models was demonstrated for typical microfluidic OoC configurations and for a liver-on-a-chip device with complex geometry. Computational modeling of microfluidically linked multiorgan MPS is far more challenging. Because of the system complexity, essentially all published mathematical models of MPS use the simplified ordinary differential equationbased models. This chapter reviewed and analyzed existing modeling approaches for simulating two advanced MPS platforms, the MIT Transwell insert-based OoC system, and the Wyss Institute microfluidic barrier OoC devices linked using a robotic autosampler. Mathematical modeling of multiorgan MPS involves not only individual OoC models but also models of various peripheral devices such as tubing, mixers, and reservoirs. Similar to PBPK, simulation of a multiorgan closed-loop MPS involves the “arterial” medium distribution to “venous” medium, and collection from individual organs. Furthermore, the reservoir model should account for total medium volume dynamics, sampling, fresh medium resupply, and the in vitro “intravenous” drug administration. In contrast, current MPS combine both medium pools into a single “mixer” or AV reservoir. We envision that next-generation mathematical models of OoC devices, and MPS will enable better resolution of the OoC barrier morphology and improved spatial resolution simulations using Q3D and multiscale models and incorporate cellular-scale PK models. Because the in vitro cell/tissue barrier geometry and morphology can be measured from experiments, it is feasible to extend the lumped transbarrier transport models to better represent the interstitial and intracellular spaces. The ultimate goal of in vitro pharmacology research is to establish complementary experimental and computational tools that can translate in vitro data to the in vivo situation. Current extrapolation or translation methods use simple models to calculate lumped in vivo parameters, such as hepatic clearance or intestinal permeability, from in vitro experimental data. Next-generation in vitro-to-in vivo translation protocols will likely use OoC experimental data to validate rather than calibrate first principles-based models. Such models could then be validated in animal experiments and used in in vivo human predictive PK simulations.

Acknowledgments This research was partially sponsored by the Wyss Institute for Biologically Inspired Engineering at Harvard University and the Defense Advanced Research Projects Agency (DARPA) under Cooperative Agreement Number W911NF-12-2-0036. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of DARPA or the US

351

352

CHAPTER 10 Computational pharmacokinetic modeling

Government. The authors would like to thank several individuals at the Wyss Institute, in particular Donald Ingber, Kit Parker, Rachelle Prantil-Baun, Richard Novak, Anna Herland, and Ben Maoz, as well as colleagues at CFDRC: Debarun Das, ZJ Chen, Teja Garimella, Yuyang Miao, and Carrie German for helpful discussions, data, and graphical illustrations. Availability of CoBi tools and models: All models have been developed using CFDRC’s CoBi tools (Available at http://medicalavatars.cfdrc.com/index.php/cobi-tools/).

References Adriani, G., Ma, D., Pavesi, A., Kamm, R.D., Goh, E.L.K., 2017. 3D neurovascular microfluidic model consisting of neurons, astrocytes and cerebral endothelial cells as a bloodbrain barrier. Lab Chip 17 (3), 448459. Agoram, B., Woltosz, W.S., Bolger, M.B., 2001. Predicting the impact of physiological and biochemical processes on oral drug bioavailability. Adv. Drug Deliv. Rev. 50 (Suppl 1), S41S67. Allen, J.W., Bhatia, S.N., 2003. Formation of steady-state oxygen gradients in vitro—application to liver zonation. Biotech. Bioeng. 82 (3), 253262. Amidon, G.L., Lennernas, H., Shah, V.P., Crison, J.R., 1995. A theoretical basis for a biopharmaceutic drug classification - the correlation of in-vitro drug product dissolution and in-vivo bioavailability. Pharm. Res. 12, 413420. Anderson, E.J., Knothe Tate, M.L., 2007. Open access to novel dual flow chamber technology for in vitro cell mechanotransduction, toxicity and pharamacokinetic studies. Biomed. Eng. Online 6, 46. Barber, R.W., Emerson, D.R., 2008. Optimal design of microfluidic networks using biologically inspired principles. Microfluid. Nanofluid. 4, 179191. Barter, Z.E., Bayliss, M.K., Beaune, P.H., Boobis, A.R., Carlile, D.J., Edwards, R.J., et al., 2007. Scaling factors for the extrapolation of in vivo metabolic drug clearance from in vitro data: reaching a consensus on values of human micro-somal protein and hepatocellularity per gram of liver. Curr. Drug Metab. 8 (1), 3345. Baudoin, R., Prot, J.M., Nicolas, G., Brocheton, J., Brochot, C., Legallais, C., et al., 2013. Evaluation of seven drug metabolisms and clearances by cryopreserved human primary hepatocytes cultivated in microfluidic biochips. Xenobiotica 43 (2), 140152. Bell, C.C., et al., 2016. Characterization of primary human hepatocyte spheroids as a model system for drug-induced liver injury, liver function and disease. Sci. Rep. 6, 25187. Bertram, R., Gram Pedersen, M., Luciani, D.S., Sherman, A., 2006. A simplified model for mitochondrial ATP production. J. Theor. Biol. 243, 575586. Bhatia, S.N., Ingber, D.E., 2014. Microfluidic organs-on-chips. Nat. Biotech. 32 (8), 760772. Bittermann, K., Goss, K.-U., 2017. Predicting apparent passive permeability of Caco-2 and MDCK cell-monolayers: a mechanistic model. PLoS One 12 (12), e0190319. Borhani, D.W., Shaw, D.E., 2012. The future of molecular dynamics simulations in drug discovery. J. Comput. Aided Mol. Des. 26, 15.

References

Bovard, D., Sandoz, A., Luettich, K., Frentzel, S., Iskandar, A., Marescotti, D., et al., 2018. A lung/liver-on-a-chip platform for acute and chronic toxicity studies. Lab Chip 18, 3814. Bowman, C.M., Benet, L.Z., 2016. Hepatic clearance predictions from in vitro-in vivo extrapolation and the biopharmaceuitcs drug disposition classification system. Drug Metab. Dispos. 44 (11), 17311735. Brown, T.D., Nowak, M., Bayles, A.V., Prabhakarpandian, B., Karande, P., Lahann, J., et al., 2019. A microfluidic model of human brain (μHuB) for assessment of blood brain barrier. Bioeng. Transl. Med. 4, e10126. Early view, 13 Jan 2019. Buckley, S.T., Fischer, S.M., Fricker, G., Brandl, M., 2012. In vitro models to evaluate the permeability of poorly soluble drug entities: challenges and perspectives. Eur. J. Pharm. Sci. 45, 235250. Burova, I., Peticone, C., Thompson, D.D.S., Knowles, J.C., Wall, I., Shipley, R.J., 2019. A parameterised mathematical model to elucidate osteoblast cell growth in a phosphateglass microcarrier culture. J. Tissue Eng. 10, 114. Byun, C.K., Abi-Samra, K., Cho, Y.-K., Takayama, S., 2014. Pumps for microfluidic cell culture. Electrophoresis. 35, 245257. Camp, J.P., Capitano, A.T., 2007. Induction of zone-like liver function gradients in HepG2 cells by varying culture medium height. Biotechnol. Prog. 23 (6), 14851491. Capulli, A.K., Tian, K., Mehandru, N., Bukhta, A., Choudhury, S.F., Suchyta, M., et al., 2014. Approaching the in vitro clinical trial: engineering organs on chips. Lab Chip 14, 31813186. Chalhoub, E., Hanson, R.W., Belovich, J.M., 2007. A computer model of gluconeogenesis and lipid metabolism in the perfused liver. Am. J. Physiol. Endocrinol. Metab. 293, E1676E1686. Chan, T.S., Yu, H., Moore, A., Khetani, S.R., Tweedie, D., 2013. Meeting the challenge of predicting hepatic clearance of compounds slowly metabolized by cytochrome P450 using a novel hepatocyte model, HepatoPac. Drug Metab. Dispos. 41 (12), 20242032. Chen, Z.J., Przekwas, A.J., 2010. A coupled pressure-based computational method for incompressible/compressible flows. J. Comput. Phys. 229 (24), 91509165. Chen, W.L.K., Edington, C., Suter, E., et al., 2017. Integrated gut/liver microphysiological systems elucidates inflammatory inter-tissue crosstalk. Biotech. Bioeng. 114 (11), 2648 59. Chen, Y., Jin, J.Y., Mukadam, S., Malhi, V., Kenny, J.R., 2012. Application of IVIVE and PBPK modeling in prospective prediction of clinical pharmacokinetics: strategy and approach during the drug discovery phase with four case studies. Biopharm. Drug Dispos. 33, 8598. Chiba, M., Ishii, Y., Sugiyama, Y., 2009. Prediction of hepatic clearance in human from in vitro data for successful drug development. AAPS J. 11 (2), 262276. Cirit, M., Stokes, C.L., 2018. Maximizing the impact of Microphysiological systems with in vitroin vivo translation. Lab Chip 2018 (18), 18311837. Clancy, C.E., An, G., Cannon, W.R., et al., 2016. Multiscale modeling in the clinic: drug design and development. Ann. Biomed. Eng. 44 (9), 25912610. Cortassa, S., Aon, M.A., 1994. Spatio-temporal regulation of glycolysis and oxidative phosphorylation in vivo in tumor and yeast cells. Cell Biol. Int. 18 (7), 687714. Crewe, H.K., Barter, Z.E., Rowland Yeo, K., Rostami-Hodjegan, A., 2011. Are there differences in the catalytic activity per unit enzyme of recombinantly expressed and

353

354

CHAPTER 10 Computational pharmacokinetic modeling

human liver microsomal cytochrome P450 2C9? A systematic investigation into intersystem extrapolation factors. Biopharm. Drug Dispos. 32 (6), 303318. Danhof, M., de Lange, E.C., Della Pasqua, O.E., Ploeger, B.A., Voskuyl, R.A., 2008. Mechanism-based pharmacokinetic-pharmacodynamic (PK-PD) modeling in translational drug research. Trends Pharmacol. Sci. 29 (4), 186191. Davies, B., Morris, T., 1993. Physiological parameters in laboratory animals and humans. Pharm. Res. 10 (7), 10931095. Deosarkar, S.P., Prabhakarpandian, B., Wang, B., Sheffield, J.B., Krynska, B., Kiani, M.F., et al., 2015. A novel dynamic neonatal blood-brain barrier on a chip. PLoS One 10 (11), e0142725. Dereli-Korkut, Z., Akaydin, H.D., Rezwanuddin Ahmed, A.H., Jiang, X., Wang, S., 2014. Three dimensional microfluidic cell arrays for ex vivo drug screening with mimicked vascular flow. Anal. Chem. 86, 29973004. Domansky, K., Inman, W., Serdy, J., Dash, A., Lim, M.H.M., Griffith, L.G., 2010. Perfused multiwell plate for 3D liver tissue engineering. Lab Chip 10, 5158. Durrant, J.D., McCammon, J.A., 2011. Molecular dynamics simulations and drug discovery. BMC Biol. 9, 71. Ebrahimkhani, M.R., Shepard Neiman, J.A., Raredon, M.S.B., Hughes, D.J., Griffith, L.G., 2014. Bioreactor technologies to support liver function in vitro. Adv. Drug Deliv. Rev. 69, 132157. Edmondson, R., Broglie, J.J., Adcock, A.F., Yang, L., 2014. Three dimensional cell culture systems and their applications in drug discovery and cell-based biosensors. Assay Drug Dev. Technol. 12 (4), 207218. Erickson, D., 2005. Towards numerical prototyping of labs-on-chip: modeling for integrated microfluidic devices. Microfluid. Nanofluid. 1, 301318. Esch, E.W., Bahinski, A., Huh, D., 2015. Organs-on-chips at the frontiers of drug discovery. Nat. Rev. Drug Discov. 14, 248260. Esch, M.B., Mahler, G.J., 2019. Body-on-a-chip systems: design, fabrication, and applications. In: Borenstein, J.T., Tandon, V., Charest, J.L. (Eds.), Elsevier Inc, Microfluidic Cell Culture Systems, second ed. pp. 323350. Esch, M.B., Mahler, G.J., Stokol, T., Shuler, M.L., 2014. Body-on-a-chip simulation with gastrointestinal tract and liver tissues suggests that ingested nanoparticles have the potential to cause liver injury. Lab Chip 14, 30813092. FDA, 1997. Guidance for Industry. Extended Release Oral Dosage Forms: Development, Evaluation, and Application of In Vitro/In Vivo Correlations. Fabre, K.M., Livingston, C., Tagle, D.A., 2014. Organs-on-chips (microphysiological systems): tools to expedite efficacy and toxicity testing in human tissue. Exp. Biol. Med. 239, 10731077. Fadda, S., Cincotti, A., Cao, G., 2012. A novel population balance model to investigate the kinetics of in vitro cell proliferation: part I. Model development. Biotechnol. Bioeng. 109 (3), 772781. Fennema, E., Rivron, N., Rouwkema, J., van Blitterswijk, C., de Boer, J., 2013. Spheroid culture as a tool for creating 3D complex tissues. Trends Biotechnol. 31, 108115. Fine, B., Vunjak-Novakovic, G., 2017. Shortcomings of animal models and the rise of engineered human cardiac tissue. ACS Biomater. Sci. Eng. 3 (9), 18841897. Gao, D., Liu, H., Lin, J.M., Wang, Y., Jiang, Y., 2013. Characterization of drug permeability in Caco-2 monolayers by mass spectrometry on a membrane-based microfluidic device. Lab Chip 13 (5), 978985.

References

Geerts, H., Spiros, A., Roberts, P., Carr, R., 2013. Quantitative systems pharmacology as an extension of PK/PD modeling in CNS research and development. J. Pharmacokinet. Pharmacodyn. 40 (3), 257265. Glatzel, T., Litterst, C., Cupelli, C., Lindemann, T., Moosmann, C., Niekrawietz, R., et al., 2008. Computational fluid dynamics (CFD) software tools for microfluidic applications  a case study. Comput. Fluids 37, 218235. Gokaltun, A., Yarmush, M.L., Asatekin, A., Usta, O.B., 2017. Recent advances in nonbiofouling PDMS surface modification strategies applicable to microfluidic technology. Technology 5 (01), 112. Griffith, L.G., Wells, A., Stolz, D.B., 2014. Engineering liver. Hepatology 60, 14261434. Hartmanshenn, C., Rao, R.T., Bae, S.-A., Scherholz, M.L., Acevedo, A., Pierre, K.K., et al., 2018. Quantitative systems pharmacology: extending the envelope through systems engineering. Comput. Aided Chem. Eng. 42, 334. Hartung, T., Zurlo, J., 2012. Alternative approaches for medical countermeasures to biological and chemical terrorism and warfare. ALTEX 29 (3), 251260. Herland, A., et al., 2019. Quantitative prediction of human drug pharmacokinetic responses enabled by fluidically coupled vascularized organ chips. Nature Biotechnology in review. Hirschhaeuser, F., Menne, H., Dittfeld, C., West, J., Mueller-Klieser, W., Kunz-Schughart, L.A., 2010. Multicellular tumor spheroids: an underestimated tool is catching up again. J. Biotechnol. 148 (2010), 315. Howell, B.A., Yang, Y., Kumar, R., Woodhead, J.L., Harrill, A.H., Clewell III, H.J., et al., 2012. In vitro to in vivo extrapolation and species response comparisons for druginduced liver injury (DILI) using DILIsymt: a mechanistic, mathematical model of DILI. J. Pharmacokinet. Pharmacodyn. 39 (5), 527541. Huang, W., Lee, S.L., Yu, L.X., 2009. Mechanistic approaches to predicting oral drug absorption. The AAPS J. 11 (2), 217224. Hubatsch, I., Ragnarsson, E.G.E., Artursson, P., 2007. Determination of drug permeability and prediction of drug absorption in Caco-2 monolayers. Nat. Protoc. 2 (9), 21112119. Huber, D., Oskooei, A., Casadevall, I., Solvas, X., deMello, A., Kaigala, G.V., 2018. Hydrodynamics in cell studies. Chem. Rev. 118, 20422079. Huh, D., Matthews, B.D., Mammoto, A., Montoya-Zavala, M., Hsin, H.Y., Ingber, D.E., 2010. Reconstituting organ-level lung functions on a chip. Science 328, 16621668. Huh, D., Hamilton, G.A., Ingber, D.E., 2011. From 3D cell culture to organs-on-chips. Trends Cell Biol. 21 (12), 745754. Huh, D., Torisawa, Y.S., Hamilton, G.A., Kim, H.J., Ingber, D.E., 2012. Microengineered physiological biomimicry: organs-on-chips. Lab Chip. 12 (12), 21562164. Hung, P.J., Lee, P.J., Sabounchi, P., Lin, R., et al., 2005. Continuous perfusion microfluidic cell culture array for high-throughput cell-based assays. Biotechnol. Bioeng. 89, 18. Ito K, Houston JB., Prediction of human drug clearance from in vitro and preclinical data using physiologically based and empirical approaches, Pharm Res. 22(1):103-112. Ito, K., Houston, J.B., 2005. Prediction of human drug clearance from in vitro and preclinical data using physiologically based and empirical approaches. Pharm. Res. 22 (1), 103112. Jamei, M., Marciniak, S., Feng, K., et al., 2009. The Simcyp® population-based ADME simulator. Expert Opin. Drug Metabol. Toxicol. 5 (2), 211223.

355

356

CHAPTER 10 Computational pharmacokinetic modeling

Jamei, M., Bajot, F., Neuhoff, S., Barter, Z., Yang, J., Rostami-Hodjegan, A., et al., 2014. A Mechanistic framework for in vitroin vivo extrapolation of liver membrane transporters: prediction of drugdrug interaction between rosuvastatin and cyclosporine. Clin. Pharmacokinet. 53, 7387. Jiang, Y., Przekwas, A.J., 1994. Implicit, Pressure-Based Incompressible Navier-Stokes Equations Solver for Unstructured Mesh, AIAA-94-0303 Jones, H.M., Rowland-Yeo, K., 2013. Basic concepts in physiologically based pharmacokinetic modeling in drug discovery and development. CPT Pharmacometrics Syst. Pharmacol. 2 (8), 112. Jusko, W.J., 2013. Moving from basic toward systems pharmacodynamic models. J. Pharm. Sci. 102 (9), 29302940. Kalra, P., Brandl, J., Gaub, T., Niederalt, C., Lippert, J., Sahle, S., et al., 2019. Quantitative systems pharmacology of interferon alpha administration: a multi-scale approach. PLoS One 14 (2), e0209587. Kannan, R., Chen, Z.J., Singh, N., Przekwas, A., Delvadia, R., Tian, G., et al., 2017. A quasi-3D wire approach to model pulmonary airflow in human airways. Int. J. Numer. Methods Biomed. Eng. 33 (7), e2838. Kannan, R.R., Singh, N., Przekwas, A., 2018. A compartment-quasi-3D multiscale approach for drug absorption, transport, and retention in the human lungs. Int. J. Numer. Methods Biomed. Eng. 34 (5), 120. Kim, H.J., Huh, D., Hamilton, G., Ingber, D.E., 2012. Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow. Lab Chip 12 (12), 21652174. Kim, M.-C., Lam, R.H.W., Thorsen, T., Asada, H.H., 2013. Mathematical analysis of oxygen transfer through polydimethylsiloxane membrane between double layers of cell culture channel and gas chamber in microfluidic oxygenator. Microfluid. Nanofluid. 15, 285296. Knight-Schrijver, V.R., Chelliah, V., Cucurull-Sanchez, L., Le Nove`rea, N., 2016. The promises of quantitative systems pharmacology modelling for drug development. Comput. Struct. Biotechnol. J. 14, 363370. Kockmann, N., Kiefer, T., Engler, M., Woias, P., 2006. Convective mixing and chemical reactions in microchannels with high flow rates. Sens. Actuators, B: Chem. 117, 495508. Kola, I., Landis, J., 2004. Can the pharmaceutical industry reduce attrition rates? Nat. Rev. Drug Discov. 3, 711715. Kondo, J., et al., 2011. Retaining cell- cell contact enables preparation and culture of spheroids composed of pure primary cancer cells from colorectal cancer. PNAS 108, 62356240. Kra¨mer, S.D., 2016. Quantitative aspects of drug permeation across in vitro and in vivo barriers. Eur. J. Pharm. Sci. 87, 3046. Kratochwil, N.A., Meille, C., Fowler, S., et al., 2017. Metabolic profiling of human longterm liver models and hepatic clearance predictions from in vitro data using nonlinear mixed-effects modeling. AAPS J. 19 (2), 534550. Kulkarni, P., Korzekwa, K., Nagar, S., 2016. Intracellular unbound atorvastatin concentrations in the presence of metabolism and transport. J. Pharmacol. Exp. Ther. 359, 2636.

References

Kwapiszewska, K., Michalczuk, A., Rybka, M., Kwapiszewski, R., Brzo´zka, Z., 2014. A microfluidic-based platform for tumour spheroid culture, monitoring and drug screening. Lab Chip 14, 20962104. Larregieu, C.A., Benet, L.Z., 2014. Distinguishing between the permeability relationships with absorption and metabolism to improve BCS and BDDCS predictions in early drug discovery. Mol. Pharm. 11, 13351344. Lee, P.J., Hung, P.J., Lee, L.P., 2007. An artificial liver sinusoid with a microfluidic endothelial-like barrier for primary hepatocyte culture. Biotechnol. Bioeng. 97, 13401346. Lee, S.H., Jun, B.-H., 2019. Advances in dynamic microphysiological organ-on-a-chip: design principle and its biomedical application. J. Ind. Eng. Chem. 71, 6577. Lewis, M.C., MacArthur, B.D., Malda, J., Pettet, G., Please, C.P., 2005. Heterogeneous proliferation within engineered cartilaginous tissue: the role of oxygen tension. Biotech. Bioeng. 91 (5), 607. Low, L.A., Tagle, D.A., 2017. Organs-on-chips: progress, challenges, and future directions. Exp. Biol. Med. 242 (16), 15731578. Ma, C.Y.J., Kumar, R., Xu, X.Y., Mantalaris, A., 2007. A combined fluid dynamics mass transport and cell growth model for a three-dimensional perfused bioreactor for tissue engineering of haematopoietic cells. Biochem. Eng. J. 35 (1), 111. Maass, C., Stokes, C.L., Griffith, L.G., Cirit, M., 2017. Multi-functional scaling methodology for translational pharmacokinetic and pharmacodynamic applications using integrated microphysiological systems (MPS). Integr. Biol. (UK) 9, 290302. Marx, U., Walles, H., Hoffmann, S., Lindner, G., Horland, R., Sonntag, F., et al., 2012. ‘Human-on-a-chip’ developments: a translational cutting-edge alternative to systemic safety assessment and efficiency evaluation of substances in laboratory animals and man? Altern. Lab. Anim. 40 (5), 235. Marx, U., Andersson, T.B., Bahinski, A., et al., 2016. Biology-inspired microphysiological system approaches to solve the prediction dilemma of substance testing. ALTEX 33 (3), 272321. Marshall, J.S., 2009. Discrete-element modeling of particulate aerosol flows. J. Comp. Physics 228, 15411561. Maxmen, A., 2011. Translational research: the American way. Nature 478, S16S18. McGonigle, P., Ruggeri, B., 2014. Animal models of human disease: challenges in enabling translation. Biochem. Pharm. 87 (1), 162171. McKee, T.J., Komarova, S.V., 2017. Is it time to reinvent basic cell culture medium? Am. J. Physiol. Cell Physiol. 312, C624C626. Molavian, H.R., Kohandel, M., Milosevic, M., et al., 2009. Fingerprint of cell metabolism in the experimentally observed interstitial pH and pO2 in solid tumors. Cancer Res. 69, 91419147. Morgan, P., et al., 2018. Impact of a five-dimensional framework on R&D productivity at AstraZeneca. Nat. Rev. Drug Discov. 17 (3), 167181. Mukhopadhyay, R., 2007. When PDMS isn’t the best. Anal. Chem. 79, 32493253. Nagar, S., Tucker, J., Weiskircher, E.A., Bhoopathy, S., Hidalgo, I.J., Korzekwa, K., 2014. Compartmental models for apical efflux by P-glycoprotein—Part 1: Evaluation of model complexity. Pharm. Res. 31, 347359. Naritomi, Y., Terashita, S., Kimura, S., Suzuki, A., Kagayama, A., Sugiyama, Y., 2001. Prediction of human hepatic clearance from in vivo animal experiments and in vitro

357

358

CHAPTER 10 Computational pharmacokinetic modeling

metabolic studies with liver microsomes from animals and humans. Drug Metab. Dispos. 29 (10), 13161324. Ng, S., March, S., Galstian, A., Hanson, K., Carvalho, T., Mota, M.M., et al., 2014. Hypoxia promotes liver-stage malaria infection in primary human hepatocytes in vitro. Dis. Models Mech. 7 (2), 215224. Nijsen, M., Wu, F., Bansal, L., et al., 2018. Preclinical QSP modeling in the pharmaceutical industry: an IQ consortium survey examining the current landscape. CPT Pharmacometrics Syst. Pharmacol. 2018 (7), 135146. Norton, K.-A., Gong, C., Jamalian, S., Popel, A.S., 2019. Multiscale agent-based and hybrid modeling of the tumor immune microenvironment. Processes 2019 (7), 37. Novak, R., et al., 2019. A robotic platform for fluidically-linked human body-on-chips experimentation. bioRxiv under review, Available at https://www.biorxiv.org/content/ 10.1101/569541v1. Olson, H., et al., 2000. Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul. Toxicol. Pharmacol. 32, 5667. Olufsen, M.S., 1999. Structured tree outflow condition for blood flow in larger systemic arteries. Am. J. Physiol. 276 (1), H257H268. Osaki, T., Sivathanu, V., Kamm, R.D., 2018. Vascularized microfluidic organ-chips for drug screening, disease models and tissue engineering. Curr. Opin. Biotechnol. 52, 116123. O’Hara, T., Hayes, S., Davis, J., Devane, J., Smart, T., Dunne, A., 2001. In vivoin vitro correlation (IVIVC) modeling incorporating a convolution step. J. Pharmacokinet. Pharmacodyn. 28, 277298. Patrachari, A.R., Podichetty, J.T., Madihally, S.V., 2012. Application of computational fluid dynamics in tissue engineering. J. Biosci. Bioeng. 114 (2), 123132. Paul, S.M., et al., 2010. How to improve R&D productivity: the pharmaceutical industry’s grand challenge. Nat. Rev. Drug Discov. 9, 203214. Pelkonen, O., Turpeinen, M., 2007. In vitroin vivo extrapolation of hepatic clearance: biological tools, scaling factors, model assumptions and correct concentrations. Xenobiotica 37 (1011), 10661089. Peskir, G., 2003. On the diffusion coefficient: the Einstein relation and beyond. Stochastic Models 19 (3), 383405. Pinero, J., Furlong, L.I., Sanz, F., 2018. In silico models in drug development: where we are. Curr. Opin. Pharmacol. 42, 111121. Poulin, P., Haddad, S., 2013. Toward a new paradigm for the efficient in vitroin vivo extrapolation of metabolic clearance in humans from hepatocyte data. J. Pharm. Sci. 102 (9), 3239 51. Poulin, P., Theil, F.P., 2000. A priori prediction of tissue: plasma partition coefficients of drugs to facilitate the use of physiologically-based pharmacokinetic models in drug discovery. J. Pharm. Sci. 89 (1), 1635. Prantil-Baun, R., Novak, R., Das, D., Somayaji, M.R., Przekwas, A., Ingber, D.E., 2018. Physiologically based pharmacokinetic and pharmacodynamic analysis enabled by microfluidically linked organs-on-chips. Annu. Rev. Pharmacol. Toxicol. 58, 3764. Przekwas, A., Makhijani, V., Athavale, M., Klein, A., Bartsch, P., 2000a. Computational simulation of bio-microfluidic processes in integrated DNA biochips. In: van den Berg, A., Olthuis, W., Bergveld, P. (Eds.), Micro Total Analysis Systems 2000. Springer, Dordrecht.

References

Przekwas, A., Friend, T., Teixeira, R., Chen, Z.J., Wilkerson, P., Oct 2006. Spatial Modeling Tools for Cell Biology, DARPA/AFRL-IF-RS-TR-2006-318 Rep. Available from: ,www.dtic.mil/cgi-bin/GetTRDoc?AD 5 ADA460852. Przekwas, A.M., Wang, D., Makhijani, V.B., Przekwas, A.J., 2000b. Microfluidic filtration chip for DNA extraction and concentration. In: Ehrfeld, W. (Ed.), Microreaction Technology: Industrial Prospects. Springer, Berlin, Heidelberg. Regehr, K.J., Domenech, M., Koepsel, J.T., Carver, K.C., Ellison-Zelski, S.J., Murphy, W. L., et al., 2009. Biological implications of polydimethylsiloxane-based microfluidic cell culture. Lab Chip 9 (15), 21322139. Riley, R.J., McGinnity, D.F., Austin, R.P., 2005. A unified model for predicting human hepatic, metabolic clearance from in vitro intrinsic clearance data in hepatocytes and microsomes. Drug Metab. Dispos. 33, 13041311. Rodgers, T., Rowland, M., 2006. Physiologically based pharmacokinetic modelling 2: predicting the tissue distribution of acids, very weak bases, neutrals and zwitterions. J. Pharm. Sci. 95 (6), 12381257. Rodgers, T., Rowland, M., 2007. Mechanistic approaches to volume of distribution predictions: understanding the processes. Pharm. Res. 24 (5), 918933. Rodgers, T., Leahy, D., Rowland, M., 2005. Physiologically based pharmacokinetic modeling 1: predicting the tissue distribution of moderate-to-strong bases. J. Pharm. Sci. 94 (6), 12591276. Rostami-Hodjegan, A., 2012. Physiologically based pharmacokinetics joined with in vitroin vivo extrapolation of ADME: a marriage under the arch of systems pharmacology. Clin. Pharmacol. Ther. 92 (1), 5061. Sarkar, U., Rivera-Burgos, D., Large, E.M., Hughes, D.J., et al., 2015. Metabolite profiling and pharmacokinetic evaluation of hydrocortisone in a perfused three-dimensional human liver bioreactor. Drug Metab. Dispos. 43, 10911099. Schuhmacher, A., Gassmann, O., Hinder, M., et al., 2016. Changing R&D models in research-based pharmaceutical companies. J. Transl. Med. 14, 105. Sengers, B.G., Heywood, H.K., Lee, D.A., Oomens, C.W., Bader, D.L., 2005. Nutrient utilization by bovine articular chondrocytes: a combined experimental and theoretical approach. J. Biomech. Eng. 127 (5), 758766. Sharma, V., McNeill, J.H., 2009. To scale or not to scale: the principles of dose extrapolation. Br. J. Pharmacol. 157, 907921. Sherwin, S.J., Formaggia, L., Peiro´, J., Franke, V., 2003. Computational modelling of 1D blood flow with variable mechanical properties and its application to the simulation of wave propagation in the human arterial system. Int. J. Numer. Methods Fluids 43 (67), 673700. Shuler, M.L., 2019. Advances in organ-, body-, and disease-on-a-chip systems. Lab Chip 19, 910. Shuler, M.L., Esch, M.B., 2010. Body-on-a chip: using microfluidic systems to predict human responses to drugs. Pure Appl. Chem. 82 (8), 16351645. Sjogren, E., Thorn, H., Tannergren, C., 2016. In silico modeling of gastrointestinal drug absorption: predictive performance of three physiologically based absorption models. Mol. Pharm. 13 (6), 17631778. Sluka, J.P., Fu, X., Swat, M., Belmonte, J.M., Cosmanescu, A., Clendenon, S.G., et al., 2016. A liver-centric multiscale modeling framework for xenobiotics. PLoS One 11 (9), e0162428.

359

360

CHAPTER 10 Computational pharmacokinetic modeling

Somayaji, M.R., 2016. Perspectives on development and regulation of therapeutic products for CED-based therapy of neurodegenerative diseases. Curr. Pharm. Biotechnol. 17, 495512. Somayaji, M.R., Das, D., Przekwas, A., 2016. Computational approaches for modeling and analysis of human-on-chip systems for drug testing and characterization. Drug Discov. Today 21 (12), 18591862. Somayaji, M.R., Das, D., Przekwas, A.J., 2018. Drug testing and characterization using human-on-chip (HoC) systems: some thoughts on the application of in vitro-in vivo correlation. Drug Discov. Today 23 (9), 1571. Sontheimer-Phelps, A., Hassell, B.A., Ingber, D.E., 2019. Modelling cancer in microfluidic human organs-on-chips. Nat. Rev. Cancer 19, 6581. Stokes, C.L., Cirit, M., Lauffenburger, D.A., 2015. Physiome-on-a-chip: the challenge of “scaling” in design, operation, and translation of microphysiological systems. CPT Pharmacometrics Syst. Pharmacol. 4, 559562. Sumaru, K., Sugiura, T., Kanamori, T., 2007. Optimal design of cell culture chip on the basis of oxygen and glucose supply to cultivated cells in the chip. Biochem. Eng. J. 36, 304309. Sun, H., Pang, K.S., 2008. Permeability, transport, and metabolism of solutes in Caco-2 cell monolayers: a theoretical study. Drug Metab. Dispos. 36 (1), 102123. Sun, W., Chen, Y., Wang, Y., Luo, P., Zhang, M., Zhang, H., et al., 2018. Analyst 143, 5431. Sung, J.H., Shuler, M.L., 2009. Prevention of air bubble formation in a microfluidic perfusion cell culture system using a microscale bubble trap. Biomed. Microdevices 11, 731738. Sung, J.H., Wang, Y.I., Kim, J.H., Lee, J.M., Shuler, M.L., 2018. Application of chemical reaction engineering principles to “body-on-a-chip” systems. AIChE J. 64 (12), 43514360. Sutherland, M.L., Fabre, K.M., Tagle, D.A., 2013. The National Institutes of Health Microphysiological Systems Program Focuses on a Critical Challenge in the Drug Discovery Pipeline Tanyeri, M., Ranka, M., Sittipolkul, N., Schroeder, C.M., 2011. A microfluidic-based hydrodynamic trap: design and implementation. Lab Chip. 11, 17861794. Thomas, P., Raghavan, S.R., Forry, S.P., 2011. Regulating oxygen levels in a microfluidic device. Anal. Chem. 2011 (83), 88218824. Tong, W.H., et al., 2016. Constrained spheroids for prolonged hepatocyte culture. Biomaterials 80 (2016), 106e120. Tostoes, R.M., Leite, S.B., Serra, M., Jensen, J., Bjorquist, P., Carrondo, M.J., et al., 2012. Human liver cell spheroids in extended perfusion bioreactor culture for repeated-dose drug testing. Hepatology 55 (4), 12271236. Tsamandouras, N., Chen, W.L.K., Edington, C.D., Stokes, C.L., Griffith, L.G., Cirit, M., 2017. Integrated gut and liver microphysiological systems for quantitative in vitro pharmacokinetic studies. AAPS J. 19 (5), 14991512. Van Meer, B.J., De Vries, H., Firth, K.S.A., van Weerd, J., Tertoolen, L.G.J., Karperien, H.B.J., et al., 2017. Small molecule absorption by PDMS in the context of drug response bioassays. Biochem. Biophys. Res. Commun. 482 (2), 323328. Visser, S.A.G., de Alwis, D.P., Kerbusch, T., Stone, J.A., Allerheiligen, S.R.B., 2014. Implementation of quantitative and systems pharmacology in large pharma. CPT Pharmacometrics Syst. Pharmacol. 3, e142.

Further reading

Vernetti, L., Gough, A., Baetz, N., et al., 2017. Functional coupling of human microphysiology systems: intestine, liver, kidney proximal tubule, blood-brain barrier and skeletal muscle. Sci. Rep. 7, 42296. Wagner, J.G., Szpunar, G.J., Ferry, J.J., 1985. Michaelis-Menten elimination kinetics: areas under curves, steady-state concentrations, and clearances for compartment models with different types of input. Biopharm. Drug Dispos. 6 (2), 177200. Wang, Y.I., Shuler, M.L., 2018. UniChip enables long-term recirculating unidirectional perfusion with gravity-driven flow for microphysiological systems. Lab Chip 18, 25632574. Wang, Z., et al., 2015. HepaRG culture in tethered spheroids as an in vitro threedimensional model for drug safety screening. J. Appl. Toxicol. 35 (8), 909917. Wikswo, J.P., Porter, A.P., 2015. Biology coming full circle: joining the whole and the parts. Exp. Biol. Med. 240, 37. Wikswo, J.P., Curtis, E.L., Eagleton, Z.E., Evans, B.C., Kole, A., Hofmeister, L.H., et al., 2013. Scaling and systems biology for integrating multiple organs-on-a-chip. Lab Chip 13, 3496. Wood, F.L., Houston, J.B., Hallifax, D., 2017. Clearance prediction methodology needs fundamental improvement: trends common to rat and human hepatocytes/microsomes and implications for experimental methodology. Drug Metab. Dispos. 45 (11), 11781188. Wu, W.-H., Punde, T.H., Shih, P.-C., Fu, C.-Y., Wang, T.-P., Hsu, L., et al., 2015. A capillary-endothelium-mimetic microfluidic chip for the study of immune responses. Sens. Actuators, B: Chem. 209, 470477. Yang, Y., Nadanaciva, S., Will, Y., Woodhead, J.L., Howell, B.A., Watkins, P.B., et al., 2015. MITOsym®: a mechanistic, mathematical model of hepatocellular respiration and bioenergetics. Pharm. Res. 32 (6), 19751992. Yu, L.X., Amidon, G.L., 1999. A compartmental absorption and transit model for estimating oral drug absorption. Int. J. Pharm. 186 (2), 119125. Zhang, B., Radisic, M., 2017. Organ-on-a-chip devices advance to market. Lab Chip 17, 23952420. Zhao, H., Caflisch, A., 2019. Molecular dynamics in drug design. Eur. J. Med. Chem. 91 (16), 414. Zuegge, J., Schneider, G., Coassolo, P., Lave, T., 2009. Prediction of hepatic metabolic clearance: comparison and assessment of prediction models. Clin. Pharmacokinet. 40 (7), 553563.

Further reading Lee, H., et al., 2014. Microvasculature: an essential component for organ-on-chip systems. MRS Bull. 39 (1), 5159. Lee, S.H., Hong, S.G., Song, J., Cho, B., Han, E.J., Kondapavulur, S., et al., 2017. Microphysiological analysis platform of pancreatic islet β-cell spheroids. Adv. Healthc. Mater. 2017, 1701111. Maass, C., et al., 2018. Establishing quasi-steady state operations of microphysiological systems (MPS) using tissue-specific metabolic dependencies. Sci. Rep. 8, 113.

361

CHAPTER

Design and engineering of multiorgan systems

12

Kasper Renggli1 and Olivier Frey2 1

ETH Zu¨rich, Department of Biosystems Science and Engineering, Basel, Switzerland 2 InSphero AG, Schlieren, Switzerland

Motivation for in vitro multiorgan systems Organ-on-chip devices, such as the examples presented in this book, allow researchers to mimic a wide range of physiological and pathological states and mechanisms of single organs in the presence or absence of stimulating agents. The grade, quality, and features of these devices for emulating the in vivo situation are largely dependent on their architecture and implemented complexity. Multiorgan systems could potentially advance in vitro assays to the next level by not only modeling the isolated effects of a single organ but also providing systemic insight into the interaction of organs in the human body (Renggli et al., 2019; Rogal et al., 2017; Ronaldson-Bouchard and Vunjak-Novakovic, 2018; Wang et al., 2018). Secreted signaling molecules of one organ, for example, insulin from the pancreas, can induce effects in other organs (liver, muscle, fat). Multiorgan systems have received growing interest in research and in the drugdevelopment process, because of their relevance to understanding complex modes of action, for instance, in the field of immuno-cancer therapy. This multitissue interplay or systemic view has typically only been available through animal models, which sometimes fail to mimic the human response and are more and more disputed (Mak et al., 2014; Matthews, 2008). The full potential of multiorgan systems can be harnessed not by simply connecting different organ-on-chip devices to monitor their function in an isolated manner, but by matching the different organs so that they can communicate and influence each other, allowing the study and modulation of interorgan pathways. The number of interconnected organs largely depends on the required model complexity and biological questions asked. It must be noted, however, that in some applications, such as off-target secondary effects of compounds, the identity of affected organs or influencing factors may be unknown, making it difficult to predict the minimal number of required models. To date, a number of meaningful examples for applications of direct organ organ interaction have been presented, illustrating the benefit these

Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00012-7 © 2020 Elsevier Inc. All rights reserved.

393

394

CHAPTER 12 Design and engineering of multiorgan systems

systems can provide to research and drug development. The systems range from 2 to 14 organs interlinked in a fluidic network. Coculture systems of human artificial liver microtissues and human neurospheres have been used to study the effects and toxicity profiles of various compounds, and a higher sensitivity than in single-tissue cultures has been observed (Materne et al., 2015). Studies have also been performed to monitor drug-induced organ toxicity in a dual-organ human liver heart system exposed to acetaminophen and a human liver cancer heart system exposed to doxorubicin (Zhang et al., 2017). An extension of the system with a lung model allowed Skardal et al. (2017) to investigate not only responses of the individual components to a panel of drugs, but also more complex integrated responses, where the functionality of one organoid influenced another organoid. Liver tumor systems have been designed to mimic the function of prodrugs and their cytostatic effect when metabolized by the liver. In one example the metabolism of tegafur, a prodrug of 5-fluorouracil, could be reproduced with consequent death of tumor cells. Death could only be observed in the presence of liver and not in cultures in a 96-well microtiter plate (Sung and Shuler, 2009). In another example, rat liver and colorectal tumor microtissues were interconnected on a chip and cultured over 8 days in the presence of cyclophosphamide. The prodrug had a significant impact on tumor growth, but only after bioactivation by the liver; the effect was only observed in the perfused and interconnected cocultures on-chip (Frey et al., 2014; Kim et al., 2015b). Imura et al. added colorectal adenocarcinoma Caco-2 cells to a permeable membrane (Imura et al., 2010) for a small intestine liver model. Human breast carcinoma MCF-7 cells were used as target of anticancer agents and estrogen-like substances for assaying intestinal absorption and hepatic metabolism. In a very similar setup, but replacing breast with lung carcinoma cells, pharmacokinetic studies have been performed with three types of substrates—epirubicin, irinotecan, and cyclophosphamide—to model the effects of orally administered or biologically active anticancer drugs (Kimura et al., 2015). The effect of hepatic metabolic function has been demonstrated in a liver heart coculture (Oleaga et al., 2018; Vernetti et al., 2016); the initially safe cyclophosphamide induced cardiotoxic effects only after metabolism. In contrast, the toxicity on the heart could be significantly reduced for the initially cardiotoxic terfenadine through metabolization. Nephrotoxicity of ifosfamide after bioactivation was demonstrated in a liver kidney biochip, with a significant reduction of viability of madin-darby canine kidney (MDCK) cells compared with cell viability in untreated cocultures or treated MDCK monocultures (Choucha-Snouber et al., 2013). In all these cases, primary aspects of in vivo cross-talk of metabolic drug activation and either their efficacy on targets or their toxicity off target could be replicated with two or three organs in a single in vitro platform. All studies emphasized the need for a functional liver model as metabolizing element and a direct fluidic interaction with the other organs, especially if short-lived metabolites are involved. A pancreatic islet liver cross-talk model developed to study

Motivation for in vitro multiorgan systems

insulin and glucose regulation (Bauer et al., 2017) demonstrated functional coupling by insulin release from the islet microtissues in response to a glucose load. Insulin secreted into the circulation stimulated glucose uptake via the liver spheroids in the model. In the absence of insulin, the liver did not consume glucose as efficiently. Metabolic interactions between the microvasculature and brain neurons can be elucidated using a neurovascular unit in which three organ chips are coupled to model influx across the blood brain barrier, the brain parenchymal compartment, and efflux across the blood brain barrier (Maoz et al., 2018). Importantly, the model illustrates that multiorgan devices can be used to study both the interactions and the roles of individual cell types. Another example is cross-talk between the gut and the liver (Chen et al., 2017; Tsamandouras et al., 2017). Human hepatocytes and Kupffer cells were modeled with intestinal absorptive cells, goblet cells, and dendritic cells to study gut liver interactions under normal and inflammatory conditions, and physiologically relevant gut liver cross-talk was demonstrated from gene expression data, showing that intestinal secretion of molecules could be associated with inhibition of hepatic enzymes. Comparing pharmacokinetic modeling with the measurement of tissue-specific phenotypic metrics indicated that gut liver microphysiological systems could be fluidically coupled with circulating common medium without compromising their functionality (Chen et al., 2017). Another functional multiorgan system simulated the in vivo female reproductive tract and the endocrine loops between the ovary, fallopian tube, uterus, cervix, and liver, with a sustained circulating flow between the tissues (Cooper et al., 2019; Xiao et al., 2017). Such a system can be used to study hormonal signaling in a manner that enables the mimicking of the menstrual cycle and pregnancy. Finally, organ organ cross-talk is also investigated in the context of differentiation from induced pluripotent stem cells (Ramme et al., 2018). Besides these illustrative examples, further multiorgan systems have been presented with the primary goal to demonstrate sustained organ function over an extended period of time. These systems constitute an important basis for future applications, and include

• A liver lung system that connects, in a single circuit, normal human





bronchial epithelial cells cultured at the air liquid interface and HepaRG liver spheroids. The system can metabolize compounds in the medium and modulate their toxicity (Bovard et al., 2018). Liver gut connected perfusion systems consisting of precision-cut rat intestinal and liver slices to mimic in vivo first-pass metabolism (Van Midwoud et al., 2010) or implemented as HepaRG spheroids and intestine barrier models using inserts (Maschmeyer et al., 2015a). Liver skin cocultures enabling systemic substance testing over a 28-day period and the exposure of substances on skin at the air liquid interface (Maschmeyer et al., 2015a; Wagner et al., 2013).

395

396

CHAPTER 12 Design and engineering of multiorgan systems

• A four-tissue system of the liver, lung, kidney, and adipose tissue that • •

• •



optimizes cellular functions by supplementing the common medium with growth factors (Zhang et al., 2009). A LEGO-like system to integrate multiorgan microphysiological devices by connecting multiple chips in series (Loskill et al., 2015). A multiorgan (intestine, liver, skin, and kidney) chip to study absorption, distribution, metabolism, and excretion by establishing reproducible homeostasis among the cocultures within 2 4 days and maintaining the functionality of all four organs over 28 days (Maschmeyer et al., 2015b; Wagner et al., 2013). An interconnected system combining intestine, liver, kidney proximal tubule, blood brain barrier (BBB), and skeletal muscle with maintained function and organ-specific processing of compounds (Vernetti et al., 2017). Four-, seven-, and ten-organ systems exhibiting accurate prediction of secreted liver protein distribution and 2-week maintenance of phenotypic markers. The systems demonstrate reliable, robust operation, and maintenance of organ phenotypic function for 3 weeks (7-organ systems) and 4 weeks (10-organ systems) in continuous interaction (Edington et al., 2018; Sarkar et al., 2017; Yu et al., 2015). A 14-chamber multiorgan system representing 13 tissues/organs with both barrier and nonbarrier tissue chambers. Five cell lines survived with high viability ( . 85%) for 7 days (Miller and Shuler, 2016).

In summary, the use of multiorgan-on-chip systems increases in various fields of applications. Illustrative studies have shown their use in (1) modeling the influence of metabolism on organs and the effects of the transformed and released metabolites on other organs; (2) modeling of on-target and off-target effects of compounds and research combining questions on efficacy and toxicity into a single experiment; (3) studying organ organ interaction via cytokines and endocrine factors; (4) modeling multiorgan diseases such as diabetes or nonalcoholic steatohepatitis, in which several organs contribute significantly to the progression of the disease; (5) systemic modeling of in vivo effects to elucidate or reverse-engineer the system to identify single-organ effects; and including (6) the immune system in multiorgan modeling.

Scope of the chapter Developing organ-on-chip devices requires the integration of several disciplines, including cell and tissue engineering, microsystems technology with microfluidics, material sciences, and molecular and biochemical engineering. In addition, these organ-on-chip devices generate large datasets that must be analyzed and modeled using advanced bioinformatics tools. Combining several organs into one system increases the demands on seamless integration of multidisciplinary

Concepts of multiorgan systems

operations and the challenges of practicality, reproducibility, and scalability. In this chapter, we present and critically discuss the design and engineering considerations for multiorgan systems and their impact on implementation and experimentation.

Concepts of multiorgan systems When culturing advanced in vitro organ models, specific culturing needs must be addressed to achieve a fully viable and equally functional state. Numerous tailored systems for single-organ model cultures have been described (Benam et al., 2015; Bhatia and Ingber, 2014; Ebrahimkhani et al., 2014; Ewart et al., 2018; Griffith and Swartz, 2006; Huh et al., 2011; Park and Shuler, 2003; Truskey, 2018; Van Der Meer and Van Den Berg, 2012; Watson et al., 2017; Wikswo, 2014). Multiorgan systems must fulfill organ-specific needs using common solutions. A logical step in establishing a multiorgan system is to simply interconnect the individual chips or perfusion chamber through tubing with media flow actuated by an external pump. The tunable flow enables the circulation of media and secreted molecules around the three-dimensional (3D) organ models to mimic in vivo blood organ interactions. In the following sections, we introduce the most common approaches (Fig. 12.1 and Table 12.1).

FIGURE 12.1 Schematic illustration of three classifications of multiorgan platforms. (A) Individual organ models or perfusion chambers are interconnected through tubing. (B) Organ models are connected on a dedicated monolithic platform. (C) Microfluidic system with plug-and-play organs that can be inserted individually and at different time points.

397

Table 12.1 Multiorgan devices categorized by organ configuration. Organ configuration

Interconnected perfusion chambers

Liver cancer

Liver intestine

Mahler et al. (2009) and Van Midwoud et al. (2010)

Monolithic design

Microfluidics with plug-in organs

Sung and Shuler (2009) and Ozkan et al. (2019)

Frey et al. (2014), Kim et al. (2015b), and Lohasz et al. (2019b) Maschmeyer et al. (2015a) and Esch et al. (2016)

Imura et al. (2010), Kimura et al. (2015), Tsamandouras et al. (2017), and Chen et al. (2017)

Liver pancreas Liver skin Liver lung Liver neurospheres Liver heart Liver kidney Liver Liver Liver BBB Liver Liver

lung kidney fat cancer heart heart muscle neuronal brain BBB intestine kidney BBB intestine skin kidney

Liver menstrual cycle Body on chip configurations BBB, Blood-brain barrier

Bauer et al. (2017) Wagner et al. (2013) and Maschmeyer et al. (2015a) Bovard et al. (2018) Materne et al. (2015) Vernetti et al. (2016) and Zhang et al. (2017) Choucha-Snouber et al. (2013) Zhang et al. (2009) Zhang et al. (2017)

Oleaga et al. (2018)

Oleaga et al. (2016) Maoz et al. (2018) Vernetti et al. (2017)

Edington et al. (2018) and Miller and Shuler (2016)

Maschmeyer et al. (2015b) and Wagner et al. (2013) Xiao et al. (2017) Ramme et al. (2018)

Concepts of multiorgan systems

Interconnection of chip perfusion chamber Interconnection of individual modules by tubing offers the integration of organ models, pumps, bubble traps, biosensors, and other modules in a simple plug-andplay mode. Early on, Zhang et al. (2009) interconnected four polydimethylsiloxane (PDMS) cell culture chips with trapping structures for individual organotypic tissues. A similar approach was presented by Choucha-Snouber et al. (2013), who combined two PDMS microfluidic monoculture chips with a pump and a reservoir tank to achieve recirculating flow. Combining tailored organ modules and biosensors with tubing can be achieved by connecting the individual PDMS modules through a common medium perfusing the entire system (Zhang et al., 2017). An external peristaltic pump connected to a bubble trap to prevent the build-up of air inside the perfusion system provides the fluidic actuation. The research group of Donald Ingber developed a similar system within a dedicated device that connects the individual modules and includes the peristaltic pump (Huh et al., 2010; Jang et al., 2013; Kasendra et al., 2018; Maoz et al., 2018; Workman et al., 2018). The individual organ modules consist of two channels on opposite sides of a membrane hosting the tissue model. This approach allows perfusing different media on the apical and basal sides of the chip. Bovard et al. created a more encased system to interconnect two organ modules with an external peristaltic pump (Bovard et al., 2018). Connecting individual culturing compartments by tubing is a simple, effective system that is commercially available (Chandorkar et al., 2017; Ramachandran et al., 2015). The dedicated, almost macroscopic, plastic perfusion chambers from Kirkstall (Quasi Vivo; kirkstall.com) and IVTech (LiveBox 1; ivtech.it) host a single Transwell insert with an inlet on the upper/apical side of the insert and an outlet on the lower/basal side of the insert. Each chamber is connected through tubing, creating a flexible network of individual organ models actuated by an external peristaltic pump. A more compact version has been engineered by Iontox (iontox.com) to interlink up to six organ types. All three commercial systems have high medium volume-to-tissue ratios because of the high dead volume in the chamber and the tubing. Another commercial system from Nortis Bio (nortisbio.com) can culture various organ models and has been used to measure and model drug excretion by the kidney. Cells are loaded onto the chip, and after maturation, the organ is perfused with the compound of interest via an external pump system (Weber et al., 2016).

Monolithic design The channels for fluidic interconnection of the individual modules can also be directly integrated into a single multiorgan platform. All parts of this monolithic type of design are fabricated at the same time from the same material, creating precise and reproducible devices with fixed flow paths and defined sizes and geometries of the organ compartments. Further, the volume ratio between the

399

400

CHAPTER 12 Design and engineering of multiorgan systems

compartments is fixed. Fluidic actuation is achieved by off- or on-chip pumps as well as by gravity-driven flow via tilting of the platform. By using pressuredriven on-chip pumps or gravity-driven flow, the total medium volume in the system can be reduced significantly. These systems feature realistic physiological environments with respect to liquid-to-cell ratios, fluid residence times, and dynamic mechanical forces. These approaches typically require the construction of the organ model onchip, including scaffold building and cell loading. Trap structures are used to position cells in the desired space to form the organ model. One general disadvantage of these monolithic systems is that the cells often cannot be harvested for endpoint analysis and remain within the chip, albeit still optically and biochemically accessible. Early adoptions combined three tissue organs (intestine, liver, and cancer) in series (Imura et al., 2010). In addition, the intestine model could be separately perfused on the apical side to mimic oral dosing. The microchip was fabricated on a glass slide stacked with PDMS sheets harboring microchannels. Kimura et al. (2015) also combined layers of PDMS with an integrated semipermeable membrane to trap the tissue models and to facilitate the interaction between two individual flow loops. Each fluidic loop harbors one integrated micropump to individually control the flow speed. The research group of Michael Shuler and James Hickman presented numerous approaches to modeling organ organ interactions in these devices (Oleaga et al., 2018; Smith et al., 2013; Sung et al., 2010). Early devices have been fabricated using silicon microtechnology processes, in which the channel structure is etched onto silicon wafers (Esch et al., 2014; Sung and Shuler, 2009). Alternatives consisted of a PDMS channel layer clamped between plexiglass frames (Sung et al., 2010). Recently developed devices use 3D printing technology to fabricate the components, which are modularly assembled after Parylene C coating (Esch et al., 2016). Alternatively, fluidic channels are manually cut from silicon gaskets and assembled in conventionally machined frames (Esch et al., 2016).

Microfluidic system with plug-and-play organs When the robust monolithic approach is combined with plug-and-play organ models that can be inserted on demand into the platform, a very flexible class of devices can be created. By integrating transferrable organ models, an assay can be designed on demand, and, furthermore, a quality control step can be introduced before assembly. The flexibility lost in the design of the fluidic part enables individual maturation times for each organ model. Simultaneously, liquid paths can be kept short, dead volumes are small, and liquid routing is precise. Microtissue models can also be networked after off-chip maturation. These approaches use gravity-driven, peristaltic, and electronically actuated pumps. Systems that can integrate transferrable barrier models have been developed by the research group of Uwe Marx and commercialized through TissUse GmbH. The first systems connected two organ models: first applications focused on the

Concepts of multiorgan systems

connection of liver microtissues and skin biopsies (Maschmeyer et al., 2015a; Wagner et al., 2013), a second configuration was developed to model type 2 diabetes (Bauer et al., 2017), and a third one included liver and neurospheres (Materne et al., 2015). The devices are based on a sandwich structure with a flexible PDMS layer for the fluidic components and pressure-driven peristaltic pumps. Another version of the chip combines four organ systems within two fluidic circuits. The two circuits can be connected through a kidney model cultivated on a semipermeable membrane (Maschmeyer et al., 2015b). Using the same fabrication approach, the group is working toward a 10-organ chip. A heart chip designed by Zhao et al. (2019) consists of an array of microwells patterned onto polystyrene sheets; two polymeric wires span the microwells where cardiac cells are seeded and electrically conditioned. Loskill et al. (2015) developed a LEGO-like plugand-play system that enables loading of different cell types and a flexible fluidic interconnection on a master chip, with tailored, flexible circulation between modules. The research group of Linda Griffith developed an open-platform configuration for multitissue interaction and analysis: two machined parts fabricated from polysulfone and acrylic sandwich an elastomeric polyurethane membrane to host the integrated fluidic channels and the peristaltic pump. Connecting liver compartments with intestinal Transwell inserts allows the study of immunocompetent liver gut cross-talk (Chen et al., 2017; Tsamandouras et al., 2017). Devices with 4, 7, and 10 organs were developed with the same approach that enables precise intra- and intertissue flow and drug distribution (Edington et al., 2018; Maass et al., 2018). Next to pneumatic pumps, electrically actuated micropumps can be integrated on-chip, and the sophisticated fluidic control, developed by DRAPER, enables flexible plug-and-play systems to be assembled on the platform (Coppeta et al., 2017). Perfusion includes recirculation within individual tissue modules, but also interconnection of the different tissue models, for example, to model the menstrual cycle (Xiao et al., 2017). Multitissue systems based on 3D microtissue spheroids have been developed using various approaches. The advantage of the spheroid model is that several organs can be realized using the same aggregation method. One method links hanging drops with microchannels to an interconnected network so that different spheroids can be aggregated on-chip and cocultured under flow in their natural environment; parallel multitissue configuration can be established in a highly modular way (Frey et al., 2014). A second approach arranged several specially designed microchambers along a microchannel, and individual preformed microtissues can be loaded into the microchambers and cultured under gravity-driven flow generated by tilting the chip back and forth (Kim et al., 2015b). The concept has also been transferred from PDMS to a robust optically accessible lowabsorbing and mass-fabricated polystyrene platform that can host 10 different microtissues connected in series (Lohasz et al., 2019b). Engineering a standing drop port at the microtissue compartments allows for automated loading of preformed microtissues or organoids in parallel. Multiple conditions can run in parallel, allowing multitissue experiments in a scalable format. Since media

401

402

CHAPTER 12 Design and engineering of multiorgan systems

compartments are open to the surrounding environment, it is possible to harvest samples at any timepoint for downstream analysis.

Building blocks for multiorgan systems Organ models The current approach to predict the effects of potential drugs and therapies includes animal studies and cellular in vitro assays. The selection, viability, and functionality of the appropriate cell model remain the greatest challenge for organ models. Although there are multiple sophisticated technical solutions to culturing and modeling multitissue cross-talk, the appropriate sourcing of cells is paramount. Many laboratories use immortalized cell lines to model health and disease because of their availability, transfectability, and ease of culturing (Kaur and Dufour, 2012). Cell lines remain the workhorses of proof-of-concept studies, but are often associated with mediocre representation of the in vivo situation (Kaur and Dufour, 2012). For multiorgan systems to be used as surrogates for animal testing or for clinical trials, mimicking the in vivo physiology must be improved (Beilmann et al., 2019). The use of primary patient-derived cells or adult stem cells enables a better representation of physiological organ function (Astashkina et al., 2012). Although relevant, these cells exhibit donor variability, have a notorious sourcing problem, require complex culturing protocols, and often display poor proliferation (Astashkina et al., 2012). Induced pluripotent stem cells may overcome some of these drawbacks (Ebert and Liang, 2012): since induced pluripotent stem cells are reprogrammed patient cells that reach a stem cell like behavior; both healthy and diseased states can be reproduced. The disadvantages of using induced pluripotent stem cells are the complex and cytokinesupplemented culturing conditions required, which often clash with common culture medium techniques. Overall, the choice of cell model depends on the research question and requires a compromise between biological authenticity, experimental reproducibility, maturation time, availability, cost, and throughput.

Suspension models Microtissues are 3D cell culture models that reproduce the morphology, mechanical, and biochemical properties of living organs (Zhang et al., 2016). Microtissues can be categorized as spheroid or organoid. Multicellular spheroids—scaffold-free and self-assembled cell clusters between 100 and 500 µm in diameter—have emerged as promising 3D model systems with tissue-like functionalities and phenotypes (Fang and Eglen, 2017). These spherical microtissues are generated from a variety of cell types and techniques, including ultralow-adhesion plates, hanging drops, bioreactors, and bioprinting at automated high-throughput (Bhise et al., 2016; Fang and Eglen, 2017; Messner et al., 2018; Moshksayan et al., 2018; Proctor et al., 2017). Organoids are self-organizing, highly complex structures

Building blocks for multiorgan systems

formed from proliferating cells, mostly stem cells, that can differentiate and grow (Clevers, 2016), and are used to investigate human development, model disease, and monitor drug effects (Clevers, 2016; Dutta et al., 2017; Fang and Eglen, 2017; Ho et al., 2018). The generation and maturation of organoids can take several weeks, and the resulting organoids exhibit organ-specific morphology and cellular organization (Rossi et al., 2018). Patient-derived cells and induced pluripotent stem cells or microdissected tissues from biopsies can also be used to construct patient-specific multi-organ-on-a-chip systems (Astolfi et al., 2016; Ebert and Liang, 2012; Ramme et al., 2018; van den Berg et al., 2019). These cell models enable personalized diagnostics of compound efficacy and safety, the investigation of disease mechanisms, and potential conduct of clinical trials on-chip (Beilmann et al., 2019; DelNero et al., 2018; Low and Tagle, 2018; Ramme et al., 2018). A major advantage of suspension cultures is their transferability between culturing, analysis, and imaging platforms, so they can be used as modular building blocks of multiorgan systems.

Plug-in models Transwell cell culture inserts enable the culturing of endothelial barriers such as skin (Chung et al., 2014; De Wever et al., 2015; Ohnemus et al., 2008; Smits et al., 2017; Van Den Bogaard et al., 2014), placenta (Aengenheister et al., 2018; Huang et al., 2016), and the blood brain barrier (Appelt-Menzel et al., 2017; Novakova et al., 2014; Paradis et al., 2016), as well as epithelial barriers such as intestine (Antunes et al., 2013; Ka¨mpfer et al., 2017; Noel et al., 2017), kidney (DesRochers et al., 2013; Van der Hauwaert et al., 2013), and lung (Blume and Davies, 2013; Dekali et al., 2014; Ren et al., 2016; Yonker et al., 2017). The integrated porous membrane acts as the basal membrane and enables cell polarization on the insert. The semipermeable nature of the membrane allows for communication between the basal and apical side of the barrier tissues through soluble signaling molecules (Ye et al., 2015). This membrane can also be coated with a hydrogel that enables growth and formation of the cellular barrier. The hydrogel can mimic the lamina propria, which forms a part of the barrier that protects internal tissues from external pathogenic microorganisms. The lamina propria recruits immune cells that further protect the internal tissue. The use of a Transwell insert arrangement enables off-chip formation and characterization of the barrier integrity using standard techniques such as transepithelial electrical resistance or dye permeability measurements and a simple retrieval of cells for downstream analysis. The Transwell element can then be inserted into a microfluidic circuit to enable supply and transport of nutrients, oxygen, and compounds of interest. Immune cells (e.g., dendritic cells) can be attached underneath the membrane to realize first-order immuno-interactions (Chandorkar et al., 2017). Some multiorgan devices employ cell models precultured on small substrates or glass slides, which are transferred from standard dishes into the perfusion systems at the beginning of the experiment (Bovard et al., 2018; Coppeta et al., 2017; Edington et al., 2018; Oleaga et al., 2018; Wielhouwer et al., 2011).

403

404

CHAPTER 12 Design and engineering of multiorgan systems

Integrated models A transition from 2D tissue cultures to in vitro 3D tissue models can also be facilitated through the use of an artificial scaffold or specially engineered microstructures. Seeded with human cells, such cultures can capture the complexity needed to model an organ in vitro (Griffith and Swartz, 2006). These 3D structures can be created using techniques such as electrospinning (Hong and Madihally, 2011; Jun et al., 2018), photolithography (Kloxin et al., 2010; Limongi et al., 2015), and 3D printing (Bhise et al., 2016; Homan et al., 2016; Lind et al., 2017). The primary goal is to design and construct tissue architecture that is morphologically similar to that of native tissue and mimics the physiological functions. Synthetic scaffolds are mostly constructed from polymers, polypeptides, or hybrid materials, since proper function in cultures and adequate cell growth require biocompatibility and if possible degradability (Bitar and Zakhem, 2014). Ideally, synthetic scaffolds should also have physical properties that allow the formation of new tissue. However, cells and substrate typically become one and cannot be separated; hence, a biodegradable scaffold that only provides initial cues for tissue formation would be ideal. These techniques can also be used to form 3D organ models inside a microphysiological system, which provides a well-controlled environment for tissue formation and maintained function over time, also through perfusion of nutrients and oxygen. Microfluidic devices have been designed to supply cells with an adequate environment for 3D organization on the surface of the microchamber or in a biomaterial matrix (Yamada et al., 2012). Most systems are based on tissues in channels (Homan et al., 2016; Trietsch et al., 2017), multilayered on membranes (Benam et al., 2016; Chen et al., 2018; Edington et al., 2018; Workman et al., 2018), or encapsulated in traps (Aeby et al., 2018; Lee et al., 2007; Loskill et al., 2017). All these chamber designs allow the use of scaffolds or hydrogels to promote cell growth, differentiation, or maintenance. Organ-on-chip devices designed as integrated models have been used to model the lung (Huh et al., 2010; Stucki et al., 2018), liver (Clark et al., 2017; Gro¨ger et al., 2016; Kane et al., 2006; Li et al., 2018; Peel et al., 2019; Prodanov et al., 2016; Vernetti et al., 2016), gut (Chen et al., 2017; Mahler et al., 2009; Trietsch et al., 2017; Workman et al., 2018), vascularization (Miller et al., 2012; Osaki et al., 2018b; Raasch et al., 2015), brain (Maoz et al., 2018; Osaki et al., 2018a), adipose tissue (Liu et al., 2019; Loskill et al., 2017), heart (Agarwal et al., 2013; Lind et al., 2017; Mathur et al., 2015; Zhao et al., 2019), and kidney (Homan et al., 2016; Jang et al., 2013; Theobald et al., 2018; Weber et al., 2016). The field has grown exponentially in the past years, and detailed descriptions of these devices are available in reviews (Ewart et al., 2018; Prantil-Baun et al., 2018; Renggli et al., 2019; Rogal et al., 2017; Ronaldson-Bouchard and Vunjak-Novakovic, 2018; Wang et al., 2018). This class of in vitro models relies on microfluidic technologies for cell and tissue cultures that enable the creation of microstructures and channels and surface modification on a single-cell scale. The multitude of devices developed

Building blocks for multiorgan systems

demonstrates the flexibility in designing organ models and how the interaction between the host microenvironment and cells can be leveraged to answer research questions. In these systems, however, cell/tissue structures and substrate merge or fuse together, a phenomenon that may reduce the spectrum of measurable endpoints without destroying the organ model and limit interconnection to other models owing to incompatibility.

Interconnections A central element of multiorgan systems is their physical interconnections. Some organ models must be realized through different technical approaches. The obvious solution is the use of tubing for a wide flexibility in how many and in which order single-organ devices are connected to a network. The network can be established after the organs-on-chips have been fabricated and gain full functionality. All organ models require compatible inlet and outlet ports to ensure tight, lasting connections. Compatibility between organ models can be guaranteed using standard fittings. Tubing is available in various formats and materials. The interconnecting tubing adds volume to the allover system, which has to be considered during the design process. This also applies to tube length and diameter that define the hydrodynamic resistance and affect flow rates and medium distribution between the organ models. As the number of organs to be connected grows, as does the connection complexity, and it becomes more challenging to ensure the same conditions between replicates with equal tubing parameters. The choice of material should be made with care: both adsorption and absorption can critically influence substance concentrations in the liquid. Tubing has a limited lifespan and tubing properties may change over time (e.g., flow-rate variation in peristaltic tubing). Further, preparation steps such as cleaning and sterilization must be considered prior to setup. Practically, connecting tubes is laborious and can be a decisive factor, depending on how fast and in what numbers multiorgan systems can be set up in parallel for larger studies. Each connection is a potential source of leakage or the introduction of air bubbles, increasing the risk of losing a complete condition, even if the problem appears in only one spot and toward the end of an experiment. However, the modularity of the setup allows the addition of bubble traps or other elements such as sensors, sampling ports, imaging windows, and reservoirs. Microfluidic chips, by definition, realize the interconnection through microfluidic channels. Using microfabrication processes, their geometry and dimensions can be precisely defined and manufactured. Miniaturization allows a substantial reduction in liquid volume and the amount of substance needed. At the same time, the fluidic layout is hard-coded and can only be modified through chip redesign and refabrication, which take time and can be costly. Once the design has been defined, many chips can be fabricated with excellent reproducibility. Further, microfluidic features such as mixers, gradients, and flow splitters can

405

406

CHAPTER 12 Design and engineering of multiorgan systems

now be integrated into the network. Miniaturization may restrict the medium volume a system can accommodate, requiring external reservoirs and a means to connect them robustly to the microchannels. Transfer of liquid between organ chips can also be realized over discrete pipetting protocols. In such approaches, minute amounts of liquids are withdrawn from one organ model and added to another. Elaborate and automated pipetting protocols enable the quasisimulation of liquid flow and communication between the organ models. As the number of organ models increases, the pipetting protocol (timing and volume transfer) may become complex, and physiological liquid distribution between the organs may become a challenge.

Flow actuation and circulation Circulating flow between organs is predominantly generated using pumps (Fig. 12.2). With the rise of microfluidic technology over the past decades, a large variety of commercial systems have been made available that can be easily connected to the organ network (Au et al., 2011; Byun et al., 2014). With peristaltic pumping schemes providing inflow and outflow, circulation flow can be readily implemented in the flow path. More sophisticated setups are required for syringe pumps and for pneumatic systems, as their basic principle relies on pumping liquid from a dead-end reservoir into the network. Because the connection of external pumping systems uses tubing and external reservoirs, liquid volume, setup and operational complexity increase substantially. Peristaltic and syringe pumps

FIGURE 12.2 Schematic illustration of flow actuation methods for multitissue systems. (A) Integration of miniature pumps into monolithic systems. (B) Off-chip pumping method using conventional systems. (C) Gravity-driven medium flow through tilting of the chip.

Building blocks for multiorgan systems

must be mechanically and sterically compatible with incubator conditions to ensure the fluid’s constant temperature and prevent long tube connections to the outside. The integration of pumps into microfluidic chips mostly relies on peristaltic pumping technology (Iverson and Garimella, 2008). The main advantages of such systems are a reduction in the overall liquid volume, compactness, and the fact that the liquid does not leave the chip. Implementation is generally realized through serial arrangements of pumping chambers and valves that contain elastic membranes. A coordinated actuation of the membranes initiates well-defined and even pulsatile flow in the microfluidic system. The deformation of the membranes is generated through pneumatic or hydraulic actuation, mechanically or piezoelectrically. The systems are connected to a control unit, typically located outside the incubator, but actuation can be multiplexed on-chip for parallel experiments. Solutions for miniature micro-electromechanical system pumps exist (e.g., Nanopump by Debiotech; debiotech.com). Because of their complex fabrication process, it is difficult to integrate them monolithically, and they must be connected via tubing. Flow can be generated using gravity. Placing two or more interconnected medium reservoirs at different heights induces flow as a result of the difference in hydrostatic pressure until liquid levels are at the same height. By using programmable tilting stages, bilateral flow between two reservoirs can be maintained over long periods of time in a very robust way (Esch et al., 2016; Kim et al., 2015a; Lohasz et al., 2019b; Miller and Shuler, 2016; Trietsch et al., 2017, 2013). The height difference, which can be modulated by the tilting angle and channel dimensions, defines the flow rate. Channel network design enables the implementation of unidirectional flow and temporal compound gradient scenarios through simple back-and-forth tilting (Lee et al., 2018; Lohasz et al., 2019a; Wang and Shuler, 2018).

Monitoring and sensors A key and central advantage of multitissue systems over animal models is the information access over several interfaces. Liquid access ports for medium sampling and exchange are generally implemented as open or closable reservoirs. Alternatively, liquid can be removed from organ compartments directly. System design defines the amount of medium that can be removed without draining the network or the compartments. Typically, only a partial exchange is possible or needed. Liquid sampling may interfere with long-term culturing if it results in a substantial reduction in volume and if the liquid removed must be replaced. Sampling quantities depend on the minimal volume required for the intended bioanalytic method and its detection limits. Optical access for intermediate or endpoint imaging requires the use of transparent materials or special imaging windows as well as appropriate placement and orientation of the organ model with respect to the imaging planes of the

407

408

CHAPTER 12 Design and engineering of multiorgan systems

microscopes. Local access ports for staining cells and tissues must also be considered. Microsystems technology allows the direct integration of sensors of different types into in vitro cell culturing systems. While such systems enable unparalleled continuous monitoring at high temporal resolution and the detection of short events, the complexity of implementation, sensor stability under harsh culture conditions, and the sensor fabrication and calibration process render the routine use of the systems challenging. Sensors can be classified as global, such as for detecting metabolites in the circulating medium, and local, which monitor tissue and cell function in a specific organ (e.g., electric sensor array reading neuron action potentials). Numerous sensor types have been implemented in organ-on-achip devices. Electrochemical sensors have been used for the in situ monitoring of 3D microtissue metabolism and secreted biomarkers (Bavli et al., 2016; Misun et al., 2016; Weltin et al., 2017; Zhang et al., 2017); surface acoustic waves are employed to investigate tissue formation (Alhasan et al., 2016; Chen et al., 2016); mechanical sensors are used to record cell contraction, primarily in cardiac tissues (Feinberg et al., 2007; Lind et al., 2017); electrical impedance is applied to assess the effects of drugs on 3D microtissues (Bu¨rgel et al., 2016; Kloß et al., 2008; Schmid et al., 2016) and to probe the transepithelial electrical resistance of a cell barrier grown on a filter membrane insert (Srinivasan et al., 2015); optical sensors measure oxygen levels in culture systems (Grist et al., 2010). A more detailed review is available elsewhere (Modena et al., 2018). Finally, deep-tissue endpoints such as transcriptomics require the independent removal of the organ models from the system, followed by a cascade of postprocess and analysis steps. Complete removal of tissue is realized through accessible organ compartments, as long as cells and tissue structures are not strongly attached. More complex methods are chip-internal cell detachment or lysis methods and cell/organ outflow. A minimal amount of material must be defined in consideration of the analysis methods, influencing initial design and scaling of the organ model and multitissue systems.

Discussion Multiorgan systems are constructed by combining the elements listed in the previous section. While the organ-on-a-chip field already merges several disciplines, requiring that all features integrate seamlessly, multiorgan systems pose another layer of complexity in technical implementation. Multiorgan systems should not only replicate physiological interactions but gain a level of feasibility that enables reproducible research and eventual transfer to routine use. In this section, we discuss several factors playing an important role in multiorgan system layout and how they can be addressed when designing new systems.

Discussion

Expanding organ-on-chips toward multiorgan systems The motivation to construct an interacting organ system is driven by a defined problem, observation, or need. Starting from a biological question enables the design of the system using a top-down approach. The functional requirements of the system are defined at an early stage, and the organ models and their functions are plotted and implemented accordingly. This approach offers the advantage of tailoring the system for a specific use-case, increasing physiological relevance and the success of in vitro to in vivo translation. This enables seamless matching of biology and technology, but if the design cannot harness existing elements, the development phase is necessarily extended for design, production, characterization, and validation activities. The results are mostly integrated, monolithic systems that emulate the specific application closely, but provide less flexibility for use in other applications, as advancing the research requires removing, replacing, or adding organ models. Multiple research teams construct multitissue systems from the bottom-up by connecting predeveloped organ-on-chip devices into higher order networks. While such an approach can rely on well-characterized models, their connection may be challenging from a technical and biological standpoint: chip design, layout, fluid connection, and perfusion speed and mechanism may not be compatible across devices, resulting in issues from tubing connections to mismatches in shear stress sensitivity between cell types. If organ-on-chip devices have a fixed predefined size, relative scaling becomes difficult and emulating correctly in vivo conditions may prove impossible. Ideally, multitissue systems should be constructed to be both modular and extendable, while preserving organ model compliance with predefined system requirements.

Physiological scaling, order, and flow distribution Correct in vivo prediction of compound effects, interorgan communication, and pharmacokinetics require the different organ models to be arranged in appropriate order and scaling. Guidelines and discussions of algometric scaling have been published (Maass et al., 2017; Wikswo et al., 2013a,b). Scaling with respect to mass and/or function requires that organ models are established and fabricated in a reproducible manner or that they are adjustable, if function is reduced in multitissue cultures because of suboptimal medium conditions. The organ-specific function may vary if the chip is operated in a multitissue setup rather than in isolated single-organ mode. Miniaturization also includes the challenge of increasing unfavorable mismatch between surface and volume. Further, the volume of the culturing medium must be large enough to provide sufficient nutrients and remove waste products over extended periods of time and small enough so that secreted molecules reach physiological levels on an acceptable time scale to exert the expected effect on

409

410

CHAPTER 12 Design and engineering of multiorgan systems

FIGURE 12.3 Schematic illustration of (A) frequent circulation with concentration changes over time and (B) systems with significant modification of medium constituents during a single pass.

other organs. This aspect can be addressed by integrating various concepts of flow distribution. Early designs were mostly single-pass systems, in which the medium was perfused once through the system from an inlet to an outlet. These designs provide fresh media constantly, and waste is removed efficiently. However, tissue crosstalk and metabolite accumulation cannot be facilitated, and a large volume of medium is used. Recirculation platforms can address these limitations but require regular replacement of media because of nutrient depletion and a build-up of waste products (Ronaldson-Bouchard and Vunjak-Novakovic, 2018; Wang et al., 2018). The distribution of flow and the relative physiological order of the organ models must be related to organ function, media flow, and liquid-to-cell ratio. All factors affect the changes in constituent concentration in the media, such as metabolism or secretion by specific cells when perfusing one of the organ models. In multiorgan systems with frequent circulation of media and smaller organ models, the rather low cell-to-liquid ratio mostly results in concentration changes that are slower than medium circulation (Fig. 12.3A). The relative placement of the organ models is less important, and the primary interest of these setups is in global changes of media constituents, while local changes cannot typically be resolved. Correct physiological organ placements become interesting in larger, macroscopy models, with higher cell-to-medium ratios, slower intertissue flow rates, and longer residence times. In such systems, the concentrations of media constituents change during single pass and between organ models, enabling pharmacokinetic modeling with single-organ resolution (Fig. 12.3B).

Interconnection versus monolithic systems Well-orchestrated fabrication and maturation of individual organ models Simple interconnection of individual organ modules by tubing offers the advantage of preparing each module prior to connection. Next to the microfabrication of the culturing chamber, the formation and maturation of the tissue can be tailored to the needs of the organ of interest. Mixing culturing methods is also possible since the individual chambers are connected only by tubing. This procedure

Discussion

enables various choices in material fabrication and adaptable flow routing schemes through interchangeable tubing ports. In addition, the integration of pumps, bubble traps, biosensors, and other modules is straightforward. In a monolithic design the integration of different organ-on-chip technologies can be difficult because they are fabricated from one master and rely on unified processes. A disadvantage of tubing-connected multiorgan systems is the rather high liquid volume-to-tissue ratio because of the higher dead volume in the tubing and the distance between modules, which can hamper intertissue communication, such as by short-lived metabolites. Moreover, tubing may lead to leakage and unspecific binding. The individual modules all require connection ports, increasing dead volume and leakage potential. The use of a monolithic design with well-defined fluidic paths integrated into the chip can be versatile if combined with plug-in tissue models. Monolithic devices are typically robustly fabricated and can be reused several times (Coppeta et al., 2017; Edington et al., 2018; Maschmeyer et al., 2015b). They minimize the dead volume through on-chip interconnection and integrated pumps. If used with plug-in organ models, the individual tissues can also be matured off-chip and timelines can be tailored by assay.

Quality control prior to interconnection (reproducibility and system yield) Plug-in models permit the production and maturation of tissues off-chip, as well as a quality control step prior to system assembly. This yields more robust and reproducible results from multitissue experiments. The success of an assembled system is significantly increased and directly scales with the number of organ models combined in the system (Rogal et al., 2017). For simple, scaffold-free models produced with high consistency in large numbers, quality control can be executed in parallel on tissues of the same batch so the measured endpoints can be destructive without compromising quality testing. Integrated models can, in principle, be tested prior to assembly as well. Testing, however, may be more complex, because the devices must be operated similarly. Testing methods should preferably not be destructive and not interfere with function, because the chip is later connected to the multitissue setup. Further, scaffold quality and mechanical functions should be tested as well.

Timing and setup (logistics, supply chain, dependency on third-party delivery) Working with quality-controlled tissue models and subsequent system assembly permits well-planned experiments. Plug-in cell models can leverage automation methods and be mass-produced. If the quality of a given tissue is not as desired, the model is simply discarded without wasting the entire organ module. This reduces the number of devices and eliminates redundancy. Furthermore, the formation and maturation of individual organ models can begin at various time points prior to assembly, enabling good control of their viability and function (Fig. 12.4).

411

412

CHAPTER 12 Design and engineering of multiorgan systems

FIGURE 12.4 Schematic illustration of a typical workflow in a modular system and the importance of timing organ-model fabrication and quality control prior to beginning multitissue experiments.

Modularity The possibility of modular organ organ configuration with regards to type and number of organs offers the advantage to increase and decrease complexity in a controlled manner. A controlled increase in complexity minimizes variability and simplifies the characterization and optimization of multiorgan systems. The higher the complexity, the more difficult it is to identify the source of changes in the system. Inversely, modular systems enable deconvolution into single parts and tracking the origin of an observed behavior.

Readout methods Developing, setting up, and operating complex multiorgan systems involves both effort and cost. Maximizing data output during and at the end of an experiment is therefore central to increasing the value of an experiment and to ensuring sufficient data for correct interpretation. Readouts can be generated on systemic and on organ-specific levels but differ by organ type: cardiac and neuron cultures, for example, can be functionally assessed over electrode arrays recording depolarization, while barrier models can be monitored using transepithelial electrical resistance measurements. Ideally, endpoints should be of clinical relevance and translatable in vivo. Liver toxicity, for example, is a preferable readout over levels and ratios of liver-specific enzymes, i.e., aspartate transaminase and alanine transaminase (AST/ALT) instead of adenosine triphosphate and lactate dehydrogenase (ATP/LDH).

Optical access Imaging cells and tissue structures can be implemented by removing the tissue or by integrating optical windows. Image quality largely depends on the system’s compatibility with imaging platforms. Integrated models must be imaged in situ, as removal mostly results in destruction of the tissue architecture. High resolution

Discussion

and optical sectioning require the use of optically clear material and a short distance from the objective. Holders should be compatible with imaging stages. On-line imaging requires transfer of the multiorgan systems from the culturing environment to the imaging system, running the entire setup directly on a cell culturing-capable microscope. External pumping must be unplugged, introducing complexity and the risk of system failure. Suspension cultures and plug-in models can potentially be removed at the end of the experiment or in some cases even during the experiment; removal allows better optical access through dedicated imaging plates and may simplify staining and preparation protocols.

Access to medium Sampling medium for downstream analysis is the simplest readout. However, closed fluid networks require the construction of sampling ports, from which medium can be withdrawn and replaced in a bubble-free manner to ensure a constant system volume. Open reservoirs allow inherent sampling access. Barrier structures ideally allow medium sampling from both sides of the barrier, enabling the perfusion of media on the apical and basal sides of the chip. This aspect can be leveraged to perfuse a common medium connecting the various organ models while maintaining tissue health and function through an organ-specific maintenance medium on the other side of the membrane. Organ models that require more complex media to connect to other organs, such as induced pluripotent stem cells, can benefit from such setups. Changes in volume and concentration resulting from sampling must be considered in long-term studies. Sampling volumes should be adjusted to the requirements of reliable analysis.

Access to cells The measurement of multiple endpoints, including transcriptomics and DNA sequencing, requires access to the tissue. Further, histology is still a central endpoint in tissue engineering because of its high translational value. Tissue removal is, therefore, of high relevance in multiorgan systems and especially in integrated cell models that typically cannot be separated from the device. Another potential issue that microtissues or large 3D tissue can encounter is the critical diffusion limit for oxygen and nutrient transport to cells, which can result in necrosis owing to the absence of vascularization (Blinder et al., 2014; Cheng et al., 2009; Gjorevski et al., 2014). Vascularization is necessary for cell survival when the number of cells for tissue formation is scaled up. While vascularization can be achieved efficiently in in vivo models by relying on preexisting blood vessels, the formation of a vascular network from endothelial cells in vitro remains a major challenge (Blinder et al., 2014; Osaki et al., 2018b; Wikswo et al., 2013a,b).

Operations complexity and robustness Merging multiple disciplines into a single system poses challenges related to complexity and robustness, which directly define experimental reproducibility and

413

414

CHAPTER 12 Design and engineering of multiorgan systems

interlaboratory variability. Multiorgan systems are typically designed in academic laboratories and rely on doctoral students and postdocs with detailed knowledge of the systems and their operation. Multitissue systems geared toward routine implementation must offer standard operation protocols. To that end, tissue chip testing centers have been established to test and characterize tissue chip platforms and to promote adoption of the technology by the broader research community. The National Institutes of Health’s National Center for Advancing Translational Sciences currently maintains three centers in the United States that are located at the Massachusetts Institute of Technology, Texas A&M University, and the University of Pittsburgh. These centers test the reproducibility of multitissue systems before examining their application and characterization. A systematic and quantitative workflow for translation of organ-on-chip data to clinical outcomes is applied using a combined experimental and computational paradigm known as quantitative systems pharmacology.

Scalability and parallelization The complexity of multiorgan systems tends to reduce the number of replicates. Long setup times for organ model fabrication and testing and convoluted connections and external pumping mechanisms prevent the parallelization of multiple conditions and limit the throughput. Scalability includes not only reliable and robust operation of the system but also scalable solutions for each individual element. Organ models must rely on scalable and reproducible cell sources and fabrication protocols, while chip and systems manufacturing should be anchored in standardized mass-fabrication processes. System assembly will ideally be plugand-play or automated. Finally, system design must be based on robust structures; tube connections are prone to leakage, and movable mechanical parts can fail under harsh incubator conditions. Manual assembly, moreover, may vary by operator.

Industry standards Compatibility of multitissue systems with industry standards simplifies technology transfer and system integration into routine processes. These standards involve liquid handling systems, laboratory protocols, readout interfaces, and conventional cell culturing equipment. Further, multiorgan systems should fulfill industry or regulatory standards if their results would be included in reports for regulatory agencies. Although the US Food and Drug Administration is open to results from microphysiological systems for filings, issues in the standardization, robustness, and reproducibility of multitissue systems are still hampering largescale adoption of the technology. Several concepts from academia have already reached the market and are optimized for use in drug screening, to replace animals or for personalized medicine applications (Zhang and Radisic, 2017).

References

Conclusion In this chapter, we provided insights into the current academic and industrial status of multiorgan-on-a-chip devices. The complexity of these in vitro models may potentially address biological questions of systemic and multidimensional character. While previously such cellular and functional complexity was only available through in vivo models, a shift toward human in vitro models is now expected. Combined with the availability of multiple biochemical, optical, and downstream analysis methods, these systems offer an in-depth view of tissues, metabolites, and therapeutics in understanding human holistic responses to disease and stimuli. However, the wider adoption and use of these systems in routine assays will largely depend on their technical implementation and practicability. The systems’ multidisciplinary nature poses several challenges, but recent advances have demonstrated how these can be effectively addressed to ultimately pave the way to patient-specific applications or clinical trials on a chip.

References Aeby, E.A., Misun, P.M., Hierlemann, A., Frey, O., 2018. Microfluidic hydrogel hangingdrop network for long-term culturing of 3D microtissues and simultaneous highresolution imaging. Adv. Biosyst. 2, 1800054. Available from: https://doi.org/10.1002/ adbi.201800054. Aengenheister, L., Keevend, K., Muoth, C., Scho¨nenberger, R., Diener, L., Wick, P., et al., 2018. An advanced human in vitro co-culture model for translocation studies across the placental barrier. Sci. Rep. 8, 5388. Available from: https://doi.org/10.1038/s41598018-23410-6. Agarwal, A., Goss, J.A., Cho, A., McCain, M.L., Parker, K.K., 2013. Microfluidic heart on a chip for higher throughput pharmacological studies. Lab Chip 13, 3599 3608. Available from: https://doi.org/10.1039/c3lc50350j. Alhasan, L., Qi, A., Al-Abboodi, A., Rezk, A., Chan, P.P.Y., Iliescu, C., et al., 2016. Rapid enhancement of cellular spheroid assembly by acoustically driven microcentrifugation. ACS Biomater. Sci. Eng. 2, 1013 1022. Available from: https://doi.org/10.1021/ acsbiomaterials.6b00144. Antunes, F., Andrade, F., Arau´jo, F., Ferreira, D., Sarmento, B., 2013. Establishment of a triple co-culture in vitro cell models to study intestinal absorption of peptide drugs. Eur. J. Pharm. Biopharm. 83, 427 435. Available from: https://doi.org/10.1016/j. ejpb.2012.10.003. Appelt-Menzel, A., Cubukova, A., Gu¨nther, K., Edenhofer, F., Piontek, J., Krause, G., et al., 2017. Establishment of a human blood-brain barrier co-culture model mimicking the neurovascular unit using induced pluri- and multipotent stem cells. Stem Cell Rep. 8, 894 906. Available from: https://doi.org/10.1016/j.stemcr.2017.02.021. Astashkina, A., Mann, B., Grainger, D.W., 2012. A critical evaluation of in vitro cell culture models for high-throughput drug screening and toxicity. Pharmacol. Ther. 134, 82 106. Available from: https://doi.org/10.1016/j.pharmthera.2012.01.001.

415

416

CHAPTER 12 Design and engineering of multiorgan systems

Astolfi, M., Pe´ant, B., Lateef, M.A., Rousset, N., Kendall-Dupont, J., Carmona, E., et al., 2016. Micro-dissected tumor tissues on chip: An ex vivo method for drug testing and personalized therapy. Lab Chip 16, 312 325. Available from: https://doi.org/10.1039/ c5lc01108f. Au, A.K., Lai, H., Utela, B.R., Folch, A., 2011. Microvalves and micropumps for BioMEMS. Micromachines 2, 179 220. Available from: https://doi.org/10.3390/ mi2020179. Bauer, S., Wennberg Huldt, C., Kanebratt, K.P., Durieux, I., Gunne, D., Andersson, S., et al., 2017. Functional coupling of human pancreatic islets and liver spheroids on-achip: towards a novel human ex vivo type 2 diabetes model. Sci. Rep. 7, 14620. Available from: https://doi.org/10.1038/s41598-017-14815-w. Bavli, D., Prill, S., Ezra, E., Levy, G., Cohen, M., Vinken, M., et al., 2016. Real-time monitoring of metabolic function in liver-on-chip microdevices tracks the dynamics of mitochondrial dysfunction. Proc. Natl. Acad. Sci. U.S.A. E2232 E2240. Available from: https://doi.org/10.1073/pnas.1522556113. Beilmann, M., Boonen, H., Czich, A., Dear, G., Hewitt, P., Mow, T., et al., 2019. Optimizing drug discovery by investigative toxicology: current and future trends. ALTEX 36, 289 313. Available from: https://doi.org/10.14573/altex.1808181. Benam, K.H., Dauth, S., Hassell, B., Herland, A., Jain, A., Jang, K.-J., et al., 2015. Engineered in vitro disease models. Annu. Rev. Pathol. Mech. Dis. 10, 195 262. Available from: https://doi.org/10.1146/annurev-pathol-012414-040418. Benam, K.H., Villenave, R., Lucchesi, C., Varone, A., Hubeau, C., Lee, H.-H., et al., 2016. Small airway-on-a-chip enables analysis of human lung inflammation and drug responses in vitro. Nat. Methods 13, 151 157. Available from: https://doi.org/10.1038/nmeth.3697. Bhatia, S.N., Ingber, D.E., 2014. Microfluidic organs-on-chips. Nat. Biotechnol. 32, 760 772. Available from: https://doi.org/10.1038/nbt.2989. Bhise, N.S., Manoharan, V., Massa, S., Tamayol, A., Ghaderi, M., Miscuglio, M., et al., 2016. A liver-on-a-chip platform with bioprinted hepatic spheroids. Biofabrication 8, 014101. Available from: https://doi.org/10.1088/1758-5090/8/1/014101. Bitar, K.N., Zakhem, E., 2014. Design strategies of biodegradable scaffolds for tissue regeneration. Biomed. Eng. Comput. Biol. 6, 13 20. Available from: https://doi.org/ 10.4137/BECB.S10961. Blinder, Y., Mooney, D., Levenberg, S., 2014. Engineering approaches for inducing blood vessel formation. Curr. Opin. Chem. Eng. 3, 56 61. Available from: https://doi.org/ 10.1016/j.coche.2013.11.003. Blume, C., Davies, D.E., 2013. In vitro and ex vivo models of human asthma. Eur. J. Pharm. Biopharm. 84, 394 400. Available from: https://doi.org/10.1016/j.ejpb.2012.12.014. Bovard, D., Sandoz, A., Luettich, K., Frentzel, S., Iskandar, A., Marescotti, D., et al., 2018. A lung/liver-on-a-chip platform for acute and chronic toxicity studies. Lab Chip 18, 3814 3829. Available from: https://doi.org/10.1039/C8LC01029C. Bu¨rgel, S.C., Diener, L., Frey, O., Kim, J.-Y., Hierlemann, A., 2016. Automated, multiplexed electrical impedance spectroscopy platform for continuous monitoring of microtissue spheroids. Anal. Chem. 88, 10876 10883. Available from: https://doi.org/ 10.1021/acs.analchem.6b01410. Byun, C.K., Abi-Samra, K., Cho, Y.-K., Takayama, S., 2014. Pumps for microfluidic cell culture. Electrophoresis 35, 245 257. Available from: https://doi.org/10.1002/ elps.201300205.

References

Chandorkar, P., Posch, W., Zaderer, V., Blatzer, M., Steger, M., Ammann, C.G., et al., 2017. Fast-track development of an in vitro 3D lung/immune cell model to study Aspergillus infections. Sci. Rep. 7, 11644. Available from: https://doi.org/10.1038/ s41598-017-11271-4. Chen, H.J., Miller, P., Shuler, M.L., 2018. A pumpless body-on-a-chip model using a primary culture of human intestinal cells and a 3D culture of liver cells. Lab Chip 18, 2036 2046. Available from: https://doi.org/10.1039/C8LC00111A. Chen, K., Wu, M., Guo, F., Li, P., Chan, C.Y., Mao, Z., et al., 2016. Rapid formation of size-controllable multicellular spheroids via 3D acoustic tweezers. Lab Chip 16, 2636 2643. Available from: https://doi.org/10.1039/C6LC00444J. Chen, W.L.K., Edington, C., Suter, E., Yu, J., Velazquez, J.J., Velazquez, J.G., et al., 2017. Integrated gut/liver microphysiological systems elucidates inflammatory inter-tissue crosstalk. Biotechnol. Bioeng. 114, 2648 2659. Available from: https://doi.org/ 10.1002/bit.26370. Cheng, G., Markenscoff, P., Zygourakis, K., 2009. A 3D hybrid model for tissue growth: the interplay between cell population and mass transport dynamics. Biophys. J. 97, 401 414. Available from: https://doi.org/10.1016/j.bpj.2009.03.067. Choucha-Snouber, L., Aninat, C., Grsicom, L., Madalinski, G., Brochot, C., Poleni, P.E., et al., 2013. Investigation of ifosfamide nephrotoxicity induced in a liver-kidney co-culture biochip. Biotechnol. Bioeng. 110, 597 608. Available from: https://doi.org/ 10.1002/bit.24707. Chung, E., Choi, H., Lim, J.E., Son, Y., 2014. Development of skin inflammation test model by co-culture of reconstituted 3D skin and RAW264.7 cells. Tissue Eng. Regen. Med. 11, 87 92. Available from: https://doi.org/10.1007/s13770-013-1113-x. Clark, A.M., Wheeler, S.E., Young, C.L., Stockdale, L., Shepard Neiman, J., Zhao, W., et al., 2017. A liver microphysiological system of tumor cell dormancy and inflammatory responsiveness is affected by scaffold properties. Lab Chip 17, 156 168. Available from: https://doi.org/10.1039/C6LC01171C. Clevers, H., 2016. Modeling development and disease with organoids. Cell 165, 1586 1597. Available from: https://doi.org/10.1016/j.cell.2016.05.082. Cooper, M., Charest, J.L., Coppeta, J., 2019. Design principles for dynamic microphysiological systems. In: Borenstein, J.T., Tandon, V., Tao, S.L., Charest, J.L. (Eds.), Microfluidic Cell Culture Systems. Elsevier, pp. 1 29. Available from: https://doi.org/ 10.1016/B978-0-12-813671-3.00001-3. Coppeta, J.R., Mescher, M.J., Isenberg, B.C., Spencer, A.J., Kim, E.S., Lever, A.R., et al., 2017. A portable and reconfigurable multi-organ platform for drug development with onboard microfluidic flow control. Lab Chip 17, 134 144. Available from: https://doi. org/10.1039/C6LC01236A. De Wever, B., Kurdykowski, S., Descargues, P., 2015. Human skin models for research applications in pharmacology and toxicology: introducing NativeSkin®, the “Missing Link” bridging cell culture and/or reconstructed skin models and human clinical testing. Appl. Vitr. Toxicol 1, 26 32. Available from: https://doi.org/10.1089/ aivt.2014.0010. Dekali, S., Gamez, C., Kortulewski, T., Blazy, K., Rat, P., Lacroix, G., 2014. Assessment of an in vitro model of pulmonary barrier to study the translocation of nanoparticles. Toxicol. Rep. 1, 157 171. Available from: https://doi.org/10.1016/j. toxrep.2014.03.003.

417

418

CHAPTER 12 Design and engineering of multiorgan systems

DelNero, P., Hopkins, B.D., Cantley, L.C., Fischbach, C., 2018. Cancer metabolism gets physical. Sci. Transl. Med. 10, eaaq1011. Available from: https://doi.org/10.1126/scitranslmed.aaq1011. DesRochers, T.M., Suter, L., Roth, A., Kaplan, D.L., 2013. Bioengineered 3D human kidney tissue, a platform for the determination of nephrotoxicity. PLoS One 8, e59219. Available from: https://doi.org/10.1371/journal.pone.0059219. Dutta, D., Heo, I., Clevers, H., 2017. Disease modeling in stem cell-derived 3D organoid systems. Trends Mol. Med. 23, 393 410. Available from: https://doi.org/10.1016/j. molmed.2017.02.007. Ebert, A.D., Liang, P.W.J., 2012. Induced pluripotent stem cells as a disease modeling and drug screening platform. J. Cardiovasc. Pharmacol. 60, 408 416. Available from: https://doi.org/10.1097/FJC.0b013e318247f642.Induced. Ebrahimkhani, M.R., Neiman, J.A.S., Raredon, M.S.B., Hughes, D.J., Griffith, L.G., 2014. Bioreactor technologies to support liver function in vitro. Adv. Drug Deliv. Rev. 69 70, 132 157. Available from: https://doi.org/10.1016/j.addr.2014.02.011. Edington, C.D., Chen, W.L.K., Geishecker, E., Kassis, T., Soenksen, L.R., Bhushan, B.M., et al., 2018. Interconnected microphysiological systems for quantitative biology and pharmacology studies. Sci. Rep. 8, 4530. Available from: https://doi.org/10.1038/ s41598-018-22749-0. Esch, M.B., Mahler, G.J., Stokol, T., Shuler, M.L., 2014. Body-on-a-chip simulation with gastrointestinal tract and liver tissues suggests that ingested nanoparticles have the potential to cause liver injury. Lab Chip 14, 3081 3092. Available from: https://doi. org/10.1039/C4LC00371C. Esch, M.B., Ueno, H., Applegate, D.R., Shuler, M.L., 2016. Modular, pumpless body-on-achip platform for the co-culture of GI tract epithelium and 3D primary liver tissue. Lab Chip 16, 2719 2729. Available from: https://doi.org/10.1039/C6LC00461J. Ewart, L., Dehne, E.-M., Fabre, K., Gibbs, S., Hickman, J., Hornberg, E., et al., 2018. Application of microphysiological systems to enhance safety assessment in drug discovery. Annu. Rev. Pharmacol. Toxicol. 58, 65 82. Available from: https://doi.org/ 10.1146/annurev-pharmtox-010617-052722. Fang, Y., Eglen, R.M., 2017. Three-dimensional cell cultures in drug discovery and development. SLAS Discov. 22, 456 472. Available from: https://doi.org/10.1177/ 1087057117696795. Feinberg, A.W., Feigel, A., Shevkoplyas, S.S., Sheehy, S., Whitesides, G.M., Parker, K.K., 2007. Muscular thin films for building actuators and powering devices. Science 317, 1366 1370. Available from: https://doi.org/10.1126/science.1146885. Frey, O., Misun, P.M., Fluri, D.A., Hengstler, J.G., Hierlemann, A., 2014. Reconfigurable microfluidic hanging drop network for multi-tissue interaction and analysis. Nat. Commun. 5, 4250. Available from: https://doi.org/10.1038/ncomms5250. Gjorevski, N., Ranga, A., Lutolf, M.P., 2014. Bioengineering approaches to guide stem cell-based organogenesis. Development 141, 1794 1804. Available from: https://doi. org/10.1242/dev.101048. Griffith, L.G., Swartz, M.A., 2006. Capturing complex 3D tissue physiology in vitro. Nat. Rev. Mol. Cell Biol. 7, 211 224. Available from: https://doi.org/10.1038/nrm1858. Grist, S.M., Chrostowski, L., Cheung, K.C., 2010. Optical oxygen sensors for applications in microfluidic cell culture. Sensors (Basel) 10, 9286 9316. Available from: https:// doi.org/10.3390/s101009286.

References

Gro¨ger, M., Rennert, K., Giszas, B., Weiß, E., Dinger, J., Funke, H., et al., 2016. Monocyte-induced recovery of inflammation-associated hepatocellular dysfunction in a biochip-based human liver model. Sci. Rep. 6, 21868. Available from: https://doi.org/ 10.1038/srep21868. Ho, B., Pek, N., Soh, B.-S., 2018. Disease modeling using 3D organoids derived from human induced pluripotent stem cells. Int. J. Mol. Sci. 19, 936. Available from: https:// doi.org/10.3390/ijms19040936. Homan, K.A., Kolesky, D.B., Skylar-Scott, M.A., Herrmann, J., Obuobi, H., Moisan, A., et al., 2016. Bioprinting of 3D convoluted renal proximal tubules on perfusable chips. Sci. Rep. 6, 34845. Available from: https://doi.org/10.1038/srep34845. Hong, J.K., Madihally, S.V., 2011. Next generation of electrosprayed fibers for tissue regeneration. Tissue Eng., B: Rev. 17, 125 142. Available from: https://doi.org/ 10.1089/ten.teb.2010.0552. Huang, X., Lu¨thi, M., Ontsouka, E.C., Kallol, S., Baumann, M.U., Surbek, D.V., et al., 2016. Establishment of a confluent monolayer model with human primary trophoblast cells: novel insights into placental glucose transport. Mol. Hum. Reprod. 22, 442 456. Available from: https://doi.org/10.1093/molehr/gaw018. Huh, D., Matthews, B.D., Mammoto, A., Montoya-Zavala, M., Hsin, H.Y., Ingber, D.E., 2010. Reconstituting organ-level lung functions on a chip. Science 328, 1662 1668. Available from: https://doi.org/10.1126/science.1188302. Huh, D., Hamilton, G.A., Ingber, D.E., 2011. From 3D cell culture to organs-on-chips. Trends Cell Biol. 21, 745 754. Available from: https://doi.org/10.1016/j. tcb.2011.09.005. Imura, Y., Sato, K., Yoshimura, E., 2010. Micro total bioassay system for ingested substances: assessment of intestinal absorption, hepatic metabolism, and bioactivity. Anal. Chem. 82, 9983 9988. Available from: https://doi.org/10.1021/ac100806x. Iverson, B.D., Garimella, S.V., 2008. Recent advances in microscale pumping technologies: a review and evaluation. Microfluid. Nanofluid. 5, 145 174. Available from: https://doi.org/10.1007/s10404-008-0266-8. Jang, K.-J., Mehr, A.P., Hamilton, G.A., McPartlin, L.A., Chung, S., Suh, K.-Y., et al., 2013. Human kidney proximal tubule-on-a-chip for drug transport and nephrotoxicity assessment. Integr. Biol. 5, 1119 1129. Available from: https://doi.org/10.1039/ c3ib40049b. Jun, I., Han, H.-S., Edwards, J., Jeon, H., 2018. Electrospun fibrous scaffolds for tissue engineering: viewpoints on architecture and fabrication. Int. J. Mol. Sci. 19, 745. Available from: https://doi.org/10.3390/ijms19030745. Ka¨mpfer, A.A.M., Urba´n, P., Gioria, S., Kanase, N., Stone, V., Kinsner-Ovaskainen, A., 2017. Development of an in vitro co-culture model to mimic the human intestine in healthy and diseased state. Toxicol. In Vitro 45, 31 43. Available from: https://doi. org/10.1016/j.tiv.2017.08.011. Kane, B.J., Zinner, M.J., Yarmush, M.L., Toner, M., 2006. Liver-specific functional studies in a microfluidic array of primary mammalian hepatocytes. Anal. Chem. 78, 4291 4298. Available from: https://doi.org/10.1021/ac051856v. Kasendra, M., Tovaglieri, A., Sontheimer-Phelps, A., Jalili-Firoozinezhad, S., Bein, A., Chalkiadaki, A., et al., 2018. Development of a primary human small intestine-on-achip using biopsy-derived organoids. Sci. Rep. 8, 2871. Available from: https://doi.org/ 10.1038/s41598-018-21201-7.

419

420

CHAPTER 12 Design and engineering of multiorgan systems

Kaur, G., Dufour, J.M., 2012. Cell lines. Spermatogenesis 2, 1 5. Available from: https:// doi.org/10.4161/spmg.19885. Kim, J.-Y., Fluri, D. a, Kelm, J.M., Hierlemann, A., Frey, O., 2015a. 96-Well formatbased microfluidic platform for parallel interconnection of multiple multicellular spheroids. J. Lab. Autom. 20, 274 282. Available from: https://doi.org/10.1177/ 2211068214564056. Kim, J.-Y., Fluri, D.A., Marchan, R., Boonen, K., Mohanty, S., Singh, P., et al., 2015b. 3D spherical microtissues and microfluidic technology for multi-tissue experiments and analysis. J. Biotechnol. 205, 24 35. Available from: https://doi.org/10.1016/j. jbiotec.2015.01.003. Kimura, H., Ikeda, T., Nakayama, H., Sakai, Y., Fujii, T., 2015. An on-chip small intestine liver model for pharmacokinetic studies. J. Lab. Autom. 20, 265 273. Available from: https://doi.org/10.1177/2211068214557812. Kloxin, A.M., Tibbitt, M.W., Anseth, K.S., 2010. Synthesis of photodegradable hydrogels as dynamically tunable cell culture platforms. Nat. Protoc. 5, 1867 1887. Kloß, D., Kurz, R., Jahnke, H.G., Fischer, M., Rothermel, A., Anderegg, U., et al., 2008. Microcavity array (MCA)-based biosensor chip for functional drug screening of 3D tissue models. Biosens. Bioelectron. 23, 1473 1480. Available from: https://doi.org/ 10.1016/j.bios.2008.01.003. Lee, D.W., Choi, N., Sung, J.H., 2018. A microfluidic chip with gravity-induced unidirectional flow for perfusion cell culture. Biotechnol. Prog. 35, e2701. Available from: https://doi.org/10.1002/btpr.2701. Lee, P.J., Hung, P.J., Lee, L.P., 2007. An artificial liver sinusoid with a microfluidic endothelial-like barrier for primary hepatocyte culture. Biotechnol. Bioeng. 97, 1340 1346. Available from: https://doi.org/10.1002/bit.21360. Li, X., George, S.M., Vernetti, L., Gough, A.H., Taylor, D.L., 2018. A glass-based, continuously zonated and vascularized human liver acinus microphysiological system (vLAMPS) designed for experimental modeling of diseases and ADME/TOX. Lab Chip 18, 2614 2631. Available from: https://doi.org/10.1039/C8LC00418H. Limongi, T., Schipani, R., Di Vito, A., Giugni, A., Francardi, M., Torre, B., et al., 2015. Photolithography and micromolding techniques for the realization of 3D polycaprolactone scaffolds for tissue engineering applications. Microelectron. Eng. 141, 135 139. Available from: https://doi.org/10.1016/j.mee.2015.02.030. Lind, J.U., Busbee, T.A., Valentine, A.D., Pasqualini, F.S., Yuan, H., Yadid, M., et al., 2017. Instrumented cardiac microphysiological devices via multimaterial threedimensional printing. Nat. Mater. 16, 303 308. Available from: https://doi.org/ 10.1038/nmat4782. Liu, Y., Kongsuphol, P., Chiam, S.Y., Zhang, Q.X., Gourikutty, S.B.N., Saha, S., et al., 2019. Adipose-on-a-chip: a dynamic microphysiological in vitro model of the human adipose for immune-metabolic analysis in type II diabetes. Lab Chip 19, 241 253. Available from: https://doi.org/10.1039/C8LC00481A. Lohasz, C., Frey, O., Bonanini, F., Renggli, K., Hierlemann, A., 2019a. Tubing-free microfluidic microtissue culture system featuring gradual, in vivo-like substance exposure profiles. Front. Bioeng. Biotechnol. 7, 72. Available from: https://doi.org/10.3389/ fbioe.2019.00072. Lohasz, C., Rousset, N., Renggli, K., Hierlemann, A., Frey, O., 2019b. Scalable microfluidic platform for flexible configuration of and experiments with microtissue multiorgan

References

models. SLAS Technol. Transl. Life Sci. Innov. 24, 79 95. Available from: https://doi. org/10.1177/2472630318802582. Loskill, P., Marcus, S.G., Mathur, A., Reese, W.M., Healy, K.E., 2015. µOrgano: a Lego®like plug & play system for modular multi-organ-chips. PLoS One 10, e0139587. Available from: https://doi.org/10.1371/journal.pone.0139587. Loskill, P., Sezhian, T., Tharp, K.M., Lee-Montiel, F.T., Jeeawoody, S., Reese, W.M., et al., 2017. WAT-on-a-chip: a physiologically relevant microfluidic system incorporating white adipose tissue. Lab Chip 17, 1645 1654. Available from: https://doi.org/ 10.1039/C6LC01590E. Low, L.A., Tagle, D.A., 2018. ‘You-on-a-chip’ for precision medicine. Expert Rev. Precis. Med. Drug Dev. 3, 137 146. Available from: https://doi.org/10.1080/ 23808993.2018.1456333. Maass, C., Stokes, C.L., Griffith, L.G., Cirit, M., 2017. Multi-functional scaling methodology for translational pharmacokinetic and pharmacodynamic applications using integrated microphysiological systems (MPS). Integr. Biol. 9, 290 302. Available from: https://doi.org/10.1039/C6IB00243A. Maass, C., Dallas, M., LaBarge, M.E., Shockley, M., Valdez, J., Geishecker, E., et al., 2018. Establishing quasi-steady state operations of microphysiological systems (MPS) using tissue-specific metabolic dependencies. Sci. Rep. 8, 8015. Available from: https://doi.org/10.1038/s41598-018-25971-y. Mahler, G.J., Esch, M.B., Glahn, R.P., Shuler, M.L., 2009. Characterization of a gastrointestinal tract microscale cell culture analog used to predict drug toxicity. Biotechnol. Bioeng. 104, 193 205. Available from: https://doi.org/10.1002/ bit.22366. Mak, I.W., Evaniew, N., Ghert, M., 2014. Lost in translation: animal models and clinical trials in cancer treatment. Am. J. Transl. Res. 6, 114 118. Maoz, B.M., Herland, A., FitzGerald, E.A., Grevesse, T., Vidoudez, C., Pacheco, A.R., et al., 2018. A linked organ-on-chip model of the human neurovascular unit reveals the metabolic coupling of endothelial and neuronal cells. Nat. Biotechnol. 36, 865 874. Available from: https://doi.org/10.1038/nbt.4226. Maschmeyer, I., Hasenberg, T., Jaenicke, A., Lindner, M., Lorenz, A.K., Zech, J., et al., 2015a. Chip-based human liver intestine and liver skin co-cultures a first step toward systemic repeated dose substance testing in vitro. Eur. J. Pharm. Biopharm. 95, 77 87. Available from: https://doi.org/10.1016/j.ejpb.2015.03.002. Maschmeyer, I., Lorenz, A.K., Schimek, K., Hasenberg, T., Ramme, A.P., Hu¨bner, J., et al., 2015b. A four-organ-chip for interconnected long-term co-culture of human intestine, liver, skin and kidney equivalents. Lab Chip 15, 2688 2699. Available from: https://doi.org/10.1039/c5lc00392j. Materne, E.-M., Ramme, A.P., Terrasso, A.P., Serra, M., Alves, P.M., Brito, C., et al., 2015. A multi-organ chip co-culture of neurospheres and liver equivalents for longterm substance testing. J. Biotechnol. 205, 36 46. Available from: https://doi.org/ 10.1016/j.jbiotec.2015.02.002. Mathur, A., Loskill, P., Shao, K., Huebsch, N., Hong, S., Marcus, S.G., et al., 2015. Human iPSC-based cardiac microphysiological system for drug screening applications. Sci. Rep. 5, 8883. Available from: https://doi.org/10.1038/srep08883. Matthews, R.A., 2008. Medical progress depends on animal models—doesn’t it? J. R. Soc. Med. 101, 95 98. Available from: https://doi.org/10.1258/jrsm.2007.070164.

421

422

CHAPTER 12 Design and engineering of multiorgan systems

Messner, S., Fredriksson, L., Lauschke, V.M., Roessger, K., Escher, C., Bober, M., et al., 2018. Transcriptomic, proteomic, and functional long-term characterization of multicellular three-dimensional human liver microtissues. Appl. In Vitro Toxicol. 4, 1 12. Available from: https://doi.org/10.1089/aivt.2017.0022. Miller, J.S., Stevens, K.R., Yang, M.T., Baker, B.M., Nguyen, D.-H.T., Cohen, D.M., et al., 2012. Rapid casting of patterned vascular networks for perfusable engineered three-dimensional tissues. Nat. Mater. 11, 768 774. Available from: https://doi.org/ 10.1038/nmat3357. Miller, P.G., Shuler, M.L., 2016. Design and demonstration of a pumpless 14 compartment microphysiological system. Biotechnol. Bioeng. 113, 2213 2227. Available from: https://doi.org/10.1002/bit.25989. Misun, P.M., Rothe, J., Schmid, Y.R.F., Hierlemann, A., Frey, O., 2016. Multi-analyte biosensor interface for real-time monitoring of 3D microtissue spheroids in hanging-drop networks. Microsyst. Nanoeng. 2, 16022. Available from: https://doi.org/10.1038/ micronano.2016.22. Modena, M.M., Chawla, K., Misun, P.M., Hierlemann, A., 2018. Smart cell culture systems: integration of sensors and actuators into microphysiological systems. ACS Chem. Biol. 13, 1767 1784. Available from: https://doi.org/10.1021/acschembio.7b01029. Moshksayan, K., Kashaninejad, N., Saidi, M., 2018. Inventions and innovations in preclinical platforms for cancer research. Inventions 3, 43. Available from: https://doi.org/ 10.3390/inventions3030043. Noel, G., Baetz, N.W., Staab, J.F., Donowitz, M., Kovbasnjuk, O., Pasetti, M.F., et al., 2017. A primary human macrophage-enteroid co-culture model to investigate mucosal gut physiology and host-pathogen interactions. Sci. Rep. 7, 45270. Available from: https://doi.org/10.1038/srep45270. Novakova, I., Subileau, E.A., Toegel, S., Gruber, D., Lachmann, B., Urban, E., et al., 2014. Transport rankings of non-steroidal antiinflammatory drugs across blood-brain barrier in vitro models. PLoS One 9, e86806. Available from: https://doi.org/10.1371/ journal.pone.0086806. Ohnemus, U., Kohrmeyer, K., Houdek, P., Rohde, H., Wladykowski, E., Vidal, S., et al., 2008. Regulation of epidermal tight-junctions (TJ) during infection with exfoliative toxin-negative Staphylococcus strains. J. Invest. Dermatol. 128, 906 916. Available from: https://doi.org/10.1038/sj.jid.5701070. Oleaga, C., Bernabini, C., Smith, A.S.T., Srinivasan, B., Jackson, M., McLamb, W., et al., 2016. Multi-Organ toxicity demonstration in a functional human in vitro system composed of four organs. Sci. Rep. 6, 20030. Available from: https://doi.org/10.1038/ srep20030. Oleaga, C., Riu, A., Rothemund, S., Lavado, A., McAleer, C.W., Long, C.J., et al., 2018. Investigation of the effect of hepatic metabolism on off-target cardiotoxicity in a multiorgan human-on-a-chip system. Biomaterials 182, 176 190. Available from: https:// doi.org/10.1016/j.biomaterials.2018.07.062. Osaki, T., Shin, Y., Sivathanu, V., Campisi, M., Kamm, R.D., 2018a. In vitro microfluidic models for neurodegenerative disorders. Adv. Healthc. Mater. 7, 1700489. Available from: https://doi.org/10.1002/adhm.201700489. Osaki, T., Sivathanu, V., Kamm, R.D., 2018b. Vascularized microfluidic organ-chips for drug screening, disease models and tissue engineering. Curr. Opin. Biotechnol. 52, 116 123. Available from: https://doi.org/10.1016/j.copbio.2018.03.011.

References

Ozkan, A., Ghousifam, N., Hoopes, P.J., Yankeelov, T.E., Rylander, M.N., 2019. In vitro vascularized liver and tumor tissue microenvironments on a chip for dynamic determination of nanoparticle transport and toxicity. Biotechnol. Bioeng. 116, 1201 1219. Available from: https://doi.org/10.1002/bit.26919. Paradis, A., Leblanc, D., Dumais, N., 2016. Optimization of an in vitro human blood brain barrier model: application to blood monocyte transmigration assays. MethodsX 3, 25 34. Available from: https://doi.org/10.1016/j.mex.2015.11.009. Park, T.H., Shuler, M.L., 2003. Integration of cell culture and microfabrication technology. Biotechnol. Prog. 19, 243 253. Available from: https://doi.org/10.1021/bp020143k. Peel, S., Corrigan, A.M., Ehrhardt, B., Jang, K.-J., Caetano-Pinto, P., Boeckeler, M., et al., 2019. Introducing an automated high content confocal imaging approach for organs-on-chips. Lab Chip 19, 410 421. Available from: https://doi.org/10.1039/ C8LC00829A. Prantil-Baun, R., Novak, R., Das, D., Somayaji, M.R., Przekwas, A., Ingber, D.E., 2018. Physiologically based pharmacokinetic and pharmacodynamic analysis enabled by microfluidically linked organs-on-chips. Annu. Rev. Pharmacol. Toxicol. 58, 37 64. Available from: https://doi.org/10.1146/annurev-pharmtox-010716-104748. Proctor, W.R., Foster, A.J., Vogt, J., Summers, C., Middleton, B., Pilling, M.A., et al., 2017. Utility of spherical human liver microtissues for prediction of clinical druginduced liver injury. Arch. Toxicol. 91, 2849 2863. Available from: https://doi.org/ 10.1007/s00204-017-2002-1. Prodanov, L., Jindal, R., Bale, S.S., Hegde, M., McCarty, W.J., Golberg, I., et al., 2016. Long-term maintenance of a microfluidic 3D human liver sinusoid. Biotechnol. Bioeng. 113, 241 246. Available from: https://doi.org/10.1002/bit.25700. Raasch, M., Rennert, K., Jahn, T., Peters, S., Henkel, T., Huber, O., et al., 2015. Microfluidically supported biochip design for culture of endothelial cell layers with improved perfusion conditions. Biofabrication 7, 015013. Available from: https://doi. org/10.1088/1758-5090/7/1/015013. Ramachandran, S.D., Schirmer, K., Mu¨nst, B., Heinz, S., Ghafoory, S., Wo¨lfl, S., et al., 2015. In vitro generation of functional liver organoid-like structures using adult human cells. PLoS One 10, e0139345. Available from: https://doi.org/10.1371/journal. pone.0139345. Ramme, A.P., Koenig, L., Hasenberg, T., Schwenk, C., Magauer, C., Faust, D., 2018. Towards an autologous iPSC-derived patient-on-a-chip. bioRxiv. 376970. Available from: https://doi.org10.1101/376970. Ren, H., Birch, N.P., Suresh, V., 2016. An optimised human cell culture model for alveolar epithelial transport. PLoS One 11, e0165225. Available from: https://doi.org/10.1371/ journal.pone.0165225. Renggli, K., Rousset, N., Lohasz, C., Nguyen, O.T.P., Hierlemann, A., 2019. Integrated microphysiological systems: transferable organ models and recirculating flow. Adv. Biosyst. 3, 1900018. Available from: https://doi.org/10.1002/adbi.201900018. Rogal, J., Probst, C., Loskill, P., 2017. Integration concepts for multi-organ chips: how to maintain flexibility?!. Future Sci. OA 3, FSO180. Available from: https://doi.org/ 10.4155/fsoa-2016-0092. Ronaldson-Bouchard, K., Vunjak-Novakovic, G., 2018. Organs-on-a-chip: a fast track for engineered human tissues in drug development. Cell Stem Cell 22, 310 324. Available from: https://doi.org/10.1016/j.stem.2018.02.011.

423

424

CHAPTER 12 Design and engineering of multiorgan systems

Rossi, G., Manfrin, A., Lutolf, M.P., 2018. Progress and potential in organoid research. Nat. Rev. Genet. 19, 671 687. Available from: https://doi.org/10.1038/s41576-0180051-9. Sarkar, U., Ravindra, K.C., Large, E., Young, C.L., Rivera-Burgos, D., Yu, J., et al., 2017. Integrated assessment of diclofenac biotransformation, pharmacokinetics, and omicsbased toxicity in a three-dimensional human liver-immunocompetent coculture system. Drug Metab. Dispos. 45, 855 866. Available from: https://doi.org/10.1124/ dmd.116.074005. Schmid, Y.R.F., Bu¨rgel, S.C., Misun, P.M., Hierlemann, A., Frey, O., 2016. Electrical impedance spectroscopy for microtissue spheroid analysis in hanging-drop networks. ACS Sens. 1, 1028 1035. Available from: https://doi.org/10.1021/acssensors.6b00272. Skardal, A., Murphy, S.V., Devarasetty, M., Mead, I., Kang, H.-W., Seol, Y.-J., et al., 2017. Multi-tissue interactions in an integrated three-tissue organ-on-a-chip platform. Sci. Rep. 7, 8837. Available from: https://doi.org/10.1038/s41598-017-08879-x. Smith, A.S.T., Long, C.J., Berry, B.J., McAleer, C., Stancescu, M., Molnar, P., et al., 2013. Microphysiological systems and low-cost microfluidic platform with analytics. Stem Cell Res. Ther. 4, S9. Available from: https://doi.org/10.1186/scrt370. Smits, J.P.H., Niehues, H., Rikken, G., Van Vlijmen-Willems, I.M.J.J., Van De Zande, G. W.H.J.F., Zeeuwen, P.L.J.M., et al., 2017. Immortalized N/TERT keratinocytes as an alternative cell source in 3D human epidermal models. Sci. Rep. 7, 1 14. Available from: https://doi.org/10.1038/s41598-017-12041-y. Srinivasan, B., Kolli, A.R., Esch, M.B., Abaci, H.E., Shuler, M.L., Hickman, J.J., 2015. TEER measurement techniques for in vitro barrier model systems. J. Lab. Autom. 20, 107 126. Available from: https://doi.org/10.1177/2211068214561025. Stucki, J.D., Hobi, N., Galimov, A., Stucki, A.O., Schneider-Daum, N., Lehr, C.-M., et al., 2018. Medium throughput breathing human primary cell alveolus-on-chip model. Sci. Rep. 8, 14359. Available from: https://doi.org/10.1038/s41598-018-32523-x. Sung, J.H., Shuler, M.L., 2009. A micro cell culture analog (µCCA) with 3-D hydrogel culture of multiple cell lines to assess metabolism-dependent cytotoxicity of anti-cancer drugs. Lab Chip 9, 1385. Available from: https://doi.org/10.1039/b901377f. Sung, J.H., Kam, C., Shuler, M.L., 2010. A microfluidic device for a pharmacokineticpharmacodynamic (PK-PD) model on a chip. Lab Chip 10, 446 455. Available from: https://doi.org/10.1039/b917763a. Theobald, J., Ghanem, A., Wallisch, P., Banaeiyan, A.A., Andrade-Navarro, M.A., Taˇskova, K., et al., 2018. Liver-kidney-on-chip to study toxicity of drug metabolites. ACS Biomater. Sci. Eng. 4, 78 89. Available from: https://doi.org/10.1021/ acsbiomaterials.7b00417. Trietsch, S.J., Israe¨ls, G.D., Joore, J., Hankemeier, T., Vulto, P., 2013. Microfluidic titer plate for stratified 3D cell culture. Lab Chip 13, 3548 3554. Available from: https:// doi.org/10.1039/C3LC50210D. Trietsch, S.J., Naumovska, E., Kurek, D., Setyawati, M.C., Vormann, M.K., Wilschut, K.J., et al., 2017. Membrane-free culture and real-time barrier integrity assessment of perfused intestinal epithelium tubes. Nat. Commun. 8, 262. Available from: https://doi.org/ 10.1038/s41467-017-00259-3. Truskey, G.A., 2018. Human microphysiological systems and organoids as in vitro models for toxicological studies. Front. Public Health 6, 1 9. Available from: https://doi.org/ 10.3389/fpubh.2018.00185.

References

Tsamandouras, N., Chen, W.L.K., Edington, C.D., Stokes, C.L., Griffith, L.G., Cirit, M., 2017. Integrated gut and liver microphysiological systems for quantitative in vitro pharmacokinetic studies. AAPS J. 19, 1499 1512. Available from: https://doi.org/10.1208/ s12248-017-0122-4. van den Berg, A., Mummery, C.L., Passier, R., van der Meer, A.D., 2019. Personalised organs-on-chips: functional testing for precision medicine. Lab Chip 19, 198 205. Available from: https://doi.org/10.1039/C8LC00827B. Van Den Bogaard, E.H., Tjabringa, G.S., Joosten, I., Vonk-Bergers, M., Van Rijssen, E., Tijssen, H.J., et al., 2014. Crosstalk between keratinocytes and T cells in a 3D microenvironment: a model to study inflammatory skin diseases. J. Invest. Dermatol. 134, 719 727. Available from: https://doi.org/10.1038/jid.2013.417. Van Der Meer, A.D., Van Den Berg, A., 2012. Organs-on-chips: breaking the in vitro impasse. Integr. Biol. 4, 461 470. Available from: https://doi.org/10.1039/c2ib00176d. Van Midwoud, P.M., Merema, M.T., Verpoorte, E., Groothuis, G.M.M., 2010. A microfluidic approach for in vitro assessment of interorgan interactions in drug metabolism using intestinal and liver slices. Lab Chip 10, 2778 2786. Available from: https://doi. org/10.1039/c0lc00043d. Van der Hauwaert, C., Savary, G., Gnemmi, V., Glowacki, F., Pottier, N., Bouillez, A., et al., 2013. Isolation and characterization of a primary proximal tubular epithelial cell model from human kidney by CD10/CD13 double labeling. PLoS One 8, e66750. Available from: https://doi.org/10.1371/journal.pone.0066750. Vernetti, L., Gough, A., Baetz, N., Blutt, S., Broughman, J.R., Brown, J.A., et al., 2017. Functional coupling of human microphysiology systems: intestine, liver, kidney proximal tubule, blood-brain barrier and skeletal muscle. Sci. Rep. 7, 42296. Available from: https://doi.org/10.1038/srep42296. Vernetti, L.A., Senutovitch, N., Boltz, R., DeBiasio, R., Ying Shun, T., Gough, A., et al., 2016. A human liver microphysiology platform for investigating physiology, drug safety, and disease models. Exp. Biol. Med. 241, 101 114. Available from: https://doi. org/10.1177/1535370215592121. Wagner, I., Materne, E.-M., Brincker, S., Su¨ßbier, U., Fra¨drich, C., Busek, M., et al., 2013. A dynamic multi-organ-chip for long-term cultivation and substance testing proven by 3D human liver and skin tissue co-culture. Lab Chip 13, 3538 3547. Available from: https://doi.org/10.1039/C3LC50234A. Wang, Y.I., Shuler, M.L., 2018. UniChip enables long-term recirculating unidirectional perfusion with gravity-driven flow for microphysiological systems. Lab Chip 18, 2563 2574. Available from: https://doi.org/10.1039/C8LC00394G. Wang, Y.I., Carmona, C., Hickman, J.J., Shuler, M.L., 2018. Multiorgan microphysiological systems for drug development: strategies, advances, and challenges. Adv. Healthc. Mater. 7, 1701000. Available from: https://doi.org/10.1002/adhm.201701000. Watson, D.E., Hunziker, R., Wikswo, J.P., 2017. Fitting tissue chips and microphysiological systems into the grand scheme of medicine, biology, pharmacology, and toxicology. Exp. Biol. Med. 242, 1559 1572. Available from: https://doi.org/10.1177/ 1535370217732765. Weber, E.J., Chapron, A., Chapron, B.D., Voellinger, J.L., Lidberg, K.A., Yeung, C.K., et al., 2016. Development of a microphysiological model of human kidney proximal

425

426

CHAPTER 12 Design and engineering of multiorgan systems

tubule function. Kidney Int. 90, 627 637. Available from: https://doi.org/10.1016/j. kint.2016.06.011. Weltin, A., Hammer, S., Noor, F., Kaminski, Y., Kieninger, J., Urban, G.A., 2017. Accessing 3D microtissue metabolism: lactate and oxygen monitoring in hepatocyte spheroids. Biosens. Bioelectron. 87, 941 948. Available from: https://doi.org/10.1016/ j.bios.2016.07.094. Wielhouwer, E.M., Ali, S., Al-Afandi, A., Blom, M.T., Olde Riekerink, M.B., Poelma, C., et al., 2011. Zebrafish embryo development in a microfluidic flow-through system. Lab Chip 11, 1815 1824. Available from: https://doi.org/10.1039/c0lc00443j. Wikswo, J.P., 2014. The relevance and potential roles of microphysiological systems in biology and medicine. Exp. Biol. Med. 239, 1061 1072. Available from: https://doi. org/10.1177/1535370214542068. Wikswo, J.P., Block, F.E., Cliffel, D.E., Goodwin, C.R., Marasco, C.C., Markov, D.A., et al., 2013a. Engineering challenges for instrumenting and controlling integrated organ-on-chip systems. IEEE Trans. Biomed. Eng. 60, 682 690. Available from: https://doi.org/10.1109/TBME.2013.2244891. Wikswo, J.P., Curtis, E.L., Eagleton, Z.E., Evans, B.C., Kole, A., Hofmeister, L.H., et al., 2013b. Scaling and systems biology for integrating multiple organs-on-a-chip. Lab Chip 13, 3496 3511. Available from: https://doi.org/10.1039/c3lc50243k. Workman, M.J., Gleeson, J.P., Troisi, E.J., Estrada, H.Q., Kerns, S.J., Hinojosa, C.D., et al., 2018. Enhanced utilization of induced pluripotent stem cell derived human intestinal organoids using microengineered chips. Cell. Mol. Gastroenterol. Hepatol. 5, 669 677. Available from: https://doi.org/10.1016/j.jcmgh.2017.12.008. Xiao, S., Coppeta, J.R., Rogers, H.B., Isenberg, B.C., Zhu, J., Olalekan, S.A., et al., 2017. A microfluidic culture model of the human reproductive tract and 28-day menstrual cycle. Nat. Commun. 8, 14584. Available from: https://doi.org/10.1038/ncomms14584. Yamada, M., Utoh, R., Ohashi, K., Tatsumi, K., Yamato, M., Okano, T., et al., 2012. Controlled formation of heterotypic hepatic micro-organoids in anisotropic hydrogel microfibers for long-term preservation of liver-specific functions. Biomaterials 33, 8304 8315. Available from: https://doi.org/10.1016/j.biomaterials.2012.07.068. Ye, D., Dawson, K.A., Lynch, I., 2015. A TEM protocol for quality assurance of in vitro cellular barrier models and its application to the assessment of nanoparticle transport mechanisms across barriers. Analyst 140, 83 97. Available from: https://doi.org/ 10.1039/C4AN01276C. Yonker, L.M., Mou, H., Chu, K.K., Pazos, M.A., Leung, H., Cui, D., et al., 2017. Development of a primary human co-culture model of inflamed airway mucosa. Sci. Rep. 7, 8182. Available from: https://doi.org/10.1038/s41598-017-08567-w. Yu, J., Cilfone, N., Large, E., Sarkar, U., Wishnok, J., Tannenbaum, S., et al., 2015. Quantitative systems pharmacology approaches applied to microphysiological systems (MPS): data interpretation and multi-MPS integration. CPT Pharmacometrics Syst. Pharmacol. 4, 585 594. Available from: https://doi.org/10.1002/psp4.12010. Zhang, B., Radisic, M., 2017. Organ-on-a-chip devices advance to market. Lab Chip 17, 2395 2420. Available from: https://doi.org/10.1039/C6LC01554A. Zhang, C., Zhao, Z., Abdul Rahim, N.A., van Noort, D., Yu, H., 2009. Towards a humanon-chip: culturing multiple cell types on a chip with compartmentalized microenvironments. Lab Chip 9, 3185 3192. Available from: https://doi.org/10.1039/b915147h.

References

Zhang, W., Zhuang, A., Gu, P., Zhou, H., Fan, X., 2016. A review of the threedimensional cell culture technique: approaches, advantages and applications. Curr. Stem Cell Res. Ther. 11, 370 380. Zhang, Y.S., Aleman, J., Shin, S.R., Kilic, T., Kim, D., Mousavi Shaegh, S.A., et al., 2017. Multisensor-integrated organs-on-chips platform for automated and continual in situ monitoring of organoid behaviors. Proc. Natl. Acad. Sci. U.S.A. 114, E2293 E2302. Available from: https://doi.org/10.1073/pnas.1612906114. Zhao, Y., Rafatian, N., Feric, N.T., Cox, B.J., Aschar-Sobbi, R., Wang, E.Y., et al., 2019. A platform for generation of chamber-specific cardiac tissues and disease modeling. Cell 176, 913 927. Available from: https://doi.org/10.1016/j.cell.2018.11.042.

427

CHAPTER

Human body-on-achip systems

13 Eva-Maria Dehne and Uwe Marx TissUse GmbH, Berlin, Germany

Introduction Microphysiological systems (MPS) have evolved greatly in sophistication during the past decade. Their increased physiological relevance enhances the translatability of assay readouts and results to the human situation (Sung et al., 2019). Various systems mimicking in vivo like tissue architecture and physiological flow conditions have already been shown to be highly valuable in basic research and various stages of the drug development process (Dehne et al., 2019). Test systems are increasingly composed of three-dimensional cocultures of several cell types, enhancing model accuracy and stability. The integration of constant medium perfusion allows a more physiologically relevant nutrient provision and removal of metabolic products (Marx et al., 2016). In addition, the shear stress exerted on the cells and other physical stimuli result in an in vivo like cellular behavior (Schimek et al., 2013; Huh et al., 2012). Finally, the integration of multiple organ or tissue models into multiorgan chips answers specialized questions on interorgan communication, drug distribution, metabolism, and the effects of metabolites on predefined organ systems (Hu¨bner et al., 2018; Chen et al., 2017; Lohasz et al., 2019; Oleaga et al., 2016). However, the true effects of a substance on humans are often a result of unpredicted multitissue cross-talk. Therefore apart from the specialized assays that are currently being adopted by the pharmaceutical industry, a systemic evaluation of compound effects is needed. The use of a model that is more predictive of the human response than current preclinical models is crucial. Human bodyon-a-chip systems (also known as human-on-a-chip) cocultivate multiple physiologically relevant organ models in a combined media circuit and aim to maintain organismal homeostasis. In the following sections, we highlight the need to go beyond single- or multiorgan chips, discuss general design principles for human body-on-a-chip devices, and outline future applications and opportunities.

Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00013-9 © 2020 Elsevier Inc. All rights reserved.

429

430

CHAPTER 13 Human body-on-a-chip systems

Why we need organismal models on a chip Multiorgan MPS aim at an in vivo like interorgan communication via soluble factors or cells trafficking between the organ compartments. Sophisticated devices modeling, for example, the human reproductive tract and a 28-day menstrual cycle, have been reported (Xiao et al., 2017). A microfluidic system supported the generation of hormone profiles controlling the female reproductive tract and peripheral tissue dynamics. The endocrine loops between organ models of the ovary, fallopian tube, uterus, cervix, and liver were simulated in a closed microfluidic circulation. Other examples of successful interorgan communication in MPS include the coupling of human pancreatic islets and liver spheroids on a chip to model type 2 diabetes. Insulin released from the islet microtissues promoted glucose uptake by the liver spheroids. As glucose concentrations decreased, insulin secretion subsided, demonstrating a functional feedback loop between the tissue models (Bauer et al., 2017). A direct cellular interaction was reported in a metastasis-on-a-chip assay studying the migration of tumor cells from a colon organoid to a liver organoid (Skardal et al., 2016). These and other systems combining a predefined set of organ models are extremely valuable in the study of specific questions. However, these systems are not designed to provide a thorough survey of the effects a substance may exert on organs other than the primary target. In many cases, drugs interact with multiple organs in the body, resulting in complex whole-body responses. Preclinical trials using laboratory animals and investigating systemic interactions have been the only option to date to detect unexpected effects on an organismal level. Organs in the human body work in an interdependent manner. Complex biological mechanisms—which are still partially unknown—orchestrate the physiological maintenance of all organs. A constant cross-talk is required to maintain homeostasis and involves cytokines and endocrine hormones. Moreover, lessstudied interactions are required for homeostasis, such as the accumulation of liver-produced albumin in the skin (Peters, 1996). Extracting one or two organs from this complex interdependent feedback and performing single-tissue assays may be possible for restricted culture periods but will provide only a limited and partial answer to the questions asked. Human body-on-a-chip systems are designed to emulate human systemic physiology through the interactions of all major organ systems. It has been postulated that at least the following 10 systems should be present and show in vivo like interaction: circulatory, endocrine, gastrointestinal, immune, integumentary, musculoskeletal, nervous, reproductive, respiratory, and urinary (Marx et al., 2016). Physiological ratios of organ sizes, an in vivo like organ arrangement, medium flow rates, and physiological fluid-to-tissue ratios should be integrated into these platforms. Such systems would permit the systemic interaction of major organ equivalents and enable the study of complex time-dependent concentration profiles of both endogenous and administered compounds. The level of accurate extrapolation of data from human body-on-a-chip devices to humans increases

Design principles

with the growing perfection of human mimicry. Hence, many design criteria and principles must be considered when developing such devices. General approaches for the engineering of a body-on-a-chip platform are described in the following section.

Design principles The design principles of body-on-a-chip devices are based on the findings and technological developments of multiorgan chips. MPS can generally be divided into two subgroups: single-pass and recirculating microfluidic platforms. The principle of a single-pass microfluidic platform is not applicable to human bodyon-a-chip devices, because the essential organ organ cross-talk and therefore establishment of systemic homeostasis are not feasible. Recirculating microfluidic chips can move the medium through the channels by a pumpless (hydrostatic pressure driven) (Lohasz et al., 2019; Chen et al., 2018; Vriend et al., 2018) or pump-driven mode (Chen et al., 2017; Coppeta et al., 2016; Maschmeyer et al., 2015). The hydrostatic pressure driven technique allows the chips to be run without specialized equipment and ensures a close-to-physiological fluid-to-tissue ratio, as no external pumps or media reservoirs are required (Miller and Shuler, 2016). Fluid flow velocities can be adjusted in a narrow range by changing the tilting angle or fluid height in the reservoirs, but once the system is constructed, adaptations of fluid split to the various organs are impossible. Therefore control over the fluidics is limited and cannot match the flexibility of the human counterpart, the cardiovascular system. The latter allows the adjustment of nutrition on demand. Furthermore, MPS operated by hydrostatic pressure exert a bidirectional flow, alternating as the medium passes the cells with each tilting of the device. This interferes with stable monodirectional physiological gradients that are crucial when mimicking human behavior. The first few examples of monodirectional perfusion in pumpless devices have been reported (Wang and Shuler, 2018). However, these techniques must still be refined and validated as applicable to body-on-a-chip devices. Pump-driven platforms should operate micropumps onchip. This enables a close-to-physiological fluid-to-tissue ratio, ensuring interorgan cross-talk via soluble factors. Furthermore, unwanted loss of substances and their metabolites from surface absorption is reduced by eliminating tubing and external pumps. The use of an on-chip pump also facilitates the integration of additional valves in front of each organ which adjust fluid flow splits dynamically, echoing the vessel contractions in the human cardiovascular system (Marx, 2014).

Size matters Design principles and downscaling rules define the accuracy of the quantitative extrapolation of body-on-a-chip in vitro data to in vivo situations in humans.

431

432

CHAPTER 13 Human body-on-a-chip systems

When the arrangement and architecture are organotypic and the downscaling is linear, the pharmacokinetics and pharmacodynamics of a drug and its metabolites are more likely to be accurate and replicate the in vivo effects. Various scaling approaches for MPS have been published (Marx et al., 2012, 2016; Wikswo et al., 2013; Stokes et al., 2015). Scaling may be based on organ function. However, the definition of which specific function to use for scaling is heavily debated and depends on the context of use. In the liver, for example, albumin production or cytochrome P450 activity may be of interest. Both factors will vary greatly depending on the cell source and may lead to different scaling assumptions. Therefore scaling based on the smallest functional units of each organ—the physiological organoids—was proposed (Marx et al., 2012). These physiological organoids can perform the tasks of the entire organ, albeit on a small scale. Recreating physiological organoids on a chip and matching their numbers should enable the cells to self-adjust their functionality to achieve homeostasis. Admittedly, most organoids generated in vitro have not yet emulated all in vivo functions to the same degree. We believe that the close interaction of organ models stabilizing organismal homeostasis will result in an adaptation of organ-specific cell mass and activity to match human functionality. In vivo like organ arrangements, liquid flows, and substance gradients are essential to physiological interorgan communication and cellular adaptation. Human body-on-a-chip systems, therefore, aim at a design compliant with physiologically based pharmacokinetics (Miller and Shuler 2016; Edington et al., 2018). The overall flow rate, medium partitioning to the organs, and fluid retention times in each organ compartment are modeled to represent those of an average human male. Human body-on-a-chip platforms should achieve a physiological fluid-totissue ratio; this has become the paramount challenge, because, ideally, that ratio should mimic the in vivo like intravascular and interstitial organ-specific situation. Because of the lack of a closed on-chip microvasculature mimicking arteriolar and venular vessels at the channel level and microcapillaries at the organoid level, systemic erythrocyte perfusion for sufficient oxygen provision is not yet possible. The existing systems must provide oxygen through the culture medium, which requires large volumes of media due to the extremely low solubility of oxygen. This leads to an artificial dilution of biological factors and drugs. Most biochemical reactions and cell signaling are concentration-dependent. Therefore low organ-specific and systemic quantities of liquid are important for interorgan communication and the detection of drug responses. The artificial dilution of a drug or its metabolites may lead to different responses or failure to detect certain effects (Wang et al., 2018). The use of microfluidic circuits generally reduces the organ-specific fluid-to-tissue ratio greatly compared to the ratio in static cultures. The systemic fluid-to-tissue ratio is still far from physiological because of the use of media reservoirs or external pumps, However, to date, hydrostatically driven chips or on-chip pumps have been the only feasible options for keeping the overall fluid volumes low (Wang et al., 2018). Another

Design principles

disadvantage of using low fluid-to-tissue ratios is the need for constant removal of metabolic products. Excretory organs, such as the kidney, perform these tasks in vivo. Complex technological solutions are required in human body-on-a-chip devices to maintain low manual media exchange frequencies. Approaches for providing constant media supply and more accurate tissue physiology are described below.

From monolayer fluidics to multi-compartmental platforms In vivo, polarized epithelia form a barrier between the intravascular fluid and organ-specific apical compartments (e.g., bile, urine, cerebrospinal fluid). Constant segregation and removal of apical fluids of varying composition are crucial for organismal homeostasis. Therefore polarized tissues are exposed to various environments on different cellular surfaces and maintain stability by active transport. Such a compartmentalized build-up of an organ equivalent with segregated flows has been shown to increase tissue performance. An increase in model accuracy has been reported when shear stress was applied to both sides of epithelial cells from kidney, intestine, and lung tissues (Kim et al., 2012; Jang et al., 2013; Huh et al., 2010). However, most multiorgan devices focus on the integration of a single microfluidic circuit to interconnect organ equivalents. Barrier organs are often integrated on porous membranes so that the cells experience a dynamic perfusion on one side and a static environment on the other (Chen et al., 2017; Coppeta et al., 2016; Maschmeyer et al., 2015). Emulation of the human situation requires the integration of multiple fluidic circuits on several layers of the device. The technical complexity of achieving this multicompartmental flow has hampered its integration into multiorgan devices. The only system reported to date that has achieved this is a four-organ chip interconnecting intestine, liver, skin, and kidney equivalents with common medium circulation (Maschmeyer et al., 2015). The kidney compartment contained a permeable membrane that was perfused by common medium circulation on one side and a secondary urinary circuit on the other side. This device, which was developed in our laboratory, was operated for 28 days at near-to-physiological fluid-to-tissue ratios. Similarly, the application of physical stimuli, such as cyclical pressure on a bone or cartilage compartment or the stretching of lung epithelial cells, would increase tissue accuracy but have, so far, only been performed in single-organ devices. In multiorgan devices, electrodes have been integrated for tissue stimulation or readout of contractile force, electrical conductivity, or transepithelial electrical resistance (Oleaga et al., 2018; Maoz et al., 2017; Zhang et al., 2017). Apart from enhancing the physiological relevance of tissues, these additions increase culture reliability by providing online readout parameters. As indicated above, the use of low systemic fluid-to-tissue ratios requires a more frequent feeding of nutrients and removal of metabolic products. The integration of an additional on-chip feeding circuit, which is operated by an independent pump and supplies nutrients to an intestinal model, would reduce manual feeding procedures. This integrated feeding strategy would require an excretory

433

434

CHAPTER 13 Human body-on-a-chip systems

kidney circuit to remove metabolic products. Marx (2014) The separation of two overlapping microfluidic circuits by a closed cell layer poses a challenge from both the biological standpoint and technical construction. Supporting structures, such as porous membranes, must be integrated to enable unobstructed fluidic passage on both sides. In practice, this requires a complex multilayer device, possibly constructed from various materials. A body-on-a-chip device that simulates more circulatory systems results in a cellular environment that is more physiologically relevant. The translation of pharmacokinetic assessments from chip experiments to the human situation also becomes more straightforward. Integrating a closed vascular system represents an important step in achieving an in vivo like cellular environment and reliable pharmacokinetic readouts.

Enhancing the circulatory system The vascular network in the human body is responsible for supplying nutrients, removing waste, and interorgan communications. In body-on-a-chip devices, these tasks are performed by the channels connecting organ compartments, although the channels cannot mimic all biological mechanisms involved in these tasks. Many organs contain specialized endothelial cells and rely on their specific functions. Endothelial cells play a major role, for example, in establishing the blood brain barrier, restricting the passage of pathogens and the diffusion of large or hydrophilic molecules into the cerebrospinal fluid. Liver sinusoidal endothelial cells, in contrast, allow a fast exchange of substances between the blood and hepatocytes and regulate the hepatic vascular tone, thereby maintaining low portal pressure despite changes in hepatic blood flow during digestion (Poisson et al., 2017). Regarding pathological conditions, liver sinusoidal endothelial cells are known to be involved in the initiation and progression of chronic liver disease (Ni et al., 2017). Endothelial cells also undertake more general tasks, such as the control of leukocyte and platelet adhesion and aggregation, secretion of angiocrine factors, and stimulation of various organ-specific cells (Bierhansl et al., 2017). MPS with a closed endothelial layer covering the channels have been reported (Schimek et al., 2013). Systems containing complex microcapillary structures inside hydrogels have also been developed (Kim et al., 2013, 2016; Wang et al., 2016). Combining these technologies with human body-on-a-chip devices will lead to higher model accuracy and enable the study of more complex physiological mechanisms, such as the migration of cells (e.g., immune cells, metastasizing tumor cells) in and out of the vascular system.

Opportunities Human body-on-a-chip devices can be applied in a broad range of fields because of their systemic nature. In the drug development arena, MPS feature three major

Challenges

opportunities: to perform trials on healthy body-on-a-chip systems, to perform patient-on-a-chip trials using systems mimicking a disease of interest, and to evaluate the individual (patient-specific) systemic effects of a substance. The first two functions would use a general cell source to model a random patient population or, possibly, a population with a specific genetic background. The third would use patient-derived material to provide an individual prediction of a therapeutic outcome. Body-on-a-chip assays using healthy organ models are envisioned to predict potential drug toxicity and aid in dose adjustment prior to entering the clinical trials stage. These assays will predict the human systemic response more accurately than standard preclinical models by mimicking human in vivo pharmacokinetics and pharmacodynamics. This use of systemic data generated in vitro may refine and reduce the need for Phase I studies on healthy individuals. Furthermore, studies on populations with whom no clinical trials can be performed—such as children or pregnant women—will be possible. Patient-on-a-chip trials may, at first, be performed prior to or in parallel with classical clinical trials and support the study of the mechanisms underlying responses to drugs. The ability to evaluate the safety and efficacy of a compound systematically and in an undirected manner will be most important, and these patient-on-a-chip trials are envisioned to shorten or even partially eliminate clinical trials. The accurate recreation of a diseased organismal state will be essential here. Therefore the in vitro induction and in vivo like progression of a disease must be established. The pathological condition of this on-chip assay has to mimic the specific organ organ cross-talk and systemic effects. The knowledge gained from preclinical animal studies may be useful when searching for ways to induce a diseased state in a healthy organism. Well-studied diseases, such as metabolic diseases and certain tumors or infections, may serve as proof-of-concept evaluations. Individualized assays using patient-derived material, such as induced pluripotent stem cell derived organ models, will enable a precise prognosis of individual drug effects. These assays may also mimic previous diseases, patient-specific parameters (such as hypertension), or the effects of drug drug interactions. Not only pharmaceuticals but also chemicals, food ingredients, or consumer health products may be tested in this way.

Challenges Only a few devices intended for systemic interconnection of all major organs exist (Miller and Shuler, 2016; Edington et al., 2018) because of the technical and biological complexity of establishing a body-on-a-chip system. General constraints of multiorgan MPS—such as the development of a common medium supporting all organ equivalents, the correct choice of material for device construction, and the generation of in vivo-relevant organ models—must be overcome.

435

436

CHAPTER 13 Human body-on-a-chip systems

The quest for a common medium supporting all organ equivalents becomes more difficult with each additional organ integrated into the device. Media can be highly specialized and vary by cell type. Primary cells, stem cells, and induced pluripotent stem cells and their progeny in particular are often supplemented in vitro with multiple growth factors. As mentioned, body-on-a-chip devices aim at a closeto-physiological fluid-to-tissue ratio, but oxygen provisioning represents a major issue. We believe that, eventually, this problem can only be solved in a physiological manner, by using whole blood containing erythrocytes. However, the use of blood-containing erythrocytes will only be possible when the vascular bed is fully closed. Elaborate systems connecting vascularized technical channels with vascularized organ models will be required. The generation of a microvascular bed inside a hydrogel has been shown by multiple groups (Kim et al., 2013, 2016; Wang et al., 2016), but the addition of parenchymal cells to generate a truly vascularized organ model is still in its infancy (Jusoh et al., 2015; Nashimoto et al., 2017). The lower throughput and higher cost of body-on-a-chip systems compared to less complex single- or multi-organ-on-a-chip devices will limit their application to later stages of the drug development process or to the study of highly complex questions during basic research. As body-on-a-chip platforms allow for a comprehensive analysis of systemic responses—which would otherwise only be feasible in animal or clinical trials—throughput and cost of assays have to be compared to those trials. Here, a marked reduction in cost and the opportunity to increase throughput without ethical constrains is feasible. An automation of assays—as discussed in Chapter 14—will further facilitate assay performance therewith increasing throughput and accuracy. Assays performed on body-on-a-chip systems enable a broad spectrum of indepth analysis techniques. General readouts, such as organ histology, may be of interest in comparing systems to animal trials. Genomics or proteomics analyses will provide a more detailed picture. Scientists must often select one endpoint to analyze per tissue because of the low number of cells in MPS. Analysis of medium samples requires validated markers of both systemic and organ-specific conditions. On-chip assays generally enable extensive online measurements that yield mechanistically relevant insights. The integration of electrodes for measuring transepithelial electrical resistance or electrical conductivity, sensor beads to measure glucose, lactate, or oxygen levels, and various optical inspection tools has been achieved. The more details we obtain about tissue performance and drug response, the higher the confidence level of body-on-a-chip devices. One major issue when designing a body-on-a-chip device is the choice of material and the techniques for production. The scale of production, tolerances, and the ease of integrating adaptations must be considered. Requirements for the underlying technical solutions multiply with the increase in biological complexity. Devices containing several materials with different properties (e.g., flexibility, gas permeability, conductivity) must be assembled in a manner that permits a fluidically tight device. More detail on this topic is given in Chapter 16, “Automation and opportunities for industry scale-up of microphysiologial systems.”

References

Conclusion A systemic model at the organismal level would be required for a comprehensive substance-testing paradigm that eliminates the need for animal testing. Human body-on-a-chip devices coculturing all relevant organ models in a combined media circuit may represent the solution. These systems are designed to study complex organ organ interactions, as well as the efficacy and potential toxicity of substances in nontarget organs. Predictive human test systems are urgently needed, especially to analyze the complex effects of biotherapeutics and living drugs. Body-on-a-chip devices using healthy tissues will predict potential drug toxicity, aid in dose adjustment, and provide information about drug effects in specific populations. Patient-on-a-chip assays may be used in basic research to study a given disease or as a support tool and may provide an alternative to clinical trials. Personalized body-on-a-chip devices containing individual-related tissues or characteristics will provide targeted information on substance effects or therapeutic outcome. Several challenges must still be overcome to achieve this, but the manifold scientific advancements of multiorgan chips will aid in their solution.

References Bauer, S., et al., 2017. Functional coupling of human pancreatic islets and liver spheroids on-a-chip: Towards a novel human ex vivo type 2 diabetes model. Sci. Rep. 7, 1 11. Bierhansl, L., Conradi, L., Treps, L., Dewerchin, M., Carmeliet, P., 2017. Central role of metabolism in endothelial cell function and vascular disease. Physiology 32, 126 140. Chen, W.L.K., et al., 2017. Integrated gut/liver microphysiological systems elucidates inflammatory inter-tissue crosstalk. Biotechnol. Bioeng. 114, 2648 2659. Chen, H.J., Miller, P., Shuler, M.L., 2018. A pumpless body-on-a-chip model using a primary culture of human intestinal cells and a 3D culture of liver cells. Lab Chip 18, 2036 2046. Coppeta, J.R., Mescher, M.J., Isenberg, B.C., Spencer, A.J., Borenstein, J.T., 2016. A portable and reconfigurable multi-organ platform for drug development with onboard microfluidic flow control. Lab Chip 17, 134 144. Dehne, E.-M., Hickman, J.J., Shuler, M.L., 2019. Biologically-inspired microphysiological systems. In: Balls, M., Combes, R., Worth, A. (Eds.), The History of Alternative Test Methods in Toxicology. Elsevier, pp. 279 284. Edington, C.D., et al., 2018. Interconnected microphysiological systems for quantitative biology and pharmacology studies. Sci. Rep. 8, 1 18. Hu¨bner, J., et al., 2018. Simultaneous evaluation of anti-EGFR-induced tumour and adverse skin effects in a microfluidic human 3D co-culture model. Sci. Rep. 8, 1 12. Huh, D., et al., 2010. Reconstituting organ-level lung functions on a chip. Science 328 (80), 1662 1668.

437

438

CHAPTER 13 Human body-on-a-chip systems

Huh, D., et al., 2012. A human disease model of drug toxicity-induced pulmonary edema in a lung-on-a-chip microdevice. Sci. Transl. Med. 4, 159ra147. Jang, K.-J., et al., 2013. Human kidney proximal tubule-on-a-chip for drug transport and nephrotoxicity assessment. Integr. Biol. (Camb) 5, 1119 1129. Jusoh, N., Oh, S., Kim, S., Kim, J., Jeon, N.L., 2015. Microfluidic vascularized bone tissue model with hydroxyapatite-incorporated extracellular matrix. Lab Chip 15, 3984 3988. Kim, H.J., Huh, D., Hamilton, G., Ingber, D.E., Links, D.A., 2012. Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow. Lab Chip 12, 2165 2174. Kim, S., Lee, H., Chung, M., Jeon, N.L., 2013. Engineering of functional, perfusable 3D microvascular networks on a chip. Lab Chip 13, 1489 1500. Kim, S., Chung, M., Ahn, J., Lee, S., Jeon, N.L., 2016. Interstitial flow regulates the angiogenic response and phenotype of endothelial cells in a 3D culture model. Lab Chip 16, 4189 4199. Lohasz, C., Rousset, N., Renggli, K., Hierlemann, A., Frey, O., 2019. Scalable, microfluidic platform for flexible configuration of and experiments with microtissue multi-organ models. SLAS Technol. 24, 79 95. Maoz, B.M., et al., 2017. Organs-on-chips with combined multi-electrode array and transepithelial electrical resistance measurement capabilities. Lab Chip 17, 2294 2302. Marx, U., 2014. Multi-organ-chip with improved life time and homoeostasis Marx, U., et al., 2012. “Human-on-a-chip” developments: a translational cutting-edge alternative to systemic safety assessment and efficiency evaluation of substances in laboratory animals and man. ATLA 40, 235 257. Marx, U., et al., 2016. Biology-inspired microphysiological system approaches to solve the prediction dilemma of substance testing. ALTEX 33, 272 321. Maschmeyer, I., et al., 2015. A four-organ-chip for interconnected long-term co-culture of human intestine, liver, skin and kidney equivalents. Lab Chip 15, 2688 2699. Miller, P.G., Shuler, M.L., 2016. Design and demonstration of a pumpless 14 compartment microphysiological system. Biotechnol. Bioeng. 113, 2213 2227. Nashimoto, Y., et al., 2017. Integrating perfusable vascular networks with a threedimensional tissue in a microfluidic device. Integr. Biol. (United Kingdom) 9, 506 518. Ni, Y., et al., 2017. Pathological process of liver sinusoidal endothelial cells in liver diseases. World J. Gastroenterol. 23, 7666 7677. Oleaga, C., et al., 2016. Multi-Organ toxicity demonstration in a functional human in vitro system composed of four organs. Sci. Rep. 6, 1 17. Oleaga, C., et al., 2018. Investigation of the effect of hepatic metabolism on off-target cardiotoxicity in a multi-organ human-on-a-chip system. Biomaterials 182, 176 190. Peters, T., 1996. All About Albumin: Biochemistry, Genetics, and Medical Applications. Academic Press. Poisson, J., et al., 2017. Liver sinusoidal endothelial cells: physiology and role in liver diseases. J. Hepatol. 66, 212 227. Schimek, K., et al., 2013. Integrating biological vasculature into a multi-organ-chip microsystem. Lab Chip 13, 3588 3598. Skardal, A., Devarasetty, M., Forsythe, S., Atala, A., Soker, S., 2016. A reductionist metastasis-on-a-chip platform for in vitro tumor progression modeling and drug screening. Biotechnol. Bioeng. 113, 2020 2032.

References

Stokes, C.L., Cirit, M., Lauffenburger, D.A., 2015. Physiome-on-a-chip: the challenge of “scaling” in design, operation, and translation of microphysiological systems. CPT Pharmacometrics Syst. Pharmacol. 4, 559 562. Sung, J.H., et al., 2019. Recent advances in body-on-a-chip systems. Anal. Chem. 91, 330 351. Vriend, J., et al., 2018. Screening of drug-transporter interactions in a 3D microfluidic renal proximal tubule on a chip. AAPS J. 20, 87. Wang, Y.I., Shuler, M.L., 2018. UniChip enables long-term recirculating unidirectional perfusion with gravity-driven flow for microphysiological systems. Lab Chip 18, 2563 2574. Wang, X., Phan, D.T.T., George, S.C., Hughes, C.C.W., Lee, A.P., 2016. Engineering anastomosis between living capillary networks and endothelial cell-lined microfluidic channels. Lab Chip 16, 282 290. Wang, Y.I., Carmona, C., Hickman, J.J., Shuler, M.L., 2018. Multiorgan microphysiological systems for drug development: strategies, advances, and challenges. Adv. Healthc. Mater. 7, 1 29. Wikswo, J.P., et al., 2013. Scaling and systems biology for integrating multiple organs-ona-chip. Lab Chip 13, 3496 3511. Xiao, S., et al., 2017. A microfluidic culture model of the human reproductive tract and 28-day menstrual cycle. Nat. Commun. 8, 1 13. Zhang, Y.S., et al., 2017. Multisensor-integrated organs-on-chips platform for automated and continual in situ monitoring of organoid behaviors. PNAS E2293 E2302. Available from: https://doi.org/10.1073/pnas.1612906114.

439

CHAPTER

14

Automation and opportunities for industry scale-up of microphysiological systems

Eva-Maria Dehne1, Hendrik Erfurth1, Ann-Kristin Muhsmann2 and Uwe Marx1 1

TissUse GmbH, Berlin, Germany Technische Universita¨t Berlin, Medical Biotechnology, Berlin, Germany

2

The automation of microphysiological systems (MPS) is still in its infancy, and only a few of the mostly research-driven platforms aim for industry scale-up. However, there is a need for automation and standardization of these systems to ensure the constant high quality of the devices and accuracy of test performance by eliminating human error. Automation is defined as the technology by which a process or procedure is performed without human assistance (Groover, 2010). Various control systems for input measurements and output control signals are used to operate programmable machines. Computer systems ensure control, sensory feedback, and information processing. Hence, the establishment of an automated process requires the interaction of mechanical engineering, electronics engineering, controls engineering, and computer science. Modern laboratory routines already incorporate a variety of automated processes, such as thermal cyclers and flow cytometers. The automation of standard routines offers savings in labor and time and improvements in quality, accuracy, and precision (Chapman, 2003). When qualifying assays basing on MPS, inter- and intralab reproducibility and robustness are desirable and will increase with the degree of automation (Rebelo et al., 2016). The increase in throughput and a corresponding reduction in overall assay time and cost are also important. Currently, the field of research believed to gain the most from MPS technology is the pharmaceutical drug development process (Marx et al., 2016). A more accurate, cost-effective, and clinically relevant testing of drugs using MPS will reduce costly late-stage failures of compounds during clinical trials. To achieve this goal, systems applicable to the various stages of the drug development process—from early high-throughput screening assays for target identification and lead selection and optimization to final highcontent systemic safety assessment and efficacy testing—must transition from academia to industry (Marx et al., 2016). To that end, industrial standardization of device production, a defined loading of tissue into MPS, automated MPS Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00014-0 © 2020 Elsevier Inc. All rights reserved.

441

442

CHAPTER 14 Automation and opportunities for industry

FIGURE 14.1 Aspects of microphysiological systems offering opportunities for automation.

culture, and automated culture analysis are required, as are user-friendly interfaces to facilitate the work of individuals with only limited training and enable an easy transfer of devices from developers to end-users (Fig. 14.1).

High-throughput versus high-content systems The traditional drug development processes often begin with compound screening during lead discovery (Hughes et al., 2011). A large number of compounds available and extra- and intracellular molecular targets require a miniaturization and automation of assays. Up to 100,000 compounds per day are tested in parallel by high-throughput screening, always ensuring redundant conditions (Mayr and Fuerst, 2008). Fewer tests must be run in parallel in the later preclinical phase, but interactions within the system observed then become much more complex. Thus the focus shifts to gathering high-content data. MPS aim to emulate the complexity of the human organism better than standard cell cultures (Wang et al., 2018). Therefore systems are designed mainly for high-content rather than high-throughput testing, but higher throughput and reliability of MPS assays are necessary to achieve industry adoption and regulatory acceptance. Consequently, a certain degree of automation of devices and processes is inevitable. As the systems become more applicable for pharmaceutical work, some MPS developers are designing solutions that require less hands-on time and allow for a scale-up of production and parallelization and automation of assays (discussed as follows). Challenges that must be overcome before MPS are ready for higher throughput testing have been covered in the literature (Lohasz et al., 2018). Some oftennamed requirements are fast and cost-effective device production, automated

History of laboratory automation

loading of tissues, automated cell culture execution (e.g., medium and substance application, incubation, fluid circulation actuation, oxygen supply), automated optical and chemical measurements, automated data storage and analysis, and a general reduction in human interaction. Feasibility and necessary steps to enable those requirements in current MPS are discussed as follows. We first cover the historical rise of these technologies, followed by an overview of the most important MPS types and their suitability for automation. This historical account and the general survey will help to explain the challenges and opportunities for automating MPS cultures. Subsequently, we detail the various aspects of automation: device production, high-throughput compatible tissue cultures, and systems for automated device operation and analysis.

History of laboratory automation Modern automated laboratory processes and routines provide the basis for automated operation of MPS. The developments conceived for clinical laboratories and pharmaceutical screening assays are applicable to designing automated MPS. The automation of laboratory practices began as early as 1875 with the introduction of a device conceived to wash filtrates unattended (Olsen, 2012). Since then, laboratory automation often followed the growth of certain industries or the demands of analytical challenges. For example the steady increase in repetitive routine analyses in clinical laboratories during the mid-1950s promoted the development of the first automated blood analyzer. This instrument, into which samples were loaded and which generated analytical output, was manufactured by Technicon. The Technicon analyzer measured the levels of urea, sugar, and calcium in a blood sample within 2.5 minutes (Olsen, 2012). Clinical laboratories have steadily continued to automate their processes to manage thousands of samples per day. Miniaturized motor and valve technology enabled the development of automated liquid-handling machines. Fully automated analytical workstations are part of standard equipment in many clinical laboratories today. Compact userprogrammable robot arms have also been adapted for numerous assays and sample-handling approaches, especially complex laboratory assays that are laborintensive when carried out manually (Boyd, 2002). Such automated procedures greatly increase the throughput of a laboratory. These developments culminated in the establishment of the first fully automated clinical laboratory built by Dr. Masahide Sasaki in Japan, where robots carried test tube racks and conveyor belts transported patient samples to various analytical workstations (Sasaki et al., 1998). By the mid-1990s, pharmaceutical companies were searching for labor-saving devices, as the number of compounds to screen during initial drug discovery was increasing. The only way to test large numbers of compounds within a reasonable time frame was to miniaturize the assays and then automate them. Much of the

443

444

CHAPTER 14 Automation and opportunities for industry

required technology was already in use in clinical laboratory analyses; liquid-handling robots in particular are still widely used in the pharmaceutical industry in screening large compound libraries, as they reduce setup time for low-volume dispensing and provide highly accurate and reproducible results that are difficult to achieve manually. Simultaneously, the increasing demand for fast and automated analyses in other fields of research led to the development of lab-on-a-chip devices. These microfluidic devices are miniaturized liquid-handling systems that incorporate interconnected fluidic channels and chambers designed to perform one or several laboratory functions. The first lab-on-a-chip analysis system was a gas chromatograph developed in 1979. The advantages of lab-on-a-chip devices include small sample and reagent volume requirements and, therefore, faster reaction times because of small diffusion distances. Their portability and disposability make lab-on-a-chip devices popular and highly researched. The development of micropumps and flow sensors generating and controlling the microfluidics and integrated fluid treatments for analysis systems were key technical advances that laid the groundwork. These technologies, together with the development of soft lithography at Bell Laboratories, led to an increase in applicability and system development (Berthier et al., 2012). Research and commercial interest increased in the mid-1990s when lab-on-a-chip devices were applied in genomics tools such as capillary electrophoresis platforms and DNA microarrays. By 2016 the number of papers on lab-on-a-chip devices published and listed in Pubmed had increased to over 1200 per year. Fabrication of lab-on-a-chip devices was initially led by the microelectronics field, and the first microdevices were made of glass or silicon, restricting research to laboratories equipped with a specialized infrastructure. The most significant breakthrough in the production of microfluidic devices was the development of soft lithography and the use of polydimethylsiloxane (PDMS) as a material for rapid prototyping (Berthier et al., 2012). Ease of use and ability to fabricate devices without expensive equipment led to widespread adoption of lab-on-a-chip devices in research laboratories. The use of PDMS also allowed the testing of microscale devices in biological contexts, as the complexity of cells and cellular systems requires rapid iteration through many designs. These technologies, designed for fabricating lab-on-a-chip devices, are also fundamental in the production of MPS devices today. Similarly, other production methods, such as lithography, or fast replication methods, such as injection molding and hot embossing, were optimized for lab-on-a-chip systems and are now widely used for MPS. We go into further detail on MPS production techniques in the corresponding section. The emergence of automated laboratory routines in clinical and pharmaceutical environments led to a marked increase in throughput and accuracy. The underlying desire for fast and reliable generation of biological data is identical in MPS research; automation of assays and miniaturization of fluidic circuits helped to achieve this goal.

The purpose of microphysiological system

The purpose of microphysiological system and their suitability for automation Grant-driven efforts have promoted the development of multiple MPS. The purpose of an MPS varies by the benefits it is expected to provide and can include applications in physiology, pharmacology, and toxicology. MPS are optimized according to their intended purpose, which influences technological features such as the mode of pumping, the format of the device, and the biological complexity. The need for throughput, content, and automation also vary. Next, we discuss major technological characteristics of MPS that are important to automation. The basic screening tool in the modern laboratory is the microtiter plate. It has a standardized footprint that can be gripped by a robot and filled by automated liquid-handling machines. The throughput of an assay is vastly increased by running up to 1536 tests in parallel. Similarly, organ-on-a-chip devices for screening are often plate-based. The fluid is often perfused one-way through open channels by passive hydrostatic pressure (Vriend et al., 2018; Kim et al., 2015; Wevers et al., 2018). This single-pass microfluidic system enables a straightforward perfusion of tissues and retrieval of the sample by automated devices. These systems operate at small media volumes, as there are no external pumping devices. Hence, the medium-to-tissue ratio is close to the physiological ratio, and xenobiotics and metabolites do not become extensively diluted. Single-tissue cultures and multitissue cocultures perfused in a row have been reported (Lohasz et al., 2018). Various cell types from different organs can be integrated, and specific questions about these organs can be answered. Therefore plate-based organ-on-a-chip devices with passive medium perfusion are optimally suited for preclinical screening tests. High-throughput assays for target identification and lead selection and optimization are suitable for plate-based organ-on-a-chip devices as well, but the biological complexity is often limited to one type of cell culture, where the cells are encapsulated in a hydrogel or various spheroids in a row, because of the strict technical boundaries of plate-based systems. To avoid this limitation, passive hydrostatic perfusion can be combined with a custom-made microfluidic network (Oleaga et al., 2016; Chen et al., 2018; Nashimoto et al., 2017; Kim et al., 2013). Here, organ models are designed in a physiologically relevant arrangement and various cell culture formats can be integrated. Some systems are even compliant with physiologically based pharmacokinetic assessments (Miller and Shuler, 2016); these systems generate organ-specific microenvironments, rendering them more physiologically relevant, and are therefore of high value at the later stages of the drug-development pipeline in preclinical studies that evaluate the pharmacokinetic/pharmacodynamic profile and toxicity or efficacy of compounds. The ability to automate liquid handling with standardized machines, however, is lost because of the custom-made footprint. On-chip micropumps are useful for MPS targeting overall small media volumes. These active pumping devices often make use of valves actuated by

445

446

CHAPTER 14 Automation and opportunities for industry

FIGURE 14.2 Microphysiological system platforms can influence the drug development process predominantly at certain stages. Plate-based devices are especially valuable in early highthroughput screening assays for target identification and lead selection. Final high-content safety assessment and efficacy testing assays also fit well into custom-made single- and multiorgan chips. Supporting clinical trials with microphysiological system-based assays will only be possible using autologous, self-sustained patient-on-a-chip devices.

pneumatics (Schimek et al., 2013; Chen et al., 2017). Automated devices tune the fluid flow velocities and peak shear rates by setting the respective frequency or gas pressure by which the valves are actuated. Furthermore, the integration of additional valves in front of each organ compartment allows for an overall adjustment of media flow regimes. Complex multiorgan cocultures with a fluidic circuit adapting to tissue requirements are then possible, suitable for investigating complex questions during preclinical trials or even for supporting and refining clinical trials (Marx et al., 2016). Human- or body-on-a-chip platforms emulating wholebody homeostasis are envisioned, to allow for so-called patient-on-a-chip assays with patient-derived materials. The complexity of tissue setup, coculture, and analysis requires at least partial automation to increase accuracy and precision. In summary, a variety of MPS are available for applications in basic research or drug development. The throughput, content, and automation needs vary by purpose, but integrating automated processes improves precision and repeatability in all systems (Fig. 14.2). The following sections discuss automation of device production, integration of biology, and operation.

Automation of device production The automation of MPS device production is a critical factor in commercial use of the systems. Both quantity and quality of the devices produced increases significantly with automation. The production tolerance of an automated procedure is typically much lower than that of a manual one. The choice of materials sets strict boundaries on the production methods that can be used and, subsequently, also influences the degree of automation that can

Automation of device production

be achieved. The main prerequisite in selecting a material is its biocompatibility, but oxygen and carbon dioxide permeation, optical transparency, softness, and affinity for unselective absorption of substances may also be factors from a biological standpoint. From a technical point of view, factors such as the availability of the material, general ease of processing, and scalability of the processing technology are important. Polymer materials have been used extensively in traditional cell culture ware and have both good cell compatibility and potential for large-scale industrial production. Most data on classical in vitro cell behavior have been derived from cells cultured on polystyrene surfaces. Polymer materials with distinct chemical, mechanical, electrical, and optical properties are now available. Similarly, techniques for surface modifications and all necessary chemistries are already in place. Hence, polymer materials are suitable for use in MPS devices. Improvements in polymer microfabrication technologies also enable the use of these materials in devices with small geometrical features. Techniques used to process polymer materials range from photodefinable technologies and various replication methods to material etching and machining (Becker and Ga¨rtner, 2008). Some materials and technologies offer advantages in producing prototype devices, whereas others are especially suitable for large-scale commercial production. Next, we detail material characteristics, production methods, and replication accuracy and scalability. The range of polymeric materials, their properties, and processibility by various techniques is vast. These materials can be grouped roughly by their response to heat and the curing process. The glass transition temperature (Tg) is the temperature more than which the material behaves like a flexible, rubbery substance as larger segments of the polymer chain begin to move more freely. The material’s molecules have relatively little mobility under Tg, and the physical properties are those of a glassy or crystalline state. The first group, thermosetting materials, has a high Tg that is close to the decomposition temperature. Thermosetting materials contain polymers that cross-link irreversibly during the curing process, forming chemical bonds. Curing is caused by the action of heat or radiation and may be promoted by high pressure or the use of a catalyst. After curing, the polymer cannot be reshaped by further application of heat; it decomposes or burns instead. Therefore thermosetting materials are especially popular in applications requiring a high temperature stability and dimensional accuracy, such as the production of master molds. Prominent examples of thermosets used in MPS production include photoresists such as SU-8 (Nashimoto et al., 2017; Esch et al., 2012). These resins are processed mainly by lithographic technologies (photolithography or stereolithography) with limited throughput and high cycle times. Because changing the design is easy, these technologies are especially useful in prototyping, where no elaborate master molds are required. There are ways of automating lithographic technologies to make the devices more robust and slightly increase device production (Carve and Wlodkowic, 2018), but these technologies are not designed for high-throughput production, making them applicable for either proof-of-concept studies using prototypes or more complex systems aiming for low-throughput and high-content analyses.

447

448

CHAPTER 14 Automation and opportunities for industry

The second group, thermoplastic polymers, can be remolded and recycled by reheating, as the curing process is completely reversible and does not involve chemical bonding. These materials soften at Tg and exhibit a large difference between Tg and the decomposition temperature. Examples include polystyrene, poly(methyl methacrylate), polycarbonate, cyclic olefin copolymer, and cyclic olefin polymer. The design and production of industrial-scale lab-on-a-chip devices, which are also mainly produced from thermoplastic materials, have greatly informed thermoplastic polymer techniques (Kong et al., 2016). Thermoplastic polymers can be used to provide the highest throughput in replication methods such as injection molding or injection compression molding. These technologies are also employed in the production of traditional cell culture ware and ensure a high replication accuracy and low cost in high-volume production. The production processes, however, are highly complex and require expensive and intricate molding tools, especially for devices such as MPS. Technologies such as injection molding are, therefore, especially useful in tried and tested systems requiring no further adaptation to system design, such as the plate-based systems used for screening (Lohasz et al., 2018). Other commercially applicable replication methods for thermoplastic materials, including microthermoforming or hot embossing, have lower requirements for master structures but also lower cycle times. These technologies are useful for medium-throughput device production and prototype design. Finally, micromachining requires the lowest cycle times and no master structures (Edington et al., 2018). These production procedures for most thermoplastic materials can be upscaled relatively easily from a micromachining or hot embossing technology to injection molding at later stages, requiring no or only small adaptations in material choice and system design. The general requirements for special equipment and machines to process thermoplastics, however, often impede the use of these materials in an academic environment. The gap between the methods and materials used in academia and in commercial industry hampers the transition of prototypes developed in academia to commercial products. These upscaling and transformation are especially difficult because of chemistries and physical parameters demanding a thorough redesign of the MPS prototype. The third group, elastomers, consists of viscoelastic polymers with long molecular chains that are physically entangled. These materials are maintained above their Tg, enabling the long chains of the materials to reconfigure themselves to distribute an applied stress, making them relatively soft and deformable. The covalent cross-linkages between the chains ensure that the elastomer will return to its original configuration after the stress is removed. Elastomers are normally thermosets but may also be thermoplastic. Elastomeric materials are generally needed in all MPS with integrated pneumatic valves or on-chip micropumps (Edington et al., 2018; Wagner et al., 2013; Coppeta et al., 2016). PDMS is the prime example of an elastomer in MPS: The replication method most often used for PDMS is elastomer casting. It has limited throughput but requires little technical equipment, explaining why PDMS is used extensively

Automation of tissue preculture and loading

in the academic environment. Other examples include synthetic rubber films that are made of fluoropolymer or polyurethane (Edington et al., 2018; Coppeta et al., 2016). These films are often processed by laser- or die-cutting. Back-end processing steps, including the integration of electrodes, surface modifications, or the bonding of the layers to form a fluidic-tight device, should also be considered when selecting materials for the device. This is especially important for MPS-integrating on-board micropumps and for systems requiring various fluidic layers or a combination of materials with, for example, different levels of flexibility or gas permeability. Automating system production can reduce the failure rate and enable future commercialization and industrial application. Material bonding, in particular, demands high-quality results to avoid bacterial contamination and ensures a fluidic tightness for prolonged culture periods, eventually even under periodic internal pressure from fluidic pumping. Technologies for bonding include adhesives, thermal pressure bonding, solvent bonding, and laser or ultrasonic welding. Processes requiring the addition of adhesives or solvents must take into account leachables and biological compatibility, while processes applying heat should adapt process parameters so that they do not damage the microfluidic network. A review highlighting various back-end processing technologies can be found elsewhere (Iannone, 2018). In short, several factors regarding the choice of material and production technique should be considered during the early phase of MPS device design. The MPS community has so far mostly elected to use polymer materials for device production because of their known and validated characteristics. The range of polymer materials, however, is vast, and the choice of material cannot be made solely by biological aspects but must also consider production accuracy, throughput, and back-end processing.

Automation of tissue preculture and loading The biological complexity of tissue models integrated into MPS varies by application, as noted above, but standardized culture procedures are a requirement for reproducible and well-characterized assays. The availability of cells of an unchanged quality, culturing under consistent conditions, and a well-defined integration of existing models into the MPS are major concerns. To enable industrial use, cell sources should be expandable to inoculate a number of devices. A high physiological relevance is desirable to provide comparability to the human situation; preclinical high-throughput screening assays often make use of established cell lines even though they may have only limited predictive power for human outcome because of their dedifferentiated or tumorigenic nature. Furthermore, one-third of all cell cultures are thought to have undergone inter- and intraspecies contamination (Hughes et al., 2007). Genetic instability (Hughes et al., 2007; Kleensang et al., 2016) and mycoplasma contamination

449

450

CHAPTER 14 Automation and opportunities for industry

(Olarerin-George and Hogenesch, 2015) also interfere with the reliability of cell lines, but the availability of the more predictive model—primary adult cells—is limited, and comparability between isolations and donors is not always guaranteed. Induced pluripotent stem cells and their differentiated progeny may hold the answer. Differentiation protocols have been established for bioreactor cultures to provide large numbers of cells (Rebelo et al., 2016; Koenig et al., 2018). These semiautomated cultures are tightly controlled and thus suited to deliver cells of consistent quality. The use of induced pluripotent stem cells is advantageous for modeling diseases and, ultimately, patients in so-called patient-on-a-chip devices (Marx et al., 2016). Integrating tissue models can be divided into on-chip cell differentiation and the external establishment and subsequent inoculation of MPS with ready-made cultures. Automated pipetting easily integrates single-cell suspensions into the devices, but the on-chip differentiation of cells requires specialized media. Therefore a simultaneous differentiation of various tissue models in one system requires the incorporation of valves or separate media circuits that can be connected after differentiation is completed (Kleensang et al., 2016; Olarerin-George and Hogenesch 2015). In contrast the external establishment of organ models can leverage existing protocols and cell culture ware. Nonetheless, the automated loading of ready-made models in a standardized manner requires specialized equipment and routines. MPS with a standardized microtiter footprint can use liquid-handling robots for loading, employing common pipette tips or wide-bore tips (e.g., for loading spheroid models). The complexity of tissue models influences the options available. Models composed of cells that are randomly distributed in a hydrogel (Wevers et al., 2018; Oleaga et al., 2016) or of self-assembled spheroids (Kim et al., 2015) can easily be formed inside an MPS. Models with predefined structures, such as the multilayers of skin, or that demand long-term differentiation protocols are predestined for the external establishment of the cell culture. Three-dimensional printing allows some degree of automation in setting up these models (Koenig et al., 2018). Forceps or tweezers are used in manual device handling for the transport of these models into the systems. Gripper concepts must be designed for an automated loading of tissues. Ideally, the gripper can be adjusted to various formats of tissue carriers, such as cell culture inserts or the ceramics used as nesting structures for bone marrow cells.

Automation of system operation When analyzing MPS device suitability for automated culture operation, it is essential to define the degree of automation. The establishment of appropriate culture conditions and the execution of the assay-handling steps may be partially automated with intermittent manual handling steps or fully automated. The degree

Automation of system operation

of automation is defined as the “proportion of automatic functions to the entire set of functions of a system or plant” (DIN IEC 60050-351, 2013), in contrast with those functions that are conducted by a human operator. To the best of our knowledge, there have been very few mainly automatic MPS maintenance systems reported so far. Automation solutions for parts of the system-operation and assay execution processes do exist: Some are dedicated developments for device maintenance (e.g., pump-control units or tilting stages for hydrostatic pressure flow control), and others are standard laboratory equipment, such as automated pipettes and robotic liquid-handling systems. Both solutions make use of standard manual or semiautomatic laboratory equipment, including incubation chambers, safety benches, microscopes, and refrigerators. The adaption of MPS platforms to standard laboratory equipment is advantageous, as most hardware already exists in laboratories, and the staff is trained in their use. Furthermore, data are highly comparable to those derived from standard static-cell cultures. The MPS research groups focusing on adapting to standard laboratory equipment typically design their systems in accordance with standardized well plate formats, such as 96- or 384-well microtiter plates, which can be used with most microscopes, liquid handlers, and analysis tools. One major challenge of automating MPS will be interconnecting all laboratory systems; this is currently achievable only by a human operator. Because some of the tasks are highly laborious or repetitive, the final aim of MPS automation should be to minimize the need for hands-on operator time.

Medium circuit flow Unlike standard cell cultures, MPS move the medium continuously through organ compartments. This ensures a physiologically relevant organ cross-talk, a constant supply of oxygen and nutrition to the organs, and shear stress on the cells. There are three main ways to propel medium circulation. Most MPS with active on-chip pumps use a peristaltic pump, which works pneumatically (Chen et al., 2017; Loskill et al., 2015). Pneumatic pressure or vacuum is applied to three membranes that deflect and deliver mechanical pressure on the liquid medium in three chambers. Actuating all three membranes in a specific rhythm generates a pulsatile fluid flow. The explicit advantage of employing pneumatics is that actuation is fast and easily controlled by electrically actuated valves. Flow characteristics can be adjusted and allow for flow curves similar to those of cardiac frequency (Schimek et al., 2013). The medium flow is typically unidirectional. Regarding automation, one disadvantage of pneumatics is the large number of tubes required for connecting the devices to the flow controller (Maschmeyer et al., 2015). This renders the transfer of systems between laboratory appliances difficult and elaborate, as all tubes must be disconnected for liquid-handling procedures in a safety cabinet or for microscopic analysis. There is also a limit to tube length because of loss of pressure. Electrically controlled on-chip pump systems have a slight advantage (Coppeta et al., 2016), as they make use of

451

452

CHAPTER 14 Automation and opportunities for industry

controlled electromagnetic pumps that utilize the same peristaltic pumping principle but employ electric cables, which have a smaller profile and are easier to plug in and remove. The second type of MPS perfusion system is passive fluid flow devices (Lohasz et al., 2018; Wevers et al., 2018; Miller and Shuler, 2016), where the organ models are arranged in line with medium reservoirs. When the system is tilted to one side, medium flows from the upper medium reservoir through all organ compartments to the lower compartment, driven by hydrostatic pressure and creating a bidirectional flow. The fluid flow is typically generated by electrically driven rockers or tilting stages, which can adjust the flow rate by varying the tilting angle, time, and frequency. The main advantage of this type of perfusion is that no pneumatic connections are required, and the tilting stage and the device itself are considerably less complex, resulting in reduced costs and production time. This setup, however, still requires the interruption of the tilting process during liquid-handling or optical measurement procedures. The third type of perfusion does not have pumps integrated into the chip. The channels are connected to an external pump that is typically a syringe pump (Hughes et al., 2007; Grix et al., 2018; DIN IEC 60050-351, 2013). While resulting in smaller MPS devices and featuring commercially available pumps, the medium flow rarely matches the physiological tissue-to-blood ratio, as a high medium volume is required. This type of MPS poses similar but even more complex automation problems as systems with active on-chip pumps, since mediumfilled tubes must be connected to the devices. Automated medium circuit control in the form of programmable pneumatic valves, fluid pumps, or tilting stages is already available for many MPS. A remaining challenge is the development of automated device unplugging or extraction from the flow controllers to allow for medium exchange, optical inspection, and other manipulations.

Oxygen supply One essential advantage of MPS over static cell cultures is the enhanced oxygen supply, especially in three-dimensional organ models. A constant oxygen-rich medium flow ideally surrounds each model. Most devices expose medium or organ models to ambient air to ensure that the medium is not deprived of oxygen over time (Chen et al., 2017; Materne et al., 2015). Other devices incorporate PDMS that is gas-permeable and allows for exchange through the material into the microfluidic layer (Lohasz et al., 2018; Olarerin-George and Hogenesch 2015). Those systems are typically opened intermittently for medium exchange and periodic gas exchange. Barrier organ models, which rely on air contact in the physiological context, may require more controlled conditions, such as gas supply. Oxygen perfusion in current MPS is generally enabled by passive diffusion. Automated MPS must ensure contact to ambient air.

Automation of system operation

Incubation and pH-stabilization Systems must be kept at body temperature (approximately 37  C) or as needed to model a disease (36 C 41  C). Most reported MPS cultures are maintained in standard laboratory incubation chambers, which also provide a carbon dioxideenriched atmosphere that allows adequate pH in the medium. However, keeping MPS in an enclosed chamber poses a problem for automation. While the medium circuit and nutrition supply can be provided via tubing from outside the chamber, substance application, sample extraction, and analytical procedures typically require the removal of the device from the incubation chamber. Systems that automatically access the devices within an incubation chamber, preferably without disturbing culture conditions, are still lacking for the various MPS platforms. There are automated incubators on the market that enable the automatic extraction of well plate-sized MPS (Kane et al., 2019), but providing tilting or plugging and unplugging of tubes and cables inside those incubators appears difficult to implement. MPS automation entails the development of a complex extraction process of devices from incubators for liquid handling or analysis. To avoid interrupting constant cell-culture conditions, MPS cultivation outside incubation chambers is most promising.

Minimizing contamination and evaporation In MPS with small liquid volumes, medium and substance evaporation poses a considerable problem. Furthermore, preventing contamination is crucial because MPS so far do not include a functional immune system. Thus most systems can be closed with a lid and are only manually opened for liquid-handling procedures under sterile conditions in a safety cabinet. Standard multiwell plate lids are used, for example, to shield organ compartments over the whole device (Lohasz et al., 2018; Edington et al., 2018). Polyester films with small holes for oxygen exchange are also used to seal organ compartments and medium reservoirs to avoid evaporation. In addition, a plate filled with phosphate-buffered saline can be placed under each device to increase humidity (Lohasz et al., 2018). Custommade screw-on lids for each organ compartment can minimize evaporation (Schimek et al., 2013). Medium changes in all current MPS are conducted in a safety cabinet at room temperature. Access to MPS with lids made of pierceable material should place the least obstacle to automation. Liquid-handling procedures can be conducted manually or with automated handlers for MPS with the appropriate format and positions of organ compartments. Common multiwell plate lids are more difficult to open automatically, but automatic delidders are already used in common liquid-handling robots and in automatic MPS cultures in a sterile enclosure (Mastrangeli et al., 2019). Lids may require a customized opening solution, such as a special gripper.

453

454

CHAPTER 14 Automation and opportunities for industry

Most MPS devices known to the authors require transport between the incubation chamber and the safety cabinet. Standard safety cabinets provide limited workspace in depth and height because of high-efficiency particulate air filters and are designed for use by a human operator. Thus liquid handling in MPS under sterile conditions may require customized, larger safety cabinets to accommodate automated liquid handlers and automatic sterility gates.

Executing handling steps The most laborious and time-consuming process in MPS handling is probably medium exchange, which should ideally be conducted as often as a human ingests nutrients. This is why some systems rely on medium reservoirs, which can sustain the culture for a longer period (Huh et al., 2010), but these systems do not meet the fluid-to-cell ratio of cellular products, impeding organ organ cross-talk or metabolite analyses. Therefore most MPS research groups omit the use of larger medium reservoirs. A maximum of two medium changes per day on business days is typical because of limited personnel resources and working hours. More frequent medium changes will require automated procedures, and some groups are adapting their MPS to standard laboratory hardware, generally using a well plate sized format (Lohasz et al., 2018). Approximately 4% of MPS devices reported in the literature are microtiter plate sized (Hegde et al., 2014). These MPS can generally be used in multichannel pipetting stations, robotic liquidhandlers, microscopes, and standard analytical systems; for liquid handling, however, the devices must be disconnected from pneumatic and electrical connections or the tilting stage, removed from the incubation chamber, and placed under sterile conditions. In most cases a lid must be opened. Dedicated solutions for accessing the medium circuit or reservoirs must be developed for MPS that are not microtiter plate sized. Devices featuring active pumping may integrate an additional microfluidic feeding circuit on the chips harboring a separate pump or a valve that can be actuated automatically, enabling a highly customizable feeding strategy (Domansky et al., 2010). From a technical standpoint, the increase in device complexity outweighs the benefits of a reduction in liquid-handling steps. From a biological standpoint an additional means to excrete metabolic “waste” products at a rate similar to that of fresh-nutrient feeding must be considered (Domansky et al., 2010). Medium extracted from MPS is often kept for further analysis and analyzed immediately or stored in appropriate containers for later analysis. Automated well plate storage solutions are commercially available and may be used if the systems can be operated with robotic liquid handlers. The same limitations as those for medium exchange still apply. The most common substances applied to organ models are medium solvents. As many solvents are volatile, they are often prepared just prior to application. Automated preparation and mixing can be carried out by liquid-handling robots, but the mixture must be applied to the MPS rapidly. Substances may be applied

Automation of system operation

directly to the organ or to the surrounding medium. The application procedure is similar to the medium exchange process and poses the same challenges to automation. Some assays may also require different modes of application, such as the application of high-viscosity creams to skin or exposure of the lung to aerosols; dedicated solutions will need to be developed for those cases. Liquid-handling procedures, such as medium exchange, sampling, and solvent application, can already be performed automatically on plate-based MPS with liquid-handling robots or multichannel pipettes. However, cell culture conditions for those procedures must be interrupted to disconnect the devices from pumping or tilting, remove them from the incubation chamber, and transport them to a safety cabinet for sterility. Regarding MPS with different footprint dedicated solutions have to be developed.

General observation of constant cell culture conditions As mentioned, most available MPS require incubation, liquid-handling procedures, and optical measurements at various locations within the laboratory. Recently, a mainly automated device was reported that features automatic incubation, media changes, cell seeding, and microscopy within a sterile Class I enclosure (Mastrangeli et al., 2019), although medium perfusion and sample storage were not shown. Furthermore, incubation must be interrupted during media changes, but microscopic analysis can be conducted in a separate incubation chamber. All other known systems, even those with a plate-based format compatible with liquid-handling robots and multichannel pipettes, require manual transportation between handling stations. Automating the transport of MPS between the incubator and the safety bench creates a challenge, as all material have to remain sterile, and possible transport ways may restrict accessibility and walkways in the laboratory. In addition, each manipulation interrupts constant culture conditions and renders the testing environment less physiological. Refined feeding regimes, more frequent sample extraction, and optical analyses add more layers of complexity. Thus fully automated MPS will require components that ensure constant incubation, medium flow, stable pH, and sterility throughout the duration of the assay while simultaneously allowing medium exchange, sampling, substance application, and tissue model-loading. To reduce manual handling, an automatic MPS must also incorporate automatic provision of the cell-culture material, medium, and other substances and sample storage (Fig. 14.3). A system fulfilling all these requirements is currently tested (Fo¨rderkatalog, 2016): Here, a robotic manipulator establishes the required constant culture conditions for the MPS developed by the Marx Laboratory at Technische Universita¨t Berlin (Becker and Ga¨rtner, 2008; Olarerin-George and Hogenesch, 2015). Medium exchange, sampling, substance application, loading, and microscopic analysis can be conducted fully autonomously; an automatic cold-storage system enables the provisioning to the systems

455

456

CHAPTER 14 Automation and opportunities for industry

FIGURE 14.3 Schematic of benefits provided by an automated MPS culture. Examples of MPS devices are provided courtesy of TissUse GmbH. MPS, Microphysiological system.

of cell culture media and substances for at least 4 days, as well as ample storage for analysis.

Automation of monitoring and sensing The significance of MPS-based experiments can increase with the use of appropriate methods for measuring and analyzing system parameters (Huh et al., 2010), especially for MPS that allow a closed-loop control or a dynamic change of system parameters based on, for example, real-time readouts. This is where an automated system displays one of its greatest strengths: to enable the automatic monitoring, analyzing, and potential intervention in the experiment executed or the operation itself. Most methods for MPS analysis have been described but not yet been made suitable for automation, indicating that the next challenge is to standardize and digitize the methods and combine them into an integrated system. The methods can be divided into online methods that can be applied directly to the operating MPS with real-time observations or offline methods that are realized in parallel or after the experiment with time-displaced observations. This section focuses on online methods, because most offline methods can be realized by simply collecting and storing samples in microtiter plates. The morphologic analysis of the organoids inside the MPS is the most obvious and common method. It can be conducted using bright-field, phase-contrast, fluorescence, or confocal microscopy. All these methods have been well-described for manual processes (Junaid et al., 2017). Where image quality, brightness, and magnification remain constant during image acquisition, all of these methods can be

Automation of monitoring and sensing

integrated into an automatic system. Automatic image-analysis software is crucial for managing the large volume of data that the optical system generates. Parameters to monitor can include the size, number, position, roundness, or surface roughness of cells or spheroidal tissue models. Automatic morphologic analysis can also be used to distinguish between organoids of different cell types (Marx et al., 2012). Furthermore, the use of an automatic monitoring system permits complex image analysis, potentially enabling the results to influence subsequent image acquisition. Peel et al. have shown how an algorithm searches for landmarks in a liver-on-a-chip device before higher magnification images are taken (Fo¨rderkatalog, 2016). The structure of on-chip microvasculature is another field that can be analyzed using optical technologies such as multiphoton microscopy (Schimek et al., 2013) or the speckle variance detected with an optical coherence tomography system (Sung et al., 2019; Kasendra et al., 2018). These methods are valuable in analyzing the biological aspects of MPS. From a technical standpoint, it is worthwhile to integrate methods that provide information on the flow through the microchannels. The flow-through capillaries or microfluidic channels throughout the device can be analyzed by optical coherence angiography, allowing the study of nutrient supply and shear stress on the cells (Thorlabs, 2019). Unlike microfluidic flow sensors, this technique can measure the fluid flow and its distribution in capillaries and is completely noninvasive. Systems featuring active medium circulation may receive feedback from optical coherence angiography and adapt their pumping strategy accordingly. A second group of analytical methods is used to gather information on substance concentration and distribution. Fluorescence staining is the most common method but is invasive and can affect cellular behavior during live tissue monitoring or even destroy the tissue during endpoint immunohistology. Hence, alternative methods should be developed. In addition to optical coherence tomography, Raman spectroscopy and visible light and near-infrared spectroscopy are especially worth mentioning. Raman spectroscopy can detect and localize substances and determine their concentration and size (Vankeirsbilck et al., 2002). This also applies to ultraviolet, visible light, and near-infrared spectroscopy, which can even sense and image oxygen (Mariampillai et al., 2008). Apart from optical methods, electrical coupling is of particular interest for barrier organs and muscle or nerve cells and has been widely adopted in custommade MPS devices. Transepithelial electrical resistance and electrical activity can be measured, and the tissue can be stimulated by electrodes integrated into the devices (Srinivasan et al., 2010). The difficulty in integrating the electrodes into the devices is balanced by the easy-to-process electrical readouts with high added value. It should be noted, however, that systems and their integrated sensors should be consumable, whereby interfaces for sensor readout and system control are not. This imposes demands on the accuracy and precision of interfaces for uninterrupted and long-term measurements. In short, adequate sensing and monitoring techniques greatly enhance the content and significance of MPS assays. The automation and integration of manually

457

458

CHAPTER 14 Automation and opportunities for industry

performed optical and electrical monitoring must overcome the challenges of missing interfaces and the need for integrated data-handling and analysis. Realtime monitoring of complex biological reactions in MPS may revolutionize in vitro studies and enhance control over and confidence in the assays.

Summary and outlook The automation of MPS is still in its infancy and, so far, only a few devices have been developed that feature aspects of large-scale industrial production or automated operation and analysis. Devices that evolved in academia and focus more on complexity than on throughput face several hurdles in transitioning to an industrial setting. First, the choice of materials sets strict boundaries on the production methods. Technologies available in academia are traditionally different from those that can be applied for larger scale production. Second, an increasing degree of biological complexity in the organ models themselves and multiorgan cocultures favors an external organ model setup. Reproducible and accurate insertion of tissue models into MPS must be established. Third, the overall layout of the MPS (e.g., footprint, mode of pumping, and means of incubation) influences the ease of automation. Multiwell plate sized systems can make use of standard laboratory automation tools, whereas custom-made devices require custom-made solutions for automation. Nonetheless, the inclusion of all standard laboratory peripherals as integral components of an automated MPS is a standing challenge. Last, the integration of automated sensing and monitoring systems will greatly increase real-time readout possibilities, enhancing the control of and confidence in MPS assays. Most technologies for the automated operation of MPS can be adopted from existing systems. The challenge is to combine these existing technologies and adapt them to the specific application. However, since the existing technologies are often only standardized to a small extent, the adaptation or combination of these involves a large development effort. The first step should therefore be to adopt existing industry standards for hardware and software components. The Standardization in Laboratory Automation 2 standard could be adopted, for example, to define the interface between laboratory devices (Thorlabs, 2019). In the next step, other standards could be developed to accelerate the development of MPS and increase their reliability and comparability. Standards for measurement accuracy or control variables for analysis and materials could be created. In the long run, automation of MPS will lead to higher accuracy and precision in assays and will facilitate intra- and interlab reproducibility. Both highthroughput assays and high-content analyses will profit from standardization. A certain degree of automation will be indispensable, especially when opting for the most complex human- and patient-on-a-chip devices. Continued industrial

References

adoption of current systems and enhanced cooperation of MPS manufacturers with end-users will gradually result in an increase of automation.

References Becker, H., Ga¨rtner, C., 2008. Polymer microfabrication technologies for microfluidic systems. Anal. Bioanal. Chem. 390, 89 111. Berthier, E., Young, E.W.K., Beebe, D., 2012. Engineers are from PDMS-land, biologists are from polystyrenia. Lab Chip 12, 1224 1237. Boyd, J., 2002. Robotic laboratroy automation. Science 295 (80), 517 518. Carve, M., Wlodkowic, D., 2018. 3D-printed chips: compatibility of additive manufacturing photopolymeric substrata with biological applications. Micromachines 9, 91. Chapman, T., 2003. Automation on the move. Nature 421, 661 666. Chen, W.L.K., et al., 2017. Integrated gut/liver microphysiological systems elucidates inflammatory inter-tissue crosstalk. Biotechnol. Bioeng. 114, 2648 2659. Chen, H.J., Miller, P., Shuler, M.L., 2018. A pumpless body-on-a-chip model using a primary culture of human intestinal cells and a 3D culture of liver cells. Lab Chip 18, 2036 2046. Coppeta, J.R., Mescher, M.J., Isenberg, B.C., Spencer, A.J., Borenstein, J.T., 2016. A portable and reconfigurable multi-organ platform for drug development with onboard microfluidic flow control. Lab Chip 17, 134 144. DIN IEC 60050-351, 2013. International Electrotechnical Vocabulary—Part 351: Control Technology (IEC 60050-351:2013) Domansky, K., et al., 2010. Perfused multiwell plate for 3D liver tissue engineering. Lab Chip 10, 51 58. Edington, C.D., et al., 2018. Interconnected microphysiological systems for quantitative biology and pharmacology studies. Sci. Rep. 8, 1 18. Esch, M.B., et al., 2012. On chip porous polymer membranes for integration of gastrointestinal tract epithelium with microfluidic ‘body-on-a-chip’ devices. Biomed. Microdevices 14, 895 906. Fo¨rderkatalog, 2016. Alternativmethoden—Verbund: HOC—Etablierung und Tauglichkeitstestung einer Pilotversuchsanlage fu¨r den Einsatz von ‘Human on a chip’ Roboter-Prototypen zur aussagekra¨ftigen Testung beliebiger Substanzen im Ersatz zu Tierversuchsanlagen. Available from: ,https://foerderportal.bund.de/foekat/jsp/ SucheAction.do?actionMode 5 view&fkz 5 031L0099A. Grix, T., et al., 2018. Bioprinting perfusion-enabled liver equivalents for advanced organon-a-chip applications. Genes (Basel) 9, 1 15. Groover, M.P., 2010. Fundamentals of Modern Manufacturing: Materials, Processes, and Systems. John Wiley & Sons, Inc. Hegde, M., et al., 2014. Dynamic interplay of flow and collagen stabilizes primary hepatocytes culture in a microfluidic platform. Lab Chip 14, 2033 2039. Hughes, P., Marshall, D., Reid, Y., Parkes, H., Gelber, C., 2007. The costs of using unauthenticated, over-passaged cell lines: how much more data do we need? Biotechniques 43, 575 586. Hughes, J.P., Rees, S.S., Kalindjian, S.B., Philpott, K.L., 2011. Principles of early drug discovery. Br. J. Pharmacol. 162, 1239 1249.

459

460

CHAPTER 14 Automation and opportunities for industry

Huh, D., et al., 2010. Reconstituting organ-level lung functions on a chip. Science 328 (80), 1662 1668. Iannone, E., 2018. Labs on Chip: Principles, Design and Technology. CRC Press. Junaid, A., Mashaghi, A., Hankemeier, T., Vulto, P., 2017. An end-user perspective on organ-on-a-chip: assays and usability aspects. Curr. Opin. Biomed. Eng. 1, 15 22. Kane, K.I.W., et al., 2019. Automated microfluidic cell culture of stem cell derived dopaminergic neurons. Sci. Rep. 9. Kasendra, M., et al., 2018. Development of a primary human small intestine-on-a-chip using biopsy-derived organoids. Sci. Rep. 8, 1 14. Kim, S., Lee, H., Chung, M., Jeon, N.L., 2013. Engineering of functional, perfusable 3D microvascular networks on a chip. Lab Chip 13, 1489 1500. Kim, J.Y., Fluri, D.A., Kelm, J.M., Hierlemann, A., Frey, O., 2015. 96-Well format-based microfluidic platform for parallel interconnection of multiple multicellular spheroids. J. Lab. Autom. 20, 274 282. Kleensang, A., et al., 2016. Genetic variability in a frozen batch of MCF-7 cells invisible in routine authentication affecting cell function. Sci. Rep. 6, 1 10. Koenig, L., Ramme, A.P., Faust, D., Lauster, R., Marx, U., 2018. Production of human induced pluripotent stem cell-derived cortical neurospheres in the DASbox ® mini bioreactor system. Appl. Note 364, 1 12. Kong, L.X., Perebikovsky, A., Moebius, J., Kulinsky, L., Madou, M., 2016. Lab-on-a-CD: a fully integrated molecular diagnostic system. J. Lab. Autom. 21, 323 355. Lohasz, C., Rousset, N., Renggli, K., Hierlemann, A., Frey, O., 2018. Scalable, microfluidic platform for flexible configuration of and experiments with microtissue multi-organ models. SLAS Technol. 1 24. Available from: https://doi.org/10.1177/ 2472630318802582. Loskill, P., Marcus, S.G., Mathur, A., Reese, W.M., 2015. µOrgano: a Lego®-like plug & play system for modular multi-organ-chips. PLoS One 10, e0139587. Available from: https://doi.org/10.1371/journal.pone.0139587. Mariampillai, A., et al., 2008. Speckle variance detection of microvasculature using sweptsource optical coherence tomography. Opt. Lett. 33, 1530 1532. Marx, U., et al., 2012. Human-on-a-chip developments: a translational cutting-edge alternative to systemic safety assessment and efficiency evaluation of substances in laboratory animals and man. ATLA 40, 235 257. Marx, U., et al., 2016. Biology-inspired microphysiological system approaches to solve the prediction dilemma of substance testing. ALTEX 33, 272 321. Maschmeyer, I., et al., 2015. A four-organ-chip for interconnected long-term co-culture of human intestine, liver, skin and kidney equivalents. Lab Chip 15, 2688 2699. Mastrangeli, M., Millet, S., van den Eijnden-van Raaij, J., Organ-on-Chip in Development: Towards a roadmap for Organs-on-Chip. Preprints, 2019, 2019030031 Available from: https://doi.10.20944/preprints201903.0031.v1. Materne, E.-M., et al., 2015. A multi-organ chip co-culture of neurospheres and liver equivalents for long-term substance testing. J. Biotechnol. 205, 36 46. Mayr, L.M., Fuerst, P., 2008. The future of high-throughput screening. J. Biomol. Screen. 13, 443 448. Miller, P.G., Shuler, M.L., 2016. Design and demonstration of a pumpless 14 compartment microphysiological system. Biotechnol. Bioeng. 113, 2213 2227.

Further reading

Nashimoto, Y., et al., 2017. Integrating perfusable vascular networks with a threedimensional tissue in a microfluidic device. Integr. Biol. (United Kingdom) 9, 506 518. Olarerin-George, A.O., Hogenesch, J.B., 2015. Assessing the prevalence of mycoplasma contamination in cell culture via a survey of NCBI’s RNA-seq archive. Nucleic Acids Res. 43, 2535 2542. Oleaga, C., et al., 2016. Multi-Organ toxicity demonstration in a functional human in vitro system composed of four organs. Sci. Rep. 6, 1 17. Olsen, K., 2012. The first 110 years of laboratory automation: technologies, applications, and the creative scientist. J. Lab. Autom. 17, 469 480. Rebelo, S.P., et al., 2016. Validation of bioreactor and human-on-a-chip devices for chemical safety assessment. In: Eskes, C., Whelan, M. (Eds.), Validation of Alternative Methods for Toxicity Testing, 856. Springer. Sasaki, M., et al., 1998. Total laboratory automation in Japan: past, present and the future. Clin. Chim. Acta 278, 217 227. Schimek, K., et al., 2013. Integrating biological vasculature into a multi-organ-chip microsystem. Lab Chip 13, 3588 3598. Srinivasan, V.J., et al., 2010. Rapid volumetric angiography of cortical microvasculature with optical coherence tomography. Opt. Lett. 35, 43 45. Sung, J.H., et al., 2019. Recent advances in body-on-a-chip systems. Anal. Chem. 91, 330 351. Thorlabs, 2019. Novel Techniques for Measuring Capillary Blood Flow Using OCT. Available from: ,https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_ id 5 6484. Vankeirsbilck, T., et al., 2002. Applications of Raman spectroscopy in pharmaceutical analysis. TrAC, Trends Anal. Chem. 21, 869 877. Vriend, J., et al., 2018. Screening of drug-transporter interactions in a 3D microfluidic renal proximal tubule on a chip. AAPS J. 20, 87. Wagner, I., et al., 2013. A dynamic multi-organ-chip for long-term cultivation and substance testing proven by 3D human liver and skin tissue co-culture. Lab Chip 13, 3538 3547. Wang, Y.I., Carmona, C., Hickman, J.J., Shuler, M.L., 2018. Multiorgan microphysiological systems for drug development: strategies, advances, and challenges. Adv. Healthc. Mater. 7, 1 29. Wevers, N.R., et al., 2018. A perfused human blood brain barrier on-a-chip for highthroughput assessment of barrier function and antibody transport. Fluids Barriers CNS 15, 23.

Further reading Ba¨r, H., Hochstrasser, R., Papenfuß, B., 2012. SiLA: basic standards for rapid integration in laboratory automation. J. Lab. Autom. 17, 86 95. Borten, M.A., Bajikar, S.S., Sasaki, N., Clevers, H., Janes, K.A., 2018. Automated brightfield morphometry of 3D organoid populations by OrganoSeg. Sci. Rep. 1 10. Available from: https://doi.org/10.1038/s41598-017-18815-8.

461

462

CHAPTER 14 Automation and opportunities for industry

Kim, H.J., Huh, D., Hamilton, G., Ingber, D.E., Links, D.A., 2012. Human gut-on-a-chip inhabited by microbial flora that experiences intestinal peristalsis-like motions and flow. Lab Chip 12, 2165 2174. Maoz, B.M., et al., 2017. Organs-on-chips with combined multi-electrode array and transepithelial electrical resistance measurement capabilities. Lab Chip 17, 2294 2302. Peel, S., et al., 2018. Introducing an automated high content confocal imaging approach for organs-on-chips. Lab Chip . Available from: https://doi.org/10.1039/C8LC00829A. Wang, X., Wolfbeis, O.S., 2014. Optical methods for sensing and imaging oxygen: materials, spectroscopies and applications. Chem. Soc. Rev. 43, 3666 3761.

CHAPTER

How to build your multiorgan-on-a-chip system: a case study

15 David Bovard and Antonin Sandoz

PMI R&D, Philip Morris Products S.A., Neuchaˆtel, Switzerland

Introduction Organ-on-a-chip (OOC) and multiorgan-on-a-chip (MOC) platforms combine an engineered microenvironment with one or several cellular models. This combination has already been shown to improve the characteristics of the biological model that resembles the human organ in vivo more closely than systems comprising only a cellular model under static conditions. Given the potential of this technology to improve in vitro models, it is tempting for laboratories, both in academia and industry, to develop “homemade” chip platforms suitable for specific testing needs. While some of the currently existing chips can have a very simple design, they all required intensive efforts to conceptualize, prototype, test, adapt, and retest. In most situations, several prototypes were constructed before a final version met the requirements successfully and was advanced to production. This chapter provides a guide through the thought process behind the development of an OOC platform to help laboratories construct their own fit-for-purpose chip. This guide largely mirrors the rationale and developmental aspects used to develop a functional lung/liver-on-a-chip system in our laboratory (Bovard et al., 2018), which will be used as a case study to assist readers in developing their chip. The development of OOC platforms requires a specialized research and development team with at least two important groups of specialists: engineers and biologists or chemists (Fig. 15.1). Because the development of in vitro models requires extensive knowledge of good cell culture practices under aseptic conditions, biologists are needed to develop a chip that is relevant and suitable for housing cellular culture models. As the development of a chip requires extensive knowledge of engineering (e.g., to design the chip, select the appropriate materials, and integrate pumps and sensors), engineers are needed to develop the chip platform. The basis for the success of the project will, therefore, be the quality of interaction and communication between the two groups (Fig. 15.1). The level and extent of interaction will depend on the planned project and its complexity. We recommend dividing the chip development into small steps, each Organ on a Chip. DOI: https://doi.org/10.1016/B978-0-12-817202-5.00015-2 © 2020 Elsevier Inc. All rights reserved.

463

464

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.1 Chip development requires good communication between engineers and biologists or chemists. Both fields of expertise can include individuals with varying specialties.

with a defined milestone. This concept, as well as the influence of agile project development methods, is explored in the present chapter, which is divided into three parts. The first part describes the development of the chip from a biological perspective, the second describes it from an engineering perspective, and the third part offers guidance on testing and evaluating chip performance. The second part is particularly geared to product design engineers as a guide through a development method adapted to biology and a transition between theory and the realization of a concrete MOC system. The methods of interaction between engineers and biologists are also described.

Selecting the appropriate models and coculture medium Choosing the appropriate biological model OOC and MOC platforms have recently emerged as a platform for testing the effects of an engineered microenvironment on a cellular model. While these models can be simple with, for example, a layer of cells exposed to medium flow, some are more complex and integrate stretchable membranes, pumps, and other technologies to mimic processes occurring in the human body and sensors to continuously monitor relevant physical and chemical parameters within the chip. The first step in the development of an OOC or MOC is deciding whether the system will be designed to study a single organ or several connected organs. This selection will depend mainly on the type of study to be performed and will determine the complexity of the chip and therefore the time required for development. OOC platforms are, in general, simpler to develop from a biological perspective, especially because the challenging selection of a coculture medium is not needed. As OOC devices are used to recreate the specific microenvironment observed within an organ (e.g., with respect to mechanical stretching, pressure, peristaltism, or shear stress) that can improve the resemblance of the cellular model to the in vivo

Selecting the appropriate models and coculture medium

organ, they are often employed to study diseases or mechanisms (Jackson and Lu, 2016; Sun et al., 2019). Their complexity is derived from the chip design, where the incorporation of a given mechanism can sometimes be challenging. Once the system has been designed, it can still be connected with other OOC devices to ultimately create a human-on-a-chip platform. This approach is currently being followed, for example, by the Wyss Institute (https://wyss.harvard.edu/technology/human-organs-on-chips/). MOCs are frequently used for toxicological assessment. The differing cellular models in the chip affect the pharmacokinetic and pharmacodynamic properties of a drug circulating in the chip in a manner similar to that of the absorbed drug in vivo (Bovard et al., 2017; Ishida, 2018; Ronaldson-Bouchard and VunjakNovakovic, 2018); MOCs therefore have the potential to improve preclinical drug testing, providing a more accurate estimate of safety and efficacy. In the second step, the appropriate cellular model(s) should be selected. In general, primary cells have a short lifespan, are more expensive to obtain, and are prone to donor variability. However, they are better physiologic mimics than immortalized cell lines, which is particularly important for mechanistic and disease studies. Instead, cell lines have the advantage of being an almost unlimited source of cells with the same characteristics yielding data that are consistent between laboratories (Montano, 2014). Despite this, and because OOC and MOC devices aim to recreate a microenvironment resembling the in vivo situation, the use of primary human cells is recommended. For each organ, several two-dimensional (2D) and three-dimensional (3D) models are currently available. While 2D models can be seeded either directly in the chip, on a coverslip, or on an insert and then moved to a chip, 3D models have various aspects and specifications that must be considered. For example, chips using lung tissues cultured at the air liquid interface (ALI) require a compartment through which the medium can circulate below the membrane and where the apical compartment can be exposed to air. Table 15.1 presents a global overview of the most commonly used models and their specific requirements; comprehensive details can be found in the literature (Ale´pe´e, 2014; Alhaque et al., 2018). When our laboratory began to develop a lung/liver-on-a-chip platform, we had to select appropriate models to mimic the functions of both organs. Several 2D lung models have been described in the literature, such as the bronchial BEAS2B cell line (Garcia-Canton et al., 2013) and the alveolar A549 cell line (Lieber et al., 1976). Because we wanted a model that mimicked the human bronchial epithelium as closely as possible, we used primary bronchial cells differentiated at the ALI to mimic lung function. The closer similarity between bronchial cells cultured at the ALI and the human bronchi with respect to morphology, functionality (Karp et al., 2002), and transcriptional profile (Pezzulo et al., 2011) ensured that this model would be suitable for toxicological (Qin et al., 2019), disease (Derichs et al., 2011), and mechanistic (Akaba et al., 2018) studies. For the liver, we selected a spheroid model composed of HepaRG cells only. HepaRG cells have a

465

466

CHAPTER 15 How to build your multiorgan-on-a-chip system

Table 15.1 Nonexhaustive list of cellular models and their requirements. Cellular model

Requirements

2D cell culture

• Cultured in a well. If the cells will be transferred, the

3D spheroids • Liver • Brain • Pancreas • Lung

• • •



3D ALI tissues



• Skin • Lung (bronchial, small-



• • • •

airway, alveolar) Nasal Buccal Gingival Gut

3D endothelial/epithelial models • Kidney • Gut • Blood brain barrier



• •



cultures can be grown on a coverslip (coated or uncoated) and then transferred into the chip Cells are submerged in medium Spheroids are maintained submerged in medium Cells within the spheroid can easily attach to the bottom of the well. An ultralow attachment coating will most likely be needed to prevent adhesion of the spheroids to the bottom of the well To improve tissue oxygenation, the ratio of surface to depth of medium in the compartment should be high Cells are cultured on a porous membrane with medium on the basolateral side and air on the apical side (ALI) Medium level in the compartment should be tightly controlled to ensure that the basal side of the membrane is in contact with the medium. Medium level should remain stable in the compartment to avoid medium covering the apical side of the tissues The compartment housing the tissue should ideally be open and easily accessible to ensure correct air circulation and to facilitate apical washes Cells are generally cultured on a porous membrane with medium on both sides of the membrane Depending on the model, apical and basolateral medium composition can change to mimic, for example, the tubular or luminal fluid of the kidney Depending on the model, both the apical and basolateral compartments require a medium flow to mimic the in vivo environment

2D, Two-dimensional; 3D, three-dimensional; ALI, air liquid interface.

good metabolic capacity (Gerets et al., 2012; Takahashi et al., 2015), closely resemble primary human hepatocytes in terms of glucose and lipid metabolism (Samanez et al., 2012), are relatively accurate in assessing drug-induced hepatotoxicity (Gunness et al., 2013), and are available in limitless quantities and as a stable phenotype. As our objective was to assess the toxicity of compounds absorbed by the lung and further metabolized by the liver, we considered HepaRG cell availability, metabolic capacity, and batch-to-batch similarity an advantage for robust toxicity testing. However, if the chip was to be used to evaluate the hepatotoxicity of particles absorbed by lung tissue, the use of primary human hepatocytes would be better, as they are more sensitive to toxicants (Bell et al., 2017).

Selecting the appropriate models and coculture medium

In the next section, we discuss one of the most challenging steps in the development of MOC platforms: the selection of a coculture medium.

Identifying and evaluating a coculture medium The human body is composed of more than 70 organs, each of which has specific functions and needs. Organ function is ensured through the delivery of nutrients via the bloodstream and specific factors produced by surrounding cells. Astrocytes, for example, secrete neuromodulators, hormones, and plastic factors needed for proper neuronal function (Verkhratsky et al., 2016). The activity of intestinal cells is modulated by factors released by the microbiome (Okumura and Takeda, 2017), and fibroblasts secrete factors affecting the proliferation rate, migration, and differentiation of surrounding cells (Yun et al., 2010). Apart from glucose, vitamins, amino acids, and oxygen that would be provided systemically in vivo by the blood, in vitro models require more specific factors for maintaining the proper function of a given cell type. In the case of bronchial and small-airway cultures, for instance, an unsupplemented medium contains the essential nutrients sufficient for cell expansion. In contrast, differentiation into ciliated and goblet cells requires exposure to air and the addition of retinoic acid and hydrocortisone (Cozens et al., 2018; Zaidman et al., 2016). Consequently, when developing an MOC system, the circulating medium must contain essential nutrients and specific factors for all cellular models integrated into the chip. The difficulty in selecting a coculture medium lies in the fact that media contain varying amounts of glucose, vitamins, amino acids, and salts, and that they can be highly enriched in factors that can be detrimental to the survival of some cell types. For example, the maintenance of 3D lung tissues with a heart model could be challenging, as the hydrocortisone needed for lung tissue maintenance is known to cause cardiac alterations (Silva et al., 2017). Similarly, there is some evidence that the hydrocortisone required for bronchial cell differentiation may also activate hepatic stellate cells, rendering the coculture of 3D lung tissues with liver spheroids containing hepatic stellate cells difficult (Wickert et al., 2004). Moreover, in most cases, the composition of commercially available cell culture media is often not disclosed by the suppliers, complicating the identification and selection of an appropriate cell culture medium. Therefore, identifying or preparing a suitable coculture medium typically requires extensive testing before the stable maintenance of the various tissue types is possible. In our opinion, there are three approaches to selecting a suitable coculture medium:

• The first option is to prepare a coculture medium based on the requirements of the tissues. Starting from a medium containing only the vital elements (amino acids, inorganic salts, vitamins, and eventually glucose), such as minimum essential medium, Dulbecco’s modified Eagle’s medium, or RPMI-1640, the medium composition can then be supplemented with a first group of

467

468

CHAPTER 15 How to build your multiorgan-on-a-chip system





compounds required by all tissues (van der Valk et al., 2010): insulin for glucose and amino acid uptake (Leney and Tavare´, 2009), transferrin as a universal iron carrier (Bogdan et al., 2016), and selenium as a cofactor for several antioxidant enzymes (Tinggi, 2008). Next, only the factors specifically required by the cellular models to maintain their key functions can be added. While this approach might require some time, it offers the opportunity to finetune the medium composition for the needs of all tissues and to reserve the option to change the medium composition if another cellular model is added to the MOC system. The second option, which we used to develop our lung/liver-on-a-chip platform, is to test whether one of the media used to maintain a cellular model would also be suitable for maintaining the other model(s). It is also possible to combine the media and to test various ratios. While this approach is faster than the first one, it does not enable easy manipulation of the medium composition. The third approach may be particularly useful when no coculture medium can be found. In this case, each cellular model is cultured in its particular culture medium. The metabolites or factors are exchanged between the organs— which are maintained in different compartments—using a dialysis membrane that would transport only some compounds between the compartments. Membranes with various molecular weight cutoffs (from 1 to 1000,000 kDa) are available. As the cellular models are cultured within their specific media, this ensures their viability and functionality. The disadvantage is that only a subset of molecules would be transported between the compartments, and the effects measured could therefore differ from those measured using a standard MOC platform. Such technology is currently used on the Dynamic Multiple Organ Plate developed by IONTOX (Ipema et al., 2014).

Once a strategy for preparing the coculture medium is selected, it should be tested on the cellular models to be used under standard culture conditions (offchip). The purpose of this step is to measure the effects of the selected coculture medium on the cellular models to ensure that the tissues will remain viable and functional over time. Notably, the cellular models that will be incorporated into the chip should be thoroughly evaluated to determine whether they retain normal morphology and tissue-specific function, even after a prolonged culture period. Therefore, assays evaluating the cellular model’s characteristics sensitively must also be selected and employed. During the development of our chip, we tested whether the HepaRG spheroids could survive once in culture with PneumaCult-ALI medium (the medium for lung epithelial cell differentiation and maintenance), undiluted or diluted with various amounts of HepaRG medium (100%/0%, 75%/25%, 50%/50%, 25%/ 75%). As the liver’s primary functions in vitro are the metabolism of xenobiotics and the secretion of albumin, we analyzed the expression of a panel of phase 1 metabolism-associated genes and the secretion of albumin at various time points

Multi-organ-on-a-chip development project

in HepaRG spheroids cultured with the candidate coculture medium. Using 100% PneumaCult-ALI medium, we observed that the HepaRG spheroids secreted higher amounts of albumin than in standard medium and that only a few metabolismassociated genes were differentially altered (Bovard et al., 2018). As the HepaRG cells retained hepatic cell function (metabolism and albumin secretion), even after 14 days in the PneumaCult-ALI medium, we selected this medium as our coculture medium. According to the hypothesis that all cellular models incorporated into a chip will alter the medium composition through the consumption of nutrients, release of metabolites, and secretion of factors, the development of a coculture medium should be facilitated by the inclusion of more cellular models within the same chip. The liver compartment would, for example, regulate albumin levels, while a pancreatic model would regulate glucose levels, removing the need to finely tune the concentrations of these compounds within the circulating medium. A fibroblast culture within each compartment and the presence of endothelial cells lining the chip circuits would ensure that factors such as fibroblast growth factors are present in the medium and made available for the tissues in the MOC device. To ensure the synthesis of these compounds, of course, the appropriate nutrients should be provided.

Multi-organ-on-a-chip development project The development of an MOC system requires a special engineering approach, because the usual methods of product development cannot always be adapted to the biological field. Traditional development methods often involve selecting the majority of solutions at the beginning of the project by dimensioning and calculating the system needs. However, in biology, it is difficult to make precise calculations that ensure the correct sizing of the system before tests are performed with the tissues. The behavior of tissues or cell cultures outside their typical environment is unpredictable. The product under development might therefore require alterations or improvement at each step, calling for flexibility. The first phase in the development of such a product uses the APTE (from the French “application des techniques d’entreprise”) development method. In this step, the project needs and core activities are identified. Next, to better envision the real use of the final product, the object is placed in its environment to identify constraints and evaluate potential risks. Finally, rational solutions are selected, and a block diagram is constructed to graph the functions and their interactions. The first step (see the “Analysis of project needs and rationalization of solutions” section) will therefore contain the following actions:

• Expression of the need • Analysis to identify the system functions • Analysis of system state change

469

470

CHAPTER 15 How to build your multiorgan-on-a-chip system

• Comparison of solutions resulting from the functional analyses • Division of functions into blocks of technical solutions • Validation of the technical solutions using the function analysis system technique (FAST) In the second step, the technical solutions identified using the APTE method are tested experimentally. To perform this step, we used two agile methods:

• Rapid application development, which enables researchers to rapidly explore



needs before a solution is created using iterative cycles inspired by the Scrum methodology (Martin, 1991). This method puts less emphasis on planning and more on adaptive processes than the plan-driven waterfall method. The Scrum development method, which was created to deliver products with the highest possible value using productive and creative means (Schwaber and Beedle, 2002). This method involves iterations, or “sprints,” lasting 1 2 weeks, where only the outcome is clearly defined. It is then up to the team working on the product to select the best approach to achieve the goal. Examples of different sprints performed are given next in the section “Project realization using agile development methods.”

Combining the APTE method with rapid application development and the Scrum method aims to provide maximum flexibility to the project in altering direction at any iteration while relying on rational and solid dematerialized product development. Moreover, the absence of preconceived ideas ensures that the final product will strictly meet the needs of the clients. Because the development of any project is a unique undertaking, the needs, constraints, and other key elements of a project are specific and cannot be covered within a chapter. In the following sections, we outline the methodology used in our laboratory and the results we obtained when developing our MOC system. The same reasoning can certainly be applied to the development of other OOC or MOC platforms.

Analysis of project needs and rationalization of solutions Statement of need At the commencement of a project, the engineer is sometimes faced with multiple demands arising from the client’s or company’s environment, such as ideas developed during brainstorming meetings and discussions with specialists from various fields. Because of the glut of information shared and the expression of many subsidiary needs, the engineer may lose sight of the fundamental needs of the project. The method presented next ensures that the main objective of the project is always kept in mind. This is also a simple way to express the needs of the project in a short and readable sentence. We identified and expressed our project needs using the “horned beast” diagram (Fig. 15.2). The left box answers the question, “who does it serve?” In our

Multi-organ-on-a-chip development project

FIGURE 15.2 Diagram used to identify the development needs of an MOC device. MOC, Multiorgan-ona-chip.

case, it was the scientists. The center box presents the product under development. The right box answers the question, “what does the system act on?” The bottom box expresses the goal of the project. Once the figure is created, the needs of the project can be expressed in one sentence that can be used at any time by the engineer to refocus. In our case, the focus sentence was “The MOC allows the scientists to circulate medium between a lung tissue cultured at the ALI and liver spheroids.”

Functional analysis of the system in its environment One of the most important elements in product development is the rationalization of solutions, with the objective of ensuring that they are selected without prejudice, not influenced by existing product specifications, and fully adapted to the current project’s needs. It can be tempting to select ready-to-use solutions from existing products, but these solutions will likely not be perfectly adapted to the needs of the project. Moreover, the development of an MOC platform requires the collaboration of individuals with varying areas of expertise and differing expectations. A technician working with the MOC system will require it to be easy to manipulate, while the research scientist will expect the system to deliver accurate and robust results. Therefore, it is crucial to ensure that the needs of all stakeholders are collected, understood, and balanced against each other to achieve optimal system functionality. To accomplish this, a functional analysis places the product under development in its environment and expresses the solutions in terms of function alone, without any associated technical solutions. This avoids the choice of ready-made solutions and makes it possible to identify solutions rationally and without prejudice. The functional analysis performed for our MOC system is presented next. The octopus diagram (Fig. 15.3) is an expression of the system in its environment and enables the identification of all the parameters and constraints of the system. Using this diagram, it is then possible to determine the main functions (MFs) that interact with the system and require action as well as the constraint factors (CFs) in the system’s environment. For example, we can see that the MOC system will have the circulation of medium between the lung and liver

471

472

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.3 Octopus diagram representing the MFs (in blue) of an MOC device and the CFs (in black) when the MOC is maintained in an incubator. CFs, Constraint factors; MFs, main functions; MOC, multiorgan-on-a-chip.

tissues as an MF and that the CFs will include the need for sterility and resistance to the incubator temperature. Once all MFs and CFs have been identified, they can be listed in a table that describes each one, its selection criteria, and a range of parameters based on antecedents or calculations. In our case, we knew that we would require circulating medium between liver spheroids and lung tissues cultured at the ALI. Because the lung tissues are cultured on inserts, where the basal side of the insert is in contact with the flowing medium, the lung tissues would potentially be exposed to shear stresses. Previous studies demonstrated that the maximum shear stress (caused by the medium circulating below the membrane) to which the lung tissues can be exposed should not exceed 0.45 dyn/cm2 (Trieu, 2014). As we knew the size of the inserts on which the lung tissues were grown and the minimum depth of medium needed below the tissues, we estimated the maximum flow to be around 800 μL/min. Such a criterion can be used later to narrow down a list of potential solutions to perform a given function. Each criterion is assigned a level of flexibility that defines to what extent the criterion can be changed. There are three levels of flexibility: “essential” (cannot be changed), “important” (should be changed only if the criterion blocks the development of the product), and “minor” (not essential for the product, and several options can be envisaged). For example, we determined that the tissues must have access to gases in the incubator, and because this condition cannot be altered, we assigned it a level of flexibility of F0. Table 15.2 represents the result of this approach in our example.

Functional analysis related to system state change The previous section described how to index the functions of an MOC system during development. Designing a system that would be transported physically

Multi-organ-on-a-chip development project

Table 15.2 System main functions (MFs) and constraint factors (CFs). MF1

Function

Criteria

Range

Flexibility

Feed and connect liver and lung tissues using a circulating culture medium

Flow rate Flow rate accuracy No fluidic bursts Controllable during the experiment Control of flow rate per circuit (not per chip plate) System operation

0 800 μL 6 3%

F0 F0 F2 F2

0 800 μL

F2

Continuously and for long periods 6 0.5 mm Equivalent or better than the referencea Equivalent or better than the referencea Equivalent or better than the referencea Equivalent to tests performed at 18 C 100% humidity

F0

Stable medium level Biological tests

MF2

Enable gas circulation for lung tissue

MF3

Enable gas circulation for liver tissue

Biological tests

CF1

Must be resistant to temperature

Biological tests

Technical tests

CF2

Must be resistant to humidity

CF3

Must protect tissues from contamination

CF4

Must be integrated into the incubator without causing disruptions Must enable lung tissue culture and retain tissue homeostasis Must enable liver tissue culture and retain tissue homeostasis

CF5

CF6

Material resistant to humidity Avoids electronic or mechanical disturbances Biological tests

F0 F0

F0

F0

F0

F0

Equivalent to tests performed at 20% humidity Equivalent or better than the referencea Equivalent or better than the referencea

F0

Biological tests

Equivalent or better than the referencea

F0

Biological tests

Equivalent or better than the referencea

F0

Technical tests and biological tests

F0

F0

(Continued)

473

474

CHAPTER 15 How to build your multiorgan-on-a-chip system

Table 15.2 System main functions (MFs) and constraint factors (CFs). Continued Function

Criteria

Range

Flexibility

CF7

Must be eco-friendly

Repairable system Long-term use

Dismountable Minimum 2 years

F2 F2

CF8

Must be easy to manipulate

CF9

Must be biocompatible

CF10

Must ensure the competitiveness of the system

Avoid disposable parts Easy to manipulate Rapid setup Non-toxic material in direct or indirect contact with the tissues Cost-efficient

Higher quality

CF11

CF12

Must be compatible with the format of tissues previously used by the laboratory Must be safe to use

24-well format

Handling of electrical and electromagnetic equipment; handling of biological materials

User tests User tests Biological tests

Comparison with competing products Comparison with competing products Plates and inserts in 24well format

European conformity (CE mark) standard biological material

F2 F1 F1

F0

F0

F0

F0

F0, Essential; F1, important; F2, minor. a The reference corresponds to the condition and results obtained when tissues are cultured in standard multi-well plates.

between locations or environments should also include an analysis of the transition, expressed in handling steps, and an analysis of the risks resulting from the transition. Such an analysis enables the identification of potential new functions required by the system to make it transportable and ensures that the previously defined functions are adapted for each handling step. To analyze the handling steps expected for our MOC platform, we prepared a chart showing the procedural steps (Fig. 15.4). As the MOC device under development will contain tissues that can be contaminated by microorganisms, we identified the first step as sterilization of the system after storage (Step 1, Fig. 15.4). Because most laboratories possess an autoclave, and because autoclaving is an efficient sterilization method, our MOC device would need to resist the conditions within an autoclave. At the

Multi-organ-on-a-chip development project

1

Sterilization of the system

2

Tissue(s) transfer and system startup

3

Transport of the system

4

Incubation

FIGURE 15.4 The four handling steps of the MOC device. MOC, Multiorgan-on-a-chip.

beginning of an experiment (Step 2, Fig. 15.4), the chip will be connected to the pump, and system operation will be initiated. As the incubator would have to be opened to connect the chip to the pump, the main risk identified for this condition was that too long of a manipulation might alter the temperature and humidity in the incubator. Therefore, the various elements that would be manipulated must be easy enough to handle rapidly. In transporting the system (Step 3, Fig. 15.4), we identified a risk of spilling medium or compounds from the chip plate onto the user or the floor and a risk of contaminating the cell cultures with microorganisms or environmental agents. These risks could be prevented by reducing or eliminating direct contact between the culture medium and the surroundings during transport, such as by designing lids that cover the chips. The final step identified when manipulating the chip was its transition from the laboratory bench to the incubator. During this action (Step 4, Fig. 15.4), the temperature and relative humidity will vary from 18 C to 37 C and from 30% to 100%, respectively. Under such conditions, moisture could condensate in the chip plate and exert adverse effects on the tissues. This could be prevented by optimizing air and humidity conditions between the chip plate and the incubator.

Rational choice of solutions identified during functional analyses In the previous sections, we identified the functions to be performed by our system and the constraints resulting from interaction of the chip with the (potentially altering) environment. The next step is to select technical solutions that enable the performance of the previously identified functions. A complete analysis of all options and their advantages and disadvantages should be performed to ensure that only the best option (and not a preconceived one) is selected. One example of this process is presented next: the choice of the pump used to circulate medium in the chip circuits and connect the two wells containing the lung and liver tissues. For our MOC, we had selected five pumping principles (Table 15.3). A list of selection criteria was then established for selecting the most appropriate solution. The criteria included the limitation in volume that may be supplied to the chip (some pumping principles imply a finite volume), the regularity of the pumping (some systems produce a continuous and regular fluid flow, while others produce an interrupted and irregular fluid flow), the estimated cost, and the pump’s lifespan (the pumps with a significant proportion of parts

475

476

CHAPTER 15 How to build your multiorgan-on-a-chip system

Table 15.3 Characteristics of pump principles driving culture medium to tissues in organ-on-a-chip devices. Criterion Noninvasive Volume limitation Pumping regularity Cost Lifespan

Peristaltic pump

Piezoelectric pump

Syringe pump

Electroosmotic pump

Diaphragm pump

1 No

2 No

1 Yes

2 No

2 No

11

2

11 1

11 1

2

11 11 1

2

1 11 1

2

11 1

that must be replaced after each experiment because of their direct contact with the culture medium were considered to have a poor service life). From these criteria, we concluded that the peristaltic pump was the best option, as it offered the following advantages:

• It is noninvasive and does not act directly on the culture medium. The

• •

pumping is actuated through tubes that can be changed before each experiment, preserving the sterility of the system and increasing the pump’s lifespan. There is no limit to the volume of the pumped fluid (the limitation is the volume of fluid in each circuit). It is relatively inexpensive.

Functional block diagram of the multi-organ-on-a-chip system The functional block diagram approach enables the presentation of all technical solutions as blocks and interconnections. Such a diagram is particularly useful for complex systems, as it ensures that the relationships between system functions are considered and understood. It also allows the identification of the various elements to develop and finalizes the expression of technical choices. In the context of the MOC system development, we constructed a block diagram representing all technical solutions that were selected in the previous step. To save time in analyzing the technical solutions, generating the block diagram and comparing the technical solutions were performed in parallel, without the need for a block diagram presenting the technical functions. For example, in the next chart (Fig. 15.5), the blocks (technical solutions) necessary to move the culture medium (technical function) are the motors and pumps. All choices were made by comparison using the same method outlined in the previous section. Our block diagram contains the following elements:

• The “controller,” “motor,” and “encoder” blocks needed to create the medium flow.

Multi-organ-on-a-chip development project

FIGURE 15.5 Block diagram of an MOC device containing various technical solutions and their interconnections. AC, Alternating current; DC, direct current; EEPROM, electrically erasable programmable read-only memory; MOC, multiorgan-on-a-chip; PWM, pulse width modulation.

• The “well lung,” “well liver,” and “reservoir” blocks needed to culture the two tissues within the chip.

• An “AC/DC converter” block needed to convert the supplied voltage into 5, 12, or 24 V.

• The “microcontroller,” “Bluetooth wireless communication module,” and “user interface” (UI) blocks needed to control the medium flow rate and flow direction. Our plan to maintain the standard size of a culture plate (to remain compatible with standard laboratory equipment) represented a constraint expressed during the functional analysis. We therefore had to design a chip of the same dimensions as a standard multiwell plate. We calculated that four circuits, each connecting two wells, would fit on a chip plate, requiring four peristaltic pumps for the pumping action.

Function analysis system technique method The FAST method can be used to express and confirm the technical solutions selected to develop a system in line with the project needs (Wixson, 1999). It can be seen as a summary of all previous steps, where the needs, constraints, functions, and solutions were identified. All of these elements are represented within

477

478

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.6 Representative scheme showing part of the function analysis system technique used in developing an MOC device. MOC, Multiorgan-on-a-chip.

the same graph, which is helpful when a change or modification to the system is required. The FAST diagram (Fig. 15.6) shows the proposed system and can be read from three directions: left to right to show the “how,” right to left to express the “why,” and top to bottom to express the chronology of the actions. The diagram is useful in outlining the various functionalities of the system as they relate to the technical choices when using agile methods later. For example, if the engineer seeks a rationalization for the inclusion of microfluidic channels in the chip plate, the diagram shows that the channels are needed to perform molecular exchanges between tissues. Conversely, the diagram also shows that to perform molecular exchanges, we implemented microfluidic channels.

Project realization using agile development methods Adaptation of agile method to product design The agile method is derived from the field of computer science, in which the development of new software is realized by iterations. After each iteration, the user can test the product and indicate the aspects of the product that require

Multi-organ-on-a-chip development project

change or further improvement. This method ensures that the product under development satisfies the needs of the user at each step and offers a high level of flexibility (Dingsøyr et al., 2010). While it emphasizes the experimental work to the detriment of activity planning, it also requires excellent communication between the stakeholders. Nevertheless, the application of such an approach to the biological field requires some adaptation, especially for the duration of the iterations. As mentioned earlier, the effect of the product on the tissues can take several days or weeks to be detected experimentally. Consequently, the next prototype would also require several weeks in preparation. In our case, the various iterations (or sprints) needed to develop the MOC system lasted 4 6 weeks, unlike the typical 2 3 weeks in the field of software development. Our sprints were performed by a biologist and an engineer. The biologist evaluated the characteristics of the tissues once inserted into a chip prototype, determined whether the conditions were optimal, and evaluated the chip (e.g., ease of use, impact on incubator). The role of the engineer was then to assess the biological data, determine the type of alteration required to improve the system, and evaluate the risk associated with the change. The biologist then validated the sprint, acting as the product owner. This collaboration between the biologist and the engineer was essential to ensure that the MOC system under development was fully adapted for tissue maintenance, easy to handle, cost-effective, and robust. In too many cases, a system is developed by a person with expertise in only one field, without taking into account the needs of the customer before the final product is delivered. Once the product is ready for delivery, however, only minor modifications can be made to the system without causing a significant increase in development costs and the time to first commercialization.

Redistribution of functional blocks into sprints As the first step of agile development, we reorganized the functional blocks obtained using the APTE method into iterations or sprints that were independent of each other and could be realized and tested individually. From Fig. 15.5, we identified three groups of functional blocks: 1. Blocks associated with the development of the pump unit 2. Blocks associated with the development of the pump controller 3. Blocks associated with the development of the chip plate We then began developing the product through iterations for each group of functional blocks.

Sprints realized for the multiorgan-on-a-chip system Fig. 15.7 shows the sprints arranged by group (see “Redistribution of functional blocks into sprints”). Each sprint typically lasted 4 6 weeks. The development of our MOC system was divided into the exploratory phase and the finalization and industrialization phase. The exploratory phase was performed using mostly proofs

479

480

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.7 Flowchart of the sprints performed to realize the project. The sprints are grouped by function (pump, controller, or chip plate). The project was divided into an exploratory phase and a finalization and industrialization phase, which began after the biological data showed that the MOC device could be used for coculturing lung and liver tissues. MOC, Multi-organ-on-a-chip.

Multi-organ-on-a-chip development project

of concept (POC), which are prototypes fabricated using rapid prototyping approaches. The advantage of this approach is that it allows the POC to be tested, validated, and improved rapidly, without focusing on precise and dedicated technical solutions that are time-consuming and provide a high degree of inertia and resistance to project changes. This rapid prototyping approach takes advantage of new technologies and platforms:

• Electronics: development boards already tested and assembled (breakout • • •

boards) and development platforms such as Arduino (Arduino.cc), Raspberry Pi, and their related firmware/software integrated development environment. Mechanics: 3D printing has been used extensively to build platforms, plates, and tools that are inexpensive and rapidly testable (Rayna and Striukova, 2016). Computer science: development of functions such as the UI is minimized at the beginning. The aim is to produce a simplified system that possesses only the essential functions needed to run the prototype. Tissue engineering: scaffolds, either 3D-printed or molded, are seeded with cells to construct a 3D architecture rapidly and artificially (Yeong et al., 2004).

At the end of the exploratory phase, the MOC system was functional and could be used for experiments. The finalization and industrialization phase then included sprints to increase the throughput and the versatility of the chip plate. It is also during this phase that various manufacturing approaches can be envisioned to reduce the cost of producing each MOC system.

Group 1: pump unit development We performed several sprints to develop the pump unit. First, a pump POC composed of only one peristaltic pump was tested to ensure that the pump motor would function properly within an incubator, that the flow rate could be adjusted, and that a moisture-proof housing could be constructed to protect the motors from moisture. Next, a pumping unit composed of four peristaltic pumps was built and tested to ensure that the heat generated by the four motors would not increase the air temperature in the incubator higher than 37 C. Because the heat emitted by multiple pump units caused the incubator to overheat, we tested new types of motors to construct an acceptable pumping unit and realized the first pump unit for production. First sprint: proofs of concept of a single pump unit in its environment. As we calculated that four circuits would fit on the same chip plate, we designed a pumping unit composed of four peristaltic pumps. Before constructing the unit, however, we had to ensure that the peristaltic pump, motors, electronics, and box protecting the electronics from moisture were adapted to our needs (Fig. 15.8). The first sprint therefore consisted of testing a peristaltic pump placed in its environment.

481

482

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.8 The MOC device and the part affected by the first sprint (inside dashed square). MOC, Multiorgan-on-a-chip.

FIGURE 15.9 View of a pump with an O-ring. (A) From left to right: the back panel of the box, the motor, the main part of the motor box, the O-ring, and the four screws. (B) To protect the pump from humidity, an O-ring was placed around the motor shaft.

Before testing the pump, we developed and implemented the peristaltic pump control electronics using electronic breakout boards to simplify the process. Next, a moisture-proof box was developed using a 3D printer to protect the motor and electronics from moisture (Fig. 15.9A). Because the moisture in the incubator could seep into the box through the motor shaft, we placed an O-ring around the shaft (Fig. 15.9B). While we observed that the box was moisture-proof, we also found that the O-ring increased the torque required by the motor to move the peristaltic pumps, resulting in higher current consumption and more heat generation. The heat emitted by one motor did not impact on air temperature in the incubator. After constructing this box with a single peristaltic pump, we tested the influence of temperature and humidity in the incubator on the proper operation of the

Multi-organ-on-a-chip development project

Table 15.4 Summary of the first sprint. Goal of the sprint

Testing one peristaltic pump, a potential moisture-proof box, and the effects of the pump unit on its environment

Technical validation criteria

• Effect of heat emitted by the pumps on the incubator • Box resistance to humidity • System accuracy and calibration

Next sprint Risk management

Test a pump unit composed of four peristaltic pumps Possible effect on incubator internal temperature; parallel development of a cooling system Sufficient for the next sprint

Results

motors and measured the flow accuracy. During this sprint, we calculated and measured the corresponding pump flow rate in microliters per minute. These measurements enabled us to calibrate the pump by adjusting the engine speed. This calibration adjustment was then programmed into the firmware. Table 15.4 presents a summary of the first sprint. Second sprint: proofs of concept of the pump unit in its environment. The second sprint consisted of a first POC pump unit composed of four peristaltic pumps to start the biological tests (Fig. 15.10). The method used in the first sprint to develop a pump unit with a single peristaltic pump was applied to the second sprint to develop a pump unit with four peristaltic pumps (Fig. 15.11A). Once again, the box containing the motors was constructed using a 3D printer to develop a prototype rapidly.

FIGURE 15.10 The MOC device and the part affected by the second sprint (inside dashed square). MOC, Multiorgan-on-a-chip.

483

484

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.11 Images of the pump unit connected to the electronics box. (A) View of the pump unit with (from left to right) the four motors (in black), the moisture-proof box (in blue), the four Orings (in black), and the four pump heads (in gray and blue). (B) The box containing the electronic components. (C) The cables running between the pump unit and the box. (D) A pump unit inside an incubator.

We then developed the electronics to control the four motors. A first electronic board was developed using component evaluation boards and derivation boards to create the first POC. The electronic card contained a microcontroller for all operations and voltage regulators and a voltage converter to supply the appropriate voltages. The motors were controlled by the microcontroller using pulse width modulation. The complete electronic board was then placed in a box that also included the various necessary connections (e.g., power supply for the electronics, motor connections). The electronics box was placed outside the incubator (Fig. 15.11B) and connected to the pump by a series of cables (Fig. 15.11C). Next, we assessed whether our pump unit was moisture-resistant by allowing the pump unit to run for several days inside an incubator (Fig. 15.11D). As in the first sprint, we measured the impact of the heat emitted from the peristaltic pump motors on the incubator. During this sprint, we realized that the use of several pumping units inside one incubator resulted in overheating, disrupting the

Multi-organ-on-a-chip development project

Table 15.5 Summary of the second sprint. Goal of the sprint

Test a pump unit composed of four peristaltic pumps and the effects of this unit on its environment

Technical validation criteria

• Effect of heat emitted by the pumps on the incubator • Box resistance to humidity • System accuracy and calibration

Next sprint Risk management Results

Resolve the problem of the incubator overheating Cooling system Sufficient for the next sprint

incubator temperature regulation and proper operation. We therefore had to identify alternative solutions. With this POC of a pump unit ready for use, it was then possible to begin biological tests in parallel (see the “Group 3: chip plate development” section) using a maximum of two pump units per incubator. Table 15.5 presents a summary of the second sprint. Fourth sprint: multiplication of pump units per incubator. We used the prototype of a pump unit developed in the second sprint to perform biological tests while the chip plate was also being developed (see the “Group 3: chip plate development” section). This version of the pump unit was functional and allowed us to perform the biological tests to confirm that our MOC device was working. Nonetheless, because of the overheating issue detected in the second sprint, only two MOC devices could be used in one incubator, a nonnegligible issue that decreased MOC throughput. To improve the platform and ensure that future consumers would be able to operate several MOC devices within one incubator, we envisioned two possible solutions: adding a cooling unit to the incubator, or constructing a new pump unit composed of different motor types that emit less heat. Because the first option would have required us to build a thermoregulated cooling unit (to avoid overcooling the incubator), it would have increased the cost, complexity, and physical footprint of the system. We therefore elected to build a new pump unit composed of more efficient motors (Fig. 15.12). The previous block of POC peristaltic pumps was operated by stepper motors. These inexpensive motors could be implemented rapidly but had poor efficiency, resulting in a significant emission of heat. We therefore decided to replace these motors with brushless motors that are more efficient but also more expensive. As the tests performed with these motors showed that almost no heat was emitted, we decided to miniaturize the electronic board and electronic components and to incorporate them into a moisture-proof box that also contained the motors (Fig. 15.13). The intent was to reduce the space occupied in the incubator and to decrease the risk of contamination from the multiple cables running between the incubator door and the incubator sealing (see the “Second sprint: proofs of

485

486

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.12 The MOC device and the part affected by the fourth sprint (inside dashed square). MOC, Multiorgan-on-a-chip.

FIGURE 15.13 View of the pump unit composed of four brushless motors with pump controllers. From left to right: the back panel of the box with the power supply connector (in blue), the four motors (in black), the electronic printed circuit board (without components) (in green), the moisture-proof box (in blue), the torque transmission adaptor (in violet), and the four pump heads with planetary reducer (in purple).

concept of the pump unit in its environment” section). Indeed, the new pump unit integrating all the electronic components required only a single power cable that could enter the incubator through the back access port. At the end of this sprint, the pump no longer affected its environment, operated more efficiently, occupied less space in the incubator, and reduced the risks of contamination. Table 15.6 presents a summary of the sprint.

Multi-organ-on-a-chip development project

Table 15.6 Summary of the fourth sprint. Goal of the sprint

To increase the number of pump units per incubator

Technical validation criteria

• Effect of heat emitted by the pumps on the incubator • Box resistance to humidity • System accuracy and calibration

Next sprint Risk management Results

Test the system with tissues Not applied in this sprint Sufficient for the next sprint

Group 2: pump controller development Once the first pump unit was designed, constructed, and validated, we began developing a system to control the peristaltic pump flow rate and pump direction (clockwise and counterclockwise) via a Bluetooth wireless connection. Third sprint: flow rate control for a single pump unit. In this sprint, we created the first UI. Wireless communication was implemented using the Bluetooth standard to transmit instructions to the firmware controlling the pump electronics, thereby controlling the flow rate easily and rapidly (Fig. 15.14). The UI was developed for a mobile device equipped with the Android system and programmed using Android Studio integrated development environment. The programming language used was Java for the logical part, XML for the UI, and SQLite for the database. We used the model-view-controller framework to program the software and the firmware because it is well adapted to the agile

FIGURE 15.14 The MOC device component affected by the third sprint. MOC, Multiorgan-on-a-chip.

487

488

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.15 Part of the code used to develop the application controlling peristaltic pump flow rate. (A) Simplified block diagram of the Java code used to control the flow rate. (B) XML code needed to see the selected flow rate and used to construct the UI. UI, User interface.

development method and the development of POCs that can be easily altered or improved at each iteration. The programming will not be discussed in detail in this chapter. Fig. 15.15A provides an overview of the XML code used to program the application’s UI, and Fig. 15.15B displays some logical elements of Java programming under the simplified block diagram. We see for the “seek bar” object the method implemented, such as “OnProgressChange,” which expresses the user’s manipulation of the “seek bar.” The corresponding firmware enabling the exchange with the application executed by Android was developed using the integrated development environment provided by Arduino, and the programming language used is C. With the first version of the UI, the user could control the flow rate of each peristaltic pump in a pumping unit at 0 500 μL/min. The wireless connection enabled pump control, even when the incubator door was closed. Thus the tissues grown inside the chip board were exposed to only minor disturbances from changes in humidity and temperature when the incubator was opened. Fig. 15.16 shows the UI. Table 15.7 presents a summary of the sprint.

Multi-organ-on-a-chip development project

FIGURE 15.16 Prototype UI.

Table 15.7 Summary of the third sprint. Goal of the sprint

Develop a UI to control the pump flow rate

Technical validation criteria

Flow accuracy Communication User-friendliness Operation without opening the incubator Control multiple pump units and flow directions Not applied in this sprint Sufficient for the next sprint

Next sprint Risk management Results

• • • •

UI, User interface.

Fourth sprint: flow direction control for multiple pump units. In the next sprint, we tested an interface to control several pump units using the same smartphone (Fig. 15.17). Following the approach described for the third sprint, we designed a

489

490

CHAPTER 15 How to build your multiorgan-on-a-chip system

UI containing a list of all pump units already connected to the smartphone via Bluetooth (Fig. 15.18A). After clicking on the pump unit to be controlled, a new page opens, allowing control of the speed of each peristaltic pump individually (Fig. 15.18B). The only change from the previous interface was the ability to change the flow direction by clicking on a button specific to each peristaltic pump. Table 15.8 presents a summary of the sprint.

FIGURE 15.17 The MOC device component affected by the fourth sprint. MOC, Multiorgan-on-a-chip.

FIGURE 15.18 Representative images of the new UI when several pump units are paired with a single smartphone (A) and the page opened once a pump is selected (B). UI, User interface.

Multi-organ-on-a-chip development project

Table 15.8 Summary of the fourth sprint. Goal of the sprint

Develop a UI to control several pump units

Technical validation criteria

• Controlling several pump units • Rapid and easy connection • Remote use

Next sprint Risk management Results

Test the application with the system Not applied in this sprint Need to improve the wireless communication

UI, User interface.

Group 3: chip plate development The last series of sprints was associated with developing the pump unit and the controller. To develop the chip plate, we first evaluated the circulation of medium within a chip prototype to ensure that medium levels in the compartment were stable. We then looked at the biocompatibility of the material used. Because the first materials tested were not biocompatible, causing the lung tissues to show signs of cytotoxicity after a few days in culture, we evaluated several materials. We then changed the design of the liver well to facilitate the manipulation of liver tissues. Finally, we designed several chip plates for various purposes to respond proactively to consumer needs. First sprint: proofs of concept of a chip plate design and biocompatibility testing. The first sprint of this block involved testing a chip plate design and assessing tissue survival with the selected biomaterial (Fig. 15.19). It allowed us to evaluate the global functioning of the system once located in its final

FIGURE 15.19 The MOC device and the part affected by the first sprint (inside dashed square). MOC, Multiorgan-on-a-chip.

491

492

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.20 Evaluation of the first chip plate prototype. (A) Design of the first chip plate prototype. (B) The blue dye used to evaluate medium flow characteristics within a circuit. (C) Measuring the temperature of medium circulating in a circuit.

environment and exposed to the expected system change (see the “Functional analysis related to system state change” section). A first chip plate POC was constructed from polycarbonate and fabricated using a machining approach (Fig. 15.20A). We first tested the ease of connecting the chip plate to the pump unit in the incubator using gloves. We ensured that the connection could be performed in only a few seconds to avoid disturbances to the incubator humidity and temperature. In the second step, we loaded blue dye into the four circuits composing a chip to assess how long it would take the medium to circulate in a circuit (Fig. 15.20B). We also tested whether the medium volume in each compartment remained stable over time and if bubbles that could disrupt medium flow were formed. Finally, we checked whether the heat emitted by the first pump unit we developed (see the “Fourth sprint: multiplication of pump units per incubator” section) would alter the temperature of the circulating medium (Fig. 15.20C). We also integrated a bronchial tissue into each lung compartment to evaluate the effects of the chip plate on tissue survival. The bronchial tissues were severely damaged after only a few days of culture within the chip plate, indicating that the material was not biocompatible (Fig. 15.21). We therefore immediately began working on the next sprint to test other materials for the chip plate. Table 15.9 presents a summary of the sprint.

Multi-organ-on-a-chip development project

FIGURE 15.21 Representative images of histologically processed bronchial tissues cultured for 5 days in a standard plate [incubator control (A)] or a polycarbonate chip plate (B). Magnification: 20 3 .

Table 15.9 Summary of the first sprint. Goal of the sprint

Test a chip plate prototype made from polycarbonate

Technical validation criteria

• Circulation of liquid in the circuits • Effect of heat emitted by the pump on culture medium temperature

• Biocompatibility of the chip plate with bronchial tissues in Next sprint Risk management Results

culture New biocompatibility tests if failure Manufacturing of new plates made from different materials Material not biocompatible

Third sprint: biocompatibility of the chip plate. Because the previous sprint demonstrated that the bronchial tissues could not survive in the chip, we tested various materials for the chip plate (Fig. 15.22). Suitable materials had to be both biocompatible and exhibit low absorbance. Some materials, such as polydimethylsiloxane, are absorb small hydrophobic molecules (Toepke and Beebe, 2006). We tested polytetrafluoroethylene, polypropylene, polysulfone, and polyetheretherketone. In biocompatibility tests, only the tissues maintained on polyetheretherketone chip plates retained morphology, functionality, and viability that were similar to those of tissues cultured in standard plates (incubator controls; Fig. 15.23). We also tested whether the liver spheroids would survive in the chip plates made from the four materials; the spheroids remained viable after a 5-day culture period. To prevent absorption of molecules from the medium by the tubes

493

494

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.22 The MOC device and the part affected by the third sprint (inside dashed square). MOC, Multiorgan-on-a-chip.

FIGURE 15.23 The chip plates constructed from different materials and morphology of tissues maintained in these chip plates for 5 days. (A) Polytetrafluoroethylene. (B) Polypropylene. (C) Polysulfone. (D) Polyetheretherketone. Magnification: 10 3 .

Multi-organ-on-a-chip development project

Table 15.10 Summary of the third sprint. Goal of the sprint Technical validation criteria Next sprint Risk management Results

Test new biocompatible materials for constructing the chip plate Biocompatibility of the material used to construct the chip plate with bronchial tissues and liver spheroids New biocompatibility test if failure Manufacturing of chip plates with new materials Polyetheretherketone showed excellent long-term results with both tissue types

connecting the chip plate to the pump unit, we used PharMed tubes (Saint-Gobain Performance Plastics, Courbevoie, France) instead of silicon tubes. In each of the sprints dedicated to culture plates, the risk of increased production time owing to the need for machining (instead of additive manufacturing) was managed by a forward-thinking ordering procedure. Table 15.10 presents the sprint summary. Fifth sprint: testing a new well design adapted for culturing liver spheroids. The fifth sprint was performed because the initial results obtained using liver spheroids (the model used to mimic the liver function in the MOC system) demonstrated that spheroids were fusing to each other, forming larger tissues with unknown cell numbers and potentially necrotic cores, which are not suitable for cell-based assays. We therefore started working on potential solution to improve the liver compartment of the chip (Fig. 15.24). To prevent the formation of these larger tissues, we added a 3D pattern to the bottom of the liver compartment to isolate the tissues. We tested two designs, one

FIGURE 15.24 The MOC device and the part affected by the fifth sprint (inside dashed square). MOC, Multiorgan-on-a-chip.

495

496

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.25 Scheme of the chip plate with the two patterns tested for covering the bottom of the liver compartment.

Table 15.11 Summary of the fifth sprint. Goal of the sprint Technical validation criteria Next sprint Risk management Results

Evaluate patterns for the bottom of the liver compartment Efficiency of well geometry to prevent spheroid aggregation Testing of different well geometry if failure Study of an alternative geometry Circular grooves are the best solution considered

composed of grooves and the other of very small cavities (Fig. 15.25). We found that the grooves were efficient at preventing spheroid aggregation and that this design could also support coverslips used with 2D cultures, as the grooves made the manipulation of the coverslips easier. Table 15.11 presents a summary of the sprint. Sixth sprint: catalog of chip plates. The MOC development project demonstrated that the chip we designed was efficient, robust, easy to manipulate, and offered excellent conditions for the coculture of bronchial tissues and liver spheroids (Bovard et al., 2018). We therefore considered the development of the chip to be complete. Because we wanted to diversify the uses of the MOC system proactively, we began developing chip plates with various well designs and well numbers (Fig. 15.26). We created a first plate with grooves at the bottom of the two compartments that composed each circuit (Fig. 15.27A). Such a chip plate would enable the coculture of any type of spheroid organ model such as pancreatic, liver, neuro, or lung spheroids. We also created a chip plate where the two wells

Multi-organ-on-a-chip development project

FIGURE 15.26 The MOC device and the part affected by the sixth sprint (inside dashed square). MOC, Multiorgan-on-a-chip.

FIGURE 15.27 Schemes of various chip plate designs created for coculturing two types of spheroid models (A), two different tissues cultured on Transwell inserts (B), or eight tissues cultured on Transwell inserts (C).

497

498

CHAPTER 15 How to build your multiorgan-on-a-chip system

Table 15.12 Summary of the sixth sprint. Goal of the sprint

Develop new chip plate design

Technical validation criteria Next sprint Risk management Results

New chip designs with potential uses Tests of the different systems under the required conditions Develop alternative designs ready for production Three new chip designs were produced

of each circuit were designed for tissue cultures on Transwell inserts (Fig. 15.27B), to enable the coculture of intestinal, small-airway, bronchial, nasal, or gingival tissues. Finally, we also designed a chip plate with eight wells per circuit (Fig. 15.27C), which could be adapted for the maturation of lung tissues, a process that requires medium replacement every 2 3 days for 28 days. Increasing the volume of the circulating medium would reduce the frequency of medium changes. Table 15.12 presents a summary of the sprint.

Final version of the multiorgan-on-a-chip system The final version of the MOC system consisted of six parts (Fig. 15.28): 1. The pump unit ensures the circulation of medium within the circuits at an adjustable medium flow rate of 10 100, 40 250, or 60 400 μL/min, depending on the pump head used (Fig. 15.29). The use of different pump heads for diverse flow rate ranges ensured the accuracy of the flow rate delivered by the peristaltic pump. The pump heads are also easy to disconnect from the pump unit, facilitating transport of the chip plate and pump head exchange (Fig. 15.30). 2. The chip plate is composed of circuits with two wells for the coculture of bronchial tissues and liver spheroids. Several chip plate designs were produced to coculture various in vitro models. The chip plate made from polyetheretherketone, the tube connectors, and the tubes connecting the chip plate to the pump unit (we used PharMed tubes) can all be autoclaved (Fig. 15.31) and reused multiple times. 3. The reservoir plate is used to increase the total volume circulating in each circuit, from 3 mL per circuit without a reservoir to 8 mL. Because the four reservoirs have a large surface, they improve gas exchange between the medium and the air in the incubator. 4. The plate holder is designed to hold a chip plate connected to a chip reservoir and facilitate the transport of the chip plate between the incubator, the laboratory bench, and the biosafety cabinet. The plate holder prevents unwanted movement and potential medium spillage during status changes.

Multi-organ-on-a-chip development project

FIGURE 15.28 Images of the complete MOC system in its final version with the pump unit (1), the chip plate (2), the reservoir plate (3), the plate holder (4), the pump controller (5), and the user manual (6). MOC, Multiorgan-on-a-chip.

FIGURE 15.29 Illustration of the three pump heads with their operating flow ranges.

499

500

CHAPTER 15 How to build your multiorgan-on-a-chip system

FIGURE 15.30 Illustration of the pump unit with the four pump heads. The arrows show the procedure to lock and unlock the pump heads from the motors.

FIGURE 15.31 Schematic showing the complete MOC system with the two parts requiring different sterilization approaches. (1) The pump unit (without the pump heads) can be sterilized using 70% ethanol. (2) The pump heads, chip plate, lids, reservoir plate, connectors, and tubes can be autoclaved. MOC, Multiorgan-on-a-chip.

5. The pump controller enables wireless control of the flow rate and flow direction of each peristaltic pump individually, even when the incubator door is closed. The UI includes a tab for each of the three pump heads (Fig. 15.32). When selecting a tab, the available flow range for the respective peristaltic pump is shown. Multiple pump units can be controlled using a single smartphone or tablet. 6. The user manual is a reference for the end user on preparing the MOC system and using it.

Testing the multiorgan-on-a-chip system

FIGURE 15.32 Final appearance of the UI once the MOC system is connected to a pump unit. The red, blue, and green tabs of the UI correspond to the three pump heads shown in Fig. 15.29. MOC, Multi-organ-on-a-chip; UI, user interface.

Testing the multiorgan-on-a-chip system At this stage of development the chip is almost ready for routine laboratory use. The appropriate cellular models have been selected by the type of experiment to be performed, and an appropriate coculture medium was identified and tested to ensure that all tissues would survive and remain functional. Next, the chip was developed with a specific design from an appropriate material, and a pump was incorporated into the system. In the third step, the MOC system is tested thoroughly to determine the stability of the tissues once cultured in the chip. It is also during this step that the cross-talk between the two tissues would be monitored.

Evaluating tissue stability During the development of the chip plate, we observed that the polyetheretherketone plate was biocompatible and enabled the culture of bronchial tissues and liver spheroids independently for 5 days. Depending on the intended uses of the chip, further testing should be performed. In our case, we planned to use the chip for repeated-exposure testing (OECD, 2008). We therefore cultured bronchial tissues and liver spheroids individually for 28 days and measured their viability,

501

502

CHAPTER 15 How to build your multiorgan-on-a-chip system

morphology, and functionality. Maintenance of the tissues in monoculture for 4 weeks generated data on the effects of the flow and the maintenance in the chip on the characteristics of both tissue types. For example, we observed that the transepithelial electrical resistance in bronchial tissues cultured in the MOC chip increased by 50% after 28 days in comparison with resistance in control tissues. As we then performed a coculture experiment with liver spheroids for 28 days and also observed an increase of 50%, we concluded that the factors released by the liver spheroids were not responsible for this increase.

Fit-for-purpose testing of the multiorgan-on-a-chip system MOC systems aim to improve in vitro evaluations of drug safety and efficacy. To ensure that the designed MOC system will fulfill this objective, the platform should be tested with drugs. Not every drug will be suitable: ideally, the test drug should be either bioactivated or inactivated by one of the models in the chip, leading to more or less cytotoxicity in the other cultured model(s). Importantly, when a liver compartment is incorporated, compound bioactivation or inactivation by the liver model’s xenobiotic metabolism can demonstrate that the “organ” is metabolically active. Most of the MOC systems currently available include a liver model and have therefore been tested with various drugs (Table 15.13). When developing and testing our lung/liver-on-a-chip system, we selected a few compounds for which pulmonary cytotoxicity was shown to be affected by the liver. We tested these compounds in both circuits containing only a bronchial tissue and circuits containing a bronchial tissue cocultured with liver spheroids. After incubation for 24 48 hours, we assessed bronchial tissue viability and compared the results between the two systems. Aflatoxin B1 cytotoxicity decreased in Table 15.13 Nonexhaustive list of compounds previously used in multiorgan-on-a-chip (MOC) systems to demonstrate organ cross-talk. MOC

Compound

Reference

Lung/Liver

Aflatoxin B1 Naphthalene Aflatoxin B1 Benzo(a)pyrene 2,5-Hexanedione Propranolol Epinephrine Doxorubicin Glucose Terfenadine

Bovard et al. (2018) Viravaidya et al. (2008) Theobald et al. (2018) Theobald et al. (2018) Materne et al. (2015) Skardal et al. (2017) Skardal et al. (2017) Kamei et al. (2017) Bauer et al. (2017) Vernetti et al. (2017)

Kidney/Liver Brain/Liver Heart/Liver

Pancreas/Liver Muscle/Liver

Acknowledgments

the presence liver spheroids, demonstrating organ cross-talk and the metabolic activity of the liver spheroids. In addition to pharmacodynamics data demonstrating tissue cross-talk, chip pharmacokinetics (modeled or experimental) should also be assessed to ensure, for example, that absorption of a compound by the first tissue would result in the delivery of a physiological dose to the second tissue. Moreover, the resulting metabolites should also be evaluated. If compound absorption and metabolism mimic the in vivo situation, and if the target organ is sufficiently sensitive to demonstrate eventual cytotoxicity, only then can the chip be considered useful for assessing drug toxicity.

Conclusion The development of an efficient, functional, and easy-to-manipulate MOC system is a relatively long process requiring numerous iterations. While the costs for developing such a system are rarely discussed, the human resources, the time allowed, and the cost of the physical components render the development process expensive. However, when considering that this technology could decrease the staggering cost of drug development—estimated at 2.6 billion dollars in 2016 (DiMasi et al., 2016), improve the efficiency/adverse effects ratio of new drugs, and increase the number of drugs approved for the market (Wong et al., 2019), the expense justifies itself. In this chapter, we showed that chip development is a long process requiring extensive exchanges between biologists and engineers. The biologist develops the tissue cultures, identifies the most suitable coculture medium, and performs tests to evaluate tissue function, morphology, and viability. The engineer adapts a development methodology designed for computer sciences to the biological sciences. The unexpected is common in the world of biology, because living organisms do not respond in a programmed way to their environment. Flexibility, patience, and perseverance are therefore essential. As seen in our case study, the results of an iteration can highlight a major problem in tissue maintenance, leading to the need for other iterations to understand and resolve the problem. The biologist must understand the physical limits of the project because not all options for creating a chip are feasible. Nonetheless, engineering and biology can combine successfully to create a platform as complex as an MOC device, pushing back the limits of scientific research and opening new possibilities.

Acknowledgments The authors thank Karsta Luettich, Anita Iskandar, Sandro Steiner, and Edanz Editing for their insightful feedbacks and edits.

503

504

CHAPTER 15 How to build your multiorgan-on-a-chip system

References Akaba, T., Komiya, K., Suzaki, I., Kozaki, Y., Tamaoki, J., Rubin, B.K., 2018. Activating prostaglandin E2 receptor subtype EP4 increases secreted mucin from airway goblet cells. Pulm. Pharmacol. Ther. 48, 117 123. Ale´pe´e, N., 2014. State-of-the-art of 3D cultures (organs-on-a-chip) in safety testing and pathophysiology. ALTEX 31, 441 477. Alhaque, S., Themis, M., Rashidi, H., 2018. Three-dimensional cell culture: from evolution to revolution. Philos. Trans. R. Soc. London, Ser. B: Biol. Sci. 373, 20170216. Bauer, S., Wennberg Huldt, C., Kanebratt, K.P., Durieux, I., Gunne, D., Andersson, S., et al., 2017. Functional coupling of human pancreatic islets and liver spheroids on-achip: towards a novel human ex vivo type 2 diabetes model. Sci. Rep. 7, 14620. Bell, C.C., Lauschke, V.M., Vorrink, S.U., Palmgren, H., Duffin, R., Andersson, T.B., et al., 2017. Transcriptional, functional, and mechanistic comparisons of stem cell derived hepatocytes, HepaRG cells, and three-dimensional human hepatocyte spheroids as predictive in vitro systems for drug-induced liver injury. Drug Metab. Dispos. 45, 419 429. Bogdan, A.R., Miyazawa, M., Hashimoto, K., Tsuji, Y., 2016. Regulators of iron homeostasis: new players in metabolism, cell death, and disease. Trends Biochem. Sci. 41, 274 286. Bovard, D., Iskandar, A., Luettich, K., Hoeng, J., Peitsch, M.C., 2017. Organs-on-a-chip: a new paradigm for toxicological assessment and preclinical drug development. Toxicol. Res. Appl. 1, 1 16. 239784731772635. Bovard, D., Sandoz, A., Luettich, K., Frentzel, S., Iskandar, A., Marescotti, D., et al., 2018. A lung/liver-on-a-chip platform for acute and chronic toxicity studies. Lab Chip 18, 3814 3829. Cozens, D., Grahame, E., Sutherland, E., Taylor, G., Berry, C.C., Davies, R.L., 2018. Development and optimization of a differentiated airway epithelial cell model of the bovine respiratory tract. Sci. Rep. 8, 853. Derichs, N., Jin, B.-J., Song, Y., Finkbeiner, W.E., Verkman, A.S., 2011. Hyperviscous airway periciliary and mucous liquid layers in cystic fibrosis measured by confocal fluorescence photobleaching. FASEB J. 25, 2325 2332. DiMasi, J.A., Grabowski, H.G., Hansen, R.W., 2016. Innovation in the pharmaceutical industry: new estimates of R&D costs. J. Health Econ. 47, 20 33. Dingsøyr, T., Dyba˚, T., Moe, N.B., 2010. Agile software development: an introduction and overview. In: Dingsøyr, T., Dyba˚, T., Moe, N.B. (Eds.), In Agile Software Development. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 1 13. Garcia-Canton, C., Minet, E., Anadon, A., Meredith, C., 2013. Metabolic characterization of cell systems used in in vitro toxicology testing: lung cell system BEAS-2B as a working example. Toxicol. In Vitro 27, 1719 1727. Gerets, H.H.J., Tilmant, K., Gerin, B., Chanteux, H., Depelchin, B.O., Dhalluin, S., et al., 2012. Characterization of primary human hepatocytes, HepG2 cells, and HepaRG cells at the mRNA level and CYP activity in response to inducers and their predictivity for the detection of human hepatotoxins. Cell Biol. Toxicol. 28, 69 87. Gunness, P., Mueller, D., Shevchenko, V., Heinzle, E., Ingelman-Sundberg, M., Noor, F., 2013. 3D organotypic cultures of human HepaRG cells: a tool for in vitro toxicity studies. Toxicol. Sci. 133, 67 78.

References

Ipema, J., LeCluyse, E., Clewell, H., Andersen, M.E., McKim, J.M., 2014. Evaluation of an Integrated Human Multi-Organ Culture Plate For Predicting Systemic Toxicity. https://iontox.com/wp-content/uploads/2015/10/DMOP-Poster-SOT2015.pdf Ishida, S., 2018. Organs-on-a-chip: current applications and consideration points for in vitro ADME-Tox studies. Drug Metab. Pharmacokinet. 33, 49 54. Jackson, E.L., Lu, H., 2016. Three-dimensional models for studying development and disease: moving on from organisms to organs-on-a-chip and organoids. Integr. Biol. 8, 672 683. Kamei, K., Kato, Y., Hirai, Y., Ito, S., Satoh, J., Oka, A., et al., 2017. Integrated heart/cancer on a chip to reproduce the side effects of anti-cancer drugs in vitro. RSC Adv. 7, 36777 36786. Karp, P.H., Moninger, T.O., Weber, S.P., Nesselhauf, T.S., Launspach, J.L., Zabner, J., et al., 2002. An in vitro model of differentiated human airway epithelia: methods for establishing primary cultures. In Epithelial Cell Culture Protocols. Humana Press, Totowa, NJ, pp. 115 137. Leney, S.E., Tavare´, J.M., 2009. The molecular basis of insulin-stimulated glucose uptake: signalling, trafficking and potential drug targets. J. Endocrinol. 203, 1 18. Lieber, M., Todaro, G., Smith, B., Szakal, A., Nelson-Rees, W., 1976. A continuous tumor-cell line from a human lung carcinoma with properties of type II alveolar epithelial cells. Int. J. Cancer 17, 62 70. Martin, J., 1991. Rapid Application Development. Macmillan Pub. Co, New York [u.a.]. Materne, E.-M., Ramme, A.P., Terrasso, A.P., Serra, M., Alves, P.M., Brito, C., et al., 2015. A multi-organ chip co-culture of neurospheres and liver equivalents for longterm substance testing. J. Biotechnol. 205, 36 46. Montano, M., 2014. Model systems. In Translational Biology in Medicine. Elsevier, pp. 9 33. OECD, 2008. Test No. 407: Repeated Dose 28-day Oral Toxicity Study in Rodents (OECD) Okumura, R., Takeda, K., 2017. Roles of intestinal epithelial cells in the maintenance of gut homeostasis. Exp. Mol. Med. 49, e338. Pezzulo, A.A., Starner, T.D., Scheetz, T.E., Traver, G.L., Tilley, A.E., Harvey, B.-G., et al., 2011. The air-liquid interface and use of primary cell cultures are important to recapitulate the transcriptional profile of in vivo airway epithelia. Am. J. Physiol. Lung Cell. Mol. Physiol. 300, L25 L31. Qin, Q., Wu, Q., Wang, Y., Xiong, R., Guo, L., Fu, X., et al., 2019. Effects of cellular differentiation in human primary bronchial epithelial cells: metabolism of 4-(methylnitrosamine)-1-(3-pyridyl)-1-butanone. Toxicol. In Vitro 55, 185 194. Rayna, T., Striukova, L., 2016. From rapid prototyping to home fabrication: how 3D printing is changing business model innovation. Technol. Forecasting Soc. Change 102, 214 224. Ronaldson-Bouchard, K., Vunjak-Novakovic, G., 2018. Organs-on-a-chip: a fast track for engineered human tissues in drug development. Cell Stem Cell 22, 310 324. Samanez, C.H., Caron, S., Briand, O., Dehondt, H., Duplan, I., Kuipers, F., et al., 2012. The human hepatocyte cell lines IHH and HepaRG: models to study glucose, lipid and lipoprotein metabolism. Arch. Physiol. Biochem. 118, 102 111. Schwaber, K., Beedle, M., 2002. Agile Software Development with Scrum. Prentice Hall, Upper Saddle River, NJ.

505

506

CHAPTER 15 How to build your multiorgan-on-a-chip system

Silva, R.A.C., Gonc¸alves, A.F., dos Santos, P.P., Rafacho, B., Claro, R.F.T., Minicucci, M. F., et al., 2017. Cardiac remodeling induced by all-trans retinoic acid is detrimental in normal rats. Cell. Physiol. Biochem. 43, 1449 1459. Skardal, A., Murphy, S.V., Devarasetty, M., Mead, I., Kang, H.-W., Seol, Y.-J., et al., 2017. Multi-tissue interactions in an integrated three-tissue organ-on-a-chip platform. Sci. Rep. 7, 8837. Sun, W., Luo, Z., Lee, J., Kim, H.-J., Lee, K., Tebon, P., et al., 2019. Organ-on-a-chip for cancer and immune organs modeling. Adv. Healthc. Mater. 8, e1801363. Takahashi, Y., Hori, Y., Yamamoto, T., Urashima, T., Ohara, Y., Tanaka, H., 2015. Threedimensional (3D) spheroid cultures improve the metabolic gene expression profiles of HepaRG cells. Biosci. Rep. 35. Theobald, J., Ghanem, A., Wallisch, P., Banaeiyan, A.A., Andrade-Navarro, M.A., Taˇskova, K., et al., 2018. Liver-kidney-on-chip to study toxicity of drug metabolites. ACS Biomater. Sci. Eng. 4, 78 89. Tinggi, U., 2008. Selenium: its role as antioxidant in human health. Environ. Health Prev. Med. 13, 102 108. Toepke, M.W., Beebe, D.J., 2006. PDMS absorption of small molecules and consequences in microfluidic applications. Lab Chip 6, 1484. Trieu, D., 2014. An Investigation into the Feasibility of Utilizing Fluidic Shear Stresses to Functionally Organize Conducting Airway Epithelium. University of Toronto ˚ ., Honegger, P., Knudsen, van der Valk, J., Brunner, D., De Smet, K., Fex Svenningsen, A L.E., et al., 2010. Optimization of chemically defined cell culture media replacing fetal bovine serum in mammalian in vitro methods. Toxicol. In Vitro 24, 1053 1063. Verkhratsky, A., Matteoli, M., Parpura, V., Mothet, J.-P., Zorec, R., 2016. Astrocytes as secretory cells of the central nervous system: idiosyncrasies of vesicular secretion. EMBO J. 35, 239 257. Vernetti, L., Gough, A., Baetz, N., Blutt, S., Broughman, J.R., Brown, J.A., et al., 2017. Functional coupling of human microphysiology systems: intestine, liver, kidney proximal tubule, blood-brain barrier and skeletal muscle. Sci. Rep. 7, 42296. Viravaidya, K., Sin, A., Shuler, M.L., 2008. Development of a microscale cell culture analog to probe naphthalene toxicity. Biotechnol. Progr. 20, 316 323. Wickert, L., Abiaka, M., Bolkenius, U., Gressner, A.M., 2004. Corticosteroids stimulate selectively transforming growth factor (TGF)-β receptor type III expression in transdifferentiating hepatic stellate cells. J. Hepatol. 40, 69 76. Wixson, J.R., 1999. Function analysis and decomposistion using function analysis systems technique. INCOSE Int. Symp. 9, 800 805. Wong, C.H., Siah, K.W., Lo, A.W., 2019. Estimation of clinical trial success rates and related parameters. Biostatistics 20, 273 286. Yeong, W.-Y., Chua, C.-K., Leong, K.-F., Chandrasekaran, M., 2004. Rapid prototyping in tissue engineering: challenges and potential. Trends Biotechnol. 22, 643 652. Yun, Y.-R., Won, J.E., Jeon, E., Lee, S., Kang, W., Jo, H., et al., 2010. Fibroblast growth factors: biology, function, and application for tissue regeneration. J. Tissue Eng. 1, 218142. Zaidman, N.A., Panoskaltsis-Mortari, A., O’Grady, S.M., 2016. Differentiation of human bronchial epithelial cells: role of hydrocortisone in development of ion transport pathways involved in mucociliary clearance. Am. J. Physiol. Cell Physiol. 311, C225 C236.