282 122 4MB
English Pages 268 [274] Year 2009
METHODS
IN
M O L E C U L A R B I O L O G Y TM
Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK
For other titles published in this series, go to www.springer.com/series/7651
High Throughput Screening Methods and Protocols, Second Edition
Edited by
William P. Janzen University of North Carolina, Chapel Hill, NC, USA and
Paul Bernasconi BASF Corporation, Research Triangle Park, NC, USA
Editors William P. Janzen UNC Eshelman School of Pharmacy Ctr Integrative Chemical Bio & Drug Disc Division of Medicinal Chem & Natural Products University of North Carolina Chapel Hill NC 27599-763 USA [email protected]
Paul Bernasconi BASF Corporation Research Triangle Park NC 27709-3528 USA [email protected]
ISBN 1064-3745 e-ISBN 1940-6029 ISBN 978-1-60327-257-5 e-ISBN 978-1-60327-258-2 DOI 10.1007/978-1-60327-258-2 Springer Dordrecht Heidelberg London New York Library of Congress Control Number: 2009929642 # Humana Press, a part of Springer ScienceþBusiness Media, LLC 2002, 2009 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Printed on acid-free paper Springer is part of Springer ScienceþBusiness Media (www.springer.com)
Preface In the 6 years since the first edition of this book, the field of high-throughput screening (HTS) has evolved considerably. In 2004, the Society for Biomolecular Screening (SBS) celebrated its 10th anniversary. The event and its timing were significant because SBS is the world’s largest association of scientists, engineers and technologists associated with HTS. While the creation of SBS did not mark the birth of HTS by any means, its foundation in 1994 helped HTS find a common voice. It provided a discussion forum and a means to define and enforce standards. In 2006, SBS became the Society for Biomolecular Sciences, underlining the expansion of the members’ interests beyond screening. Like any new technology, HTS went through growth stages. During the initial hype phase of the 1980s and 1990s, HTS, together with chemistry and genomics, was predicted to solve all of the pharmaceutical industry’s pipeline problems. A fundamental change in drug discovery was afoot: the time-consuming physiology or medicinal chemistry experiments would be replaced by a numbers’ game, made possible by screening large, combinatorially generated compound libraries against numerous genomically identified targets. While this approach did (and continues to) deliver, it fell short of the expected revolution, exposing it to criticism from within and outside the industry (1). Learning from its mistakes, the HTS profession entered a period of change marked by an increased integration. The once stand-alone HTS groups matured into an essential, integrated component of the discovery effort. Contrary to the fears of many of our colleagues, HTS did not replace hypothesis-driven research but rather expanded it. In addition, because an HTS campaign is inherently expensive, more effort was expended to insure the quality of the hypothesis. Finally, compounds discovered by HTS enabled the testing of marginal hypotheses, thereby increasing the serendipity role in discovery. To reach this maturity level, the HTS field had to learn to ‘‘play well with others.’’ Of course, robotics, automation engineering, and data handling remain the hallmarks of HTS. But to be truly useful, HTS had to be integrated with the other discovery disciplines: genomics, molecular biology, cell biology, enzymology, pharmacology, and chemistry. Successful discovery starts long before and continues long after an HTS campaign. It also became clear that a large number of tests is not a replacement for quality components. Long gone are the days in which a marginally active target, or a target in a marginally relevant physiological state, is screened against large collections of compounds of questionable quality, diversity, or purity. Success is measured less by the number of compounds screened or by the hit rate and more by the quality of the chemical series entering the clinical pipeline. As a reward, HTS researchers can now point to several marketed drugs whose birth place was a well in a microtiter plate (2). For example, in the breast cancer therapeutic indication alone, three drugs have been introduced, which originated from an HTS campaign and are worth mentioning here: (1) IressaTM, an ATP-competitive inhibitor of the epidermal growth factor receptor tyrosine kinase; (2) sorafenib tosylate or NexavarTM, a specific inhibitor of the kinase Raf-1; and (3) tipifarnib, or ZarnestraTM, an inhibitor of protein farnesyl transferases. v
vi
Preface
With technology comes training. The preface of the first edition described how, at that time, ‘‘Nearly every scientist working in HTS had a unique story for how they came to be there,’’ and that ‘‘All that is changing. Training programs are beginning to appear and the techniques created in HTS are being used more and more frequently in laboratories outside the field.’’ Six years later, reality surpassed even the most optimistic predictions. Over 55 academic screening centers have been created (3), which provide both HTS services and training. Universities have become a major player in this field, educating researchers who, in the past, had to rely on extramural institutions to learn the trade. The National Institute of Health Roadmap, created in 2002, has completed its first phase and created the Molecular Libraries Screening Center Network (MLSCN) as part of the Molecular Libraries Initiative (4). These 10 HTS centers were established as a pilot program to apply HTS techniques in academic research with the overarching goal to ‘‘expand the availability and use of chemical probes to explore the function of genes, cells, and pathways in health and disease and to provide annotated information on the biological activities of compounds contained in the central Molecular Libraries Small Molecule Repository in a public database’’. Historically, serendipity and keen observation of natural events have been the main source for these tools. HTS now allows the systematic search for such probes. In addition, HTS allows a better understanding of the specificity of these compounds, an essential characteristic for their usefulness. While much has changed, the core principles of HTS have largely remained unchanged. Each organization is structurally unique, but all retain key elements: an assay must be developed, a chemical library must be assembled and managed, a screen must be performed, and data must be analyzed. Each of these functions is discussed in this volume. While assembling this new edition, we made a few choices. First, we wanted to remain true to the mission of the first edition: to serve as an introduction to HTS for scientists who are just entering the field, as well as providing enough details to be useful for scientists in established HTS operations. Second, while the HTS field regularly sees the introduction of new screening technologies, we wanted to give the lion’s share of the volume to the well established methods. They are most likely to be widely used by the intended reader. Third, we wanted to give a detailed treatment of the activities that are immediately related to HTS: compound library management, data handling, and robotics. Finally, we purposely left out ancillary methods: natural compound selection, chemical diversity assessment, orthogonal assays, and ADME-Tox issues. These essential tools would have been underserved in this volume. The reader will encounter terminology that is unique to HTS and has unique connotations in this industry. To assist with this problem, the Society of Biomolecular Sciences has assembled a glossary (5). We encourage both experienced ‘‘screeners’’ and those new to the field to review these definitions. We hope this manual will be of use to you and would like to acknowledge the authors who contributed to this manual: not only are they experts in their field, they are also great teachers who wanted to share their knowledge and enthusiasm for HTS. Chapel Hill, NC Research Triangle Park, NC
William P. Janzen Paul Bernasconi
Preface
vii
References 1. Landers, P. (2004) Testing machines were built to streamline research – but may be stifling it. Wall Street Journal, 24 Feb 04 2. Fox, S., Farr-Jones, S., Sopchak, L., Boggs, A., Nicely, H. W., Khoury, R. and Biros, M. (2006) Highthroughput screening: update on practices and success. Journal of Biomolecular Screening, 11: 864–869 3. Society for Biomolecular Sciences website: http://www.sbsonline.org 4. NIH Roadmap website: http://nihroadmap.nih.gov 5. http://www.sbsonline.org/links/terms.php
Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Design and Implementation of High-Throughput Screening Assays . . . . . . . . . . . Ricardo Macarro´ n and Robert P. Hertzberg 2 Creation of a Small High-Throughput Screening Facility . . . . . . . . . . . . . . . . . . . . Tod Flak 3 Informatics in Compound Library Management . . . . . . . . . . . . . . . . . . . . . . . . . . Mark Warne and Louise Pemberton 4 Statistics and Decision Making in High-Throughput Screening . . . . . . . . . . . . . . . Isabel Coma, Jesus Herranz, and Julio Martin 5 Enzyme Assay Design for High-Throughput Screening . . . . . . . . . . . . . . . . . . . . . Kevin P. Williams and John E. Scott 6 Application of Fluorescence Polarization in HTS Assays. . . . . . . . . . . . . . . . . . . . . Xinyi Huang and Ann Aulabaugh 7 Screening G Protein-Coupled Receptors: Measurement of Intracellular Calcium Using the Fluorometric Imaging Plate Reader . . . . . . . . . . . . . . . . . . . . . . . . . . . . Renee Emkey and Nancy B. Rankl 8 High-Throughput Automated Confocal Microscopy Imaging Screen of a Kinase-Focused Library to Identify p38 Mitogen-Activated Protein Kinase Inhibitors Using the GE InCell 3000 Analyzer. . . . . . . . . . . . . . . . . . . . . . . . . . . . O. Joseph Trask, Debra Nickischer, Audrey Burton, Rhonda Gates Williams, Ramani A. Kandasamy, Patricia A. Johnston, and Paul A. Johnston 9 Recent Advances in Electrophysiology-Based Screening Technology and the Impact upon Ion Channel Discovery Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andrew Southan and Gary Clark 10 Automated Patch Clamping Using the QPatch . . . . . . . . . . . . . . . . . . . . . . . . . . . Kenneth A. Jones, Nicoletta Garbati, Hong Zhang, and Charles H. Large 11 High-Throughput Screening of the Cyclic AMP-Dependent Protein Kinase (PKA) Using the Caliper Microfluidic Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . Leonard J. Blackwell, Steve Birkos, Rhonda Hallam, Gretchen Van De Carr, Jamie Arroway, Carla M. Suto, and William P. Janzen 12 Use of Primary Human Cells in High-Throughput Screens . . . . . . . . . . . . . . . . . . Angela Dunne, Mike Jowett, and Stephen Rees Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ix
v xi 1 33 55 69 107 127
145
159
187 209
225
239 259
Contributors JAMIE ARROWAY • GlaxoSmithKline, Collegeville, PA, USA ANN AULABAUGH • Chemical and Screening Sciences, Wyeth Research, Collegeville, PA, USA STEVE BIRKOS • Nanosyn, Durham, NC, USA LEONARD J. BLACKWELL • Wyeth Pharmaceuticals, Sanford, NC, USA AUDREY BURTON • Scynexis, Inc., Research Triangle Park, NC, USA GARY CLARK • BioFocus DPI, Saffron Walden, Essex, UK ISABEL COMA • Molecular Discovery Research, Glaxo SmithKline, Tres Cantos, Madrid, Spain ANGELA DUNNE • Screening and Compound Profiling Department, GlaxoSmithKline, Harlow, Essex, UK RENEE EMKEY • Amgen Inc., Cambridge, MA, USA TOD FLAK • BioAutomatix Consulting, Alameda, CA, USA NICOLETTA GARBATI • Department of Biology, Psychiatry CEDD, Glaxo SmithKline SpA, Verona, Italy RHONDA HALLAM • Independent Consultant JESUS HERRANZ • Molecular Discovery Research, Glaxo SmithKline, Tres Cantos, Madrid, Spain ROBERT P. HERTZBERG • Molecular Discovery Research, GlaxoSmithKline, Collegeville, PA, USA XINYI HUANG • Chemical and Screening Sciences, Wyeth Research, Collegeville, PA, USA WILLIAM P. JANZEN • Assay Development and Compound Profiling, Division of Medicinal Chemistry and Natural Products, Center for Integrative Chemical Biology and Drug Discovery, Eshelman School of Pharmacy, University of North Carolina, Chapel Hill, NC, USA PATRICIA A. JOHNSTON • Discovery Programs, Cellumen, Inc., Pittsburgh, PA, USA PAUL A. JOHNSTON • Department of Pharmacology, University of Pittsburgh School of Medicine, Pittsburgh, PA, USA KENNETH A. JONES • Lundbeck Research, Inc., Paramus, NJ, USA MIKE JOWETT • Screening and Compound Profiling Department, GlaxoSmithKline, Harlow, Essex, UK RAMANI A. KANDASAMY • BASF Corporation, Research Triangle Park, NC, USA CHARLES H. LARGE • Department of Biology, Psychiatry CEDD, Glaxo SmithKline SpA, Verona, Italy • Molecular Discovery Research, GlaxoSmithKline, Collegeville, ´ RICARDO MACARRoN PA, USA JULIO MARTIN • Molecular Discovery Research, Glaxo SmithKline, Tres Cantos, Madrid, Spain
xi
xii
Contributors
DEBRA NICKISCHER • Thermo Fisher Scientific, Pittsburgh, PA, USA LOUISE PEMBERTON • Exelgen Ltd., Bude, Cornwall, UK NANCY B. RANKL • BASF Corp., Research Triangle Park, NC, USA STEPHEN REES • Screening and Compound Profiling Department, GlaxoSmithKline, Harlow, Essex, UK JOHN E. SCOTT • Department of Pharmaceutical Sciences and BRITE, North Carolina Central University, Durham, NC, USA ANDREW SOUTHAN • Ion Channel Biology, BioFocus DPI, Saffron Walden, Essex, UK CARLA M. SUTO • Independent Consultant O. JOSEPH TRASK, JR: • Cellular Imaging Technologies, Duke University Center for Drug Discovery, Durham, NC, USA GRETCHEN VAN DE CARR • Nanosyn, Durham, NC, USA MARK WARNE • Exelgen Ltd., Bude, Cornwall, UK KEVIN P. WILLIAMS • Department of Pharmaceutical Sciences and BRITE, North Carolina Central University, Durham, NC, USA RHONDA G. WILLIAMS • BD Diagnostics – Diagnostic Systems, TriPath, Burlington, NC, USA HONG ZHANG • Lundbeck Research, Inc., Paramus, NJ, USA
Chapter 1 Design and Implementation of High-Throughput Screening Assays ´ and Robert P. Hertzberg Ricardo Macarron Abstract HTS is at the core of the drug discovery process, and so it is critical to design and implement HTS assays in a comprehensive fashion involving scientists from the disciplines of biology, chemistry, engineering, and informatics. This requires careful analysis of many variables, starting with the choice of assay target and ending with the discovery of lead compounds. At every step in this process, there are decisions to be made that can greatly impact the outcome of the HTS effort, to the point of making it a success or a failure. Although specific guidelines should be established to ensure that the screening assay reaches an acceptable level of quality, many choices require pragmatism and the ability to compromise opposing forces. Keywords: HTS process, Assay technology, Biochemical assays, Cellular assays, HTS quality, HTS validation.
1. Introduction to the HTS Process In most pharmaceutical and biotechnology companies, highthroughput screening (HTS) is a central function in the drug discovery process. This has resulted from the fact that there are increasing numbers of validated therapeutic targets being discovered through advances in human genomics, and increasing numbers of chemical compounds being produced through highthroughput chemistry initiatives. Many large companies have over 100 targets in their pipeline at any given time, and lead compounds must be found to progress these targets. In some cases we know enough about the target and can apply knowledge-based approaches to hit discovery such as focused screening and structure-based design. However in many cases, particularly for more novel targets, there is limited knowledge about the types W.P. Janzen, P. Bernasconi (eds.), High Throughput Screening, Methods and Protocols, Second Edition, vol. 565 ª Humana Press, a part of Springer Science+Business Media, LLC 2009 DOI 10.1007/978-1-60327-258-2_1, Springerprotocols.com
1
2
´ and Hertzberg Macarron
of compounds that may interact with the protein. As such, pharmaceutical companies often rely on HTS as the primary engine driving lead discovery. The HTS process is a subset of the drug discovery process and can be described as the phase from Target to Lead. This phase can be broken down in the following steps: Target Choice
Reagent Procurement Screening Collections
Assay Development and Validation HTS Implementation Data Capture, Storage and Analysis Leads
It is critically important to align the target choice and assay method to ensure that a biologically relevant and robust screen is configured. The assay must be configured correctly so that compounds with the desired biological effect will be found if they exist in the screening collection. The assay must demonstrate low variability and high signal to background so that false negatives and false positives are minimized. The screen must have sufficient throughput and low cost to enable screening of large compound collections. To meet these requirements, organizations must ensure that communication between therapeutic departments, assay development groups, and screening scientists occurs early – as soon as the target is chosen – and throughout the assay development phase. Reagent procurement is often a major bottleneck in the HTS process. This can delay the early phases of assay development – e.g., when active protein cannot be obtained – and also delay HTS implementation if scale-up of protein or cells fails to produce sufficient reagent to run the full screen. For efficient HTS operation, there must be sufficient reagent available to run the entire screening campaign before HTS can start. Otherwise, the campaign will need to stop halfway through and the screening robots will have to be reconfigured for other work. Careful scheduling between reagent procurement departments and HTS functions is critical to ensure optimum use of robotics and personnel. Modern HTS laboratories have borrowed concepts from the manufacturing industry to smooth the flow of targets through the hit discovery process (e.g., supply chain management, constrained workin-progress, and statistical quality control) and these ideas have begun to pay off with higher productivity and shorter lead times. Successful HTS implementation is multidisciplinary and requires close alignment of computational chemists directing the synthesis or the acquisition of compound collections, sample
Design and Implementation of High-Throughput Screening Assays
3
management specialists maintaining and distributing screening decks, technology specialists responsible for setting up and supporting HTS automation, biologists and biochemists with knowledge of assay methodology, IT personnel capable of collecting and analyzing large datasets, and medicinal chemists capable of examining screening hits to look for patterns that define lead series. Through the marriage of these diverse specialties, therapeutic targets can be put through the lead discovery engine called HTS and lead compounds will emerge.
2. Choice of Therapeutic Target There are three major considerations for choosing a therapeutic target destined for HTS: target validity (i.e., disease relevance), chemical tractability, and screenability. Disease relevance is the most important consideration and also the most complex. Since there is an inverse relationship between target novelty and validity, organizations should choose a portfolio of targets, which span the risk spectrum. Some targets will have a high degree of validation but low novelty (fast follower targets) and others will be highly novel but poorly linked to disease. Target validity can be assessed with genetic approaches and/or compound-based experiments. Genetic approaches such as gene knockouts or RNAi can be time-consuming and sometimes lead to false conclusions but can be performed without the need for expensive screening. Compound-based target validation approaches require taking a risk with less-validated targets and spending money to screen for tool compounds, followed by cell-based or in vivo experiments. Both approaches have their advantages and disadvantages, and most organizations use a combination. However, many fail to fully analyze the economics of this equation. Efforts to reduce the cost and increase the success rate of HTS can shift the equation in favor of running screens for targets on the less-validated end of the spectrum. While disease relevance should be the primary consideration when choosing a target, one should also consider technical factors important to the HTS process. Chemical tractability considerations relate to the probability that drug-like compounds capable of producing the therapeutically relevant effect against a specific target are present in the screening collection and can be found through screening. Years of experience in HTS within the industry have suggested that certain target classes are more chemically tractable than others, including G protein-coupled receptors (GPCRs), ion channels, nuclear hormone receptors, and kinases. On the other side of the spectrum, targets that work via protein–protein interactions have a lower probability of being successful in HTS campaigns. One reason for this is the fact that compound libraries often do not contain compounds of sufficient
4
´ and Hertzberg Macarron
size and complexity to disrupt the large surface of protein–protein interaction that is encountered in these targets. Natural products are one avenue that may be fruitful against protein–protein targets, since these compounds are often larger and more complex than those in traditional chemical libraries. The challenge for these targets is finding compounds that have the desired inhibitory effect and also contain drug-like properties (e.g., are not too large in molecular weight). Recently, several groups have had success with protein–protein interactions by screening for small fragments that weakly inhibit the interaction and building them up to produce moderate-sized potent inhibitors. Certain subsets of protein–protein interaction targets have been successful from an HTS point of view. For example, chemokine receptors are technically a protein–protein interaction (within the GPCR class) and there are several examples of successful lead compounds for targets in this class (1). Similarly, certain integrin receptors that rely on small epitopes (i.e., RGD sequences) have also been successful at producing lead compounds (2). There may be other classes of tractable protein–protein interactions that remain undiscovered due to limitations in compound libraries. Based on the thinking that chemically tractable targets are easier to inhibit, most pharmaceutical companies have concentrated much of their effort on these targets and diminished work on more difficult targets. While this approach has some merits, one should be careful not to entirely eliminate target classes that would otherwise be extremely attractive from a biological point of view. Otherwise, the prophecy of chemical tractability will be self-fulfilled, since today’s compound collections will not expand into new regions and we will never find leads for more difficult, biologically relevant targets. There is clearly an important need for enhancing collections by filling holes that chemical history has left open. The challenge is filling these holes with drug-like compounds that are different from the traditional pharmacophores of the past. This is critical if we are to increase HTS success rates (proportion of targets which give starting points for medicinal chemistry) from the current 60% (3,4) to 80% or higher. A final factor to consider when choosing targets is screenability – the technical probability of developing a robust and high-quality screening assay. The impact of new assay technologies has made this less important, since there are now many good assay methods available for a wide variety of target types (see Section 3). Nevertheless, some targets are more technically difficult than others. Of the target types mentioned above, GPCRs, kinases, proteases, nuclear hormone receptors, and protein–protein interactions are often relatively easy to establish as screens. Ion channels are more difficult, although new technologies are being developed, which make these more approachable from an HTS point of view (5). Enzymes other than kinases and proteases must be considered on a case-by-case basis depending on the nature of the substrates involved.
Design and Implementation of High-Throughput Screening Assays
5
The reductionist approach in which a single target is hypothesized to be important for disease carries some risks. Have you chosen an irrelevant or intractable target? What if a combination of targets is required to elicit the desired biological effect? An alternative approach gaining favor is the use of phenotypic and/or pathway assays for hit discovery. Phenotypic assays, sometimes called ‘‘black box’’ assays, measure a cellular property in response to test compound. Examples include secretion of protein factors, chemotaxis, apoptosis, and cell shape change. Pathway assays are more precise in that protein properties within a cellular pathway are measured. Examples include intracellular protein phosphorylation and cellular trafficking. Often a combination of these approaches can be used to turn a ‘‘black box’’ into a ‘‘gray box.’’ An advantage of phenotypic and pathway assays is the fact that multiple targets are screened at once, providing multiple chances for compounds to ‘‘find’’ the most tractable and biologically relevant target(s) in the cell. However, phenotypic assays are more difficult to configure and more expensive to run, and hit deconvolution to define the specific target(s) of your hits is time-consuming and complex. Furthermore, provision of relevant cells is difficult but recent advances in human stem cells are beginning to alleviate this problem. All of these factors must be considered on a case-by-case basis and should be evaluated at the beginning of a Target-to-Lead effort before making a choice to go forward. Working on an expensive and technically difficult assay must be balanced against the degree of validation and biological relevance. While the perfect target is chemically tractable, technically easy, inexpensive, and biologically relevant, such targets are rare. The goal is to work on a portfolio that spreads the risk among these factors and balances the available resources.
3. Choice of Assay Method There are usually several ways of looking for hits of any given target. The first and major choice to make is between a biochemical and a cell-based assay. By biochemical we understand an assay developed to look for compounds that interact with an isolated target in an artificial environment. This was the most popular approach in the early 1990 s, the decade in which HTS became a mature and central area of drug discovery. This bias toward biochemical assays for HTS was partly driven by the fact that cellbased assays were often more difficult to run in high throughput. However, advances in technology and instrumentation for cellbased assays that translated to commercial products around the early 2000 s, together with disappointments in the success rates of molecular-based hit discovery campaigns, changed the tilt toward cell-based HTS. Among these advances are the emergence of
6
´ and Hertzberg Macarron
HTS-compatible technology to measure G protein-coupled receptor (GPCR) (6) and ion channel function (5), confocal imaging platforms for rapid cellular and subcellular imaging, and the continued development of reporter gene technology. In a recent survey (3), HTS labs reported a 50/50 split between biochemical and cell-based assays in 2006, with a projection to be 60% cell-based, 40% biochemical in 2008. For most drug discovery programs, both types of assays are required for hit discovery and characterization and subsequent lead optimization. Everything being equal (technical feasibility, cost, and throughput), cell-based assays are often preferred for HTS because compounds tested will be interacting with a more realistic mix of protein target conformations in their physiological milieu, i.e., with the right companions (proteins, metabolites, etc.) at the right concentration. Additionally, cell-based assays tend to avoid some common artifacts in biochemical assays such as aggregators (7). On the other hand, cell-based assays may identify hits that do not act on the target or the pathway of interest and may miss hits of interest that do not penetrate the cell membrane. If a cell-based assay is chosen for primary screening, a biochemical assay will often be used as a secondary screen to characterize hits and guide lead optimization. A wide variety of assay formats is now available at relatively affordable prices to cope with most needs in the HTS labs. The following sections provide a very succinct summary of some of the most popular choices. A recent comprehensive review by Inglese and coworkers is recommended for further reading (8). 3.1. Biochemical Assay Methods
While laborious separation-based assay formats such as radiofiltration and ELISAs were common in the early 1990 s, most biochemical screens today use simple homogeneous ‘‘mix-and-read’’ formats. This is particularly true for HTS assays run in industrial labs that are conducted on high-density microtiter plates (384 or 1536 wells). The most common assay readouts used in biochemical assay methods for HTS are optical, including absorbance, fluorescence, luminescence, and scintillation. Among these, fluorescence-based techniques are amongst the most important detection approaches used for HTS (9). Fluorescence techniques give very high sensitivity, which allows assay miniaturization, and are amenable to homogeneous formats. One factor to consider when developing fluorescence assays for screening compound collections is wavelength; in general, short excitation wavelengths (especially those below 400 nm) should be avoided to minimize interference produced by test compounds. Although fluorescence intensity measurements have been successfully applied in HTS, this format is mostly applied to a narrow range of enzyme targets for which fluorogenic substrates are available. A more widely used fluorescence readout is time-resolved fluorescence resonance energy transfer (TR-FRET) (10). This is a
Design and Implementation of High-Throughput Screening Assays
7
dual-labeling approach based upon long-range energy transfer between fluorescent Ln3+ complexes and a suitable resonance energy acceptor. These approaches give high sensitivity by reducing background and a large number of HTS assays have now been configured using TR-FRET. This technique is highly suited to measurements of protein–protein interactions and has also been tailored to detect important metabolites such as cAMP. Another versatile fluorescence technique is florescence polarization (FP), which can be used to measure bimolecular association events (10). Immobilized metal-ion affinity-based FP (IMAP, Molecular Devices) (11) is a variation of FP that can be applied to test activity of kinases and other enzymes. Radiometric techniques such as scintillation proximity assay (SPA, GE) (12) used to be very common in the 1990 s. Despite advances in imaging and bead technology that enabled faster readouts and reduced the occurrence of optical interferences, radiometric assays have several disadvantages including safety and limited reagent stability. In recent years, these techniques have been displaced by fluorescence assay technologies; current estimates from various surveys of HTS laboratories indicate that radiometric assays presently constitute around 5% of all screens performed. Other technologies able to circumvent technical hurdles for niche difficult assays are amplified luminescence proximity (AlphaScreen, Perkin Elmer) (13), electrochemiluminescence (ECL, Meso Scale Discovery) (14), fluorescence correlation spectroscopy (FCS), and other confocal techniques (9). Label-free assays are a diverse set of techniques of growing interest and demand. Many of the methods are modern adaptation to the high-throughput environment of well-established technologies such as mass spectroscopy or calorimetry. An overview of the commercial solutions in place and their principles has been recently published by Rich and Myszka (15). 3.2. Cell-Based Assay Methods
As recently as the mid-1990s, most cell-based assay formats were not consistent with HTS requirements. However, as recent technological advances have facilitated higher throughput functional assays, cell-based formats now make up a reasonable proportion of screens performed today. One of the most important advances in cell-based assay methodology is the development of the FLIPR1 (MDS Analytical Technologies), a fluorescence imaging plate reader with integrated liquid handling that facilitates the simultaneous fluorescence imaging of 384 samples to measure intracellular calcium mobilization in real time (6). This format is now commonly used for GPCR and ion channel targets. Based on the success of the FLIPR1, several additional cell-based assays for GPCRs were developed. One useful technology uses the photoprotein aequorin to measure intracellular calcium levels. When aequorin binds to calcium, it oxidizes coelenterazine with the
8
´ and Hertzberg Macarron
emission of light, which can be easily measured on a suitable plate reader. Another important cell-based assay method involves the measurement of intracellular cAMP levels, which allows the screening of Gi- and Gs-coupled GPCRs. Technologies for cAMP measurement include the older RIA, ELISA, and SPA methods as well as recent techniques such as TR-FRET, amplified luminescence proximity (AlphaScreen1), and enzyme fragment complementation (EFC, HitHunterTM), which are less expensive and have higher throughput (16). Significant advances in ion channel screening have occurred over the past decade (17). Calcium-sensing dyes read on FLIPR1 are commonly used to measure channels that conduct calcium, while voltage-sensing dyes are used to track changes in membrane potential. An important advance in high-throughput ion channel assays was the development of FRET-based voltage-sensing dyes, where a pair of molecules exhibit FRET, which is disrupted when the membrane is depolarized. Ion flux assays using nonradioactive tracers analyzed by atomic absorbance spectroscopy (AAS) can now be run in HTS format using recently available instrumentation. And the standard for measuring ion channel activity, patch clamp measurements, has been facilitated by the development of automated instrumentation such as Ionworks Quattro, PatchXpress, and QPatch (18). While these technologies are remarkable, further improvements are necessary before patch clamp measurements can be used for primary HTS. Cellular phenotypes and pathways are now routinely measured using a variety of techniques amenable to HTS. The reporter gene assay is the oldest and most well-studied method, which allows the discovery of compounds that modulate a pathway resulting in changes in gene expression (19). This method offers certain advantages relative to other cell-based assays, in that it requires fewer cells, is easier to automate, and can be performed in 1536 well plates. Descriptions of miniaturized reporter gene readouts include luciferase – undoubtedly, the most popular reporter gene (20) – secreted alkaline phosphate, and beta-lactamase. However, reporter gene assays are of relatively low resolution since they measure effects on an entire pathway at once. Recent advances in cellular imaging have allowed HTS of higher resolution phenomena such as intracellular protein redistribution, GPCR internalization, and other cellular pathway events. Methods using protein complementation assays and bioluminescence resonance energy transfer (BRET) combined with cellular imaging can be very useful (8). Another cellular phenotype that is commonly measured in HTS is protein secretion. Classical methods to measure protein secretion such as RIA and ELISA are being replaced by improved techniques such as AlphaLISATM (21) and MultiArray1 or MultiSpot1 electrochemiluminescence-based solutions (14).
Design and Implementation of High-Throughput Screening Assays
3.3. Matching Assay Method to Target Type
9
Often, one has a choice of assay method for a given target type (Table 1.1). To illustrate the various factors that are important when choosing an assay type, let us consider the important GPCR target class. GPCRs can be screened using cell-based assays such as FLIPR, aequorin, and reporter gene or biochemical formats such as SPA and FP. One overriding factor when choosing between functional or binding assays for GPCRs is whether one seeks to find agonists or antagonists. Functional assays are much more amenable to finding agonists than are binding assays, while antagonists can be found with either format. FLIPR assays are relatively easy to develop, but this screening method is more labor-intensive (particularly with respect to cell culture requirements) and more difficult to automate than reporter gene assays. In contrast, the need for longer term incubation time for reporter gene assays (4–6 h vs. min for FLIPR) means that cytotoxic interference by test compounds may be more problematic. On the plus side, reporter gene readouts for GPCRs can sometimes be more sensitive to agonists than FLIPR. Aequorin offers some advantages of FLIPR while being easier to run and less expensive. Regarding biochemical assays for GPCRs, SPA remains a common format since radiolabeling is often facile and nonperturbing. However, fluorescence assays for GPCRs such as FP and FIDA are becoming more important. Fluorescent labels are more stable, safer, and often more economical than radiolabels. However,
Table 1.1 The most important assay formats for various target types are shown Assay formats Target type
Biochemical
Cell-based
GPCRs
SPA, FP, FIDA
FLIPR, reporter gene, aequorin, TR-FRET, AlphaScreen, EFC, cell imaging
Ion channels
SPA, FP
FLIPR, FRET, AAS, automated patch clamp
Nuclear hormone receptor
FP, TR-FRET, SPA, AlphaScreen
Reporter gene, cell imaging
Kinases
FP, TR-FRET, SPA, IMAP
Cellular phosphorylation, cell imaging
Protease
FLINT, FRET, TR-FRET, FP, SPA
Reporter gene, cell imaging
Other enzymes
FLINT, FRET, TR-FRET, FP, SPA, absorbance
Protein– protein
TR-FRET, FRET. BRET, SPA, ECL, AlphaScreen
BRET, cell imaging, reporter gene
10
´ and Hertzberg Macarron
while fluorescent labeling is becoming easier and more predictable, these labels are larger and thus can sometimes perturb the biochemical interaction (in either direction). These examples illustrate some of the trade-offs one needs to consider when choosing an assay type. In general, one should choose the assay format that is easiest to develop, most predictable, most relevant, and cheapest to run. These factors, however, are not always known in advance. And even worse, they can be at odds with each other and thus must be balanced to arrive at the best option. Additional important quality considerations include compound interference issues and assay variability. It makes little sense to run a cheap and easy assay that is variable or overly sensitive to inhibition. In some cases it makes sense to parallel track two formats during the assay development phase and choose between them based on which is easiest to develop and most facile. Finally, in addition to these scientific considerations, logistical factors such as the number of specific readers or robot types available in the HTS lab and the queue size for these systems must be taken into account.
4. Assay Development and Validation
4.1. Critical Biochemical Parameters in HTS Assays
The final conditions of an HTS assay are chosen following the optimization of quality without compromising throughput, while keeping costs low. The most critical points that must be considered in the design of a high-quality assay are biochemical data and statistical performance. Assay optimization is often required to achieve acceptable HTS performance while keeping assay conditions within the desired range. This usually significantly improves the stability and/or the activity of the biological system studied and has therefore become a key step in the development of screening assays (22). The success of an HTS campaign in finding hits with the desired profile depends primarily on the presence of such compounds in the collection tested. But it is also largely dependent on the ability of the researcher to engineer the assay in accordance with that profile while reaching an appropriate statistical performance. A classical example that illustrates the importance of the assay design is how substrate concentration determines the sensitivity for different kinds of enzymatic inhibitors. If we set the concentration of one substrate in a screening assay at 10 times Km, competitive inhibitors of that enzyme–substrate interaction with a Ki greater than one-eleventh of the compound concentration used
Design and Implementation of High-Throughput Screening Assays
11
in HTS will show less than 50% inhibition and will likely be missed – i.e., competitive inhibitors with a Ki of 0.91 mM or higher would be missed when screening at 10 mM. On the other hand, the same problem will take place for uncompetitive inhibitors if substrate concentration is set at one-tenth of its Km. Therefore, it is important to know what kind of hits are sought in order to make the right choices in substrate concentration; often, one chooses a substrate concentration that facilitates discovery of both competitive and uncompetitive inhibitors. In this section, we describe the biochemical parameters of an assay that have a greater influence on the sensitivity of finding different classes of hits and some recommendations about where to set them.
4.1.1. Enzymatic Assays
The sensitivity of an enzymatic assay to different types of inhibitors is a function of the ratio of substrate concentration to Km (S/Km).
4.1.1.1. Substrate Concentration l
Competitive inhibitors: for reversible inhibitors that bind to a binding site that is the same as one substrate, the more of that substrate present in the assay, the less inhibition observed. The relationship between IC50 (compound concentration required to observe 50% inhibition of enzymatic activity with respect to an uninhibited control) and Ki (inhibition constant) is (23): IC50 ¼ ð1 þ S=KmÞ Ki
As shown in Fig. 1.1, at S/Km ratios less than 1 the assay is more sensitive to competitive inhibitors, with an asymptotic limit of IC50 ¼ Ki. At high S/Km ratios, the assay becomes less suitable for finding this type of inhibitors. l Uncompetitive inhibitors: if the inhibitor binds to the enzyme–substrate complex or any other intermediate complex but not to the free enzyme, the dependence on S/Km is the opposite to what has been described for competitive binders. The relationship between IC50 and Ki is (23): IC50 ¼ ð1 þ Km=SÞ Ki High substrate concentrations make the assay more sensitive to uncompetitive inhibitors (Fig. 1.1). l Noncompetitive (allosteric) inhibitors: if the inhibitor binds with equal affinity to the free enzyme and to the enzyme–substrate complex, the inhibition observed is independent of the substrate concentration. The relationship between IC50 and Ki is (23):
´ and Hertzberg Macarron 12
10
8 IC50/Ki ratio
12
Competitive inhibitor Uncompetitive inhibitor Non-competitive inhibitor
6
4
2
0 0.1
1 S/Km ratio
10
Fig. 1.1. Variation of IC50/Ki ratio with the S/Km ratio for different types of inhibitors. At [S] ¼ Km, IC50 ¼ 2Ki for competitive and uncompetitive inhibitors. For noncompetitive inhibitors IC50 ¼ Ki at all substrate concentrations.
IC50 ¼ Ki l
Mixed inhibitors: if the inhibitor binds to the free enzyme and to the enzyme–substrate complex with different affinities (Ki1 and Ki2, respectively), the relationship between IC50 and Ki is (24): IC50 ¼ ðS þ KmÞðKil þ S=Ki2Þ
In summary, setting the substrate(s) concentration(s) at the Km value is an optimal way of ensuring that all types of inhibitors exhibiting a Ki close to or below the compound concentration in the assay can be found in an HTS campaign. Nevertheless, if there is a specific interest in favoring or avoiding a certain type of inhibitor, then the S/Km ratio would be chosen considering the information provided above. For instance, many ATP-binding enzymes are tested in the presence of saturating concentrations of ATP to minimize inhibition from compounds that bind to the ATP-binding site. Quite often the cost of one substrate or the limitations of the technique used to monitor enzymatic activity (Table 1.2) may preclude setting the substrate concentration at its ideal point. As in many other situations found while implementing an HTS assay, the screening scientist must consider all factors involved and look for the optimal solution. For instance, if the sensitivity of a detection technology requires setting S ¼ 10 Km to achieve an acceptable signal to background, competitive
Design and Implementation of High-Throughput Screening Assays
13
Table 1.2 Examples of limitations to substrate concentration imposed by some popular assay technologies. These limitations also apply to ligand in binding assays or other components in assays monitoring any kind of binding event Assay technology
Limitations
Fluorescence
Inner filter effect at high concentrations of fluorophore (usually >1 mM)
Fluorescence polarization
>30% substrate depletion required
Capture techniques (ELISA, SPA, FlashPlate, BET, others)
Concentrations of the reactant captured must be in alignment with the upper limit of binding capacity
Capture techniques and anyone monitoring binding
Nonspecific binding (NSB) of the product or of any reactant to the capture element (bead, plate, membrane, antibody, etc.) may result in misleading activity determinations
All
Sensitivity limits impose a lower limit to the amount of product detected
inhibitors with a Ki greater than one-eleventh of the compound concentration tested will not likely be identified and will limit the campaign to finding more potent inhibitors. In this case, working at a higher compound concentration would help to find some of the weak inhibitors otherwise missed. If this is not feasible, it is better to lose weak inhibitors while running a statistically robust assay, rather than making the assay more sensitive by lowering substrate concentration to a point of unacceptable signal to background. The latter approach is riskier since a bad statistical performance would jeopardize the discovery of more potent hits (see Section 4.3). 4.1.1.2. Enzyme Concentration
The accuracy of inhibition values calculated from enzymatic activity in the presence of inhibitors relies on the linear response of activity to the enzyme concentration. Therefore, an enzyme dilution study must be performed in order to determine the linear range of enzymatic activity with respect to enzyme concentration. As shown in Fig. 1.2 for valyl-tRNA synthetase, at high enzyme concentrations there is typically a loss of linearity due to substrate depletion, protein aggregation, or limitations in the detection system. If the enzyme is not stable at low concentrations, or if the assay method does not respond linearly to product formation or substrate depletion, there could also be a lack of linearity in the lower end. In addition, enzyme concentration marks a lower limit to the accurate determination of inhibitor potency. IC50 values lower than one-half of the enzyme concentration cannot be measured;
14
´ and Hertzberg Macarron 2,500
Product formed (CPM)
2,000
1,500
1,000
500
0 0
1
2
3
4
5
[VRS] nM
Fig. 1.2. Protein dilution curve for valyl-tRNA synthetase. The activity was measured after 20 min incubation following the SPA procedure described (25).
this effect is often referred to as ‘‘bottoming out.’’ As the quality of compound collections improves, this could be a real problem since SAR trends cannot be observed among the more potent hits. Obviously, enzyme concentration must be kept far below the concentration of compounds tested in order to find any inhibitor. In general, compounds are tested at micromolar concentrations (1–100 mM) and as a rule of thumb, it is advisable to work at enzyme concentrations below 100 nM. On the other hand, the assay can be made insensitive to certain undesired hits (such as inhibitors of enzymes added in coupled systems) by using higher concentrations of these proteins. In any case, the limiting step of a coupled system must be the one of interest, and thus the auxiliary enzymes should always be in excess. 4.1.1.3. Incubation Time and Degree of Substrate Depletion
As described above for enzyme concentration, it is important to assess the linearity vs. time of the reaction analyzed. HTS assays are often end-point and so it is crucial to select an appropriate incubation time. Although linearity vs. enzyme concentration is not achievable if the end-point selected does not lie in the linear range of the progress curves for all enzyme concentrations involved, exceptions to this rule do happen, and so it is important to check it as well. To determine accurate kinetic constants, it is crucial to measure initial velocities. However, for the determination of acceptable inhibition values, it is sufficient to be close to linearity. Therefore, the classical rule found in Biochemistry textbooks of working at or below 10% substrate depletion [e.g. (26)] does not necessarily apply to HTS assays. Provided that all compounds in a
15
100
50
80
40
60
30
40
20 % S depleted uninhibited reaction
20
% Inhibition observed
% S depleted
Design and Implementation of High-Throughput Screening Assays
10
% S depleted with 50 % enzyme inhibition % I observed
0 0
1000
2000 3000 Time (arbitrary units)
4000
0 5000
Fig. 1.3. Theoretical progress curves at S ¼ Km of an uninhibited enzymatic reaction and a reaction with an inhibitor at its IC50 concentration. The inhibition values determined at different end-points throughout the progress curve are shown as well. Initial velocities are represented by dotted lines.
collection are treated in the same way, if the inhibitions observed are off by a narrow margin, it is not a problem. As shown in Fig. 1.3, at 50% substrate depletion with an initial substrate concentration at its Km, the inhibition observed for a 50% real inhibition is 45%, an acceptable error. For higher inhibitions the errors are lower (e.g., instead of 75% inhibition, 71% would be observed). At lower S/Km ratios the errors are slightly higher (e.g., at S ¼ 1/10 Km, a 50% real inhibition would yield an observed 42% inhibition, again at 50% substrate depletion). This flexibility to work under close-to-linearity but not truly linear reaction rates makes it feasible to use certain assay technologies in HTS – e.g., fluorescence polarization – that require a high proportion of substrate depletion in order to produce a significant change in signal. Secondary assays configured within linear rates should allow a more accurate determination of IC50 s for hits. In reality, the experimental progress curve for a given enzyme may differ from the theoretical one depicted here for various reasons such as non Michaelis–Menten behavior, reagent deterioration, product inhibition, and detection artifacts. In view of the actual progress curve, practical choices should be made to avoid missing interesting hits. 4.1.1.4. Order of Reagent Addition
The order of addition of reactants and putative inhibitors is important to modulate the sensitivity of an assay for slow binding and irreversible inhibitors.
16
´ and Hertzberg Macarron
A preincubation (usually 5–10 min) of enzyme and test compound favors the finding of slow-binding competitive inhibitors. If the substrate is added first, these inhibitors have a lower probability of being found. In some cases, especially for multisubstrate reactions, the order of addition can be engineered to favor certain uncompetitive inhibitors. For instance, a mimetic of an amino acid that could act as an inhibitor of one aminoacyl-tRNA synthetase will exhibit a much higher inhibition if preincubated with enzyme and ATP before addition of the amino acid substrate. 4.1.2. Binding Assays
Although this section is focused on receptor binding, other binding reactions (protein–protein, protein–nucleic acid, etc.) are governed by similar laws, and so assays to monitor these interactions should follow the guidelines hereby suggested.
4.1.2.1. Ligand Concentration
The equation that describes binding of a ligand to a receptor, developed by Langmuir to describe adsorption of gas films to solid surfaces, is virtually identical to the Michaelis–Menten equation for enzyme kinetics: BL ¼ Bmax L=ðKd þ LÞ where BL ¼ bound ligand concentration (equivalent to v0), Bmax ¼ maximum binding capacity (equivalent to Vmax), L ¼ total ligand concentration (equivalent to S), and Kd ¼ equilibrium affinity constant also known as dissociation constant (equivalent to Km). Therefore, all equations disclosed in Section 4.1.1.1 can be directly translated to ligand-binding assays. For example, for competitive binders IC50 ¼ ð1 þ L=KdÞ Ki Uncompetitive binders cannot be detected in binding assays; functional assays must be performed to detect this inhibitor class. Allosteric binders could be found if their binding modifies the receptor in a fashion that prevents ligand binding. Typically, ligand concentration is set at the Kd concentration as an optimal way to attain a good signal (50% of binding sites occupied). This results in a good sensitivity for finding competitive binders.
4.1.2.2. Receptor Concentration
The same principles outlined for enzyme concentration in Section 4.1.1.2 apply to receptor concentration, or concentration of partners in other binding assays. In most cases, especially with membrane-bound receptors, the nominal concentration of receptor is not known but can be determined by measuring the proportion of bound ligand at the Kd. In any case, linearity of response (binding) with respect to receptor (membrane) concentration should be assessed.
Design and Implementation of High-Throughput Screening Assays
17
In traditional radiofiltration assays, it was recommended to set the membrane concentration so as to reach at most 10% of ligand bound at the Kd concentration, i.e., the concentration of receptor present should be below one-fifth of Kd (27). Although this is appropriate to get accurate binding constants, it is not absolutely required to find competitive binders in a screening assay. Some formats (FP, SPA in certain cases) require a higher proportion of ligand bound to achieve acceptable statistics, and receptor concentrations close or above the Kd value have to be used. Another variable to be considered in ligand-binding assays is nonspecific binding (NSB) of the labeled ligand. NSB increases linearly with membrane concentration. High NSB leads to unacceptable assay statistics, but this can often be improved with various buffer additives (see Section 4.2). 4.1.2.3. Preincubation and Equilibrium
As discussed for enzymatic reactions, a preincubation of test compounds with the receptor would favor slow binders. After the preincubation step, the ligand is added and the binding reaction should be allowed to reach equilibrium in order to ensure a proper calculation of displacement by putative inhibitors. Running binding assays at equilibrium is convenient for HTS assays, since one does not have to carefully control the time between addition of ligand and assay readout as long as the equilibrium is stable.
4.1.3. Cell-Based Assays
The focus of the previous sections has been on cell-free systems. Cell-based assays offer different challenges in their setup with many built-in factors that are out of the scientist’s control. Nevertheless, some of the points discussed above apply to them, mutatis mutandi. One of the most important considerations is cell type. The most physiologically relevant cells are primary human cells, but these are very difficult and expensive to procure. Recent advances in stem cell science are beginning to facilitate the provision of cells for HTS that are closer to the primary human cell. However, recombinant cells remain the most commonly used cell type for HTS. Important considerations when developing cell-based assays include the following (22, 28): l Cell culture details should be well documented and reproducible. Most problems with cell-based assays can be traced to problems with the cells. l
Consider using cryopreserved cells as an assay source to reduce variability and improve screening scheduling logistics.
l
Adherent cells or suspension cells can be used, and the choice is based on the cell type and the assay readout method. In general, try to mimic the physiological conditions as much as possible while considering assay logistics.
18
´ and Hertzberg Macarron
4.2. Assay Optimization
l
Either stable cell lines or transient transfection can be used. Expression levels of the recombinant protein(s) should be confirmed. Extremely high expression levels should generally be avoided.
l
Consider using modified baculovirus (BacMam virus) gene delivery technology for transient expression of target proteins in mammalian cells (29).
l
When using stable cell lines, use early passages to avoid cells losing their responsiveness.
l
Lower numbers of cells are preferred for cost reasons, but at least 1000 cells per well should generally be used to minimize stochastic single-cell events. The response observed should be linear with respect to the number of cells.
l
Pay attention to cell clumps which can cause variability.
l
Preincubation of cells with compounds should be considered when applicable (e.g., assays in which a ligand is added).
l
Optimal incubation time should be selected in accordance with the rule of avoiding underestimation of inhibition or activation values (see Section 4.1.1.3). All other factors being equal, shorter incubation times minimize cytotoxic interference problems.
l
Cell-based assays tend to be more sensitive to DMSO than biochemical assays. Determine the DMSO sensitivity of the assay and configure the protocol to remain well below this level.
l
Use standard inhibitors and/or activators during the screening run to confirm the desired signal is observed.
l
Pay attention to edge effects, which occur commonly in cellbased assays due to problems with incubators or uneven cell distribution of cells in the well. Incubating seeded plates at room temperature before placing them in the incubator can help this problem (30).
In vitro assays are performed in artificial environments in which the biological system studied could be unstable or exhibiting an activity below its potential. The requirements for stability are higher in HTS campaigns than in other areas of research. In HTS runs, diluted solutions of reagents are used throughout long periods of time (typically 4–12 h) and there is a need to keep both the variability low and the signal to background high. Additionally, several hundreds of thousands of samples are usually tested, and economics often dictates one to reduce the amount of reagents required. In this respect, miniaturization of assay volumes has been in continuous evolution, from tubes to 96-well plates to 384-well plates to 1536 and beyond. Many times, converting assays from
Design and Implementation of High-Throughput Screening Assays
19
low-density to high-density formats is not straightforward. Thus, in order to find the best possible conditions for evaluating an HTS target, optimization of the assay should be accomplished as part of the development phase. HTS libraries contain synthetic or natural compounds that in most cases are dissolved in DMSO. The tolerance of the assay to DMSO must be considered. Typically, compounds are stored at concentrations ranging between 1 and 30 mM. Test compound concentrations in primary screening are in the 1–30 mM range. Therefore, DMSO concentrations from 0 to 10% are tested. It is critical to work at DMSO concentrations in a region of minimal variation, as otherwise compound effects can be obscured by variability in the addition of compound stocks (typically the smallest volume in the assay mix and thus the most sensitive liquid handling step). If significant decrease in activity/binding is observed at the standard solvent concentration – typically 0.5–1% (v/v) DMSO – lower test compound concentrations may be required. In some cases the detrimental effect of solvent can be circumvented by optimizing assay conditions. In all cases, key biochemical parameters (e.g., Km) should be checked in the final assay conditions (DMSO concentration) before starting the screening campaign. The stability of reagents should be tested using the same conditions intended for HTS runs, including solvent concentration, stock concentration of reagents, reservoirs, and plates. Quite often signal is lost with time not because of degradation of one biological partner in the reaction but because of its adsorption to the plastics used (reservoir, tips, or plates) (Fig. 1.4). Addition of detergents below their critical micellar concentration (CMC) and/or carrier proteins (e.g., BSA) is a common technique to minimize this undesirable phenomenon. These assay components can also aid in reducing nonspecific enzymatic inhibition caused by the aggregation of test compounds (7). The number of factors that can be tested in an optimization process is immense. Nevertheless, initial knowledge of the system (optimal pH, metal requirements, sensitivity to oxidation, etc.) can help to select the most appropriate ones. Factors to be considered can be grouped as follows: l Buffer composition l
pH
l
Temperature
l
Ionic strength
l
Osmolarity
l l l
Monovalent ions ðNaþ ; K þ ; C1 Þ Divalent cations Mn2þ ; Mg2þ ; Ca2þ ; Zn2þ ; Cu2þ ; Co2þ
Rheological modulators (glycerol, polyethylene glycol)
´ and Hertzberg Macarron 20
Enzyme activity (mOD/min)
20
15
10
5
0 0
10
20 30 40 Preincubation time (min)
50
60
Fig. 1.4. Example of loss of signal in an enzymatic reaction related with adsorption of enzyme (or substrate) to plasticware. The data are from a real assay performed in our lab. Stability of reagents was initially measured using polypropylene tubes and 384-well polystyrene plates, without CHAPS (circles). Once HTS was started, using polypropylene reservoirs and polystyrene 384-well plates (triangles), a clear loss of signal was observed. Addition of 0.01% (w/v) CHAPS not only solved the problem but also improved the enzyme activity (squares). Reactions were initiated at 10, 30, and 50 min after preparation of diluted stocks of reagents that remained at 4C before addition to the reaction wells.
l
Polycations (heparin, dextran)
l
Carrier proteins (BSA, casein)
l
Chelating agents (EDTA, EGTA)
l
Blocking agents (PEI, milk powder)
l
Reducing agents (DTT, b-mercaptoethanol)
l
Protease inhibitors (PMSF, leupeptin)
Detergents (Triton, Tween, CHAPS). Cell-based assays are usually conducted in cell media of complex formulation. Factors to be considered in this case are mainly medium, supplier, selection and concentration of extra protein (human serum albumin, BSA, gelatin, and collagen). One also needs to take into account cell density, plate type, plate coatings, incubation time, temperature, and atmosphere. Since cell-based assays generally have more variables than biochemical assays, extreme care must be taken when documenting and reproducing the cell culture and assay conditions. Besides analyzing the effect of factors individually, it is important to consider interactions between factors because synergies and antagonisms can commonly occur (31). Full-factorial or partialfactorial designs can be planned using several available statistical l
Design and Implementation of High-Throughput Screening Assays
21
packages (e.g. JMP, Statistica, Design Expert). Experimental designs result in quite complex combinations as soon as more than four factors are tested. This task becomes rather complicated in high-density formats when taking into consideration that more reliable data are obtained if tests are performed randomly. Therefore, an automated solution is necessary because manually running an experiment of this complexity would be extremely difficult. Several commercial packages exist that integrate design of experiments and necessary liquid handling steps to conduct the experiments. A good example is AAO (automated assay optimization) developed by Beckman Coulter (Fullerton, CA) in collaboration with scientists from GlaxoSmithKline (32). An example of the outcome of one assay improved in our lab using this methodology is shown in Fig. 1.5. The paper by Taylor et al. (32) describes examples of assay optimization through AAO for several types of targets and assay formats. A typical optimization process starts with a partial-factorial design including many factors (20). The most promising factors are then tested in a full-factorial experiment to analyze not only main effects but also two-factor interactions. These experiments are done with two levels per factor (very often one level is the absence of the ingredient and the other is the presence at a fairly typical concentration). Finally, titrations of the more beneficial factors are conducted in order to find optimal concentrations of every component. Usually the focus of optimization is on activity (signal or signal to background), but statistical performance should also be taken into account when doing assay optimization. Though this is not feasible when many factors and levels are scrutinized without replicates, whenever possible duplicates or triplicates should be run and the resulting variability measured for every condition. Some buffer ingredients make a reproducible dispensement very difficult, and so should be used only if they are really beneficial (e.g., glycerol). For some factors it is critical to run the HTS assay close to physiological conditions (e.g., pH) in order to avoid missing interesting leads for which the chemical structure or interaction with the target may change as a function of that factor. 4.3. Statistical Evaluation of HTS Assay Quality
The quality of an HTS assay must be determined according to its primary goal, i.e., to distinguish accurately hits from nonhits in a vast collection of samples. In the initial evaluation of assay performance, several plates are filled with positive controls (signal; e.g., uninhibited enzyme reaction) and negative controls or blanks (background; e.g., substrate without enzyme). Choosing the right blank is sometimes not so obvious. In ligand–receptor-binding assays, the blanks referred to as NSB controls are prepared traditionally by adding an excess of
´ and Hertzberg Macarron
22
Totals NP Plot 4200
H B
Expected Normal Value
3200
C A G BH BG DH FH AC CF FG AH DF AE DE AG BE GH E CE EG AF CD DG EF CH AB BD CG BF EH F AD
2200 1200 200 –800 –1800 –2800 –3800
D BC
– 0 – –3 00 200 100 0 0 0
1 3 5 7 6 4 8 9 10 2 1 00 000 000 000 000 000 000 000 000 000 100 0 0 Standardized Effects
A 25000
Predicted activity (CPM)
20000 +H +B +C +H +B –C 15000
+H –B +C +H –B –C –H +B +C –H +B –C
10000
–H –B +C –H –B –C 5000
0 0
5000
10000
15000
20000
25000
30000
Observed activity (CPM)
B Fig. 1.5. Example of optimization of a radiofiltration assay using Beckman Coulter’s AAO program and a Biomek 2000 to perform the liquid handling. The target was to increase activity of this enzyme, bacterial biotin-ligase, aiming to improve assay quality and reduce costs. The initial partial-factorial test included 20 factors, 8 of which were identified as positive. The test shown in this figure used these eight factors and was designed as a two-level full-factorial experiment with duplicates. Five hundred and twelve samples were generated. (A) The probability plot resulting from the statistical analysis of experimental data showed three factors being positive (H, B and C) although the interaction of B and C was
Design and Implementation of High-Throughput Screening Assays
23
unlabelled (cold) ligand; the resulting displacement could be unreachable for some specific competitors that would not prevent nonspecific binding of the labeled ligand to membranes or labware. A better blank could be prepared with membranes from the same cell line not expressing the receptor targeted. Though this is not always practical in the HTS context, it should be at least tested in the development of the assay and compared with the NSB controls to which they should be ideally pretty close. A careful analysis of these control plates allows identifying errors in liquid handling or sample processing. For instance, an assay with a long incubation typically produces plates with edge effects due to faster evaporation of the external wells even if lids are used, unless the plates are placed in a chamber with humidity control. Analysis of patterns (per row, per column, per quadrant) helps to identify systematic liquid-handling errors. Obvious problems must be solved before evaluating the quality of the assay. After troubleshooting, random errors are still expected to happen due to instrument failure or defects in the labware used. They should be included in the subsequent analysis of performance (removing outliers is a misleading temptation equivalent to hiding the dirt under the rug). The analysis of performance can be accomplished by several means. Graphical analysis helps to identify systematic errors (e.g., Fig. 1.6). The statistical analysis of raw data involves the calculations of a number of parameters, starting with mean (M) and standard deviations (SD) for signal and background, and combinations of these are as follows: l Signal to background S=B ¼ Msignal =Mbackground S/B provides an indication of the separation of positive and negative controls. It can be useful in early assay development to understand the potential of an assay format or to validate reagents under development. But it is a poor indicator of assay quality as it is independent of variability (33). l
Coefficient of variation of signal and background CV ¼ 100 SD=M ð%Þ A relative measure of variability provides a good indication of variability. Variability is a function of the assay stability and the precision of liquid handling and detection instruments.
Fig. 1.5. (continued) negative. D showed significant negative effect, while the other four factors had statistically marginal or no effect. (B) Applying the statistical model, the correlation between observed and predicted values was very good. The presence of H ¼ CHAPS 0.03% (w/v) (+H, –H) is clearly positive. In the absence of B ¼ 125 mM Bicine (+B squares, –B triangles) and C ¼ 125 mM TAPS (+C, –C), the enzyme was less active. The original conditions yielded 5,000 CPM vs. 25,000 CPM with the optimized buffer (backgrounds were 100 CPM in all cases).
´ and Hertzberg Macarron
24
Activity (mOD/min)
18 16
16–18
14
14–16 12–14
12
10–12
10
8–10
8
6–8
6
4–6
4
2–4 0–2
2 0
Row
22
24
18
M
20
Column
16
I
14
10
12
6
8
2
4
A E
14
14
12
12
12 10 8 6 4
Activity (mOD/min)
14
Activity (mOD/min)
Activity (mOD/min)
A 16
10 8 6 4 2
2 0
8 6 4 2
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
10
0 A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
1
2
3
Column
Row
Quadrant
B
C
D
4
Fig. 1.6. Graphical analysis of a 384-well plate of positive controls of an enzymatic reaction monitored by absorbance (continuous readout). The plate was filled using a pipettor equipped with a 96-well head and indexing capability. (A) 3-D plot of the whole plate showing that four wells (I1, I2, J1, and J2) had a dispensement problem. The corresponding tip may have been loose or clogged. Analysis by columns (B), rows (C), and quadrants (D) reveals that the fourth quadrant was receiving less reagent.
l
Signal to noise S ¼ Msignal Mbackground Þ=SDbackground
This classic expression of S/N provides an incomplete combination of signal window and variability. Its original purpose was to assess the separation between signal and background in a radio signal (33). It should not be used to evaluate performance of HTS assays. Another parameter referred to as S/N by some authors is
Design and Implementation of High-Throughput Screening Assays
25
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2ffi SDsignal þ SDbackground S=N ¼ Msignal Mbackground
This second expression provides a complete picture of the performance of an HTS assay but as discussed below, the field has converged in using Z0 as the standard measure of HTS assay quality.
l
Z0 factor Z0 ¼ 1 3 ðSDsignal þ SDbackground Þ=Msignal Mbackground j
Since its publication in 1999 (33), the Z0 factor has been widely accepted by the HTS community as a very useful way of assessing the statistical performance of an assay (34). Z0 is an elegant combination of signal window and variability, the main parameters used in the evaluation of assay quality. The relationship between Z0 factor and S/B is not obvious from its definition but can be easily derived as Z 0 ¼ 1 0:03 S=Bj CV Signal þ CV Background ðjS=Bj 1Þ
The value of Z0 factor is a relative indication of the separation of the signal and background populations. It is assumed that there is a normal distribution for these populations, as it is the case if the variability is due to random errors. Z0 factor is a dimensionless parameter that ranges from 1 (infinite separation) to 0.4, although in practice the majority of our assays demonstrate Z0 > 0.6. A Z0 of 0.4 is equivalent to having an S/B of 3 and a CV of 10%. Low variability allows for a lower S/B, but a minimum of 2 is usually required, provided that CVs are rarely below 5%. Figure 1.7 shows Z0 at work in three different scenarios. Full analysis of the corresponding data is collected in Table 1.3. Z0 should be evaluated during assay development and validation, and also throughout HTS campaigns on a per plate basis to assess the quality of dispensement and reject data from plates with errors. Chapter 5 describes in more detail the different tools used to assess statistical performance in HTS campaigns. 4.4. Assay Validation
Once an assay optimized to find compounds of interest passes its quality control with a Z0 greater than 0.6 (or whatever is the applied acceptance criteria), a final step must be done before starting an HTS campaign. The step referred to here as assay validation consists of testing a representative sample of the screening collection in the same way HTS plates will be treated; i.e., on the same robotic system using protocols identical to the HTS run. The purposes of this study are to
´ and Hertzberg Macarron Blanks
Positive controls
70 60 50 # cases
40 30
Plate 1 Plate 2 Plate 3
20 10
15
13
Enzymatic activity (mOD/min)
14
12
11
9
10
7
8
6
4
5
2
3
0
0 1
26
Fig. 1.7. Distribution of activity values (bins of 0.5 mOD/min) for three 384-well plates half-filled with blanks and half-filled with positive controls of an enzymatic reaction monitored by absorbance (continuous readout). Z0 factors were 0.59 for plate 1, 0.42 for plate 2, and 0.10 for plate 3. A complete analysis of performance is shown in Table 1.3.
Table 1.3 Statistical analysis of data from the three plates described in Fig. 1.7 Parameter
Plate 1
Plate 2
Plate 3
Msignal (mOD/min)
10.09
7.77
5.84
SDsignal (mOD/min)
0.84
0.81
0.96
Mbckg (mOD/min)
0.30
0.69
0.74
SDbckg (mOD/min)
0.51
0.57
0.57
S/B
34
11
8
SW (mOD/min)
9.80
7.08
5.09
S/N1
19
12
9
10.0
7.1
4.6
CVsignal (%)
8%
10%
16%
CVbckg (%)
173%
82%
77%
Z0 factor
0.59
0.42
0.10
S/N
1
2
S=N ¼ ðMsignal Mbackground Þ=SDbackground qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 S=N ¼ ðMsignal Mbackground Þ= ðSDsignal Þ2 þðSDbackground Þ2
Design and Implementation of High-Throughput Screening Assays
27
l
obtain production data on assay performance;
l
assess interferences from screening samples;
l
evaluate the reproducibility of results obtained in a production environment;
l
estimate the hit rate and determine optimal sample concentration.
A dramatic example of how the test of a pilot collection helps to detect interferences is shown in Fig. 1.8. This target, HCV RNA-dependent RNA polymerase, has been found to be 140 120
Number of cases
100 80
Without BSA With 0.05 % BSA
60 40 20
>100
90
100
80
70
60
50
40
30
20
0
10
–10
0.3 strong patterns
4.2.3. Screening Data Correction
l
VEP of 0.2–0.3 medium patterns
l
VEP < 0.2 smooth patterns, or lack of pattern
In some cases, after detecting the patterns, we can correct the experimental conditions to make some of these patterns disappear or decrease in intensity. However, this is not always possible and sometimes we need to correct the original data, in order to make the best decisions about which compounds are active. The basic idea of correction is to take the distance to the surface as a new value of compound activity. In practice, this means adding or subtracting a quantity to the original data, depending on whether the well is positioned in a high or low area of the surface. Corrrc ¼ Uncorrrc þWeightrc ðmedianðpatternÞPatternrc Þ8r;c where r and c are row and column identifiers, respectively. This weight function will take values near to 1 for all the low activity values and near to 0 for high activity. The use of a weight function compensates for the lack of linearity in the response.
4.3. Temporal Patterns
The use of automation in HTS is the main cause of systematic errors across time, associated with a certain position in several consecutive plates. For example, these problems relate to compound dispensing, obstructed pipettes, contaminated wells, etc. The presence of temporal systematic errors principally affects falsepositive findings, because inactive compounds are misclassified as active compounds. The statistical techniques that we use to detect this kind of systematic error are different from those used to detect the spatial
90
Coma, Herranz, and Martin
patterns, but the mathematical foundation is similar, since we use smoothers based on medians as defined by Tukey (24), which can estimate the tendency of data coming from a temporal series. This inter-plate analysis is based on studying the data of each position on the plate, and each well, as an independent temporal series. In other words, in an HTS run of p plates with n samples each, we study n temporal series with p point each. The idea is to adjust a very robust temporal series to find strong trends in one of these series, describing what happens in each position of the plate across time. Systematic error level (SEL) can be defined by applying a robust smoother. Normally, we use estimators such as the 11RH or the 15RH (24), which are stronger than those advised by Tukey (4253H, 3RSSH), since our objective is to find a strong trend that is maintained in several consecutive plates and thus identify it as a systematic error. If a position on a plate does not present a systematic error, and assuming that the majority of the compounds are inactive, we can hope that the data will be randomly distributed below or above 0% of the activity. In this case, when we fit to a robust smoother, the trend of the series (i.e. SEL) should be close to 0%. Figure 4.9 shows different examples of the evolution of the activity in a well across time. Figure 4.9A shows a high systematic error, where all values are around 100% of activity in a sequence of 100 consecutive plates. We imagine that this will be the case in the majority of the wells. However, Fig. 4.9B shows a position of the plate where the values are around 20% of activity, and this could be classified as a low systematic error. Finally, Fig. 4.9C shows a position of the plate without systematic error, during 100 consecutive plates, when the values are randomly below or above 0% activity, and the fitted trend is near to 0% in all cases.
400
450
500
Plate Sequence
550
Activity
–20 0
20 40 60 80 100
20 40 60 80 100
C Systematic Error Level
–20 0
Activity
–20 0
Activity
B Systematic Error Level
20 40 60 80 100
A Systematic Error Level
600
650
700
Plate Sequence
750
400
450
500
550
Plate Sequence
Fig. 4.9. Evolution of the response values in a well along a set of consecutive plates, and the systematic error level (SEL). (A) A well with a high systematic error (SEL 100%). (B) A well with a medium systematic error (SEL 45%). (C) A well without systematic error (SEL 0%). Dots correspond to experimental activity. Curve line is the calculated activity upon fitting of the temporal sequence to the SEL robust smoother.
Statistics and Decision Making in High-Throughput Screening
91
After calculating the SEL for all the plate positions, the data analysis is focused on detecting errors in the set of hits. 4.4. Summary and Conclusions
5. Statistically Guided Selection of Hits in HTS
The presence of spatial and temporal patterns in HTS data can be a significant problem in some assays. Exploratory techniques based on repeated running median has been found to be an effective tool for detecting and correcting these systematic errors. The method has been found to be very flexible and robust for dealing with them in all HTS where they have been used. We have validated the method, and it shows a great capacity to detect false positives and recover false negatives.
Besides a sound scientific rationale, an appropriate compound collection and a high-quality execution, the process of data mining is a key to success of every screening campaign. At the end of the day, HTS is a number game. In order to make the screening valuable, the vast amount of data and information that is gathered in any HTS blitz has to be conveniently processed, interpreted and transformed into real and meaningful information and knowledge. We assume here that in primary HTS every compound is tested at the same concentration and just once. Screening in replicates (11) or at different compound concentrations, such as quantitative HTS (qHTS) described by Inglese et al. (26), can certainly help reduce the identification as hits of compounds associated with assay errors. It should be noted though that some claims of diminishing the burden in false negatives and false positives are based on generic assumptions, such as fixed activity threshold or 3 and 6 SD of the mean of actives being commonly used cut-offs. As will be reviewed below, the field has moved away from simple cut-offs, and high-quality assays are routinely run. Therefore, some of these claims should be downplayed. The decision point relates to which compounds from a single shot test will be pushed forward as positives in the screen. Although the selection of hits can be guided biologically (e.g. potency and profiling of positives through secondary assays for specificity, selectivity or enablers) and chemically (e.g. clustering into representative chemotypes and deselecting intractable structures), the first question is merely statistical, i.e. where to set the threshold of activity that best segregates true positives from true negatives? On the other hand, the screening scientist struggles with a logistical constraint: the number of samples selected cannot surpass a limit dictated by the maximum number of samples that can be reasonably (that is in a timely and resource-efficient manner) prepared and tested through the subsequent assays.
92
Coma, Herranz, and Martin
It is not uncommon for primary positives in HTS simply to be selected on the basis of potency above a particular cut-off of activity that accommodates logistics. The ultimate consequence is that some putatively weak, but still valuable, true hits may have to be abandoned. However, the assignment of potency from a single shot experiment is rather risky. First, reliability of the activity value depends much upon the quality of the assay and the distribution of activity of the sample population tested. In other words, the same threshold of activity does not have the same reliability meaning in all assays. Second, the actual concentration of the compound in the assay might significantly differ from the nominal value (27), so apparently weak compounds may turn out to be potent if they were actually tested as a trace. In all, a hit selection process that minimises the rates of false positives and negatives, regardless of their level of activity, would eventually optimise the use of limited screening resources. False positives are annoying. False negatives are unacceptable, because they are usually abandoned forever. Highly potent positives that eventually turn out to be false are worthless, disappointing and can negatively bias the selection. On the other hand, true weak positives are valid for SAR and provide a starting point in a hit-to-lead chemical programme. Our preference is giving the highest chances of picking up weak hits, accepting the risk of progressing with them false positives that will be unveiled in subsequent stages. The process of hit identification involves three steps: (1) validation, removal and adjustment of screening data (see Section 4), (2) ranking compounds by activity and (3) setting a meaningful threshold to declare positive compounds. Below, we will describe statistical approaches to address the two last steps. 5.1. What Is a Statistical Cut-Off?
Currently, the most commonly used method for hit selection in HTS experiments is statistical significance (or p-value) for testing no mean difference and in particular the mean – k SD method and its variants, where SD is the standard deviation of the negative reference (i.e. background controls or inactive samples) and k is a multiplying scalar. Alternatively, methods of clustering have been described on the basis of finding two statistically significant clusters of samples, namely active and inactive samples (28). The statistical significance approach addresses the question of controlling the rate of false positives, also known as type I error or . The value of k is chosen so that the false-positive rate (i.e. ) can be kept below a certain level. The higher the value of k, the more stringent the cut-off we set to lower the rate of false positives, but the higher the rate of false negatives (also known as type II error or b), and vice versa. Hence, it is a hard challenge to make both rates low at the same time, and there is always a trade-off between minimising the two types of error.
Statistics and Decision Making in High-Throughput Screening
93
A simple way of estimating the probability of making true assignments for negative compounds at one particular threshold of compound activity is the calculation of the power of an assay (19) (Fig. 4.10). The power parameter is calculated as the complementary probability of the type II error (i.e. 1 b) of the assay:
Fig. 4.10. Meaning of power of discrimination for an assay, alpha- and beta-errors.
0
1
B jS C j C 1 ffiC Power ¼ 1 ¼ B @rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi A ðÞ ðSD2S Þ ðSD2C Þ nS þ nC where is the cumulative distribution function of the standard normal distribution (i.e. (x) represents the area under the standard normal distribution between minus infinity and x), n is the number of replicates, SD is the standard deviation, is the mean of activity in the assay, S is the population of samples and C is the population of controls or inactive samples. Zhang et al. (29) and Fogel et al. (30) have developed predictive models of hit confirmation rates based merely on the primary HTS data obtained from compound testing in singlet. Recently, a new statistical parameter (SSMD, strictly standardised mean difference) has been introduced that also contemplates falsenegative rate in an attempt to achieve a balanced control of both (31). The idea of power is underlying SSMD, which can also be applied to quality control of HTS assays. We will focus later in this chapter on the description of methodologies that estimate the lowest threshold with statistical significance for the call of true positives. Although there may be labs where HTS is run in simultaneous multiple testing (e.g. qHTS (11, 26)), we will not discuss statistical approaches based on replicate testing (2, 11) because these data are not ordinarily available in routine HTS.
94
Coma, Herranz, and Martin
As a result of an HTS campaign, we obtain the distributions of three populations: signal controls, background controls and samples. The distribution of sample population in an HTS experiment can be modelled as a composition of one major population of inactive compounds (>95%) with a single mean and experimental variability, and a combination of many other minor populations corresponding to several active compounds (95% vs. 95%) of the whole population, we can conceptually approach the problem through two different routes: (1) sample population is just one single homogenous population containing outlier observations and (2) sample populations can be modelled as an overlapping sum of normal distributions from inactive samples (>95%) and hits ( 1 represents negative cooperativity (the inhibitor binds weaker to the receptor–probe complex than to the free receptor), while a < 1 represents positive cooperativity (the inhibitor binds stronger to the complex than to the free receptor). R is the receptor, L is the probe, and I is the inhibitor.
140
Huang and Aulabaugh
Fraction probe bound
1.0 0.8 0.6 0.4 0.2 0.0
0
1000
2000
3000
4000
5000
Compound (nM)
Fig. 6.7. Theoretical plots of competitive (.), uncompetitive (r), and noncompetitive (a ¼ 1 (.), 3 (*), and 0.33 (&)) inhibitors. The receptor and probe concentrations are 53.7 and 10 nM, respectively. The Kd for the probe is 20 nM. The Ki for compounds in cases of competitive and uncompetitive mechanisms is 100 nM, while the Ki values in cases of noncompetitive mechanism are 100 and 100a nM.
fraction of ligand bound with increasing inhibitor concentration. In contrast, an uncompetitive compound increases the fraction of bound ligand, resulting in a higher polarization value. The displacement of probe by a noncompetitive compound is more complex and is dependent upon the degree of cooperatively. When the compound reduces (a > 1), increases (a < 1), or has no effect on probe affinity (a ¼ 1), the resulting fraction bound will decrease, increase, or not change, respectively (Fig. 6.7). 2.4. Limitations of Steady-State FP Measurements
As with any technique, fluorescence polarization has its share of limitations. General FP complications include interactions between the fluorescent probe and the compound, compound aggregation, scattered light, sample turbidity, plate or buffer polarization, compound fluorescence, fluorescence quenching due to various factors, and instrument detector saturation. Probe–target (or compound– target) interactions may also not fully mimic the native interactions, in cases including labeled probes having altered affinity/binding mode relative to unlabeled counterparts; using mutant enzymes in place of active enzymes; multiple enzyme conformations; and enzymes with multiple substrates. In addition, nonideal FP data can result from steady-state FP measurements of systems that possess multiple fluorescence lifetimes and/or multiple rotational correlation times and thus complex exponential decays, which may be accurately determined only by time-resolved FP measurements. Finally, because the FP assay signal window does not change significantly with the G factor and the starting polarization value is in practice set arbitrarily, the standard S/B parameter no longer indicates the robustness of the assay. For FP assays, the meaningful statistical parameters are assay signal window (mP) and Z0 .
FP-Based HTS Assays
2.5. Summary
141
Fluorescence polarization is a powerful technique for the study of biomolecular interactions in solution and has been widely used in biochemical high-throughput screens. This chapter reviewed three representative steady-state FP applications and good practices in the development and execution of FP-based HTS assays. After hits are obtained from FP-based HTS assays, a good practice is to always confirm the hits in a secondary assay. For direct FP competition-binding assays, an orthogonal functional assay may be used. For FP detection assays, a second functional assay that employs a different detection method is recommended.
3. Notes 1. Fluorescence polarization measurements on commercial fluorescence plate readers follow one-photon excitation, which has a maximum anisotropy of 0.4. Anisotropy values higher than 0.4 indicate misalignment in the instrument or the presence of scattered light. Excitation with two photons or multiple photons by picosecond or femtosecond laser sources uses different photoselection processes, which can lead to a maximum anisotropy greater than 0.4 (21). 2. The common practice is to set the instrument to 27 mP for 1 nM fluorescein and to subtract the background fluorescence (buffer only) from the IS and IP values as the background fluorescence is often polarized. We recommend adjusting the G factor of the instrument such that the probe mP is between 50 and 100 mP instead. Empirical results have shown that a FP assay signal window (probe bound minus probe free) does not vary significantly with the G factor. A higher initial mP for the probe can avoid situations in screens where polarization values are close to zero or turn negative. When the assay fluorescence intensity S/B is low ( 50, background effects are minimized and background subtraction is not required.
142
Huang and Aulabaugh
Acknowledgments The authors would like to thank Ray Unwalla of Wyeth Research for molecular modeling of E-76 and E-76 based probes; Rebecca Shirk and Belew Mekonnen of Wyeth Research for collaboration on the TF/FVIIa; Shannon Stahler, Nina Kadakia, Gary Kalgaonkar, William Martin, Mariya Gazumyan, Pedro Sobers, and Jim LaRocque of Wyeth Research for contribution to the NR project; and Richard Harrison of Wyeth Research for critical review of the chapter. References 1. Perrin, F. (1926) Polarisation de la lumie`re de fluorescence. Vie moyenne des mole´cules dans l’e´tat excite. J. Phys. Radium V, Ser.6 7, 390–401. 2. Jameson, D. M. (2001) The seminal contributions of Gregorio Weber in modern fluorescence spectroscopy. In: New Trends in Fluorescence Spectroscopy, Springer-Verlag, Heidelberg, pp. 35–53. 3. Jameson, D. M., and Sawyer, W. H. (1995) Fluorescence anisotropy applied to biomolecular interactions. Methods in Enzymol. 246, 283–300. 4. Checovich, W. J., Bolger, R. E., and Burke, T. (1995) Fluorescence polarization-a new tool for cell and molecular biology. Nature 375, 141–144. 5. Terpetschnig, E., Szmacinski, H., and Lakowicz, J. R. (1997) Long-lifetime metalligand complexes as probes in biophysics and clinical chemistry. Methods in Enzymol. 278, 295–321. 6. Hill, J. J., and Royer, C. A. (1997) Fluorescence approaches to study of protein-nucleic acid complexation. Methods in Enzymol. 278, 390–416. 7. Kakehi, K., Oda, Y., and Kinoshita, M. (2001) Fluorescence polarization: Analysis of carbohydrate-protein interactions. Anal. Biochem. 297, 111–116. 8. Lakowicz, J. R. (1999) Fluorescence anisotropy. In: Principals of Fluorescence Spectroscopy, second edition (Lakowicz, J. R.), Kulwer Academic/Plenum Publishers, New York, pp. 291–319. 9. Perrin, F. (1929) La fluorescence des solutions. Induction mole´culaires. Polarisation et dure´e d’e´mission. Photochimie. Ann. Phys. Ser. 10 12, 169–275. 10. Perrin, F. (1931) Fluorescence. Dure´e e´le´mentaire d’e´mission lumineuse. Confe´rences
11.
12.
13.
14.
15.
16.
17.
18.
19.
d’Actualite´s Scientifiques et Industrielles XXII, 2–41. Lakowicz, J. R. (1999) Introduction to fluorescence. In: Principals of Fluorescence Spectroscopy, second edition (Lakowicz, J. R.), Kulwer Academic/Plenum Publishers, New York, pp. 1–23. Cantor, C. R., and Schimmel, P. R. (1980) P. R. Biophysical Chemistry Part II: Techniques for the Study of Biological Structure and Function, W. H. Freeman, pp. 454–465. Lakowicz, J. R. (1999) Time-dependent anisotropy decays. In: Principals of Fluorescence Spectroscopy, second edition (Lakowicz, J. R.), Kulwer Academic/Plenum Publishers, New York, pp. 321–345. Mackman, N. (2004) Role of tissue factor in homeostasis, thrombosis and vascular development. Arterioscler. Thromb. Vasc. Biol. 24, 1015–1022. Shirk, R. A. and Vlasuk, G. P. (2007) Inhibitors of FactorVIIa/Tissue Factor. Arterioscler. Thromb. Vasc. Biol. 27, 1895–1900. Dennis, M. S., Eigenbrot, C., Skelton, N. J., Ultsch, M. H., Santell, L., Dwyer, M. A., O’Connell M. P., and Lazarus, R. A. (2000) Peptide exosite inhibitors of factor VIIa as anticoagulants. Nature 404, 465–470. Huang, X. (2003) Fluorescence polarization competition assay: The range of resolvable inhibitor potency is limited by the affinity of the fluorescent ligand. J. Biomol. Screening 8, 34–38. Huang, X. (2003) Equilibrium competition binding assay: Inhibition mechanism from a single dose response. J. Theor. Biol. 225, 369–376. Turconi, S., Shea, K., Ashman, S., Fantom, K., Earnshaw, D. L., Bingham, R. P., Haupts, U. M., Brown, M. J. B., and Pope, A. (2001)
FP-Based HTS Assays Real experiences of uHTS: A prototypic 1536-well fluorescence anisotropy-based uHTS screen and application of well-level quality control procedures. J. Biomol. Screening 6, 275–290. 20. Wu, G., Yuan, Y., and Hodge, C. N. (2003) Determining appropriate substrate conversion for enzymatic assays in high-
143
throughput screening. J. Biomol. Screening 8, 694–700. 21. Lakowicz, J. R., Gryczynski, I., Gryczynski, Z., and Danielsen, E. (1992) Time-resolved fluorescence intensity and anisotropy decays of 2,5-diphenyloxazole by two-photon excitation and frequency-domain fluorometry. J. Phys. Chem. 96, 3000–3006.
Chapter 7 Screening G Protein-Coupled Receptors: Measurement of Intracellular Calcium Using the Fluorometric Imaging Plate Reader Renee Emkey and Nancy B. Rankl Abstract G protein-coupled receptors (GPCRs) are the target of approximately 40% of all approved drugs and continue to represent a significant portion of drug discovery portfolios across the pharmaceutical industry. As a result, GPCRs are the focus of many high-throughput screening (HTS) campaigns. Historically, ligand-binding assays were used to identify compounds that targeted GPCRs. Current GPCR drug discovery efforts have moved toward the utilization of functional cell-based assays for HTS. Many of these assays monitor the accumulation of a second messenger such as cAMP or calcium in response to GPCR activation. Calcium stores are released from the endoplasmic reticulum when Gaq-coupled GPCRs are activated. Although Gai- and Gas-coupled receptors do not normally result in this mobilization of intracellular calcium, they can often be engineered to do so by expressing a promiscuous or a chimeric Gaprotein, which couples to the calcium pathway. Thus calcium mobilization is a readout that can theoretically be used to assess activation of all GPCRs. The fluorometric imaging plate reader (FLIPR) has facilitated the ability to monitor calcium mobilization in the HTS setting. This assay format allows one to monitor activation and inhibition of a GPCR in a single assay and has been one of the most heavily utilized formats for screening GPCRs. Key words: GPCR, FLIPR, Heterotrimeric G proteins, Calcium mobilization, HTS.
1. Introduction In the early days of GPCR drug discovery, there was limited choice with respect to assay formats and researchers relied primarily on radioligand-binding assays. The advances in understanding GPCR biology coupled with the introduction of new assay technologies have permitted the implementation of functional cell-based assays W.P. Janzen, P. Bernasconi (eds.), High Throughput Screening, Methods and Protocols, Second Edition, vol. 565 ª Humana Press, a part of Springer Science+Business Media, LLC 2009 DOI 10.1007/978-1-60327-258-2_7, Springerprotocols.com
145
146
Emkey and Rankl
in HTS campaigns (1). Perhaps the most widely used functional assay for screening GPCRs has been the measurement of calcium mobilization. This assay is applicable for Gaq-coupled GPCRs because this signaling pathway culminates in the release of calcium stores from the endoplasmic reticulum into the cytosol. Gai- and Gas-coupled GPCRs do not normally signal through this pathway, but they can be engineered to do so by expressing chimeric (2, 3) or promiscuous (4) Ga proteins. The basic principle of the calcium mobilization assay is to load cells with a calcium-sensitive dye, which fluoresces when bound to calcium; therefore, an increased fluorescence signal is indicative of activation of the target GPCR. The utilization of this assay in HTS applications was enabled by the development of the FLIPR (Molecular Devices Corporation, Sunnyvale, CA) in the mid-1990s (5). Other instruments with similar capabilities such as the FDSS (Hamamatsu Corporation, Hamamatsu City, Japan) and the CellLux (Perkin Elmer, Waltham, MA) have been released in recent years. The latest TM FLIPR model, FLIPRTETRA , has an integrated 96-, 384-, or 1536-well pipettor and is able to accommodate multiple reagent reservoirs allowing one to perform multiple additions to the assay plate. The earlier FLIPR models used an argon ion laser, but the TM FLIPRTETRA uses LEDs to excite the microtiter plate. The image of each well is captured simultaneously by a cooled charge-coupled device (CCD) camera, which is capable of updating images once per second. This is critical for measurement of the rapid calcium response that is typically observed for GPCRs. The kinetic data obtained on the FLIPR essentially provides a fingerprint for activation of the target GPCR. It enables one to extract more information regarding the activity of a test compound as compared to a single-read endpoint assay. The kinetic profile of an agonist can vary slightly depending on the GPCR and/or the cell background. The agonist-induced increase in intracellular calcium may be transient and return to baseline levels relatively quickly (Fig. 7.1A). Alternatively, in some instances, the level of intracellular calcium remains elevated for an extended period of time following agonist treatment (Fig. 7.1B). Knowing what the kinetic profile of a true agonist looks like can help the researcher distinguish between compounds that have agonist activity and those that are likely false positives (Fig. 7.1C,D). The most obvious false positive is a fluorescent compound that elicits a very distinct kinetic profile characterized by an extremely rapid increase in fluorescence signal, which remains relatively unchanged over time (Fig. 7.1D). The experimental workflow for a calcium mobilization assay on the FLIPR is shown in Fig. 7.2. Briefly, cells expressing the GPCR of interest are plated the day prior to the assay, the cells are incubated with a cell-permeable calcium-sensitive dye and then the plate is placed on the FLIPR for the assay. Here we describe the
Screening G Protein-Coupled Receptors
147
Fig. 7.1. Representative kinetic profiles of a calcium mobilization assay. CHO-K1 cells stably expressing a promiscuous Ga protein and a Gas-coupled GPCR were subjected to a calcium mobilization assay using a standard two-addition format 1 performed on the FLIPRTETRA . The first addition was either assay buffer (A and B) or test compound (C and D). The fluorescence signal was captured for 3 minutes and then a known agonist was added to the cell plate (A–D) and the fluorescence signal was monitored for an additional 2 minutes. (A) A representative kinetic profile of a GPCR that exhibits a transient increase in intracellular calcium upon treatment with an agonist. (B) An example of a GPCR that has a prolonged response to agonist. (C) An example of a compound that results in an increase in intracellular calcium, but with different kinetics than a known agonist. (D) A representative kinetic profile of a fluorescent compound. The sharp spike in fluorescence signal at the beginning of the kinetic profiles in A–C is an artifact of the clear FLIPR tips entering the well prior to the first addition.
Cell Line Generation
Plate Cells
Load Cells with Dye
Assay on FLIPR
Data Analysis
I
II
III
IV
V
Fig. 7.2. Experimental flow scheme for a calcium mobilization assay on FLIPR.
method for a two-addition calcium mobilization assay on the TM FLIPRTETRA for a Gas-coupled GPCR that was stably coexpressed with the promiscuous G protein, Ga16, in CHO-K1 cells. Recommendations regarding alternative conditions and parameters that may be applicable to other GPCRs and/or cell lines are also discussed.
148
Emkey and Rankl
2. Materials 2.1. Cell Culture
1. Cell culture medium for CHO-K1 cells: Ham’s F-12 (Mediatech Inc., Manassas, VA) supplemented with 10% characterized fetal bovine serum (FBS, Hyclone, Ogden, UT). 2. Solution of trypsin (0.05%) and ethylenediamine tetraacetic acid (EDTA) (0.53 mM) from Mediatech, Inc. 3. Phosphate-buffered saline (PBS) from Mediatech, Inc. 4. T175 tissue culture flasks (Greiner Bio-One, Monroe, NC). 5. 384-well black-walled, clear-bottom tissue culture-treated microtiter plates (Greiner Bio-One). 6. Multidrop (Thermo Scientific, Waltham, MA) or equivalent liquid dispenser for 384-well format. 7. FuGENETM 6 transfection reagent from Roche (Indianapolis, IN).
2.2. Calcium Mobilization Assay
1. Assay buffer: 20 mM HEPES, 11.1 mM glucose, 1.8 mM CaCl2, 1 mM MgCl2, 2.5 mM NaCl, 5 mM probenecid, and adjust pH to 7.4. Store at 4C and warm to room temperature before use. 2. 500 mM probenecid (MP Biomedicals, Solon, OH) made in 1 N NaOH and stored at room temperature. 3. Fluo-4/AM cell permeant from Invitrogen (Carlsbad, CA) is solubilized to a 1 mM stock solution with 100% DMSO by sonicating for 10 minutes. This stock solution is stable at –20C. 4. The 2 mM Fluo-4/AM dye-loading solution is prepared on the day of assay as follows. Mix the necessary volume of 1 mM Fluo-4/AM with an equal volume of 20% Pluronic F-127 in DMSO (Invitrogen). Add this solution to the appropriate amount of assay buffer. The final dye-loading solution consists of 2 mM Fluo-4/AM and 0.04% Pluronic F-127 in assay buffer and should not be stored and used the next day. 5. 384-well clear FLIPR tips from Axygen (Union City, CA).
3. Methods The method presented here was used to enable a calcium mobilization assay for a family of Gas-coupled GPCR using the FLIPRTETRATM . The signaling of this GPCR was redirected to the calcium mobilization pathway by generating a stable cell line
Screening G Protein-Coupled Receptors
149
(see Note 4.1, item 1) that coexpresses the GPCR and Ga16(see Note 4.1, item 2). Chinese hamster ovary (CHO-K1) cells (ATCC, Manassas, VA) were used for these studies. These cells are adherent, but nonadherent cells can also be used for these types of assays (see Note 4.1, item 3). 3.1. Cell Line Generation
Several methods exist to introduce the target GPCR and Ga proteins (if needed) into cells. The most commonly used methods for transfection include calcium phosphate, electroporation, or newer lipid transfection reagents such as FuGENETM 6. The example presented here utilizes FuGENETM 6. 1. The optimal DNA:FuGENETM 6 ratio is determined by performing several small-scale transfections in which the concentration of a plasmid encoding green fluorescent protein (GFP) and FuGENETM 6 is varied. The GFP signal is monitored and used to determine the DNA:FuGENETM 6 ratio for optimal transfection efficiency. 2. Seed cells in a 6-well plate at a density of 1 104cells per well in 3 mL of growth medium for 24 hours prior to transfection. Maintain cells in an incubator set at 37C/5% CO2. Transfect cells with 6 mL of FuGENETM 6 + 1.5 mg of DNA according to the manufacturer’s instructions (see Note 4.1, item 4). 3. Harvest cells 48 hours posttransfection and seed a T175 flask containing growth medium supplemented with the appropriate concentration of antibiotic encoded by the transfected plasmid. Maintain cells under antibiotic selection and monitor for the formation of drug-resistant colonies. 4. Harvest the cells once colonies have developed. This cell population is called the stable ‘‘pool.’’ A clonal line is isolated by single-cell sorting cells into a 96-well plate with the selection media (see Note 4.1, item 5). Other methods, such as seeding cells into 100 mm dishes and isolating colonies with cloning rings, may also be used. Wells are monitored over 2 weeks and wells with single colonies are harvested for testing and expansion. Each colony is tested in the FLIPR for correct response to agonist. Those clones with a positive response are chosen for further characterization with known agonists and antagonists. 5. Once a clonal cell line is chosen, expand the cells and make a liquid nitrogen cell bank for long-term storage. Cells are frozen at a density of 1 107cells/mL in 90% FBS/10% DMSO.
6. Cells are maintained in culture by splitting every 3 days with a seeding density of 3 106cells in a T175 flask (see Note 4.1, item 6).
150
Emkey and Rankl
3.2. Cell Plating
1. Cells for the calcium mobilization assay should be in exponential growth for optimal response. 2. Harvest cells with trypsin–EDTA (see Note 4.2, item 1). 3. Seed cells in a 384-well black-walled, clear-bottom plate (seeNote 4.2, item 2) at a density of 10,000 cells per well (see Note 4.2, item 3) in 50 mL of growth medium. 4. Leave plates in a single layer at room temperature for at least 60 minutes (see Note 4.2, item 4). 5. Incubate plates overnight in a 37C/5% CO2 incubator.
3.3. Loading Cells with Dye
1. Prepare dye-loading solution (2 mM Fluo-4/AM, 0.04% Pluronic F-127 in assay buffer) (see Note 4.3, items 1–2). 2. Completely remove media from the cell plate (see Note 4.3, item 3). This can be done by manually inverting the plate and flicking the media out of the plate or by using a plate washer to aspirate the media from the wells (see Note 4.3, item 4). 3. Add 20 mL per well of the dye-loading solution and place the plate into a 27C incubator in a single layer for 60 minutes (see Note 4.3, item 5). 4. After 60 minutes, remove dye and replace with 20 mL per well of assay buffer (see Note 4.3, item 6).
3.4. Assay on FLIPR
TM
These instructions are for an assay performed on the FLIPRTETRA instrument in 384-well format using LED excitation at 470–495 nm and emission at 515–575 nm. These instructions can be adapted to other instruments that can read fluorescent signals in real time and have liquid handling capabilities such as those mentioned in Section1. Prior to running the assay, a protocol file needs to be written on the FLIPR. The protocol file tells the instrument where the plates are located on the deck, the type of plates being used, the type of pipettor head (96, 384, or 1536), the excitation and emission wavelengths, and the height and speed for the integrated pipettor (see Note 4.4, item 1). The protocol also defines the length of the assay, when to add the reagents, and the frequency that images are captured during the run. The protocol for the example presented here consists of two additions to the cell plate (see Note 4.4) and has the following settings: a. Ten reads are collected with a 1-second read interval prior to the first addition. b. First addition: add 20 mL of compound. The dispense speed is 20 mL/second and the height is 20 mL. Images are collected at 1-second interval for 3 minutes.
Screening G Protein-Coupled Receptors
151
c. Second addition: add 20 mL of agonist with a dispense speed of 20 mL/second and a height of 40 mL. Images are collected at 1-second interval for 2 minutes. 1. After dye loading is complete, place the cell plate in the read TM position on the FLIPRTETRA . Wait for 5 minutes before running the assay to allow the cells to stabilize. Reading the plate immediately after dispensing buffer will result in a drift in the response. 2. Place the reservoirs containing the reagents to be added to the plate in the appropriate positions on the FLIPR. The reservoir for the first addition consists of compound at twice the desired final concentration. The reservoir for the second addition contains agonist at three times its EC80 concentration. 3. Place pipet tips in tip-loading position on the FLIPR (see Note 4.4, item 2). 4. Perform a signal test to determine the dye-loading efficiency of the cells and the variability across the plate. Set the LED camera gain to 100 and the excitation intensity at 70 with the CCD camera exposure held constant at 0.4 seconds. The CCD camera takes a picture of the entire plate and the images are converted to a numerical readout called relative fluorescent units (RFU). The maximal saturation of the LEDs in the TM FLIPRTETRA is reported to be 9000 RFU; however, the authors’ experience is that the saturation limit is 6000 RFU. Above 6000 RFU a saw-tooth pattern is visible in the data, which can result in false positives. Once the signal test is performed, the gain and excitation should be adjusted so that the average RFU across the plate is 1000. In addition to viewing the RFUs, it is possible to see the image of the plate. The image is extremely helpful in identifying smudges or lint that may be causing high or low RFU values in specific wells. It also helps to visualize the cell monolayer and patterns resulting from cell plating, dye, or buffer additions. The overall %CV for a 384-well plate under optimal culturing and assaying conditions should be 50%
Number tested Number Confirmed
Number tested Number Confirmed
% of total 32,891
100
474
1.44
270
100
163
60.37
163
100
156
95.71
The InCell 3000 Raven software like many HCS imaging software provides a means to ‘‘post hoc’’ analyze potential artifacts such as noise, debris, fluorescent compound interference, and cell loss from suspected adherence issues and/or cytotoxicity as a result of morphological alterations in the cell. The Raven software provides a method to assess how captured images correlate with the numerical metadata from the images for the output parameters shown in Table 8.1 such as Nuc:Cyt ratio, cell number, and the NPasNC parameter, which indicates the number of nuclei that were found to pass the intensity and size filters. Since the InCell 3000 was set up to capture a minimum of 100 cells/well or a maximum of two frames/well, whichever came first, compounds that generated data from wells of cell objects less than 100 were flagged as either cytotoxic or effecting cell adherence. We found the NPasNC parameter to be useful in determining cytotoxic or cell adherence issues in the IC50 dose– response data sets at higher compound concentrations such as 50 and 16.7 mM. After reviewing the images we in fact confirmed that compounds with an abnormal NPasNC parameter showed a reduction in cell number as a result of dose-dependent cell adherence or cytotoxicity (Fig. 8.15A). There are two distinct populations from data scatter plot of the mean cytoplasmic intensity per well versus average nuclear intensity per well. The plate control well data from anisomycin-treated cells and untreated or media-only wells are clearly separated into the two populations. Neither control populations exhibited cytoplasmic intensity thresholds above 2000 and/ or nuclear intensity thresholds above 3000. Although the majority of compound-treated wells were also within these defined threshold ranges, there were a considerable number of compounds that
High-Throughput Automated Confocal Microscopy Imaging Screen
181
Fig. 8.15. Nuclear trafficking module secondary analysis for cytotoxicity and fluorescence. The nuclear trafficking analysis module provides data on a number of parameters that can be used to identify potential interference due to compound fluorescence or off-target compound effects such as cytotoxicity or disruption of cell adherence (Table 8.1). These data were exported from the Raven software to Spotfire1 and visualized in a variety of scatter plots: (A) the NPasNC parameter was plotted for the different plate controls versus compound concentration to assess cytotoxicity or a reduction in cell adherence; (B) a scatter plot of the average cytoplasmic intensity/well versus average nuclear intensity/well in the target channel (EGFP) was used to identify fluorescent compounds. (C) Representative images from compound wells exhibiting a dose-dependent cytotoxicity or reduction in cell adherence; (D) representative images from compound wells exhibiting a dose-dependent fluorescence.
exceeded the cytoplasmic intensity threshold above 2000 and/ or nuclear intensity threshold above 3000. After reviewing the images we confirmed that at high concentrations (50 or 16.7 mM), the majority of these compounds were either fluorescent and/or affected the MK2–EGFP fluorescence signal (Fig. 8.15B). Only 13 (8.33%) of the 156 active compounds from the 10-point IC50 dose–response curve showed NPasNC data with less than 100 cells/ well, which may be a result of reduced cell adherence and/or cytotoxicity. However, a higher number (25, 16%) of the 10-point IC50 dose–response compounds were considered fluorescent as a result of the cytoplasmic intensity threshold output parameter above 2000 and/or nuclear intensity threshold above 3000 (Table 8.3).
182
O. Joseph Trask et al.
Table 3 Summary of follow up active hit compounds with secondary image analysis parameters measured to determine false positive from high-levels of fluorescence in nucleus and/or cytoplasm, and number of cytotoxic or cytotoxic-like compounds Secondary Analysis Parameters Total number tested
163
%
>50 mM
7
4.29
13
Table 11.1 HTS statistics Total number of points
261,432
Total number of samples
115,158
Number of samples (3s–/3s+)
111,497
96.8%
Number of samples (3s+/6s+)
1,375
1.2%
Number of samples (>6s+)
224
0.19%
6s cutoff
13.34%
Boolean samples
508
.8 17
.2 14
.7 10
1 7.
6 3.
1 0.
.5 –3
.0 –7
6
1
0. –1
4.
7.
7
40 35 30 25 20 15 10 5 0
–1
Frequency
Boolean samples noted in Table 11.1 are those samples where replicate measurements fall in both the active and inactive populations. Figure 11.3 is a Gaussian distribution of the sample results around the center of the assay or those compounds without effect. Figure 11.4 is the distribution of samples >3s with a standard deviation determined among replicate points. Figure 11.5 is a plot of the minimum and maximum values for two replicate points that conform to a linear progression.
–1
234
% Inhibition
Fig. 11.3. Gaussian distribution of the data around the center of the assay (no effect). The percent inhibition of the assay points was binned within intervals of 2.96% and plotted as a histogram with the frequency of the range on the y-axis.
High-Throughput Screening of the Cyclic AMP-Dependent Protein Kinase
235
100 90
% Inhibition (Average)
80 70 60 50 40 30 20 10 0 0
400
800
1200
1600
Compounds Fig. 11.4. Distribution and grey area of the outlier samples. The average and the standard deviation of duplicate assay points were plotted with rank order potency up to 100% inhibition of PKA activity. The 6s cutoff for the screen was 13.3% inhibition.
100 90 80
max value
70 60 50 40 30 20 10 0 0
30
60
90
min value Fig. 11.5. Plot of minimum and maximum outliers of the sample duplicates. Scatter diagram with linear R2 highlights the precision of the duplicate outliers.
236
Blackwell et al.
4. Notes 1. All solutions should be prepared in filtered water that has a resistance of 18.2 -cm and a total organic content of less than five parts per billion. 2. The peptide is solubilized in 40% DMSO and water (v/v) to a concentration of 1 mM. 3. The fluorescently labeled peptide is quantitated by measuring the absorbance of the fluorescein fluorophore at A492 using an extinction coefficient of 79,000 cm–1M–1. The concentrated peptide is diluted in 50 mM sodium carbonate, pH 9 prior to measuring the absorbance. Following quantitation, the peptide is examined on the Caliper to determine the degree of purity. The accepted level of purity, where the contaminant does not interfere with product peak assignment, is >95%. 4. The Caliper instrument can be run continuously for 5 days to shorten machine preparation and prolong the life of the microfluidic chip. All buffer solutions are stable throughout this time period. Only the chip needs to be refreshed daily with separation buffer. 5. Percent inhibition calculation
1
PSR compound PSR 100 ðPSR 0 PSR 100 Þ 100:
6. Z0 calculation:
1 ð3 ðstdev0 þ stdev100 ÞÞ= absolute value of average0 average100 :
References 1. O’Neill, L.A. (2006) Targeting signal transduction as a strategy to treat inflammatory diseases. Nat. Rev. Drug Discov. 5(7), 549–563. 2. Hennessy, B.T., Smith, D.L., Ram, P.T., Lu, Y., Mills, G.B. (2005) Exploiting the PI3K/ AKT pathway for cancer drug discovery. Nat. Rev. Drug Discov. 4(12), 988–1004. 3. Vlahos, C.J., McDowell, S.A., Clerk, A. (2003) Kinases as therapeutic targets for heart failure. Nat. Rev. Drug Discov. 2(2), 99–113. 4. Druker, B.J., Tamura, S., Buchdunger, E., Ohno, S., Segal, G.M., Fanning, S., Zimmermann, J., Lydon, N.B. (1996) Effects of a selective inhibitor of the Abl tyrosine kinase on the growth of Bcr-Abl positive cells. Nature Med. 2(5), 561–566.
5. Warner, G., Illy, C., Pedro, L., Roby, P., Bosse, R. (2004) AlphaScreen kinase HTS platforms. Curr. Med. Chem. 11(6), 721–730. 6. Sportsman, J.R., Gaudet, E.A., Boge, A. (2004) Immobilized metal ion affinitybased fluorescence polarization (IMAP): advances in kinase screening. Assay Drug Dev. Technol. 2(2), 205–214. 7. Koresawa, M., Okabe, T. (2004) Highthroughput screening with quantitation of ATP consumption: a universal non-radioisotope, homogeneous assay for protein kinase. Assay Drug Dev. Technol. 2(2), 153–160. 8. Dunne, J., Reardon, H., Trinh, V., Li, E., Farinas, J. (2004) Comparison of on-chip and off-chip microfluidic kinase assay formats. Assay Drug Dev. Technol. 2(2), 121–129.
High-Throughput Screening of the Cyclic AMP-Dependent Protein Kinase 9. Cheng, H.C., Kemp, B.E., Pearson, R.B., Smith, A.J., Misconi, L., Van Patten, S.M., Walsh, D.A. (1986) A potent synthetic peptide inhibitor of the cAMP-dependent protein kinase. J. Biol. Chem. 261(3), 989–992. 10. Janzen, W., Bernasconi, P., Cheatham, L., Mansky, P., Popa-Burke, I., Williams, K., Worley, J., Hodge, N. (2004) Optimizing
237
the Chemical Genomics Process. In: Darvas, F., Guttman, A., and Dorman, G. (eds.), Chemical Genomics. Marcel Dekker, New York, pp. 59–100. 11. Zhang, J.H., Chung, T.D., Oldenburg, K.R. (1999) A simple statistical parameter for use in evaluation and validation of high throughput screening assays. J. Biomol. Screen. 4, 67–73.
Chapter 12 Use of Primary Human Cells in High-Throughput Screens Angela Dunne, Mike Jowett, and Stephen Rees Abstract Traditionally, the objective of high-throughput screening (HTS) has been to identify compounds that interact with a defined target protein as the starting point for a chemistry lead optimisation programme. To enable this it has become commonplace to express the drug target in a recombinant expression system and use this reagent as the source of the biological material to support the HTS campaign. In this chapter we describe an alternative HTS methodology with the objective of identifying compounds that mediate a change in a defined physiological end point as a consequence of compound activity in human primary cells. Rather than screening at a defined molecular target, such ‘‘phenotypic’’ screens permit the identification of compounds that act at any target protein within the cell to regulate the end point under study. As an example of such a screen we will describe an HTS campaign to identify compounds that promote the production of the cytokine interferon-a from human blood peripheral mononuclear cells (PBMCs) isolated from whole blood. We describe the procedures required to obtain and purify human PBMCs and the electrochemiluminescence-based assay technology used to detect interferon-a and highlight the challenges associated with this screening paradigm. Keywords: High-Throughput Screen, Primary cells, PBMCs, Electrochemiluminescence, MesoScale Discovery, Interferon-a.
1. Introduction During the past 15 years, high-throughput screening (HTS) has become a central engine of drug discovery. As described in this volume, pharmaceutical and biotechnology companies, and increasingly academic institutions, have established the infrastructure to screen large libraries of chemically diverse molecules against drug targets, using automated robotic screening platforms (1, 2). In parallel with the development of HTS automation and instrumentation, a huge range of bioassay technologies have been W.P. Janzen, P. Bernasconi (eds.), High Throughput Screening, Methods and Protocols, Second Edition, vol. 565 ª Humana Press, a part of Springer Science+Business Media, LLC 2009 DOI 10.1007/978-1-60327-258-2_12, Springerprotocols.com
239
240
Dunne, Jowett, and Rees
enabled, which share a number of common features; the assays are typically homogeneous, amenable to assay in sub-100 ml assay volumes, tolerant to compound solvents such as dimethyl sulphoxide (DMSO) and are relatively cheap and simple to configure (3). Importantly, almost all HTS assays rely upon the use of recombinant protein or recombinant cell lines as the source of biological target due to the ability to generate a virtually limitless supply of material of consistent and high quality (4). Following hit identification, recombinant assays are usually complemented by downstream native tissue phenotypic assays used to profile hit or lead compounds for efficacy and mechanism of action (MOA). These assays typically rely upon the determination of compound activity in a cellular model of disease using either human primary cells or animal tissue in which the target protein is expressed in the native environment (5, 6). If the compounds are active in the phenotypic assay, the programme may progress towards the clinic; however, if the compounds are inactive, the compound series may be declared of no further interest. As a consequence, many years of effort may be wasted in optimizing molecules that ultimately have no activity in the disease-relevant phenotypic assay. For this reason it is attractive to move the phenotypic assay to an earlier point in the programme to avoid wasted work. To run the HTS using a native tissue phenotypic assay maximizes the possibility of identifying hits with the desired phenotypic activity and use recombinant assays to identify the MOA and to profile off-target activities. A phenotypic HTS enables the direct assessment of compound action on a pathway, rather than a defined target (7). This allows the scientist to probe all molecular targets on the pathway of interest and increases the likelihood of identifying compounds with the desired mechanism of action. However, it leads to a major question regarding the need to identify the mechanism of action of that molecule. There are two approaches to this issue: first, all hits can be profiled in recombinant assays against targets suspected to be of interest to identify the MOA and subsequent compound optimisation can be performed in recombinant assays. Second, if knowledge of the MOA is not required, or if it is not possible to identify the MOA, then all subsequent activities could be performed using the phenotypic assay as the primary assay. In this chapter we explore the use of phenotypic assays involving primary human cells for hit discovery and discuss the issues to be addressed to enable this screening paradigm using a HTS as the example to identify activators of interferon-a (IFN-a) production from human peripheral blood mononuclear cells (PBMCs).
Use of Primary Human Cells in High-Throughput Screens
241
1.1. Issues to Be Addressed to Enable a Phenotypic Assay HTS
There are a number of practical issues to be addressed prior to running a phenotypic HTS. First, the supply of primary human or animal tissue is limited. Cryopreserved Primary cells can be purchased from a number of vendors and many cell lines including human lung fibroblasts, chondrocytes and neuronal cells are available. However, primary human tissue remains difficult to obtain and for this reason we have run our first phenotypic screens using human blood cells. Second, the range of assay technologies available for HTS in native tissue is relatively limited. Phenotypic assays often rely upon the detection of the level of expression of a surface protein or the determination of the concentration of a secreted analyte in the culture media using ELISA (enzyme-linked immunosorbent assay) for detection. The development of miniaturized ELISA technologies such as electrochemiluminescence (MesoScale Discovery) (8) or AlphaLISA (Perkin-Elmer) (9) enables the performance of these assays in 96- and 384-well microtitre plates, thus making these assays HTS compatible. The third issue is often organisational and comes from a perception that phenotypic assays cannot be run for HTS due to the logistical reasons mentioned here or a belief that the HTS department will not run such an assay.
1.2. HTS to Detect IFNa Production by Human PBMCs
Toll-like receptors (TLR) are a family of at least 10 single-membrane spanning receptors, expressed in immune cells, that play a key role in mediating the innate immune response to the presence of pathogens (10). The activation of these receptors promotes leucocyte recruitment to the site of infection and causes the release of pro-inflammatory cytokines including IFN-a to cause the induction of the immune response to combat the presence of the foreign antigen (11). For this reason, TLR agonists are of interest as pro-inflammatory therapeutics to fight pathogen infection and as vaccine adjuvants (10–13). One such molecule that has been described is the imidazoquinoline compound Resiquimod (R-848) (Fig. 12.1). Resiquimod induces the production of IFN-a and a number of other pro-inflammatory cytokines from cultured human PBMCs. While the precise mechanism of action remains unclear, it has been NH2 N
N
O
N OH
Resiquimod Fig. 12.1. Structure of the TLR agonist Resiquimod (R-848).
242
Dunne, Jowett, and Rees
demonstrated to act as an agonist at both the TLR-7 and TLR-8 receptors (14). Rather than develop a recombinant HTS to identify agonists of TLR receptors, we elected to develop a screen to identify compounds capable of mediating the production of the physiologically relevant end point, IFN-a, from human PBMCs through any mechanism of action. This required the establishment of a robust supply chain for the collection and preparation of human PBMCs and the identification of an assay technology amenable to IFN-a detection with HTS performance characteristics. 1.3. MesoScale Discovery (MSD) Assay Platform
MesoScale Discovery (http://www.meso-scale.com) assay technology allows the performance of ELISA assays within 96- or 384well microtitre plates using electrochemiluminescence (ECL) detection (8). ECL is a non-isotopic, homogeneous and sensitive assay technology that allows the detection of analytes within the media of cultured cells (Fig. 12.2). MSD assays are performed using microtitre plates, which contain an electrode built into the
Measured signal is light LIGHT
*Ru(bpy)32+
Ru(bpy)32+
TPA–
Ru(bpy)32+
TPA
–H+
TPA– +
Detection antibody Analyte Capture antibody
Counter electrode
Working electrode
Dielectric
Fig. 12.2. Schematic representation of the electrochemiluminescence-based IFN-a detection assay. 384-well MSD plates are coated with capture antibody. Following the binding of analyte to this antibody, two ruthenium-labelled detection antibodies are added to the assay plate to form an ELISA sandwich at the base of the plate. Following the addition of MSD Read Buffer, an oxidation reaction occurs, which results in the generation of light. Light intensity is proportional to the concentration of captured analyte (see Section 3 and www.meso-scale.com for further details).
Use of Primary Human Cells in High-Throughput Screens
243
base of each well of the assay plate. To establish an MSD assay for the detection of human IFN-a, 384-well MSD plates, pre-coated with a goat anti-rabbit immunoglobulin, were coated with a rabbit polyclonal antibody to human IFN-a. Following the addition of cell culture media containing IFN-a, the cytokine is captured by the antibody. Captured analyte is detected following the addition of two monoclonal anti-IFN-a antibodies previously labelled with Ruthenium. We used two detection antibodies that recognize different epitopes on IFN-a. During assay development we determined that the signal window was enhanced through the use of an equimolar ratio of the two antibodies compared to the use of each antibody alone (data not presented). This is unusual; typically a single detection antibody is used. Following addition of MSD Read Buffer, the level of analyte is detected by reading the assay plate in the MSD Sector Imager and an electric current applied. This promotes the oxidation of Ruthenium with the resulting generation of a chemiluminescent signal, which is detected in the reader (Figs. 12.3 and 12.4). 1.4. Regulations Regarding the Use of Human Cells
DAY 1
It is necessary to consider whether there are any regulatory procedures that need to be adopted regarding the use of human tissue. Our screen was run in the United Kingdom and we briefly describe the regulatory issues encountered to perform this work. In 2004
Isolate PBMCs Re-suspend in Culture Media 50µl/well cells (4x10e5/ml) onto compound plate 48 hr incubation @ 37°C, 5% CO2 Add 20µl/well (1:8000)anti IFN-α polyclonal antibody to MSD plate Incubate at 4°C overnight
DAY 2
Harvest cell supernatant in compound plate by centrifugation (3 mins at 300g)
DAY 3
Wash MSD plate with 50μl/well PBS (Tecan Power Washer)
Transfer 20µl/well cell supernatant from compound plate to washed MSD plate using Cybiwell Add 20µl/well secondary mouse monoclonals sulpho-tagged antibodies Seal plate Incubate at 4°C overnight
DAY 4
Remove seal and aspirate contents (Tecan Power Washer) Add 30µl 2X MSD read buffer, incubate 10mins at Room Temp and read on Sector Imager
Fig. 12.3. Flow chart describing the assay protocol developed for the PBMC IFN-a production electrochemiluminescence HTS (see Section 3 for details).
244
Dunne, Jowett, and Rees
ASSAY 1
MONDAY
TUESDAY
WEDNESDAY
Isolation of PBMCs
Preparation of MSD plates
Isolation of PBMCs
ASSAY 2
IFNα MSD assay
Preparation of MSD plates
THURSDAY
FRIDAY
Read plates
IFNα MSD assay
Read plates
Fig. 12.4. Chart outlining weekly work pattern required to support the PBMC IFN-a production electrochemiluminescence HTS (see Section 3 for details).
the Human Tissue Act came into force, which regulates research with human tissue (15). An institute wanting to undertake such work must apply to the Human Tissue Authority to obtain a license that describes the type of work and the mechanism of how it will be conducted. The license requires among other things the following: First, records of all scientists performing the work are kept, where the material is stored and by what manner, a description of all equipment used including service and maintenance schedules, the generation of appropriate safety documentation including risk assessments and standard operating procedures, records of all staff training, a description of what is the material going to be used for and finally all disposal records. It is a legal requirement that human tissue has an audit trail starting with when the sample was obtained through to disposal. This license will describe the consent process under which tissue is taken ensuring that the donor understands why the tissue is being taken and for what purpose it will be used. The material can be used only for the purpose for which it was taken. Finally, the vaccination status of employees handling human tissue should be considered.
2. Materials 2.1. PBMC Preparation
1. Human blood was obtained from healthy volunteers by the GSK Blood Donation Unit 2. RPMI 1640 Media (Gibco, Paisley, Scotland) 3.
L-Glutamine
(100 ) (Gibco, Paisley, Scotland).
4. Penicillin/streptomycin (Gibco, Paisley, Scotland) 5. Foetal bovine serum (FBS) (Low Endotoxin) (Invitrogen, Paisley, Scotland)
Use of Primary Human Cells in High-Throughput Screens
245
6. Cell culture media: 10% foetal bovine serum (FBS), 2% penicillin/streptomycin and 1% L-glutamine in RPMI 1640 media. Stored at 4C for up to 4 weeks 7. Leucosep tubes pre-filled with ficol-histopaque (Greiner, Kremsmunster, Austria) 8. Centrifuge 5810 (Eppendorf, Hamburg, Germany) 9. Human recombinant IFN-g (Peprotech, Rocky Hill, NJ) 10. Citrate buffer (Baxter HealthCare, Glendale, CA) 11. Phosphate-buffered saline (PBS) (Gibco, Paisley, Scotland) 12. Controlled-rate freezer (Planer, Sunbury-On-Thames, UK) 13. Dimethyl sulphoxide (DMSO) (Sigma-Aldrich, St Louis, MO). 14. Freezing media (10% DMSO/90% FBS) 15. Cryovials (Corning, Corning, NY) 16. 140C freezer (ThermoScientific, Waltham, MA) 2.2. Antibody Labelling
1. Rabbit polyclonal anti-IFN-a (Carrier Free) (Stratech Scientific, Tonbridge, UK) 2. Mouse monoclonal anti-hIFN-a (MMHA-2 Carrier Free) (Stratech Scientific, Tonbridge, UK). Diluted to 2 mg/ml in PBS 3. Mouse monoclonal anti-hIFN-a (MMHA-11 Carrier Free) (Stratech Scientific, Tonbridge, UK). Diluted to 2 mg/ml in PBS 4. MSD SULPHO-TAG NHS-Ester (MesoScale Discovery, Gaithersburg, MD) 5. PD-10 columns (SEPHADEX G-25 M) (GE Healthcare, Bucks, UK) 6. Biorad protein assay kit (BioRad, San Ramon, CA) 7. Tube rotator (Stuart Scientific, Stone, UK)
2.3. MSD Assay
1. MSD Sector Imager 6000 Reader (MesoScale Discovery, Gaithersburg, MD) 2. MSD Read Buffer T 4 (MesoScale Discovery, Gaithersburg, MD). Dilute 4 stock to 2 with water 3. GAR-Coated Standard MA6000 384 plates (MesoScale Discovery, Gaithersburg, MD) 4. Plate seal (Weber Labelling, Arlington Heights, IL) 5. Water (Sigma-Aldrich, St Louis, MO)
2.4. Compound Plates
1. For HTS, compounds were supplied as 0.5 ml of 1 mM stock solutions in 100% DMSO in 384-well clear microtitre plates (Greiner, Kremsmunster, Austria). Compounds were
246
Dunne, Jowett, and Rees
supplied in all columns of the plate except column 6 and 18. Column 6 contained 0.5 ml of DMSO (low control) and column 18 contained 0.5 m1 of 100 mM resiquimod (high control). 2. Resiquimod was prepared by GSK Medicinal Chemistry and supplied at 10 mM in 100% DMSO. 2.5. Automation Used for HTS
1. 384-Well Tecan Power Washer (Tecan Trading AG, Zurich, Switzerland) 2. 384-Well Multidrop (ThermoScientific, Waltham, MA) 3. 384-Well Cybiwell (Cybio, Jena, Germany) 4. Plate Incubator (ThermoScientific, Waltham, MA) 5. Cedex Cell Counter (Innovatis, AG, Bielefeld, Germany) 6. Class II Cell Culture Cabinet (ThermoScientific, Waltham, MA) 7. Spectrophotometer (Perkin-Elmer, Waltham, MA)
3. Methods 3.1. Blood Collection
3.2. Isolation of Peripheral Blood Mononuclear Cells (PBMCs)
1. Collect blood by vein puncture into 15% citrate buffer (blood anticoagulant) by blood volume (9 ml citrate for 60 ml of blood). See Notes 1 and 2. 1. Add 30 ml blood to 50 ml leucosep tubes pre-filled with 15 ml histopaque 1077. 2. Centrifuge for 20 minutes at 1000g at room temperature. 3. Pour off enriched mononuclear fraction (upper phase) into second 50-ml centrifuge tube. Rinse out walls of leucosep tube with PBS, add to centrifuge tube, top up to 50 ml with PBS. 4. Centrifuge at 300 g for 10 minutes at room temperature. 5. Discard supernatant and wash cell pellet once in PBS and once in culture media. 6. Resuspend cell pellet in culture media and determine cell number on Cedex Cell Counter. 7. Dilute cells in culture media to 4 105/ml.
8. Store at 4C for a maximum of 4 hours before use in assay.
3.3. Cryopreservation of Human PBMCs
See Notes 3and 4. 1. Prepare Freezing Media and store at 4C.
Use of Primary Human Cells in High-Throughput Screens
247
2. Follow PBMC preparation method (Section 3.2) until Step 5. Resuspend the cell pellet in freezing media at a cell density of 4 107/ml.
3. Immediately aliquot cells into 1 or 5 ml Cryovials.
4. Transfer vials to a Controlled-Rate Freezer and freeze using the following programme: l Start temperature 5C l
Hold at 5C for 7 minutes
l
Cool 1C per minute to 5C
l
Cool 3C per minute to 12C
l
Cool 5C per minute to 14C
l
Cool 7.5C per minute to 20C
l
Cool 6.5C per minute to 25C
l
Hold at 25C for 2 minutes
l
Warm 3C per minute to 20C
l
Hold at 20C for 2 minutes
l
Cool 1C per minute to 50C
Cool 10C per minute to 130 C. 5. Transfer frozen vials to 140C freezer for storage. We have found that cells can be stored for a maximum of 6 months without any loss of viability. l
3.4. Labelling of IFN-a Monoclonal Antibody with MSD Sulpho-TAG
1. Dilute both mouse monoclonal antibodies to 2 mg/ml in PBS. 2. Dilute Sulpho-TAG NHS-Ester in DMSO to 10 nmol/ml immediately before use. 3. Add MSD Sulfo-TAG NHS ester solution to the antibody preparation to give a ratio of 20:1 molar excess of ester solution and mix. 4. Wrap the tubes in foil and mix on Tube Rotator at room temperature for 2 hours. 5. Prepare the G25M Sephadex PD-10 column by filling with PBS and allow to drain by gravity. Repeat three times before loading antibodies. 6. Add antibody label mix (Step 3) to the column and elute by gravity. 7. Elute from column using PBS. Collect eluate into 500 ml fractions. 8. Determine the concentration of labelled protein in each elute using Biorad Protein Assay following the instructions therein. 9. Labelled antibody can be stored at 4C at a concentration of 2 mg/ml for 6 months.
248
Dunne, Jowett, and Rees
3.5. IFN-a MSD Assay Protocol
See Notes 5 and 6. Day 1: 1. Using a ThermoLab 384-well Multidrop add 50 ml of PBMCs in culture media 4 105/ml into each well of a 384-well compound plate (hereinafter referred to as the compound plate). All plates should be lidded. 2. Incubate at 37C/5% CO2for 48 hours in a Heraeus incubator. Day 2: 3. Add 20 ml of diluted (1:8000 in culture media) anti-IFN-a polyclonal antibody to each well of a 384-well GAR Coated Standard MSD plate using a ThermoLab 384-well Multidrop (hereinafter referred to as the MSD plate). 4. Incubate at 4C overnight. Day 3: 5. Using a 384-well Tecan Power Washer remove the antibody solution from the MSD plate, wash each well twice in 50 ml PBS. 6. In parallel, centrifuge the compound plate for 3 minutes at 300g (1200 rpm). 7. Using a 384-well Cybiwell transfer 20 ml of cell supernatant from the compound plate to the MSD plate. 8. Add to the MSD plate 20 ml of the two mouse monoclonal sulpho-tagged antibodies (from 3.4). Cover plates using a Plate Seal. 9. Incubate at room temperature overnight in the dark. Day 4: 10. Remove the plate seal. Using a 384-well Tecan Power Washer aspirate the solution from the MSD plate. 11. Using a ThermoLab 384-well Multidrop add 30 ml of 2 MSD Read Buffer. 12. Incubate for 10 minutes at room temperature. 13. Read plate on MSD Sector Imager.
3.6. Preparation of Human PBMCs
Our objective was to screen 1.2 M compounds in 384-well microtitre plates with a throughput of 180 plates/ experiment with two screening experiments each week. Assay development data indicated that it was necessary to take blood, prepare PBMCs and add these to compound plates on the day of donation. To support the HTS we elected to take blood donations of 200 ml to allow for recycling of volunteers. We found that each donation generated sufficient cells to screen around twenty 384 well assay plates. Thus we had to establish a supply chain that enabled collection of blood
Use of Primary Human Cells in High-Throughput Screens
249
Plate Z’
from nine volunteers on each day of assay with 18 volunteers required each week. Following collection each donation has to be processed individually as it is not possible to mix blood from separate donors due to surface antigen cross-reactivity. This required multiple parallel processing of samples. The most significant challenge for this HTS was the effect of donor variability on assay performance. A requirement of any HTS assay is that the assay has a high signal window, usually defined as a Z0 of greater than 0.4 (16), which is consistent across plates and across days. In an HTS supported using a recombinant reagent it is possible to generate a reagent that enables consistent assay performance throughout the screening campaign. This is not possible in a phenotypic HTS. We observed differences in the ability of PBMCs from different donors to produce IFN-a in response to Resiquimod, which caused significant differences in assay window, biological activity cut-off and hit rate throughout the screen (Fig. 12.5). This led us to treat each donation as an individual batch within the HTS, with data processed on a donation-by-donation basis. We found that 20% of the donations failed to give a robust response to the standard compound Resiquimod and plates from these donors failed in the assay. The consequence of this was a high plate failure rate (20%) in the HTS.
Donor
Fig. 12.5. Variation in assay performance is donor dependent. PBMCs were prepared from 13 donors to support assay development. Data show the plate Z0 obtained in the IFN-a assay using PBMCs prepared from each donor (range ¼ 0–0.73). Z0 was calculated according to Zhang et al. (16) using the response obtained from 16 wells of a 384-well plate containing a maximal concentration of Resiquimod against 16 wells of a 384-well plate containing DMSO alone (numbers are actual Z0 from each experiment; data are the mean of three experiments).
250
Dunne, Jowett, and Rees
While this HTS used cells prepared on the day of assay, we later found that PBMCs could be cryopreserved for subsequent use (17, 18). Cryopreservation of cells allowed us to decouple blood preparation from the screening assay and led to a simpler and more flexible work pattern. In addition, the use of cryopreserved cells allows cells to be performance tested such that cells from donors that do not show a robust response to Resiquimod can be discarded ahead of screening. A number of factors were optimized ahead of HTS including cell density, incubation times, antibody concentrations, plate types, screening concentration, assay stability across screen batches, pharmacological validation and solvent tolerance. The experiments required to develop the PBMC assay were no different to those required for any cell-based HTS with the exception that all experiments had to be repeated on blood samples taken from multiple donors to account for the effects of donor variability (3). As an example of this we studied the tolerance of the assay to the compound solvent DMSO in multiple donors. This was determined by monitoring the ability of resiquimod to promote the production of IFN-a over a range of DMSO concentrations. In most donors, assay performance was not affected by DMSO concentrations of up to 1%; however, in a minority of donors the assay window collapsed at concentrations of DMSO above 0.5% (Fig. 12.6). In the final assay conditions the standard agonist Resiquimod had a pEC50 of 7.5 0.25 in agreement with other reports (11) and the assay gave a Z0 of 0.55 0.32 (n ¼ 84). Compounds were screened at a final assay concentration of 10 mM in 1% DMSO (see Notes 7–21).
A
B MSD Raw Count
MSD Raw Count
3.7. Assay Development
% [DMSO]
% [DMSO]
Fig. 12.6. Sensitivity of the IFN-a assay to the solvent DMSO. The ability of an EC100concentration of Resiquimod to promote IFN-a production by human PBMCs prepared from two donors (A and B) was determined in the presence of the indicated concentrations of DMSO. Each data point represents the mean SEM of quadruplicate determinations.
Use of Primary Human Cells in High-Throughput Screens
Prior to committing to HTS a small validation screen is performed. At GSK we have constructed a validation compound set containing 9855 compounds dispensed into 28 384 well assay plates, which is representative of compounds drawn from the GSK screening collection. This set is screened on three independent occasions to determine the performance of the assay during extended screen runs and the ability of the assay to reproducibly identify the same active molecules. A number of observations were made during the validation screen: 1. Each validation set was screened using cells prepared from separate donors. As expected we saw data variation between donors (data not shown). 2. Using a statistical activity cut-off (compounds with activities greater than 3 standard deviations above the sample mean) the calculated hit rate for the screen was 1.5% with a cut-off of 3% of the Resiquimod response (Fig. 12.7). This was not altogether surprising as we typically see low hit rates in agonist screens. As a consequence we elected to progress compounds from the HTS that exhibit activities greater than 10% of the Resiquimod response.
50
TLR_VAL_0006 ACT
3.8. HTS Assay Validation
251
40
30
20
10
0
0
20
40
60
80
100
120
TLR_VAL_0008 ACT Fig. 12.7. Assay validation. The GSK validation compound set (9855 compounds) was screened at 10 mM compound concentration in the IFN-a assay. Compounds were screened against PBMCs prepared from two donors. Data show the correlation of activity between donor 1 and donor 2.
3. There is little correlation between active molecules in different screens (Fig. 12.7). During assay validation we routinely observed molecules that were active in specific donors. As we
252
Dunne, Jowett, and Rees
are interested only in identifying molecules with activities across multiple donors we elected to continue with the screen. 4. In contrast to a recombinant HTS in which plates are failed if the Z0 is below 0.4, a number of additional QC criteria were put in place for the PBMC HTS to account for donor variation. We elected to fail all plates where the Resiquimod response was less than 2000 raw counts and we passed any plates failed on Z0 if that plate contained hits displaying activities greater than 10% of the Resiquimod response. 3.9. Primary Screen
We screened 1,212,006 compounds at 10 mM final assay concentration across 3388 384-well compound plates. Other than the logistical issues caused by donor variation the HTS ran as predicted for a recombinant HTS using MSD detection. We saw a range of plate Z0 throughout the screen (Fig. 12.8). The mean Z0 for plates that passed quality control was 0.40 0.23 and as predicted, the plate failure rate was 20% with all failures being due to the absence of a robust IFN-a with certain donors. Using a cut-off of 10% of the Resiquimod response, we identified 2480 active compounds; a hit rate of 0.2% (Fig. 12.9).
A
B
Robust Z’
Number of Records
800 700 600 500 400 300 200 100 0 0.95
Robust Z’
Fig. 12.8. HTS quality control statistics. Z0 is calculated for each plate according to the signal window between column 6 (DMSO) and column 18 (Resiquimod) as described in Section 2.4. (A) Z0 for each screen plate plotted against each data set (one data set corresponds to an assay on PBMCs prepared from a single blood donation). (B) Binned Z0 for all plates screened during the HTS. The average Z’ for all plates that passed QC was 0.4 0.23.
3.10. Concentration– Response Determinations
One thousand nine hundred and ninety-two active molecules were progressed to concentration–response testing. We generated concentration–response curves on four experimental occasions with each compound being tested against cells prepared from four donors. As anticipated some compounds were similarly active
Use of Primary Human Cells in High-Throughput Screens
B
% Agonism Binned Response vs Frequency
Count
% Activity (of Resiquimod)
A
253
Data Set
% Agonism Binned Response
Fig. 12.9. HTS activity rates. (A) Data show percent activity of each compound on each screening plate. Using a cut-off of 10% of the Resiquimod response (normalised to 100%), the activity rate in the screen was 0.2%. (B) Activity distribution of all molecules identified in the HTS with activities greater than 10% of the response to Resiquimod. Two thousand four hundred and eighty active compounds were identified. Numbers represent the number of compounds in each activity bin.
against all donors, whereas others appeared to show donor-specific activity (Fig. 12.10). As our objective was to identify molecules with clinical efficacy in broad patient populations, we elected to progress molecules that had activity against all donors tested and did not progress apparently donor-selective molecules. As a consequence this screen identified 17 molecules for
B
% Response
% Response
A
Concentration
Concentration
Fig. 12.10. Representative concentration–response curves for two hits from the HTS. HTS hits were screened against PBMCs prepared from four donors. Ten-point concentration–response curves were generated for each compound. Data are presented as a percentage of the maximum response to resiquimod in each donor. Each curve represents a single potency determination. Compound A generated reproducible potency determinations in all assays. Compound B was inactive on two of the four test occasions and displayed donor dependent efficacy.
254
Dunne, Jowett, and Rees
progression with pEC50values in the range of 4.3–7.3. Following an assessment of the data, chemistry has been initiated on a number of series.
4. Conclusions There are many factors to be considered prior to a phenotypic HTS. First, a phenotypic screen is molecular target independent and allows the screener to identify molecules that regulate the disease-relevant end point. Second, a phenotypic screen may be considered for targets for which recombinant expression is difficult or for which the pharmacology of the target changes when expressed in a recombinant system. Third, in situations where a recombinant assay has failed to identify hits, there is a possibility that screening against the target in the native environment may facilitate hit identification. While it may be possible to determine the MOA of hits from a phenotypic screen, it is likely that hits will be identified for which the MOA is unknown. Hence prior to committing to such a screen, one should consider whether knowledge of the MOA is a requirement for progression and if so an experimental plan has to be constructed to allow determination. In that regard, phenotypic HTS can be regarded as the natural heir to tissue strip pharmacology through which all drugs were identified prior to the 1980s and the dawn of the recombinant era. Perhaps the major achievement of this work has been to show that it is possible to alter the paradigm of hit identification and run an HTS using primary human tissue. The PBMC HTS successfully identified a number of chemical series that regulate IFN-a production from human PBMCs and we have since completed a number of other high- and low-throughput screens to identify modulators of cytokine release from human PBMCs. While it is possible to obtain sufficient human blood to support HTS, the availability of most other tissue types in sufficient quantity for HTS remains a challenge. We have run smaller screens using hepatocytes, chondrocytes and neuroblastoma cells. The ability to run HTS using other primary tissue will depend upon the development of assay technologies that reduce the cell requirement or the development of alternative sources of biological material such as the enablement of terminally differentiated stem cells. However, it is clear that phenotypic screening offers an exciting alternative to recombinant screens that may enhance the success of early hit identification.
Use of Primary Human Cells in High-Throughput Screens
255
5. Notes 1. Rinse steps are included to collect as many cells as possible. 2. Samples from different donors are kept separate throughout the isolation and assay protocol. 3. Cryopreserved PBMCs were not used in the HTS described here. We have since found that this method can be used to prepare cells in advance of screening. 4. Samples from different donors must be kept separate throughout the freezing procedure. 5. See Figs. 12.3 and 12.4 for overview of HTS protocol. 6. For the HTS, plates were processed in 30-plate batches with two scientists. 7. Capture antibody is spotted directly onto the electrode in the assay well. MSD will supply plates where antibody has been spotted as a catalogue item or as a custom service (8). Alternatively, MSD will supply base plates for the customer to perform this exercise. Plate spotting requires specialist expertise and equipment and, in our experience, is difficult to perform reproducibly. 8. In the work described here we purchased plates from MSD in which a goat anti-rabbit immunoglobulin (GAR) had been spotted onto the electrodes (8). We used this to capture the IFN-a polyclonal capture antibody. 9. Following spotting of capture antibody, plates are stable for up to 1 year when stored at 4C. 10. Plates spotted with capture antibody may need to be blocked with protein to prevent non-specific binding. This was not required in the assay described here. If required this can be performed using 1% milk powder reconstituted in PBS. The requirement for blocking should be determined during assay development. 11. Excess blocking reagent should be removed by washing in PBS. We typically use a 384-well plate washer to do this. 12. When transferring reagents to the MSD assay plate, care should be taken not to damage electrode in the bottom of the plate. 13. It is critical to define cell plating density during assay development and the cell number per well used should be minimized to conserve cells. 14. The addition of antibiotics to the culture media is advised to avoid bacterial contamination of the samples.
256
Dunne, Jowett, and Rees
15. To minimize the number of steps in the assay, the detection antibody mixture should be added directly to the assay plate containing the analyte. However, assay performance may be enhanced if plates are washed before addition of the detection antibody. 16. It is a requirement that the detection antibody recognizes a different epitope to the capture antibody. 17. Following the addition of MSD Read Buffer to the plates, the optimal final volume should be 35 ml. At volumes of less than this, assay performance decreases as the camera in the Sector Imager is unable to detect the assay signal. For this reason, plates are sealed to prevent evaporation prior to reading. The camera height or the read time cannot be adjusted without engineer intervention. 18. Unbound detection antibody not washed away prior to the addition of the MSD Read Buffer will generate a background signal. Assay performance can be improved by washing of the plate prior to the addition of read buffer. The use of the MSD Sector Imager to read plates is a requirement of this assay. MSD assay plates are not compatible with other readers. 19. As work with human tissue carries potential health risks, all work should be contained. In our laboratory, specific equipment is used for PBMC work and not for other purposes. All tissue culture is performed within a Class II Safety Cabinet. 20. A robust data-handling process should be established ahead of screening to facilitate the identification of any quality-control failures prior to the commitment of large numbers of plates for screening to minimize waste. 21. Screen data from each PBMC batch were analysed separately to account for donor variation.
6. Acknowledgements The authors would like to acknowledge the expertise of the members of the Biological Reagents and Assay Development, Screening and Compound Profiling, Discovery Technology Group and the Infectious Diseases Centre of Expertise for Drug Discovery for their work to enable phenotypic HTS at GlaxoSmithKline: Ken Grace, Barbara Hebeis, Ketaki Shah, David Gray, Sian Lewis, Rupal Kapadia, Shie Chang, Claire Purkiss, Jason Signolet, Anesh Sitaram, Peter Morley, Lucy Reynell, Elena Sciamanna, Mike Sowa, Gavin Harper, Karen Amaratunga, Carolyn O’Malley,
Use of Primary Human Cells in High-Throughput Screens
257
Michael Wilson. Finally we would like to thank the GSK Blood Donation Unit and the 200 blood donors who made this work possible. References 1. Posner, B. A. (2005) High-throughput screening-driven lead discovery: meeting the challenges of finding new therapeutics. Curr. Op. Drug Disc. Dev. 8, 487–494. 2. Gribbon, P. and Andreas, S. (2005) Highthroughput drug discovery: What can we expect from HTS? Drug Disc. Today 10, 17–22. 3. Walters, W. P. and Namchuck, M. (2003) Designing screens: How to make your hits a hit. Nat. Rev. Drug Disc. 2, 259–266. 4. Moore, K. and Rees, S. (2001) Cell-based versus isolated target screening: How lucky do you feel? J. Biomol. Scr. 6, 66–74. 5. Horrocks, C., Halse, R., Suzuki, R., and Shepherd, P. A. (2003) Human cell systems for drug discovery. Curr. Op. Drug Disc. Dev. 6, 570–575. 6. Clemons P. A. (2004) Complex phenotypic assays in high-throughput screening. Curr. Op. Chem. Biol. 8, 334–338. 7. Rossi, C., Padmanaban, D., Ni, J., Yeh, L.A., Glicksman, M., and Waldner, H. (2007) Identifying drug-like inhibitors of myelinreactive T cells by phenotypic high-throughput screening of a small-molecule library. J. Biomol. Scr. 12, 481–489. 8. See http://www.meso-scale.com for literature describing the theory and application of electrochemiluminescence detection. 9. See http://las.perkinelmer.com/ for literature describing the theory and application of AlphaLisa detection. 10. Gay, N. J. and Gangloff, M. (2007) Structure and function of Toll receptors and their ligands. Ann. Rev. Biochem. 76, 141–165.
11. Uematsu, S. and Akira, S. (2007) Toll-like receptors and type-1 interferons. J. Biol. Chem. 282, 15319–15324. 12. Weeratna, R. D., Makinen, S. R., McCluskie, M. J., and Davis, H. L. (2005) TLR agonists as vaccine adjuvants: Comparison of CpG ODN and Resiquimod (R-848). Vaccine 23, 5263–5270. 13. Gerondakis, S., Grumont, R. J., and Banerjee, A. (2007) Regulating B-cell activation and survival in response to TLR signals. Immunol. Cell Biol. 85, 471–475. 14. Jurk, M., Heil, F., Vollmer, J., Schetter, C., Krieg, A. M., Wagner, H., Lipford, G., and Bauer, S. (2002) Human TLR7 or TLR8 independently confer responsiveness to the anti-viral compound R-848. Nat. Immunol. 3, 499–504. 15. Human Tissue Act (2004) available from http://www.opsi.gov.uk. 16. Zhang, J.-H., Chung, D. Y., and Oldenberg, K. R. (1999) A simple statistical parameter for use in evaluation and validation of high throughput screening assays. J. Biomol. Scr. 4, 67–73. 17. Kunapuli, P., Zheng, W., Weber, M., Solly, K., Mull, R., Platchek, M., Cong, M., Zhong, Z., and Strulovici, B. (2005) Application of division arrest technology to cell-based HTS: comparison with frozen and fresh cells. Assay & Drug Dev. Technol. 3, 17–26. 18. Smith, J. G., Joseph, H. R., Green, T., Field, J. A., Wooters, M., Kaufhold, R. M., Antonello, J., and Caulfield, M. J. (2007) Establishing acceptance criteria for cell-mediatedimmunity assays using frozen peripheral blood mononuclear cells stored under optimal and suboptimal conditions. Clinical & Vaccine Immunol. 14, 527–537.
INDEX Apatchi-1TM ................................................................... 192 See also Reader ASDIC............................................100, 101, 102, 103, 104 Assay AlphaLISA ........................................................... 8, 241 automated patch clamp......................... 9, 192, 209–222 See also Electrophysiology Binding .............9, 13–29, 118, 129, 131, 132, 139, 141 BRET............................................................................ 9 cell-based.....................5–6, 7, 8, 9, 17–18, 20, 145, 153 CE, see Capillary Electrophoresis (CE) coupled ..................8, 14, 112, 113, 122, 145, 146, 147, 148, 156 ECL .................................................................. 7, 9, 242 See also Electrochemiluminescence (ECL) Efflux ................................................................ 154, 191 ELISA .......................................................... 8, 241, 242 end-point ............................................................ 14, 113 enzymatic ................................11, 28, 73, 108, 109, 137 FCS............................................................................... 7 See also Fluorescence Correlation Spectroscopy (FCS) FIDA ............................................................................ 9 See also Fluorescence Intensity Distribution Analysis (FIDA) FLINT .......................................................... 9, 137, 141 See also Fluorescence intensity (FLINT) Format............6, 7, 9, 10, 21, 23, 75, 76, 109, 110, 111, 112, 113, 115, 119, 123, 145, 190, 191, 205 FP..................9, 17, 114, 121–122, 130, 131, 134, 135, 136, 137, 138, 139, 140, 141 n.2 See also Fluorescence Polarization (FP) FPIA ......................................................... 114, 115, 116 See also Fluorescence Polarization Immunoassay (FPIA) FRET................................................8, 9, 113, 114, 190 See also Fluorescence Resonance Energy Transfer (FRET) HCS.................................................. 160, 161, 175, 177 See also High Content Screening (HCS) incubation time .............9, 14, 18, 20, 27, 43, 119, 120, 123 n.5, 166, 185 n.13, 250 microfluidic ....................................................... 112, 226 phenotypic..................................................... 5, 240, 241 PolarScreen ............................................................... 138 radioligand binding................................... 145, 190, 211
A AAO ........................................................................... 21, 22 See also Automated Assay Optimization (AAO) Absorbance......................6, 8, 9, 24, 26, 49, 112–113, 115, 117, 236 n.3 Activator .....................18, 42, 110, 116, 117, 164, 196, 240 Active ...............2, 37, 70, 72, 73, 82, 83, 85, 89, 92, 94, 96, 97, 99, 100, 109, 110, 131, 132, 135, 137, 138, 140, 175, 176, 177, 178, 179, 180, 181, 182, 198, 210, 226, 234, 240, 251, 252, 253 ADME-Tox Affinity binding ...................................................................... 140 km ............................................................................... 16 Agonist EC50................................................................. 250 Akt .................................................................................. 109 See also Enzyme ALA Scientific See also Manufacturer Alembic Instruments ...................................................... 199 See also Manufacturer Alkaline phosphatase ...................................................... 113 See also Enzyme Allosteric ligand.............................................................. 131 AlphaLISA ................................................................. 8, 241 See also Assay AlphaScreen........................................................7, 8, 9, 226 See also Reader Amlodipine ..................................................................... 189 See also Prescription Drugs Amphotericin-B.............................................................. 194 AMP kinase .................................................................... 110 See also Enzyme Analysis software ........................................ 47, 53, 168, 220 Analyst AD ..................................................... 132, 133, 136 See also Reader Anisomycin ....................160, 164, 165, 166, 167, 168, 170, 172, 173, 174, 175, 178, 179, 180, 183 n.4, 184 n.8 Anisotropy ..............70, 128, 129, 130, 132, 133, 134, 135, 136, 141 n.1 Antagonist IC50 Anthropomorphic arm.......................................... 35, 36, 39 See also Robot Antibiotic selection......................................................... 149
259
HIGH THROUGHPUT SCREENING
260 Index
Assay (continued) SPA...............................................7, 8, 9, 13, 14, 17, 34 See also Scintillation proximity assay (SPA) Transcreener PDE .................................................... 138 TR-FRET...................................6, 7, 8, 9, 77, 114, 116 Assay development..............2, 10, 23, 25, 34, 70, 108, 109, 110, 113, 115, 119, 120, 122, 131, 164, 165, 166, 170, 203, 228, 243, 248, 249, 250, 255 n.10, 255 n.13 Assay plate ..................40, 66, 67, 72, 74, 76, 80, 117, 130, 141, 146, 177, 196, 222, 228, 229, 230, 231, 232, 242, 243, 248, 251, 255 n.12, 256 n.18 ATP ......................12, 16, 109, 221 n.5, 226, 227, 228, 229 Automated Assay Optimization (AAO), see AAO Automated patch clamp.............................. 9, 192, 209–222 See also Assay The Automation Partnership ........................................... 46 See also Manufacturer Average .........47, 49, 51, 70, 76, 79, 81, 88, 101, 103, 111, 122, 123, 131, 154, 162, 163, 172, 174, 177, 179, 180, 181, 195, 201, 216, 221, 230, 232, 235, 252 See also Statistical analysis
B Barcode label BDTM Calcium Assay Kit .............................................. 153 See also Dye Beckman Coulter ..........................21, 22, 36, 170, 228, 231 See also Manufacturer Binding Bmax ........................................................................... 16 competitive........................................10, 11, 16, 17, 139 equilibrium................................................................ 134 uncompetitive...................................................... 16, 139 See also Assay Bioluminescence Resonance Energy Transfer (BRET)......................................................... 8, 9 See also Bioluminescence Resonance Energy Transfer (BRET) Biomek.............................................................. 22, 228, 231 Biomek 2000..................................................................... 22 See also Robot Bmax ......................................................................... 16, 154 Boltzmann’s equation ............................................. 216, 217 Boolean ........................................................................... 234 BRET.............................................................................. 8, 9 See also Assay
C Calcium....................................................................... 3, 153 See also Dye Calcium-activated potassium channel ............................ 196 See also Ion Channel
Calcium Green-1............................................................ 153 See also Dye Calcium Sensitive Dye............................ 146, 153, 154, 190 See also Dye Caliper BioSciences See also Manufacturer CAMP ............................................................................ 7, 8 Campaign....................2, 10, 12, 13, 19, 25, 34, 47, 69, 71, 83, 84, 85, 91, 94, 95, 96, 97, 99, 104, 136, 153, 160, 175, 198, 205, 249 Capillary Electrophoresis (CE) ...................................... 226 Carbamazepine ............................................................... 189 CCD cooled charge coupled device camera (CCD) ........146, 151, 161, 170 n.1–171 n.1, 201 Cell-based assay ......................5, 6, 7, 8, 9, 17, 18, 20, 145, 153, 185 n.13 Cellectricon..................................................................... 202 See also Manufacturer CE, see Assay Chemical library................................................................ 49 Chemiluminescent Nitrogen Detector (CLND)........................................................ 228 Cherry picking ...................................................... 46, 66, 67 Chloride channel .................................................... 188, 196 See also Ion Channel CHO cells....................................................... 154, 212, 218 CLND ............................................................................ 228 Clone selection................................................................ 197 Coefficient of variation of signal and background (CV) ........23, 25, 97, 122, 151, 184 n.12 Coelenterazine .................................................................... 7 Competitive ............................10, 11, 12, 16, 17, 114, 116, 119, 121, 135, 139, 140 Compound chemical library ........................................................... 49 collection .................................2, 3, 4, 6, 10, 14–15, 29, 55, 91, 117, 205, 206, 242, 251 concentration ..........................6, 10, 11, 12, 13, 14, 19, 66, 79, 83, 91, 92, 112, 139, 151, 156, 174, 175, 176, 177, 179, 180, 181, 185, 209, 215, 228, 251 fluorescent...............................116, 117, 137, 140, 146, 147, 156, 180, 181 identifier..............................................47, 48, 49, 59, 62 interference .............6, 9, 111, 112, 114, 115, 116, 123, 124, 136, 180 library ............................42, 55–68, 119, 136, 167, 175, 178, 205, 206 registration .................................................................. 59 storage ............................................................. 44, 46, 57 tracking
HIGH THROUGHPUT SCREENING
Index 261
Concentration ...............................6, 10, 11, 12, 13, 14, 15, 16, 17, 19, 21, 27, 65, 66, 75, 79, 83, 91, 92, 111, 112, 113, 115, 116, 118, 119, 120, 132, 133, 134, 135, 136, 137, 139, 140, 141, 149, 151, 152, 154, 155, 156, 157, 160, 164, 165, 167, 168, 169, 170, 172, 174, 176, 177, 178, 179, 180, 181, 184, 185, 188, 194, 209, 212, 214, 215, 217, 218, 219, 220, 222, 226, 228, 229, 241, 250, 252, 253 Control software system ................................................... 36 Conventional patch clamp ...................................... 192, 206 See also Electrophysiology Cooled Charge Coupled Device Camera (CCD)...........146, 151, 161, 171, 184 n.11, 201 See also CCD CV.........................................................23, 25, 97, 151, 184 See also Statistical analysis CyBi1-Lumax ................................................................ 154 See also Reader Cyclic AMP-dependent protein kinase.................. 225–236 See also Enzyme Cystic fibrosis transmembrane conductance regulator (CFTR)......................................................... 196 See also Ion Channel Cytokine..................................160, 185 n.14, 241, 243, 254 Cytotoxicity.......................................177, 180, 181, 183 n.2
D DAPI ................................................................161, 184 n.9 See also Dye Database............39, 47, 48, 49, 50, 51, 52, 65, 68, 177, 232 Data handling .....................34, 47–48, 50, 54, 75, 256 n.20 Daughter plate ..........................................66, 213, 228, 231 Deacetylase...................................................... 108, 110, 226 See also Enzyme DecisionSite ...................................................................... 47 Diazepam ........................................................................ 188 Dimethyl sulfoxide (DMSO) ............18, 19, 44, 66, 67, 95, 109, 119, 120, 122, 134, 148, 149, 164, 167, 168, 170, 171, 172, 175, 185 n.13, 212, 214, 215, 218, 227, 228, 231, 240, 245, 246, 247, 249, 250, 252 See also DMSO tolerance Dispenser ................................................................ 148, 169 See also Robot DMSO tolerance ..................109, 119, 134, 164, 168, 171, 172, 185 DNA .......................................114, 127, 149, 161, 182, 212 Dye BDTM Calcium Assay Kit ........................................ 153 calcium .................................................................. 3, 153 calcium green–1 ........................................................ 153 calcium sensitive................................146, 153, 154, 190 DAPI ..........................................................161, 184 n.9 fluo-4/AM ........................................ 148, 150, 153, 154 fluo-8/NW Calcium Assay Kit ................................ 153
fluorescein ......................115, 116, 130, 141 n2, 235 n.3 Hoechst 33342..................160, 161, 162, 166, 169, 170 membrane potential-sensitive oxonol dye ................ 190 oxonol dye ................................................................. 190 voltage-sensing dyes...................................................... 8 Dynaflow1 ...................................................................... 202 Dynamic................................................38, 39, 77, 116, 188
E E-76, 132, 135, 136 EC50, 138, 160, 164, 167, 168, 171, 172, 173, 192 See also Statistical analysis EC100, 250 See also Statistical analysis ECL ........................................................................ 7, 9, 242 See also Assay Edge effect ..........................................18, 23, 153, 169, 170 EDTA.....................20, 120, 122, 124, 148, 150, 211, 226, 227, 228, 229, 231 Electrochemiluminescence (ECL) ...............7, 8, 239, 241, 242, 244 See also ECL Electronic lab notebook.................................................... 58 Electrophysiology automated patch clamp......................... 9, 192, 209–222 See also Assay conventional patch clamp ................................. 192, 206 See also Assay gigaseal.............................................. 192, 210, 218, 222 high resistance seal.................................... 192, 200, 210 perforated patch clamp See also Assay planar patch clamp ............................ 198, 201, 202, 212 PPC...........................................193, 195, 198, 203, 206 See also Population Patch Clamp (PPC) Electroporation ............................................................... 149 ELISA ......................................................6, 8, 13, 241, 242 See also Assay End-Point ................................................... 14, 15, 113, 121 See also Assay Enzymatic assay ............................11, 28, 73, 108, 109, 137 See also Assay Enzyme alkaline phosphatase ................................................. 113 Akt .................................................................................. 109 AMP kinase .............................................................. 110 cyclic AMP-dependent protein kinase ............. 225–236 deacetylase................................................. 108, 110, 226 FVIIa.................................131, 132, 133, 134, 135, 136 histone acetylase........................................................ 226 isomerase................................................................... 108 kinase ..............3, 4, 7, 9, 107, 108, 109, 110, 114, 115, 118, 120, 137, 138, 160, 175, 179, 226, 227, 228, 229, 230, 231
HIGH THROUGHPUT SCREENING
262 Index
Akt (continued) ligase............................................................ 22, 108, 114 phosphatase...............108, 109, 110, 113, 116, 117, 226 PKA .................................................................. 225–236 protease .............4, 9, 20, 107, 108, 110, 111, 113, 114, 120, 131, 226 recombinase............................................................... 108 RNA polymerase......................................................... 27 valyl-tRNA synthetase.......................................... 13, 14 Enzyme fragment complementation .............................. 8, 9 Equilibrium......................................................... 16, 17, 134 Error handling ................................................................ 38, 52 See also Statistical analysis Essen Instruments........................................................... 194 See also Manufacturer Experimental design ................................................... 21, 85
F False negative......................................2, 29, 82, 83, 91, 92, 100, 115 False positive.............2, 29, 82, 91, 92, 113, 115, 116, 131, 136, 138, 146, 151, 156, 182, 184 FBS .......................148, 149, 165, 166, 167, 168, 169, 173, 175, 212, 244, 245 FCS..................................................................................... 7 See also Assay FIDA .................................................................................. 9 See also Assay FLINT ................................................................ 9, 137, 141 See also Assay FLIPR...................7, 8, 9, 42, 43, 146, 147, 148, 149, 150, 151, 152, 153, 154, 155, 156 See also Reader FlipTip .................................................................... 200, 201 See also Reader Fluo-4/AM ............................................. 148, 150, 153, 154 See also Dye Fluo-8/NW Calcium Assay Kit ..................................... 153 See also Dye Fluorescein ........................115, 116, 130, 141, 227, 236 n.3 See also Dye Fluorescence................6, 7, 9, 13, 15, 42, 50, 77, 112, 113, 114, 115, 116, 117, 122, 127–142, 146, 147, 153, 154, 155, 157, 161, 164, 181, 182, 190, 191, 195, 196, 205 See also FCS Fluorescence Correlation Spectroscopy (FCS)................... 7 Fluorescence Intensity Distribution Analysis (FIDA) ....... 9 See also FIDA Fluorescence intensity (FLINT)...............6, 128, 132, 133, 134, 137, 141, 162 See also FLINT
Fluorescence Polarization (FP)......................... 13, 15, 114, 121, 122, 127–142 Boltzmann equation.......................................... 216, 217 See also FP Fluorescence Polarization Immunoassay (FPIA)................................... 114, 115, 116, 121 See also FPIA Fluorescence Resonance Energy Transfer (FRET) .................................6, 8, 113, 114, 190 See also FRET Fluorescent compound..........................116, 117, 137, 146, 147, 156, 180, 181 See also Compound Fluorometric Imaging Plate Reader (FLIPR) ............................................ 9, 145–157 See also FLIPR Fluorophore acceptor .............................................. 113, 114 Flyion GmbH See also Manufacturer FlyScreen1.............................................................. 200, 201 See also Reader FP............................7, 9, 17, 114, 121, 122, 127, 128, 130, 135, 137, 138, 140 See also Assay FPIA ............................................................... 114, 115, 116 See also Assay Freezer..................................................................... 245, 247 FRET.............................................................. 8, 9, 113, 114 See also Assay FVIIa............................................... 131, 132, 134, 135, 136 See also Enzyme FVIIa/E-76 complex ...................................................... 132
G GABAA.......................................................................... 204 See also Ion Channel Gaussian distribution.......................................... 70, 99, 234 See also Statistical analysis GFP ..........................................................149, 159, 183 n.6 Gigaseal.............................................192, 210, 218, 222 n.2 See also Electrophysiology Gleevec............................................................................ 225 See also Prescription Drugs G-protein coupled receptor (GPCR) ....................... 3, 4, 6, 7, 8, 9, 145, 146, 147, 148, 149, 152, 153, 155, 156 See also GPCR Graphic User Interface (GUI)............................ 58, 78, 179 See also GUI Green Fluorescent Protein (GFP)............149, 160, 183 n.6 See also GFP GUI..................................................................... 58, 78, 179 See also Graphic User Interface (GUI)
HIGH THROUGHPUT SCREENING
Index 263
H
I
HCS................................................ 160, 161, 175, 177, 180 See also Assay HERG ..................196, 199, 203, 210, 211, 212, 213, 218, 219, 220, 221 n.1 See also Ion Channel High Content Screening (HCS) .................. 160, 161, 175, 177, 180 See also HCS High-Performance Liquid Chromatography (HPLC) ........................................................ 228 See also HPLC High resistance seal ........................................ 192, 200, 210 See also Electrophysiology High Throughput Screening (HTS).......................... 1–29, 33–54, 69–103, 107–125, 189, 190, 211, 225–236 See also HTS Hill slope......................................................................... 110 Histone acetylase ............................................................ 226 See also Enzyme Hit false negative ...........................2, 29, 71, 82, 83, 91, 92, 100, 115 positive ................21, 23, 26, 71, 89, 91, 92, 93, 96, 97, 100, 135, 138, 149 n.4, 156, 163, 228 Hoechst 33342................160, 161, 162, 166, 169, 170, 171 See also Dye Hotel ..................................................................... 36, 43, 44 See also Robot HPLC ............................................................................. 228 See also High-Performance Liquid Chromatography (HPLC) HTS ........... 1, 2, 3, 4, 5, 6, 7, 8, 10, 11, 12, 14, 15, 17, 18, 19, 20, 21, 23, 25, 27, 28, 29, 33, 34, 35, 37, 39, 40, 46, 47, 48, 54, 65, 71, 74, 76, 77, 80, 82, 83, 84, 88, 89, 91, 92, 93, 94, 95, 96, 99, 100, 103, 104, 108, 109, 110, 112, 119, 120, 121, 122, 123 n.3, 124 n.6, 129, 130, 134, 137, 146, 153, 164, 174, 175, 189, 190, 191, 196, 204, 205, 206, 229, 232, 239, 240, 242, 244, 248, 249, 250, 252, 253, 256 HTS:campaign................................2, 3, 10, 12, 18, 25, 33, 34, 40, 71, 83, 84, 85, 94, 96, 99, 104, 136, 146, 153, 203, 205 Hudson Control Group.................................................... 37 See also Manufacturer Human Tissue Act.......................................................... 244 Humidity.........................................................23, 161, 167, 169, 170 Hyperpolarisation activated cyclic nucleotide-gated channel.......................................................... 196 See also Ion Channel
IC50 ........................11, 12, 15, 16, 132, 135, 138, 180, 182 Imatinib mesylate Inactivated state block ............................................ 189, 191 InCell ...................................................................... 159–185 Incubation time...................9, 14, 18, 20, 27, 43, 119, 120, 123, 166, 185 n.13, 250 Incubator...................18, 42, 43, 44, 45, 53, 149, 150, 153, 169, 170, 213, 218, 222 n.5, 246, 248 Inhibitor competitive................................10, 11, 12, 16, 119, 139 IC50 ................11, 12, 15, 110, 121, 138, 164, 171, 230 Interface patch clamp...................................................... 192 Interferon ........................................................................ 240 Intracellular calcium............................................ 7, 145–157 Intra-Plate....................................................... 77, 78, 81, 85 See also Statistical analysis Inventory management ............................................... 58, 64 Ion channel calcium-activated potassium ..................................... 196 CFTR........................................................................ 196 chloride ............................................................. 188, 196 cystic fibrosis transmembrane conductance regulator (CFTR)......................................................... 196 GABAA.................................................................... 204 hERG......................196, 199, 203, 210, 211, 212, 213, 218, 219, 220, 221 hyperpolarisation activated cyclic nucleotide-gated . 196 KCNQ2/3................................................................. 196 Kv1........................................................................ 4, 196 Nav1..............2, 210, 211, 213, 214, 215, 216, 217, 221 TASK3...................................................................... 196 TRP........................................................................... 188 See also Transient Receptor Potential (TRP) two-pore domain ...................................................... 196 VGSC ....................................................................... 210 See also Voltage-gated Sodium channel (VGSC) IonWorks1 HT....................194, 195, 196, 197, 198, 203, 204, 205, 210 See also Reader IonWorks1 QuattroTM .................................................. 205 IonWorks............................................................................ 8 See also Reader Iressa See also Tipifarnib Isomerase ........................................................................ 108 See also Enzyme
K KCNQ2/3....................................................................... 196 See also Ion Channel Ki relationship between IC50 and Ki.......... 11, 12, 16, 120
HIGH THROUGHPUT SCREENING
264 Index
Kinase............110, 114, 115, 118, 120, 137, 138, 160, 175, 179, 226, 227, 228, 231 See also Enzyme Kinetic constants...................................................................... 14 read.................................................................... 117, 120 Km ..............................................10, 11, 12, 15, 16, 19, 138 Kv1.............................................................................. 4, 196 See also Ion Channel
L Label ....................................................................... 131, 247 Laboratory...........................................56, 68, 121, 256 n.19 Labware....................................................................... 23, 36 LED........................................................................ 150, 151 Library.....................42, 46, 49, 56, 57, 58, 60, 61, 66, 119, 136, 167, 175, 176, 178, 179, 205, 206 Lidocaine ........................................................................ 188 See also Prescription Drugs Ligand, allosteric....................................................... 16, 131 Ligase ...................................................................... 108, 114 See also Enzyme Light Emitting Diode (LED)................................ 150, 151 See also LED Linear translation.............................................................. 36 Liquid handling ................7, 19, 21, 22, 23, 28, 64, 75, 83, 121, 150, 172, 174, 175, 176, 199, 200, 221, 228, 230, 231, 232 See also Robot Luminescence .............................6, 7, 8, 50, 127, 154, 241, 242, 243, 244
M MAD ................................................................................ 88 Manufacturer ALA Scientific alembic Instruments.................................................. 199 the Automation Partnership....................................... 46 Beckman Coulter ....................21, 22, 36, 170, 228, 231 Caliper BioSciences cellectricon ................................................................ 202 Essen Instruments..................................................... 194 Flyion GmbH Hudson Control Group.............................................. 37 molecular devices ..................7, 42, 132, 137, 146, 153, 194, 195, 198, 210, 226 multichannel systems ................................................ 203 Nanion Technologies................................................ 198 REMP....................................................................... 467 Sophion Bioscience........................... 192, 198, 211, 218 Velocity11 ................................................................... 36 Master buffer .......................................................... 227, 229 Master plate .............................................................. 66, 228
Mean .........................23, 70, 72, 78, 80, 88, 91, 92, 93, 94, 96, 97, 98, 100, 102, 110, 250, 252 See also Statistical analysis Mean Absolute Distance (MAD) .................................... 88 Microfluidic .................................................... 112, 225–236 See also Assay Microtitre plate.................................72, 241, 242, 245, 248 Minimum Significant Ratio (MSR)................................. 73 See also MSR Molecular Devices ......................7, 42, 132, 137, 146, 153, 194, 195, 198, 210, 226 Molecular Libraries Screening Center Network ............ 108 Mother plate ................................................................... 228 MSR.................................................................................. 73 See also Statistical analysis Multichannel Systems..................................................... 203 See also Manufacturer Multidrop......................................148, 169, 170, 228, 229, 230, 232, 246, 248 See also Robot
N Nanion Technologies...................................................... 198 See also Manufacturer Natural products ................................................................. 4 Nav1.2...........................................210, 211, 213, 214, 215, 216, 217 See also Ion Channel Nav1.7............................................................................. 196 See also Ion Channel NexavarTM See also Prescription Drugs Nifedipine ....................................................................... 189 NIH ........................................................................ 108, 122 Normal distribution ..............25, 70, 73, 80, 93, 94, 96, 97, 98, 99, 101, 102 See also Statistical analysis Normalized values................................................... 123, 233 See also Statistical analysis Norvasc1 ........................................................................ 189 See also Prescription Drugs NR................................................................... 131, 136, 137 See also Receptor Nuclear receptor (NR), see NR Nuclear trafficking ................161, 162, 164, 165, 166, 167, 168, 170, 173, 175, 181, 184 n.10
O Oocyte............................................................................. 203 Operation..............2, 35, 36, 48, 52, 53, 63, 64, 74, 75, 87, 165, 166, 171, 178, 179, 226, 232 OpusXpress1 6000A ...................................................... 203 See also Reader
HIGH THROUGHPUT SCREENING
Index 265
Outlier.................23, 47, 50, 51, 73, 78, 79, 85, 97, 98, 99, 100, 101, 233, 235 See also Statistical analysis Oxonol dye...................................................................... 190 See also Dye
P Patchbox.................................................................. 201, 202 See also Reader Patch clamp automated ............................................. 9, 192, 209–222 See also Electrophysiology; Assay gigaseal...................................................................... 210 high resistance seal.................................................... 210 planar.........................................194, 198, 201, 202, 212 voltage steps .............................................. 194, 214, 215 Patchliner# ..................................................... 200, 201, 205 See also NPC16 Patchliner# Patchliner NPC–16 ........................................................ 198 See also Reader PatchPlate ....................................................................... 197 PatchXpress......................................................................... 8 See also Reader PBS .................148, 169, 170, 245, 246, 247, 248, 255 n.10 Peptide ......................114, 226, 227, 228, 229, 232, 236 n.2 Percent inhibition calculation...................................236 n.5 Perforated patch clamp See also Electrophysiology Pharmacophore ................................................................... 4 Phenotypic assay ................................................. 5, 240, 241 See also Assay Phosphatase ....................108, 109, 110, 113, 116, 117, 226 See also Electrophysiology Phosphorylated product.......................................... 115, 226 Phosphorylation ..............................5, 9, 160, 226, 227, 232 Photina1 ......................................................................... 154 Photoprotein ............................................................... 7, 154 Pipetting workstation ................................................. 34, 35 See also Robot PKA ........................................................................ 225–236 See also Enzyme Planar patch clamp..........................194, 198, 201, 202, 212 See also Electrophysiology Plate 1536-well .................................................... 40, 146, 190 384-well ..............18, 20, 26, 40, 42, 66, 111, 117, 132, 136, 148, 150, 151, 156, 159, 161, 164, 165, 168, 169, 170, 173, 175, 176, 177, 179, 183 n.4, 193, 197 96-well ....................................18, 40, 66, 149, 160, 164 barcode ............................................................ 39, 46, 47 cherry picking........................................................ 46, 67 daughter ................................................ 66, 67, 228, 231 hotel ................................................................ 36, 43, 44
microtiter .............................................. 6, 118, 146, 148 mother................................................................. 66, 228 plastic ................................................ 161, 164, 170, 183 transport system .................................................... 42, 43 Plate gripper.................................................... 35, 36, 42, 44 See also Robot Plate reader ...................7, 8, 40, 41, 42, 43, 116, 117, 131, 132, 137, 141 n.1 Plate sealer .................................................................. 44, 45 See also Robot Plate washer ................................................ 43, 44, 150, 255 See also Robot Polarization...........................................7, 13, 15, 114, 121, 122, 127–142 PolarScreen ..................................................................... 138 See also Assay Population Patch Clamp (PPC)................... 193, 195, 198, 203, 206 See also PPC Port-a-Patch ........................................................... 201, 202 See also Reader Positive....................21, 22, 23, 24, 26, 71, 80, 92, 96, 135, 139, 149, 156, 163, 228 Power of an assay .............................................................. 93 See also Statistical analysis PPC.................................................193, 195, 198, 203, 206 See also Electrophysiology Precision radius......................................................... 73, 100 See also Statistical analysis Preincubation ............................................16, 17, 18, 20, 27 Prescription drugs amlodipine ................................................................ 189 gleevec ....................................................................... 225 See also Imatinib mesylate iressa lidocaine .................................................................... 188 nexavarTM See also Sorafenib tosylate norvasc1 ................................................................... 189 See also Amlodipine tegretol1 ................................................................... 189 See also Carbamazepine valium See also Diazepam zarnestraTM See also Tipifarnib Price-Supplier Score ......................................................... 61 Primary cells.................................................... 201, 240, 241 Probenecid .............................................................. 148, 154 Process.......................................2, 3, 19, 21, 23, 35, 38, 46, 49, 51, 52, 53, 58, 60, 61, 64, 65, 66, 68, 71, 74, 75, 77, 78, 81, 82, 91, 92, 103, 109, 129, 131, 141, 154, 184–185, 189, 193, 228, 230, 244, 256
HIGH THROUGHPUT SCREENING
266 Index
Product/Sum ratio (PSR) ...............226, 228, 229, 230, 280 See also PSR Protease.........................9, 20, 110, 111, 114, 120, 131, 226 See also Enzyme
Q QC ......................2, 25, 72, 74, 75, 76, 77, 79, 82, 93, 131, 137, 252, 256 QPatchTM HT........................................................ 198, 200 See also Reader QPlate .............................................198, 199, 216, 218, 219 QSAR ............................................................................... 71 Quality Assurance and Quality Control...............2, 25, 72, 74, 75, 76, 77, 79, 82, 93, 131, 137, 252, 256 Quality control (QC) spatial uniformity correction..................... 151, 156, 157 troubleshooting ............................................... 23, 35, 71 See also QC Quencher......................................................... 113, 114, 137
R R-848 .............................................................................. 241 Radiolabel ligand .................................................... 145, 190 Radioligand..................................................... 145, 190, 211 Radioligand binding ....................................... 145, 190, 211 See also Assay Raw data .............................................23, 49, 50, 76, 94, 96 Reader absorbance............................................................. 24, 49 AlphaScreen..................................................7, 8, 9, 226 Analyst AD ............................................... 132, 133, 136 Apatchi-1TM ............................................................. 192 automated patch clamp......................... 9, 192, 209–222 CyBi1-Lumax .......................................................... 154 FLIPR.........................9, 146, 147, 148, 149, 150, 151, 152 n.1, 154, 155, 156 See also Fluorometric Imaging Plate Reader (FLIPR) FlipTip .............................................................. 200, 201 FlyScreen1........................................................ 200, 201 IonWorks ...................................................................... 8 IonWorks1 HT................................................ 194, 195 IonWorks1 QuattroTM ............................................ 205 OpusXpress1 6000A................................................ 203 Patchbox............................................................ 201–202 Patchliner NPC-16................................................... 198 PatchXpress...........................8, 198, 199, 200, 205, 210 Port-a-Patch ............................................................. 202 QPatchTM HT.................................................. 198, 200 Robocyte ................................................................... 203 Reagent addition .......................15, 121, 123, 135, 154, 155, 157 management.................................................... 60, 61, 62 selection....................................................................... 58
Receptor concentration ........................................................ 16, 17 GPCR .....................3, 4, 6, 7, 8, 9, 145, 146, 147, 149, 155, 157 NR............................................................. 131, 136, 137 relationship between IC50 and Ki.......... 11, 12, 16, 120 TLR .......................................................... 241, 242, 251 Recombinase ................................................................... 108 Registration...............................................57, 58, 59, 60, 65 Relative fluorescence units (RFU).......................... 133, 151 See also RFU REMP............................................................................... 46 See also Manufacturer Resiquimod .....................241, 246, 249, 250, 251, 252, 253 RFU ........................................................................ 133, 151 RNA polymerase............................................................... 27 Robocyte ......................................................................... 203 See also Reader Robot anthropomorphic arm ..................................... 35, 36, 39 Biomek 2000, 22 cherry picking........................................................ 46, 66 dispenser............................................................ 148, 169 hotel ...................................................................... 43–44 linear translation ......................................................... 36 liquid handling............7, 19, 21, 22, 23, 28, 64, 75, 83, 121, 150, 172, 174, 175, 176, 199, 200, 221, 228, 230, 231, 233 Multidrop........................148, 169, 170, 228, 229, 230, 232, 246, 248 pipetting workstation............................................ 34, 35 plate gripper .............................................. 35, 36, 42, 44 plate sealer............................................................. 44, 45 plate washer...................................43, 44, 150, 255 n.11 shelf.......................................43, 44, 56, 57, 58, 65, 206 washer ..................................................... 40, 43, 44, 150 Robust statistics ......................................78, 97, 98, 99, 100 Row marker............................................................. 227, 232 R-score ........................................................................ 73, 94 See also Statistical analysis Running median procedure .............................................. 87 See also Statistical analysis
S Sample integrity................................................................. 64, 65 See also Compound Sample management................................................... 57, 64 SAR............................................................................. 14, 92 See also Structure Activity Relationship (SAR) SB203580..............................164, 165, 166, 167, 168, 169, 172, 174, 175 SBS See also Society for Biomolecular Sciences (SBS)
HIGH THROUGHPUT SCREENING
Index 267
Scheduling ............................................2, 17, 38, 39, 42, 46 Scheduling software.................................................... 42, 46 Scintillation proximity assay (SPA).........7, 8, 9, 13, 14, 17, 34, 112 See also SPA Screening Quality Control ............................. 74, 77, 79, 82 SDI................................................69, 95, 99, 100, 101, 102 See also Statistical analysis Second messenger SEL............................................................................. 90, 91 See also Statistical analysis Separation buffer............................................. 227, 232, 236 Shelf ......................................43, 46, 56, 57, 58, 64, 65, 206 See also Robot Signal to background ........2, 12, 13, 18, 21, 23, 72, 97, 154 See also Statistical analysis Signal Window ...........24, 25, 73, 134, 135, 138, 140, 141, 165, 171, 172, 174, 175, 243, 249, 252 See also Statistical analysis Society for Biomolecular Sciences (SBS) See also SBS Software ..............34, 35, 36, 38, 39, 42, 46, 47, 48, 49, 51, 53, 57, 58, 62, 76, 77, 78, 79, 81, 82, 84, 103, 151, 156, 161, 164, 167, 170, 171, 176, 177, 179, 180, 193, 198, 200, 201, 202, 212, 220, 231, 232 Solution...........7, 8, 12, 21, 34, 42, 46, 54, 57, 58, 64, 107, 114, 118, 120, 122, 124, 127, 128, 129, 130, 148, 150, 153, 161, 169, 170, 171, 183, 190, 192, 193, 197, 199, 200, 202, 204, 211 Sophion Bioscience................................. 192, 198, 211, 218 See also Manufacturer Sorafenib tosylate SPA.....................................................7, 8, 9, 13, 14, 17, 34 See also Assay SPC............................................................................. 74, 82 See also Statistical analysis SSMD............................................................................... 93 See also Strictly standardised mean difference (SSMD) Standard Deviation...............23, 51, 70, 72, 73, 78, 92, 93, 95, 96, 97, 98, 99, 101, 102, 122, 137, 231, 233, 234, 251 See also Statistical analysis Standard Deviation of Inactives ................................. 95, 99 See also SDI Statistical analysis average....................................................................... 231 CV....................................................................... 97, 184 distribution.................................................................. 96 EC100....................................................................... 250 EC50.................138, 160, 164, 167, 172, 183, 192, 250 error...........................................................75, 89, 92, 94 Gaussian distribution.......................................... 99, 234 IC50 ..............................11, 12, 15, 110, 121, 138, 164, 171, 230
Intra-Plate.................................................77, 78, 81, 85 MAD .......................................................................... 88 mean........................................................49, 50, 70, 76, 87, 92, 96 MSR, see Minimum Significant Ratio (MSR) normal distribution ......................................... 70, 73, 96 normalised values ........................................................ 94 outlier .................................................................. 73, 233 power of an assay ........................................................ 93 precision radius ................................................... 73, 100 R-score .................................................................. 73, 94 running median procedure.......................................... 87 SDI, see Standard Deviation of Inactives (SDI) SEL....................................................................... 90, 91 See also Systematic Error Level (SEL) signal to background.....................2, 12, 13, 18, 21, 23, 72, 97, 154 Signal Window .............24, 25, 73, 134, 135, 138, 141, 145, 171, 172, 175, 243, 249, 252 spatial uniformity correction..................... 151, 156, 157 SPC....................................................................... 74, 82 See also Statistical Process Control (SPC) SSMD......................................................................... 93 standard deviation.................23, 51, 69, 70, 72, 73, 78, 92, 93, 95, 96, 97, 98, 99, 101, 122, 137, 231 threshold .......................49, 51, 73, 80, 91, 94, 97, 161, 176, 180, 197, 226, 233 VEP....................................................................... 88, 89 See also Variance Explained by the Patterns (VEP) Z’ calculation............................................................. 236 Z’ factor........................................................... 25, 26, 47 Statistical Cut-Off ..........................92, 94, 97, 98, 100, 103 Statistical evaluation ................................................... 21, 71 Statistical Process Control (SPC)......................... 71, 74, 82 See also SPC Staurosporine .................................................. 227, 230, 231 Storage ..................................2, 40, 44, 46, 57, 64, 75, 149, 213, 222, 247 See also Compound Storage (PBMC)...........................2, 37, 40, 44, 46, 57, 64, 75, 149, 213, 222, 240, 241, 243, 244, 246, 247, 248, 249, 252, 254, 256 Strictly standardised mean difference (SSMD)................ 93 See also SSMD Structure Activity Relationship (SAR) ...................... 14, 92 See also SAR Sub-cellular distribution ................................................. 159 Substrate ...................................4, 6, 10, 11, 12, 13, 15, 20, 73, 109, 111, 114, 116, 119, 120, 124, 138, 160, 193, 195, 198, 199, 200, 202, 205, 206, 226, 227, 229, 232 Subtract bias.................................................................... 156 Systematic Error Level (SEL) .................................... 90, 91 See also SEL
HIGH THROUGHPUT SCREENING
268 Index T
V
Target.........2, 3, 5, 9, 22, 27, 108, 109, 110, 116, 131, 132, 135, 139, 140, 146, 149, 152, 160, 162, 172, 181, 185, 187, 189, 191, 196, 206, 211, 226, 240, 254 Target-to-Lead effort ......................................................... 5 Target type...................................................................... 4, 9 TASK3, 196 Tegretol1 ........................................................................ 189 See also Prescription Drugs Temperature..............18, 19, 20, 28, 42, 44, 47, 50, 75, 83, 87, 120, 121, 124, 133, 135, 148, 150, 154, 161, 169, 170, 188, 229, 230, 232, 246, 248 Termination buffer ................................. 227, 229, 230, 232 Threshold................49, 51, 72, 73, 80, 82, 91, 93, 97, 161, 176, 177, 179, 180, 181, 184, 197, 226, 233 See also Statistical analysis Time-resolved fluorescence resonance energy transfer (TR-FRET)........................6, 8, 9, 77, 114, 116 See also TR-FRET Tipifarnib Tissue culture flask ................................................. 148, 169 Titration..................21, 115, 119, 132, 133, 134, 135, 136, 138, 139, 228 TLR ................................................................ 241, 242, 251 See also Receptor Toll-like receptor (TLR), see TLR Tracking.................................................................... 65, 161 See also Compound Transcreener PDE .......................................................... 138 See also Assay Transfection ......................................18, 148, 149, 152, 153 Transient Receptor Potential (TRP).............................. 188 Transport system................................................... 37, 42, 43 TR-FRET.............................................6, 8, 9, 77, 114, 116 See also Assay Troubleshooting.................................................... 23, 35, 71 TRP channel ................................................................... 188 See also Ion Channel Two-pore domain ion channel ....................................... 196 See also Ion Channel
Valium See also Prescription Drugs Valyl-tRNA synthetase............................................... 13, 14 See also Enzyme Variance Explained by the Patterns (VEP)................ 88, 89 See also VEP Velocity11 ......................................................................... 36 See also Manufacturer Vendor ........34, 35, 38, 40, 42, 43, 46, 48, 52, 54, 228, 241 VEP............................................................................. 88, 89 See also Statistical analysis Verapamil........................................................................ 188 VGSC ............................................................................. 210 See also Ion Channel Vmax ................................................................................. 16 Voltage-dependent block................................................ 221 Voltage-gated sodium channel (VGSC) ................ 210, 211 See also VGSC Voltage-sensing dye............................................................ 8 See also Dye Voltage steps ................................................... 194, 214, 215
U Uncompetitive...............................11, 12, 16, 119, 139, 140 Use-dependent block ...................................................... 189
W Washer ..........................40, 43, 44, 150, 243, 246, 248, 255 See also Robot 96-well Plate ...................................................... 18, 40, 149, 160 See also Format 384-well Plate ............18, 20, 24, 26, 40, 66, 111, 117, 151, 161, 165, 174, 175, 176, 177, 179, 183, 197, 226, 229, 249, 255 n.11 See also Format 1536-well format ............................................................ 192
X Xenopus oocyte ............................................................... 203
Z ZarnestraTM Z’ calculation................................................................... 236 Z’ factor....................................................................... 25, 26 See also Statistical analysis See also Prescription Drugs