Biophysical Characterization of Proteins in Developing Biopharmaceuticals [2 ed.] 0444641734, 9780444641731

Citation preview

BIOPHYSICAL CHARACTERIZATION OF PROTEINS IN DEVELOPING BIOPHARMACEUTICALS SECOND EDITION

Edited by

DAMIAN J. HOUDE Biomolecular Discovery, Relay Therapeutics, Cambridge, MA, United States

STEVEN A. BERKOWITZ Consultant, Sudbury, MA, United States

Elsevier Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States Copyright © 2020 Elsevier B.V. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-444-64173-1 For information on all Elsevier publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Susan Dennis Acquisition Editor: Anita Koch Editorial Project Manager: Sara Pianavilla Production Project Manager: Surya Narayanan Jayachandran Cover Designer: Matthew Limbert Typeset by TNQ Technologies

Contributors David A. Keire Food and Drug Administration, St. Louis, MO, United States

Yves Aubin Centre for Biologics, Regulatory Research Division, Evaluation, Biologics and Genetic Therapies Directorate, Health Products and Food Branch, Health Canada, Ottawa, ON, Canada Steven A. Berkowitz United States

Francis Kinderman Amgen, Thousand Oaks, California, United States Lee Makowski Department of Bioengineering and Department of Chemistry and Chemical Biology, Northeastern University, Boston, MA, United States

Consultant, Sudbury, MA,

George M. Bou-Assaf Analytical Development, Biogen, Cambridge, MA, United States

John P. Marino National Institute of Standards and Technology, Institute for Bioscience and Biotechnology Research, Rockville, MD, United States

Mark L. Brader Drug Product Analytical Development, Moderna, Cambridge, MA, United States

Alan G. Marshall Ion Cyclotron Resonance Program, National High Magnetic Field Laboratory, Florida State University, Tallahassee, FL, United States; Department of Chemistry & Biochemistry, Florida State University, Tallahassee, FL, United States

Richard K. Burdick Burdick Statistical Consulting, LLC, Colorado Springs, CO, United States John F. Carpenter Department of Pharmaceutical Sciences, Center for Pharmaceutical Biotechnology, University of Colorado Anschutz Medical Campus, Aurora, CO, United States Stephen J. Demarest Eli Lilly Biotechnology Center, San Diego, CA, United States

A.J. Miles Institute of Structural and Molecular Biology, Birkbeck College, University of London, London, United Kingdom

Ertan Eryilmaz Takeda, Cambridge, Massachusetts, United States

John S. Philo Alliance Protein Laboratories, San Diego, CA, United States

Verna Frasca Field Applications Manager, Malvern Panalytical, Northampton, MA, United States

Angelika Reichel Coriolis Pharma, Munich, Germany Deniz B. Temel Bristol-Myers Squibb, Devens, Massachusetts, United States

Darron L. Freedberg Structural Biology Section, Laboratory of Bacterial Polysaccharides, Silver Spring, MD, United States

B.A. Wallace Institute of Structural and Molecular Biology, Birkbeck College, University of London, London, United Kingdom

John P. Gabrielson Elion Labs, A Division of KBI Biopharma, Inc., Louisville, CO, United States Andrea Hawe Germany

Coriolis

Pharma,

Daniel Weinbuch Germany

Coriolis Pharma, Munich,

William F. Weiss IV Bioproduct Research and Development, Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, IN, United States

Munich,

Damian J. Houde Biomolecular Discovery, Relay Therapeutics, Cambridge, MA, United States

Sarah Zölls

xi

Coriolis Pharma, Munich, Germany

Prefaces for the second edition Critical to the development of any successful therapeutic drug is our ability to identify and manufacture the drug such that its beneficial therapeutic effect can be safely delivered to the patient. In the case of protein biopharmaceuticals, these large, heterogeneous (complex) and marginally stable molecules are often very sensitive to their micro-environment. This makes the process of developing and manufacturing a protein therapeutic extremely challenging. Throughout this entire process a protein biopharmaceutical must maintain its complex and delicate structure (or conformation) to realize its beneficial therapeutic attributes, while avoiding the potential harmful effects in failing to achieve this goal. When we set out to write the first edition of this book, our goal was to provide a general resource that specifically dealt with the many challenges associated with the testing and characterization of the higher order structure and biophysical properties of protein biopharmaceuticals from a practical point of view to support its safety and beneficial therapeutic activity. As stated in the book’s first preface we wanted to keep the reader focused on obtaining a pragmatic understanding and knowledge of the utility of biophysical tools and how they are used to meet these challenges by understanding what information can realistically be extracted from these tools. While we felt we had initially achieved our goal, the progression of time inevitably led to better and improved

scientific developments and to the realization that there was room for improvements. As a result, in writing this second edition we have undertaken the job of updating old information, correcting mistakes, improving clarity and the introduction of new topics that were not covered in the first edition. Therefore, we gathered our co-authors once again, invited a few new ones, and tasked ourselves with the goal to achieve these objectives. In so doing, all original chapters have been updated, corrected and enhanced, while new chapters have been added. Globally, the format of the book has remained the same, consisting of three sections. Section I, which deals with the complexity of proteins and the relevance of biophysical methods in the biopharmaceutical industry. It has for the most part been altered to remove errors and achieve clarity. Section II, which discusses the biophysical tools and techniques most commonly used in the biopharmaceutical industry to characterize protein therapeutic molecules has similarly been altered, but has also been enhanced by the addition of a new chapter (Chapter 14) dedicated to the area of chromatography and electrophoresis. The tools in this chapter, which we did not cover in the first edition of the book (with the exception of size-exclusion chromatography), are typically not thought of or classified as biophysical tools. Nevertheless, an important objective in adding this chapter is to bring more attention

xiii

xiv

PREFACES FOR THE SECOND EDITION

to their unrealized linkage as effective biophysical characterization tools without getting too deep into the details of their inner workings (which are extensively covered in many excellent books and review articles that are solely dedicated to these two enormously important techniques). Overall, however, Section III of the book has experienced the most significant change and expansion via the addition of four new chapters that cover the following: • Chapter 15, which deals with the biophysical characterization of complex biopharmaceuticals; • Chapter 16, which deals with the rigor of statistical analysis; • Chapter 17, which deals with biopharmaceutical developability; • Chapter 18, which deals with technical decision making.

Finally, we would like to point out that in writing this second edition we have made a particular effort, wherever possible, to better link and cross-reference information in each chapter to bring more cohesion to the book as appose to just providing the reader with a collection of isolated chapters. We think his cohesion is in particular made apparent by the four additional chapters in Section III (described above). In the end, we and our coauthors hope we have further enhanced the initial objective of the first edition of the book, of enlightening the reader to the challenges, tools and inner workings of the task associated with the biophysical characterization of protein biopharmaceuticals. An integral part of today’s modern and challenging world of developing lifesaving drugs.

C H A P T E R

1

The complexity of protein structure and the challenges it poses in developing biopharmaceuticals Steven A. Berkowitza, Damian J. Houdeb a

Consultant, Sudbury, MA, United States; bBiomolecular Discovery, Relay Therapeutics, Cambridge, MA, United States

1.1 The basics of protein higher order structure (HOS) Proteins are an important class of large biological molecules that are classified more generally as macromolecules or polymers. However, given their biological origin, these unique molecules are often referred to as biomacromolecules or biopolymers. They are truly complex, particularly when compared to synthetic (man-made) polymers and even other types of biopolymers, e.g., DNA. One of the main reasons for this complexity arises from their basic building blocks, which in synthetic polymer chemistry are referred to as monomer units. In the case of most synthetic polymers, the chemical composition consists typically of only one type of monomer (although some synthetic polymers called copolymers or block-copolymers are composed of two or possibly more different monomer units). Proteins made in nature via a process called translation utilizing the genetic code are composed of not one, two, or even three different monomer units, but rather are composed of as many as 20 different “naturally” occurring monomer units called amino acids. These 20 amino acids (or proteinogenic amino acids, which does not include the other know, but rare proteinogenic amino acids selenocystine or pyrrolysine) are referred to as the standard amino acids. Although not all proteins contain all 20 amino acids, most do. The presence of such a large diversity in chemical composition, in virtually every protein, is a key element for their structural complexity, which in turn gives rise to their diverse functionality. Indeed, this chemical complexity, coupled with the large number of amino acid units or residues (N) present in proteins (that can number in the thousands), and the uniqueness of the amino acids linear sequential arrangement (which in protein chemistry is called the primary (1
) structure, see Fig. 1.1A), enables a

Biophysical Characterization of Proteins in Developing Biopharmaceuticals, Second Edition https://doi.org/10.1016/B978-0-444-64173-1.00001-9

3

Copyright © 2020 Elsevier B.V. All rights reserved.

4

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

FIG. 1.1

(A) The linear sequential ordering of amino acids (represented by the rectangular black dashed boxes) in a protein is referred to as its primary structure. The extreme left amino acid corresponds to the amino-terminus, while the extreme right amino acid corresponds to the carboxyl-terminus end of the protein chain. The gray shaded area corresponds to the peptide bonds that link all the amino acid units in a protein, yielding the polypeptide backbone (or chain) indicated by the red (gray in print version) dotted rectangle. (B) An illustration of the planar structure of two adjacent amide planes (each resulting from the double bond character, due to resonance, of the peptide bond shown as black dashes), corresponding to the light blue (light gray in print version) shaded areas in (A), where the bottom amide plane is formed from the peptide bond between the carboxyl group of amino acid 1 (containing R1) and the amino group of amino acid 2 (containing R2) and the top amide plane is formed from the peptide bond formed between the carboxyl group of amino acid 2 and the amino group of amino acid 3 (containing R3). Due to steric issues, angular rotation around CaN (expressed by F, phi) and CCa (expressed by J, psi) bonds are limited. (C) A representation of a common secondary structure, the a-helix. The small rectangle outlined in black dashes corresponds to a small section of the helical arrangements of the amide planes, shown in (B).

staggering number of different possible proteins, 20N, to be made. Given the enormous array of different proteins that can be made, the cell has exploited this diversity in protein structure to create proteins to perform nearly every functional and structural role needed for its existence.

I. Proteins and biophysical characterization in the biopharmaceutical industry

1.1 The basics of protein higher order structure (HOS)

5

In proteins, the amino acid units are linked together through a unique chemical bond called the peptide bond, which is also referred to as the amide link, see Fig. 1.1A. The collection of these peptide bonds in a given protein form a common element found in all proteins called the polypeptide backbone or chain, see Fig. 1.1A. A unique feature of the peptide bond is the planar structure that it forms between the carbonyl oxygen, carbon and the a-carbons (Ca or alpha carbon) of one amino acid and the amide nitrogen, hydrogen and a-carbons of an adjacent amino acid. The resulting planar feature of these linked atoms arises as a result of the partial double bond character that exists between the carbonyl carbon (C) and the amide nitrogen (N) atoms due to the presences of resonance structures, see Fig. 1.1B. This planar structure and its attributes play an important role in a protein’s structure, as its presences confines the polypeptide backbone to only certain configurations, via steric effects, which restricts the angular range of bond rotation around the CaeN (expressed by F, phi) and C-Ca (expressed by J, psi) bonds. These restrictions have been summarized in a 2-dimensional graphical plot called the Ramachandran plot, developed by Ramachandran and others in 1963 [1]. Such a plot graphically shows how certain structural features of proteins can only exist within a limited range of angles characterized by J and F, e.g., a-helix, see Fig. 1.1C. These restrictions play an important role in the development of protein’s spatial structure or higher order structure (HOS).

1.1.1 The levels of protein HOS In developing protein biopharmaceuticals and in studying proteins in general, the most important concept is “structure”. In the previous section, we briefly discussed the most basic component of a protein’s structure, its linear sequence of amino acids, or primary structure. However, the focus of this book is concerned with a protein’s three-dimensional (3D) or spatial structure, also referred to as its conformation or HOS. Ultimately, when considering the structures of proteins, it is the HOS in concert with its primary structure (which also includes all the primary chemical bond modifications that occur to its amino acid units, see Section 1.1.4) that enables a protein to properly function or, as we will also discuss in latter sections, malfunction. In terms of protein HOS, there are three different levels that have been defined. These three levels include: secondary (2
), tertiary (3
), and quaternary (4
) structure, see Fig. 1.2. The first two structural levels are concerned with a single polypeptide chain, while the latter is associated with protein structures that involve the interaction of two or more polypeptide chains. A protein’s 2
structure refers to the local folding patterns of a protein’s polypeptide chain, in which the a-helix (see Fig. 1.2A), the b-sheet, turns, and random coils are the most prominent resulting structural elements that are formed. These local folded elements can further participate in higher levels of folding that involve an array of secondary structural elements that give rise to the final 3D structure of a protein referred to as 3
structure of a protein; see Fig. 1.2B. The summation of 2
, 3
and (if present) 4
structure, along with its entire 1
structure, is what gives a protein its unique structure, chemical and physical properties and therefore its unique function. Indeed, it is this relationship between structure and function that is the genesis of the protein “structure-function” concept, which states that a protein’s structure determines its function.

I. Proteins and biophysical characterization in the biopharmaceutical industry

6

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

FIG. 1.2 Illustration of the three levels of a protein’s HOS. (A) Representative secondary structural element, as illustrated by a ribbon representative structure of an a-helix. (B) A cartoon representation of the folding of all the secondary structural elements in a polypeptide chain, which gives rise to the polypeptide’s tertiary structure. (C) A cartoon representation of the quaternary structure of a protein, which arises when the final protein structure involves the association of more than one polypeptide chain to form the final folded protein structure (also see Fig. 1.3).

Although the folding and interactions of the secondary structural elements can give rise to an enormous array of different protein tertiary structures, each with unique properties and functions, it’s not uncommon to find that the 3
structure of a protein often consists of one or more commonly folded patterns called motifs, super-secondary structures, or complex folds [2e4]. These commonly folded structures contain several folded secondary elements involving only a portion of the entire polypeptide chain of a protein, which can blur some of the distinction between a protein’s 2
and 3
structure. Hence, one might look at motifs, super-secondary structures, or complex folds as “local 3
structure”, while referring to the 3
structure of the entire protein molecule as its “global 3
structure”. Another structural element that further subclassifies the structural level of a protein between what we call a protein’s 2
and 3
structure is the concept of domain [5,6]. Domains are typically a much larger collection of folded structural elements than motifs, supersecondary structures, or complex folds. In terms of the global structure of a protein, domains correspond to one or more independent compact region of a protein’s polypeptide chain, as

I. Proteins and biophysical characterization in the biopharmaceutical industry

1.1 The basics of protein higher order structure (HOS)

7

FIG. 1.3 Different representations of the HOS of a monomeric IgG1 antibody. The two heavy chains are colorcoded in blue (light gray in print version) and gray, while the two light chains are both color-coded in red (dark gray in print version). (A) A ribbon model of an IgG1 antibody (PDB: 1HZH). The black circle corresponds to the variable domain on one of the IgG1 light chain (VL). (B) A simplified cartoon of the monomeric IgG1 antibody indicating the various sections of individual domains present. The black lines linking the various interchain domains correspond to areas where covalent linkages exist (disulfide bonds) between different polypeptide chains in the IgG1 molecule. The black circle corresponds to the same VL domain in the IgG1 molecule as shown in (A). (C) A spacefilling structural model of the monomeric IgG1 antibody. The black circled region again corresponds to the same VL domain in the IgG1 antibody as shown in (A). (D) A linear depiction of a monomeric IgG1 structure showing all the various covalent linkages (disulfide bonds) present in the IgG1 antibody. Those disulfide bonds present within the same polypeptide chain are referred to as intrachain disulfide bonds, while those disulfide bonds that link two different polypeptide chains are referred to as interchain disulfide bonds.

indicated by the black circles shown in Fig. 1.3AeC. Proteins containing two or more domains are frequently referred to as multidomain proteins. In these proteins, the domains are chemically linked by short sections of the polypeptide chain that are typically highly flexible, called a “linker”, but nevertheless exist as stable and independent folded units. In certain cases, common domain structures can also be found in proteins much like that observed with motifs, super-secondary structures, or complex folds. What is interesting about these folded elements is that there is a certain amount of change in the 1
structure that can be tolerated while still arriving at, effectively, the same folded structure. This observation explains the common presence of similar secondary, supersecondary, and even domain structures seen in different proteins with different sequences. Hence, the formation of these basic folding elements can display some level of discrepancy in terms of the required or allowable amino acid sequence variations and still give rise to

I. Proteins and biophysical characterization in the biopharmaceutical industry

8

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

the same functioning protein. This feature plays an important role in biological evolution, in generating HOS building blocks, and in controlling and regulating groups of proteins that perform very similar functions in different biochemical pathways [7e9]. Nevertheless, it is important to mention that in proteins, there exist many sequence regions where even a slight change, i.e., one amino acid change or a minor chemical modification (e.g., oxidation, deamidation), can significantly alter a protein’s structure and therefore its function [10,11]. For many proteins, however, the unique folded state of its polypeptide chain is not the last step in attaining a final overall 3D structure. Many proteins are composed of more than one polypeptide chains, which may be identical or nonidentical, giving these proteins an added level of structural complexity, 4
structure; see Fig. 1.2C. When referring to a protein’s 4
structure, a lack of clarity or confusion can unfortunately arise. An example is illustrated in Fig. 1.3. In this figure, a monomeric intact IgG1 antibody is shown. However, this protein could be referred to as a protein dimer (made of two identical protein units) or a protein tetramer made of four separate polypeptide chains, which in this case are chemically cross linked via covalent bonds called disulfide bonds (which is the most common primary bond used in nature to cross-link parts of polypeptides). Such a choice of descriptive words unfortunately can lead to some confusion. As a result, some care should be taken when describing the basic structure of a protein. In the case of the 4
structure of IgG1 molecule, as shown in Fig. 1.3, the use of a tetramer in the context of its 4
structure would be correct. However, in the context of a complete functioning unit (in its lowest complete form) the molecule is a monomer.

1.1.2 Stabilizing the HOS of proteins In all three levels of a protein’s HOS (i.e., 2
, 3
, and 4
), various changes in the conformation of the polypeptide chain(s) occur as a protein folds to reach its final native structure. These changes are typically accompanied by an increase in overall structural order, which imparts a significant reduction in the protein’s entropy that by itself is highly unfavorable, in terms of the overall free-energy change. However, as a protein folds, various weak noncovalent (secondary) bonds form via ionic, dipoles (hydrogen bonds), nonpolar (hydrophobic effect), and van der Waals interactions. These weak bonds involve the interactions of amino acid side chains, as well as elements of the polypeptide backbone, particularly the amide hydrogen. While individually these interactions are weak, during the folding process their large number and the cooperative way they form provide the necessary enthalpic and entropic driving forces (release of structured water via the hydrophobic effect) to override the large unfavorable decrease in entropy that occurs as a protein folds into its native (more ordered) conformation. The stabilization of the folded protein, however, is only marginal. Comparing the level of stabilization against the average thermal energy content of a protein molecule (which is equal to kT, where k ¼ Boltzmann constant and T ¼ temperature) and the distribution of this energy, in terms of the amount of thermal energy per molecule, a variety of these weak secondary bonds can be broken as a function of time. Such spatial and temporal rupturing of these weak secondary bonds enables a protein to display dynamic structural properties in its conformation (sometimes referred to as protein breathing). This dynamic property can play an important role in a protein’s function [12e15] and stability [16,17]. This dynamic property, however, can also constitute a weakness for protein biopharmaceuticals, given the

I. Proteins and biophysical characterization in the biopharmaceutical industry

1.1 The basics of protein higher order structure (HOS)

9

wide range of stressful environments an average biopharmaceutical must endure during its biosynthesis, purification, formulation, packaging/storage, patient handling, and its administration. Hence, in searching for a good therapeutic biopharmaceutical, scientists look for molecules with high stability, such that the dynamic properties of the protein do not result in loss of activity or adverse structural changes. Proteins that have such attributes are said to have good developability properties. In addition to weak secondary bonds, stabilization of the HOS of a protein can be achieved through primary bonds formed between folded elements within a protein. As already mentioned, the most common such bond is the disulfide bond, see Fig. 1.3D. Although the number of disulfide bonds found in a given protein typically amounts to only a few such bonds per protein molecule (and may not even exist within some proteins), they often play important roles in a protein’s overall structure-function and stability [18]. Disulfide bonds can occur both within a single polypeptide chain (where they are referred to as intrachain disulfide bonds; see Fig. 1.3C) and between two different polypeptide chains in the same protein (where they are referred to as interchain disulfide bonds; see Fig. 1.3C). Disulfide bonds also occur between two different protein molecules where they function to stabilize large complex multiprotein supramolecular structures [19]. Unfortunately, however, the formation of disulfide bonds can go astray leading to altered HOS structures or aggregates via disulfide scrambling or exchange between other disulfide bonds or free cysteine residues in the same protein or different proteins. These modes of protein degradation [20e27] are another reason why the biopharmaceutical scientist need to constantly scrutinize the structure of the biopharmaceutical during its development.

1.1.3 Dynamics properties of a Protein’s HOS The HOS of virtually all proteins is primarily held together by a large array of relatively weak bonds. In the context of a protein’s thermal energy content, these bonds can break enabling various levels of fluctuations within a protein’s HOS that can span an enormous time range, from 1015 s to tens of seconds and even longer [12,28]. Again, the fluctuations in a protein’s conformation essentially occur because of the opening or breaking of various weak secondary bonds. The extent of these fluctuations in terms of amplitude and location is very dependent on many factors, e.g., environmental conditions, the strength of each secondary bond, the distribution of these bonds within the protein, as well as the distribution of thermal energy within the protein. Variations in these (and other) factors will determine the location of which secondary bonds will break in a protein’s HOS and therefore, the nature of the conformational change(s) and the population of protein molecules in a specific conformation as a function of time. While these changes are for the most part contained to a region where the secondary bond(s) break, changes might also extend to other areas of the protein, via allosteric effects. Due to the random nature of the thermal energy fluctuations within a protein, a range of different conformations and populations of different conformational states will exist at any one time. For the most part, the extent of change in a protein’s HOS are typically not that large and are often reversible allowing the altered protein structure to return to its more stable conformations. Consequently, in solution proteins exist as an ensemble of different conformations, rather than as a single fixed unique conformation. This ensemble is limited and controlled by the

I. Proteins and biophysical characterization in the biopharmaceutical industry

10

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

interplay of the overall structure of the protein and its physicochemical environment. However, under appropriate conditions, involving some form of stress or subtle changes in a protein’s chemical structure, changes in conformation may cause a protein to display different physicochemical properties. In the case of a protein biopharmaceutical, changes in its physicochemical properties could alter the drug’s ability to bind with its therapeutic target or enable it to bind to different materials it encounters, e.g., various container closure surfaces [29e33]. Other possible adverse events include the formation of aggregates that are nonfunctional and/or even more concerning, immunogenic [34e36]. It should be noted that the formation of aggregates and their associated link to loss of protein function and/or immunogenicity corresponds to one of the most common forms of protein degradation that is closely monitored in the biopharmaceutical industry.

1.1.4 Finer structural alteration of proteins Once a protein is synthesized, or as it is being synthesized, additional primary structural changes can occur in vivo. In most cases, these changes are due to additional enzymatic processing reactions involving a multitude of potential chemical modifications to various amino acids, as well as changes involving cleavage or cross-linking reactions. These reactions may or may not play an important role in the normal function/activity of a protein, but rather may represent alterations that play out to the determent of the cell or even the organism due to an immunogenic response. Generally, most modifications are confined to the protein’s surface. However, modifications can also occur to the protein’s interior due to the dynamic properties of its structure (which exposes these buried internal areas) or during its synthesis when these normally buried internal areas had not had a chance to properly fold. Such alterations can lead to changes in the local or global HOS of the protein. In general, these modifications are referred to as posttranslational modifications (PTMs). Principally, PTMs occur in vivo and the number of different PTMs that a protein can experience is quite large [37]. In eukaryotes, one of the more common (and biopharmaceutically relevant) PTMs is glycosylation. This modification involves the enzymatic addition of carbohydrate (also called glycan or sugar) units to a protein at specific asparagine (where they are called N-linked glycan) or serine or threonine (where they are called O-linked glycan) amino acid [38]. While most PTMs occur in vivo (inside the cell), PTMs can also occur in vitro (outside the cell). These latter PTMs, however, typically represent forms of protein degradation that occur due to direct physical or chemical interactions (e.g., oxidation, deamidation, glycation, etc.) and are also of great concern in the biopharmaceutical industry as they are often linked to instability leading to lose of drug activity and adverse effects [3,39e44].

1.2 The search for how proteins attain their correct HOS: the protein folding problem In the 1950s and 1960s, biophysical research led scientists to the realization that a protein’s HOS is effectively dictated by its primary sequence. Christian Anfinsen was the key scientist who formalized this idea, and in 1972 was awarded the Nobel Prize in chemistry for his contributions [45]. In the scientific literature, this idea has been frequently referred to as the

I. Proteins and biophysical characterization in the biopharmaceutical industry

1.2 The search for how proteins attain their correct HOS: the protein folding problem

11

“Anfinsen dogma” or the “thermodynamic hypothesis”. The folding path a protein takes to achieve its correct functional HOS is intrinsically dictated by its 1
structure (which may also include PTMs). How the folding process advances so efficiently, in combination with the way a protein is synthesized in vivo, in the specific physicochemical environment within the cell, has fascinated scientists for many years [46]. This fascination stems from the realization that proteins achieve their correct HOS within a matter of milliseconds to seconds! In the 1960s, Cyrus Levinthal prosed the following interesting and simple problem concerning protein folding. For a protein consisting of 100 amino acids in an initially unfolded state, how long would it take this protein to find, through a completely random process, its appropriate native HOS given its physicochemical environment [28]? This problem is nicely restated in the words of Amit Kessel and Nir Ben-Tal in their book “Introduction to Proteins: Structure, Function and Motion” [47] as follows: Assuming that the protein folding process involves the free sampling of all possible conformations of the protein (i.e., of each residue independently), and that each residue has at least three states, then the folding of a 100-residue protein is excepted to sample 3100 ¼ 5  1047 conformations. Now if we assume that it takes a protein 1 picosecond to sample a single conformation, then the time it takes to sample all possible conformations in order to find the right one should be 3100  1012 s ¼ 5  1035 s ¼ 1.6  1028 years. This period of time is about 1018 times longer than the age of the universe!!

This simple problem proposed by Levinthal is called “Levinthal’s Paradox” and was a significant driving force for the generating what is called “the protein folding problem”. Clearly, the nature of protein folding is nowhere as simple as starting with the completely synthesized and unstructured (denatured or random coil) form of a protein, which is then allowed to undergo a completely random sampling process of conformational space. Protein folding must proceed via a process that is enormously more efficient, but how!!? Answers to this problem appear to lie within the idea of a “funnel-shaped folding energy landscape” [48e52], see Fig. 1.4, which might possibly take advantage of the way proteins are made in vivo along with a concept of “divide and conquer”. In this process a protein proceeds to fold through a hierarchy of subassembly units called a “foldon” [53,54]. These units can fold somewhat independent of each other in parallel to form relatively local higher order structures that can eventually collapse into the final native HOS of the protein. In general, the funneling process of protein folding is likely not as simple as that portrayed in Fig. 1.4A. Rather, it is expected to be more complex and treacherous, as indicated in Fig. 1.4B. In the latter scenario, a folding protein could encounter conformational states that are not as optimally folded as its native state and contain high activation energy barriers that inhibit its search to find the most stable conformation. Hence, the protein in these states would find itself trapped, due to the high energy of activation needed to transition the misfolded state back into a more unfolded state so it can find its more stable and native form. Although these misfolded protein forms may be encountered at very low levels under normal conditions, the situation could escalate under stressed conditions, such as forcing a cell to produce a large quantity of one protein in a very short period. For such a situation, a higher frequency of misfolded or metastable folded protein states could be encountered leaving the biopharmaceutical scientist with a more difficult purification process that results in a lower protein drug yield.

I. Proteins and biophysical characterization in the biopharmaceutical industry

12

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

FIG. 1.4 A graphical view of the three-dimensional funnel-shaped energy landscape for protein folding. The top of each funnel corresponds to the completely unfolded protein. The bottom of each funnel plot corresponds to the fully folded protein molecule in its native state, which under closer scrutiny actually consists of a large array of slight different energetically folded states (conformations) that differ in most cases by a small amount of free energy thus enabling the native protein to exist in solution as an ensemble of different conformations. (A) A folding process free of situations where it can be trapped in incomplete or partial folded state. (B) A folding process that enables partially folded proteins to be potentially trapped due to the presence of smaller shaped folding funnels with relatively large energy of activation that must be overcome in order to escape and find its final native state.

1.2.1 In vivo production of proteins: revisiting the protein folding problem Another unique attribute of proteins is the complex manner with which they are made in vivo. Protein synthesis involves a complex array of cellular machinery, the main component of which is the ribosome. In vivo, proteins are synthesized from the N-terminus to the C-terminus in a sequential manner at a rate of 50e300 amino acids/min [55,56]. The specific ordering and chemical coupling of the amino acids for a given protein is achieved by a process called translation, which controls the protein synthesis process dictated by the genetic coding information stored in a specific messenger RNA (m-RNA). As the nascent protein chain is synthesized and exposed to the cell matrix, it can begin to fold. However, it should be noted that the first 50e60 amino acids in the growing polypeptide are initially limited to some extent in their ability to freely fold, due to the physical restrictions (steric hindrance) of the environment within the ribosome [57]. This idea of concurrent, in vivo, protein synthesis and folding are referred to as cotranslational protein folding [58] and likely plays an important role in the folding of newly synthesized polypeptide.

I. Proteins and biophysical characterization in the biopharmaceutical industry

1.2 The search for how proteins attain their correct HOS: the protein folding problem

13

The importance of cotranslational protein folding most likely arises because only the growing polypeptide chain that has advanced beyond the ribosome tunnel will be able to fully participate in the folding process. This allows only a portion of the growing protein chain to fold without the interference from other parts of the protein that has either not been synthesized or is located in the ribosome tunnel. As a result, this should improve the efficiency of the sequential folding of the local higher order structural elements characterized as foldon units to proceed in a more orderly manner. Such foldon units most likely correspond to local HOS elements that are present in the final native protein. Nevertheless, as these local higher order structural elements are formed, they must search out and undergo higher levels of folding as the protein chain continues to grow. As a result, these various hierarchy of folded structural elements are probably not arranged or packed optimally (as they are in the protein’s final native state) until the entire protein is fully synthesized and release from the ribosome. Once this happens, what remaining loose arrangement of folded (or partially folded) structural elements that still exist must collapse into the final native structure of the protein. This final consolation of folded or partially folded structural elements most likely proceeds through the interactions of key amino acid side chains to make the final functional protein (notwithstanding any additional changes in HOS resulting from PTMs and the formation of quaternary structures). Consequently, cotranslational protein folding constrains and to some extent guides the overall protein folding process. By limiting the number of folding pathways (specifically bad folding pathways, which would significantly increase the amount of time required to find the correct native conformation) available to a protein, relative to the situation where folding only begins once the protein is fully synthetized and release from the ribosome, could make the role of cotranslational protein folding truly an important attribute in the successful folding of a protein.

1.2.2 In vivo production of proteins: avoiding and eliminating folding errors via the use of chaperones In vivo, there are mechanisms involving other proteins, called chaperones, that help proteins that are folding avoid the situation of being misfolded. This task is achieved via the chaperone’s ability to assist a folding protein to avoid folding traps by participating in the protein folding process through proteineprotein interactions [59e63]. In addition to chaperones, there also exists in vivo cellular machinery whose function is to identify the presences of misfolded proteins and eliminate them via proteolytic hardware existing within the cell [64]. However, these systems are not perfect, and failure to remove or prevent these erroneously folded proteins from accumulating within the cell can alter the cell, causing adverse effects that could eventually lead to its death. In the case of producing a protein biopharmaceutical, once a misfolded protein is released into the cell culture media, it then becomes the problem for the process scientist to develop appropriate purification strategies to remove the misfolded protein from the final protein drug product. If these erroneously folded proteins are not removed, they could lead to adverse effects when the final drug product is administered to a patient. Hence biophysical analysis of the biopharmaceutical’s HOS again becomes an important activity in developing protein biopharmaceuticals with minimal levels of these misfolded forms in the final drug product.

I. Proteins and biophysical characterization in the biopharmaceutical industry

14

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

1.3 Surprises in the world of protein folding: intrinsically disordered or unstructured proteins (an apparent challenge to the protein StructureeFunction paradigm) Within the past two decades, it has been realized that many proteins, especially in eukaryotes or multicellular organism, exist within the cell with no defined HOS [28]. Rather, these proteins appear to be disordered or unstructured, approaching what might be called a random coil structure, anomalous to what is frequently seen with synthetic polymers or denatured proteins. However, when these proteins interact with their target molecule(s) they commonly appear to take on a level of organized HOS. Hence, this structural disorder is transient in many cases and a disorder-to-order transition occurs during their functioning (i.e., interacting with their binding target). Such behavior could play an important role in allowing these proteins to bind to an array of different partner molecules by taking advantage of the plasticity of their polypeptide chain’s flexibility [28]. This process is liable to be modulated by other factors within the cell, which control and regulate the binding partners they interact with. Indeed, the level of disordered proteins is higher in eukaryotes or multicellular organism, in comparison to prokaryotes, where high levels of signaling and regulation is required. This unique class of proteins has been referred to as intrinsically disordered proteins (IDPs) or intrinsically unstructured proteins (IUPs) [65,66]. The existence of these IDPs would appear to present a challenge to the paradigm of structureefunction discussed earlier in this chapter. With the realization of the existence of IDPs, many of the large random coil-like regions of proteins consisting of 20e30 or more amino acids in length are now being referred to as intrinsically disordered or unstructured regions (IDRs or IURs) [66]. These structural elements are commonly seen as linkers between ordered protein regions such as domains where they are thought to also play important roles in providing protein flexibility, allowing proper folding or to facilitate domainedomain interactions or domain binding to functioning binding targets. At present, IDPs have not made their way into the biopharmaceutical industry, although it is probably only a matter of time until such a protein drug will appear.

1.4 Proteins and the biopharmaceutical industry: problems and challenges Although proteins can be chemically synthesized external to the cell, their high cost (which is a function of protein size), as well as their overall complexity leads to poor economics for building a viable commercial drug business via this approach. Over the years, however, scientists have figured out how to get cells to produce significantly large amounts of a specific protein, by manipulating their DNA via recombinant DNA technology. The development of this capability was the key in enabling the biopharmaceutical drug industry to flourish. Overall this process of making protein biopharmaceutical differ greatly from the classical process used to make simple organic drug molecules called pharmaceuticals, see Fig. 1.5. Cellular and molecular biologist can now produce protein biopharmaceuticals at concentrations expressed in the culture media volume as great as 10 g/L [67]. Nevertheless, the challenges of doing this

I. Proteins and biophysical characterization in the biopharmaceutical industry

1.4 Proteins and the biopharmaceutical industry: problems and challenges

15

FIG. 1.5

A simple comparison illustrating differences in the process of making a pharmaceutical versus making a biopharmaceutical: (A) Coarse outline of the sequential chemical reactions for making a pharmaceutical, using aspirin as an example and (B) a coarse outline of the basic steps for making a biopharmaceutical, which consists of first synthesizing a piece of DNA containing the correct nucleotide sequence code for making the desired biopharmaceutical’s polypeptide chain(s), the insertion of this DNA into an initial small collection of cells (the microscale factories for making the biopharmaceutical) using recombinant DNA technology, the large-scale growth of these cells during which the cell’s internal protein synthesizing nano-machine (the ribosome, a complex cellular organelle composed of many proteins and several pieces of RNA) are directed to synthesize the target biopharmaceutical, illustrated here as either interferon beta-1a (IFNb) or a monoclonal antibody (mAb). Note that the space-filling molecular models of aspirin, IFNb and mAb have all been displayed roughly on the same arbitrary scale to help provide the reader with an approximate perspective on how they would relatively compare to each other on the basis of size. The dashed circle highlighting part of the structure of IFNb corresponds to the carbohydrate-containing portion of this biopharmaceutical that plays a dominant role in giving rise to its microheterogeneity shown in chapter 2 in Figure 2.3 when coupled with other posttranslational modifications (PTMs). Reproduced with permission from Berkowitz SA [116].

successfully are significant. Forcing a cell to produce unusually large amounts of a single protein presents unique problems to the cell. Particularly in terms of making sure that all the protein molecules are properly folded and have consistent physical, chemical, and biological properties. Hence, to achieve this goal requires the constant and diligent monitoring and characterization of the protein biopharmaceutical’s HOS. The process of finding, developing, and obtaining regulatory approval of a protein biopharmaceutical that is made using recombinant DNA technology proceeds through a sequence of key activities or basic phases of activity that is outlined in Fig. 1.6. Success of

I. Proteins and biophysical characterization in the biopharmaceutical industry

16

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

FIG. 1.6 A coarse overview of the basic areas and sequence of major activities involved in commercializing a protein biopharmaceutical. The relative length of each block arrow is roughly associated with the length of time typical spent at each stage, from research through commercialization. Overall, the cost in this process can easily be in excess of one or more billion dollars and can require more than an decade to develop [68]. These numbers can vary significantly from company to company and from drug to drug. As a result they are only approximate.

this process is highly dependent on biophysical characterization work associated with monitoring the consistency of the physicochemical properties of the protein drug, confirming the absence of changes in the drug’s HOS (which might give rise to small unwanted subpopulations of altered molecules), and in assessing the potential impact that PTMs might have on a drug’s structure. During the first part of this chapter we have dealt with the very basic properties of protein structure. In the remaining sections, we will discuss how these properties are responsible for many of the potential problems that are of great concern to a large range of biopharmaceutical scientists. In addition, we will briefly look at some of the more novel types of protein biopharmaceuticals that have and are being developed that further challenge the task of biophysically characterizing these complex drugs.

1.4.1 Impact of PTMs on the HOS of protein biopharmaceuticals The complex chemical composition of proteins, consisting of 20 chemically different naturally occurring building blocks (i.e., amino acids) effectively empower the cell with the needed components (chemistry set) to make the necessary array of proteins it needs to properly function. However, these amino acids also offer a range of chemically different targets that can undergo chemical changes, via direct chemical reactions or through the participation of various enzymatic reactions. The chemical changes that a protein biopharmaceutical can incur offer opportunities to alter the HOS of these molecules, impacting the consistency of manufacturing or worse, cause adverse events when administrated to a patient. As mentioned in section 1.1.4, these chemical changes, whether they occur in vivo or in vitro, are collectively referred to as PTMs. Many PTMs play roles in the biological function of a protein in vivo, while others are a result of normal degradation or aging. Hence, the idea that a given protein exists as a single defined unique chemical entity is misleading. In fact, nearly all protein biopharmaceuticals exist as a collection of highly similar variant forms. The range of these variants and their amounts in the final biopharmaceutical drug product is determined by the nature of the cell-line used, the cell culture conditions (e.g., raw materials and holdtimes within the bioreactor), the resolution properties of the purification process, as well as our ability to detect and characterize them. The collection of highly similar proteins, variant forms or proteoforms [69] of effectively the same protein that characterize a biopharmaceutical

I. Proteins and biophysical characterization in the biopharmaceutical industry

1.4 Proteins and the biopharmaceutical industry: problems and challenges

17

is referred to as microheterogeneity and is a unique property of these drugs. In making a protein biopharmaceutical, the attributes of microheterogeneity need to be carefully characterized, measured and controlled. In so doing, microheterogeneity becomes a fingerprint or a signature of the protein drug that is linked to its therapeutic behavior. However, it should be noted that manufacturers of biopharmaceuticals cannot “exactly” duplicate this fingerprint or signature microheterogeneity on a lot-to-lot basis. Nevertheless, within the concept of consistency of manufacturing, microheterogeneity needs to be contained to within an established and reasonable level of variation that is bounded by the biopharmaceutical’s specifications. The task of establishing this level of variation permitted in a biopharmaceutical’s microheterogeneity is a collaboration between the drug manufacturer and regulatory agencies (who will eventually review and approve the drug). Although, in the end it is the regulators who have the final say on what is acceptable, in terms of necessary specifications that defines this level of variation. These specifications are commonly associated with critical quality attributes (CQAs) and are attributes that are directly related to the structural characteristics that define the chemical, physical, and biological properties of the protein drug that impact the protein’s therapeutic activity. As mentioned, PTMs occur both inside (in vivo) and outside (in vitro) of the cell. Key factors that control in vivo PTMs include the following: cell line, culture media, and growth conditions. In the case of in vitro PTMs, once the protein biopharmaceutical is excreted into the culture media, the following are just a few key factors that can affect the protein drug product: temperature, pH, contact surfaces, light, metal contamination, released enzymes in the culture media or resulting from contamination, and sample handling (e.g., shaking, freezing, and thawing) [70e75]. From beginning to end, there is a host of environmental challenges that the protein biopharmaceutical must endure without altering its primary structure or HOS. In the case of the former, scientists in the biopharmaceutical industry have extensively used mass spectrometry, MS, as a key analytical tool to detect and characterize PTMs. This dependence on MS is primarily due to the change in mass that accompanies nearly all chemical reactions and the high mass accuracy and resolution of most commercially available MS instruments. However, it is worth noting that PTMs that involve isomerization reactions and or isobaric mass transitions can occur, yielding no obvious or little mass change and may require more sophisticated techniques to detect their presence [76e78]. In some cases, these isobaric mass modifications can be chromatographically separated or fragmented differentially via tandem MS [79]. In addition, although MS can detect, quantitate, and localize PTMs within a protein, the impact of a PTM on a protein’s HOS is largely unknown. Therefore, the application of biophysical analysis and characterization, as discussed in this book, in combination with bioassays plays an important role in attempting to assess the impact of PTMs in the biopharmaceutical industry.

1.4.2 Impact of changes in noncovalent interactions (secondary bonds) on the HOS of protein biopharmaceuticals Due to the high level with which a protein depends on an ensemble of weak secondary bonds or interactions to maintain its HOS and dynamic nature of its conformation, a protein can find itself potentially trapped in an altered conformation (metastable or intermediate state) without any change in its primary structure. A protein is particularly vulnerable to

I. Proteins and biophysical characterization in the biopharmaceutical industry

18

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

these HOS changes when placed under stress conditions. Under such stress conditions the normally native protein can adopt a nonnative but energetically stable conformation (albeit not as stable as its native conformation). When this partially unfolded protein state is accompanied with a relatively large kinetic energy activation barrier, when the stress is removed the partially unfolded protein can find itself trapped; see Fig. 1.4B. In this situation, the protein would encounter significant difficulty in returning to its more stable native state. As a result, proteins can undergo an alteration in HOS “without” the need of requiring 1
structure change. The ability to detect these types of HOS changes by MS would be very difficult since the change in conformation would occur without any change in mass! Such changes in the HOS of a protein can in terms of mass spectroscopy be considered silent HOS changes. Thus, for the biopharmaceutical scientists to detect and quantify such changes a battery of biophysical tools are often required.

1.4.3 A more detail discussion concerning protein biopharmaceutical aggregations and its influence on HOS Earlier in this chapter (Sections 1.1.2 and 1.1.3), it was mentioned that one of the most concerning properties linked to biopharmaceutical proteins is their ability to self-associate and form aggregates. While proteins can aggregate through many different mechanisms [80], in general, protein aggregation can be crudely classified to arise from two basic properties of a protein’s stability, its colloidal, and conformational stability. Aggregation resulting from the attractive complex nature of the “normal” stable surface properties of a protein is a characteristic associated with the protein’s colloidal properties and is related to its “colloidal stability”. Aggregates formed via these properties are general referred to as “colloidal aggregates”. However, due to the dynamic properties of proteins, these molecules can undergo a range of fluctuations in its HOS, especially under stress conditions. Thus, changes in a protein’s conformational properties that expose buried (hydrophobic) chemical groups prone to self-associate with other similar or different chemical groups on another protein molecule are related to a protein’s “conformational stability”. Aggregates formed via these properties are generally referred to as “conformational aggregates”. In some cases, aggregates that are formed may not neatly arise from one form of stability or the other. Rather they might arise through a hybrid combination of both [81,82]. In considering the unique self-associating properties of protein drugs, additional concerns surface due to the presence of “macromolecular crowding” [83]. These concerns play a in two areas which include: 1) the impact of the in vivo conditions of macromolecular crowding on the self-association process, and 2) the impact of in vitro conditions required in developing high protein biopharmaceutical concentration formulations [84,85]. In the former case, the direct administration of a protein drug into the blood stream of a patient is thought to lead to its rapid dilution, alleviating the adverse effect of a drug’s own high concentration. However, the effect of macromolecular crowding resulting from the presence of other proteins in the blood and lymph system along with the changes in the environmental solution conditions, relative to the vialed-protein drug product can enhance self-association. As a result, better methods for assessing these situations are needed [86,87], see chapter 15. In the case of developing high protein drug formulations for subcutaneous (SC) injection (which enables patient

I. Proteins and biophysical characterization in the biopharmaceutical industry

1.4 Proteins and the biopharmaceutical industry: problems and challenges

19

convenience, avoiding the costly and inconvenient process of administrating large amounts of protein drugs intravenously, IV), an additional question arises concerning the effects of macromolecular crowding on protein drug self-association. This is because upon SC injection a high protein concentration is deliberately created within the injected area that can remain at a high concentration for a relatively long time due to the slow passive ability of the drug protein to find its way into the patient’s blood or lymph system in comparison to intravenous injection (IV). In addition, the formulated SC delivered protein drug required for these injections must also be stable and unaffected at these very high protein concentrations within its container closure, e.g., vial or prefilled syringe, for several years [88], see chapter 15. In general, the characterization and assessment of protein self-association is by no means an easy task, especially at high protein concentrations. The bulk of the biophysical tools available to detect and characterize protein self-association have been developed for use on dilute protein solutions, often referred to as ideal solution conditions. Here, the details of macromolecular physical chemistry are simplified and much better understood. Tackling the solution behavior of proteins at concentrations as high as several 100 mg/mL forces the biopharmaceutical scientist into experimental space where their ability to interpret acquired data is extensively lacking, due to the poorly understood complexity of this situation and the absence of a completely well-developed theory. As a result, conducting useful biophysical characterization work is a real challenge resulting in biopharmaceutical scientists resorting to either very empirical methodologies [89,90], or to the extrapolation of data from very low concentrations to high concentrations [91e93] to make primitive and risky assessments, again see Chapter 15 for further discussion on this topic.

1.4.4 The novelty of different classes of protein biopharmaceuticals that create unique questions and challenges in characterizing and monitoring their HOS Some protein biopharmaceuticals that have and are being developed contain unusual or “unnatural” constructions and/or properties. In these cases, unique questions, problems, and challenges arise concerning their physicochemical properties and HOS. Two of the main types of “unusual” or “unnatural” drug candidates involve: (1) the covalent coupling of a biopharmaceutical to another protein or chemical compound to create a fusion [78] or conjugated (e.g., pegylated proteins [94,95], antibody drug conjugate, ADC [96], and XTEN technology [97] protein biopharmaceutical); and (2) very large assembly of proteins such as virus or virus-like particles (VLPs) or nanoparticle that serve as drug delivery systems [98], see chapter 15. The following sections illustrate some of these novel biopharmaceuticals that give rise to unique questions that bring into play the importance of biophysical measurements concerning HOS. 1.4.4.1 Example 1: Fc fusion proteins The fusion of an Fc (fragment crystallizable) part of an antibody (typically an IgG1 antibody) with that of another pharmaceutically relevant protein through recombinant technology, results in an Fc fusion protein. The reason for undertaking this fusion process is that in an antibody, the Fc portion has been shown to be responsible for increasing the antibody’s circulation time [99e101]. Thus, by expressing a pharmaceutically relevant protein fused with

I. Proteins and biophysical characterization in the biopharmaceutical industry

20

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

an Fc fragment, it is hoped a similar effect of increased circulation time will be achieved. As an example, fusing an Fc to the blood-clotting protein’s Factor VIII (FVIII-Fc) and Factor IX (FIX-Fc) have been shown to reduce the clearance of these factors, while retaining the correct biological activity [102e105]. This should enable patients suffering from Hemophilia A/B to reduce the number of drug infusions. However, the fusion of two relatively large proteins (each > 50 kDa) beckons the questions, “does the fusion of these two proteins cause any significant changes to either protein that would lead to an alteration in their corresponding HOS?” and, “would that fusion impact the functionality of either part of the molecule or potentially lead to an adverse effect?”. While the answer to these questions will be proteindependent, for FVIII-Fc and FIX-Fc, the answer is no, as revealed by a battery of biophysical studies [106e108]. 1.4.4.2 Example 2: PEGylated proteins and antibody drug conjugates (ADCs) The chemical coupling of a polyethylene glycol polymer (PEG) to a protein biopharmaceutical yields a pegylated protein drug [94,95,109]. As observed with Fc fusion protein, pegylated proteins show a significant reduction in the clearance of the administrated modified protein drug from the body [109], significantly increasing their therapeutic value. However, just as in the case of Fc fusion proteins the conjugation of the PEG molecule to a protein beckons the same question concerning the impact of this modification on the HOS of the modified protein (as discussed in the previous section) and for the same reasons. Nevertheless, to date there are currently more than a dozen pegylated biopharmaceuticals approved and marketed [110] and more in development [111]. Similar situation exists for ADCs where a small generally toxic drug (called a payload) is covalently bound to a mAb that uniquely bind to a specific cell (typically a cancerous cell [96]). In this situation, the normally nonspecific toxic small molecule drug is turned into a highly specific targeting drug by taking advantage of the high specificity of the mAb. The mAb essentially guides (carries) the toxic pharmaceutical payload to its cellular targets (e.g., cancer cell) where it’s internalized by the target cell and activated by its cleavage from the ADC causing highly specific cell (e.g. cancer cell) killing. In constructing these ADCs careful biophysical characterization studies are required to assess the impact of the small pharmaceutical and its load (drug to antibody ratio, DAR) on the mAb [112] to insure its overall HOS [113] (binding specificity) is not compromised. 1.4.4.3 Example 3: viruses, VLPs The formation of very large (MDa) multisubunit protein complexes, such as a virus (e.g., used in gene therapy), VLP, nanoparticles (e.g., lipid nanoparticles, LNPs, liposomes, and exosomes) which are used as drug delivery systems present unique challenges when trying to biophysically characterize these complex biopharmaceuticals, which may be only partially composed of protein material (or even in some cases may not contain any proteins, e.g., nonprotein biopharmaceuticals composed of specific nucleic acid molecules incased in LNPs or Liposomes). A critical challenge in developing these very large biopharmaceuticals is in assessing their homogeneity and overall structure. Due to their large size, the common workhorse tool for acquiring this information, size-exclusion chromatograph (SEC, see Chapter 7), may have limited utility because of limitations in its separation range (requiring SEC columns with pore sizes that are often larger than what is commercially commonly found to be

I. Proteins and biophysical characterization in the biopharmaceutical industry

References

21

available, i.e., >1000 Å). In this case, alternative analytical tools such as asymmetric flow field flow fractionation (AF4, Chapters 10 and 15), analytical ultracentrifugation (AUC, Chapter 9), and nanoparticle tracking analysis (NTA, Chapter 10) and flow cytometry (Chapter 10) can be important biophysical tools capable of filling in this gap. Some of these tools, such as AUC can also provide additional characterization information (chemical composition), which is unique to these classes of very large complex biopharmaceuticals, e.g., the level of virus drug particles that are filled, partially filled empty in terms of nucleic acid material [114].

1.5 Conclusion In this chapter, the authors have provided, in broad strokes, brief discussions on the fundamental structural properties of proteins. Since a protein’s structure dictates its function, these properties empower proteins with important functional roles for maintaining the cascade of activities that characterize all living systems. However, these same structural properties also create a heavy but required characterization workload for the biophysical scientist working in the biopharmaceuticals industry. It is our hope that the reader is in a better position to understand the unique challenges the biopharmaceutical industry encounters in striving to bring these protein drugs to the market place. These challenges are significantly more daunting compared with those typically encountered in the pharmaceutical area, where the drug product corresponds to small organic molecules with much simpler, rigid and homogenous structure. The developments that have occurred since the discovery of the structure of DNA [115], a little over a half century ago that ushered in the molecular biology era, has culminated in the last four decades with the successful commercial development of today’s growing biopharmaceutical industry. Concurrent with this development has been the advancement of the bioanalytical sciences, which has led to the development of better instruments and methods. Not only are we able to better understand how these complex molecules work, but we are also capable of characterizing them to better assure their safety and consistency of manufacturing. In the area of biophysical characterization, significant innovative developments in newer biophysical tools and techniques continue to occur along with the improvements in the older and traditional biophysical tools and techniques. This situation makes our ability to characterize the HOS of biopharmaceuticals on a routine level truly impressive. However, today knowing what we do know versus knowing what we don’t know can be very sobering! It seems, to the authors, that the more we discover the less we realize we know. Although still lacking in our ability to characterize (biophysically) these fascinating protein drugs, the good news is we are moving forward, learning to do a better job in participating in the overall task of making more effective and safer drugs!

References [1] Ramachandran GN, Ramakrishnan C, Sasisekharan V. Stereochemistry of polypeptide chain configurations. J Mol Biol 1963;7:95e9. [2] Creighton TE. The folded conformations of globular proteins. In: Proteins: structures and molecular properties. New York (NY): W.H. Freeman and Co.; 1992.

I. Proteins and biophysical characterization in the biopharmaceutical industry

22

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

[3] Creighton TE. The biophysical chemistry of nucleic acids & proteins. Helvetian Press; 2010. [4] Kessel A, Ben-Tal N. Primary structure [Chapter 2]. In: Introduction to proteins: structure function and motion. Boca Raton (FL): CRC Press; 2010. [5] Kessel A, Ben-Tal N. Secondary structure. In: Introduction to proteins: structure function and motion. Boca Raton (FL): CRC Press; 2010. [6] Kyte J. Structure in protein chemistry. New York (NY): Garland Publishing, Inc.; 2006. [7] Kessel A, Ben-Tal N. Tertiary structure. In: Introduction to proteins: structure function and motion. Boca Raton (FL): CRC Press; 2010. [8] Creighton TE. The physical and chemical basis of molecular biology. Helvetian Press; 2010. [9] Creighton TE. The folded conformations of globular proteins. In: Proteins: structures and molecular properties. New York (NY): W.H. Freeman and Co.; 2010. [10] Houde DJ, Bou-Assaf GM, Berkowitz SA. Deciphering the biophysical effects of oxidizing sulfur-containing amino acids in interferon-beta-1a using MS and HDX-MS. J Am Soc Mass Spectrom 2017;28(5):840e9. [11] Bobst CE, Abzalimov RR, Houde D, Kloczewiak M, Mhatre R, Berkowitz SA, Kaltashov IA. Detection and characterization of altered conformations of protein pharmaceuticals using complementary mass spectrometry-based approaches. Anal Chem 2008;80(19):7473e81. [12] Henzler-Wildman K, Kern D. Dynamic personalities of proteins. Nature 2007;450:964e72. [13] Smock RG, Gierasch LM. Sending signals dynamically. Science 2009;324:198e203. [14] Teilum K, Olsen JG, Kragelund BB. Functional aspects of protein flexibility. Cell Mol Life Sci 2009;66:2231e47. [15] Tokuriki N, Tawfik DS. Protein dynamism and evolvability. Science 2009;324:203e7. [16] Kamerzell TJ, Middaugh CR. The complex inter-relationships between protein flexibility and stability. J Pharm Sci 2008;97:3494e517. [17] Kamerzell TJ, Ramsey JD, Middaugh CR. Immunoglobulin dynamics, conformational fluctuations, and nonlinear elasticity and their effects on stability. J Phys Chem 2008;112:3240e50. [18] Liu H, May K. Disulfide bond structures of IgG molecules: structural variations, chemical modifications and possible impacts to stability and biological function. mAbs 2012;4:17e23. [19] Chen PL, Wang M, Ou WC, Lii CK, Chen LS, Chang D. Disulfide bonds stabilize JC virus capsid-like structure by protecting calcium ions from chelation. FEBS Lett 2001;500:109e13. [20] Toichi K, Yamanaka K, Furukawa Y. Disulfide scrambling describes the oligomer formation of superoxide dismutase (SOD1) proteins in the familial form of amyotrophic lateral sclerosis. J Biol Chem 2013;288:4970e80. [21] Trivedi MV, Laurence JS, Siahaan TJ. The role of thiols and disulfides on protein stability. Curr Protein Pept Sci 2009;10:614e25. [22] Ho SC, Wang T, Song Z, Yang Y. IgG aggregation mechanism for CHO cell lines expressing excess heavy chains. Mol Biotechnol 2015;57(7):625e34. [23] Sakurai K, Nakahata R, Lee YH, Kardos J, Ikegami T, Goto Y. Effects of a reduced disulfide bond on aggregation properties of the human IgG1 CH3 domain. Biochim Biophys Acta 2015;854(10 Pt A). [24] Nag M, Bera K, Basak S. Intermolecular disulfide bond formation promotes immunoglobulin aggregation: investigation by fluorescence correlation spectroscopy. Proteins 2015;83(1):169e77. [25] Moritz B, Stracke JO. Assessment of disulfide and hinge modifications in monoclonal antibodies. Electrophoresis 2017;38(6):769e85. [26] Morgan GJ, Usher GA, Kelly JW. Incomplete refolding of antibody light chains to non-native, protease-sensitive conformations leads to aggregation: a mechanism of amyloidogenesis in patients. Biochemistry 2017;56(50):6597e614. [27] Serebryany E, Woodard JC, Adkar BV, Shabab M, King JA, Shakhnovich EI. An internal disulfide locks a misfolded aggregation-prone intermediate in cataract-linked mutants of human gD-crystallin. J Biol Chem 2016;291(36):19172e83. [28] Kessel A, Ben-Tal N. Protein structural dynamics. Introduction to proteins: structure function and motion. Boca Raton (FL): CRC Press; 2010. [29] Bee JS, Randolph TW, Carpenter JF, Bishop SM, Dimitrova MN. Effects of surfaces and leachables on the stability of biopharmaceuticals. J Pharm Sci 2011;100(10):4158e70. [30] Majumdar S, Ford BM, Mar KD, Sullivan VJ, Ulrich RG, D’Souza AJ. Evaluation of the effect of syringe surfaces on protein formulations. J Pharm Sci 2011;100:2563e73.

I. Proteins and biophysical characterization in the biopharmaceutical industry

References

23

[31] Sharma B. Immunogenicity of therapeutic proteins. Part 2: impact of container closures. Biotechnol Adv 2007;25:318e24. [32] Sharma B. Immunogenicity of therapeutic proteins. Part 3: impact of manufacturing changes. Biotechnol Adv 2007;25:325e31. [33] Sharma B. Immunogenicity of therapeutic proteins. Part 1: impact of product handling. Biotechnol Adv 2007;25:310e7. [34] Bucciantini M, Giannoni E, Chiti F, Baroni F, Formigli L, Zurdo J, et al. Inherent toxicity of aggregates implies a common mechanism for protein misfolding diseases. Nature 2002;416:507e11. [35] Filipe V, Hawe A, Schellekens H, Jiskoot W. Aggregation and immunogenicity of therapeutic proteins. In: (editors: Wang W, Roberts CJ.) Aggregation of therapeutic proteins. John Wiley and Sons: Hoboken (NY), pp. 400e33. [36] Rosenberg AS. Effects of protein aggregates: an immunologic perspective. AAPS J 2006;8:E501e7. [37] Walsh CT. Posttranslational modification of proteins: expanding nature’s inventory. Englewood , CO: Roberts and Co.; 2006. 490pp. [38] Varki A, Cummings RD, Esko JD, Freeze HH, Stanley P, Bertozzi CR, et al. Essentials of glycobiology. Cold Springs Harbor (NY): Cold Springs Harbor Laboratory Press; 2009. [39] Kessel A, Ben-Tal N. Protein structure. Introduction to proteins: structure function and motion, CRC Press, Boca Raton (FL). [40] Creighton TE. Biosynthesis of proteins. Proteins: structures and molecular properties, W.H. Freeman and Co., New York (NY). [41] Walsh CT, Garneau-Tsodikova S, Gatto Jr GJ. Protein posttranslational modifications: the chemistry of proteome diversifications. Angew Chem Int Ed 2005;44:7342e72. [42] Farley AR, Link AJ. Identification and quantification of protein posttranslational modifications. Methods Enzymol 2009;463:725e63. [43] Walsh G, Jefferis R. Post-translational modifications in the context of therapeutic proteins. Nat Biotechnol 2006;24:1241e52. [44] Anfinsen CB. Principles that govern the folding of protein chains. Science 1973;181:223e30. [45] Dill KA, MacCallum JL. The protein-folding problem, 50 years on. Science 2012;338:1042e6. [46] Levinthal C. How to fold graciously. Mossbauer Spectrosc Biol Syst 1969;67:22e6 [Proceedings]. [47] Bryngelson JD, Onuchic JN, Socci ND, Wolynes PG. Funnels, pathways, and the energy landscape of protein folding: a synthesis. Proteins 1995;21:167e95. [48] Dill KA, Chan HS. From Levinthal to pathways to funnels. Nat Struct Biol 1997;4:10e9. [49] Dill KA, Ozkan SB, Shell MS, Weikl TR. The protein folding problem. Annu Rev Biophys 2008;37:289e316. [50] Onuchic JN, Wolynes PG. Theory of protein folding. Curr Opin Struct Biol 2004;14:70e5. [51] Onuchic JN, Wolynes PG, Luthey-Schulten Z, Socci ND. Toward an outline of the topography of a realistic protein-folding funnel. Proc Natl Acad Sci USA 1995;92:3626e30. [52] Panchenko AR, Luthey-Schulten Z, Wolynes PG. Foldons, protein structural modules, and exons. Proc Natl Acad Sci USA 1996;93:2008e13. [53] Englander SW, Mayne L, Krishna MM. Protein folding and misfolding: mechanism and principles. Q Rev Biophys 2007;40:287e326. [54] Fedorov AN, Baldwin TO. Protein folding and assembly in a cell-free expression system. Methods Enzymol 1998;290:1e17. [55] Netzer WJ, Hartl FU. Recombination of protein domains facilitated by co-translational folding in eukaryotes. Nature 1997;388:343e9. [56] Cabrita LD, Dobson CM, Christodoulou J. Protein folding on the ribosome. Curr Opin Struct Biol 2010;20:33e45. [57] Fedorov AN, Baldwin TO. Cotranslational protein folding. J Biol Chem 1997;272:32715e8. [58] Beissinger M, Buchner J. How chaperones fold proteins. Biol Chem 1998;379:245e59. [59] Gottesman S, Wickner S, Maurizi MR. Protein quality control: triage by chaperones and proteases. Genes Dev 1997;11:815e23. [60] Lee S, Tsai FT. Molecular chaperones in protein quality control. J Biochem Mol Biol 2005;38:259e65. [61] Saibil H. Chaperone machines for protein folding, unfolding and disaggregation. Nat Rev 2013;13:630e42.

I. Proteins and biophysical characterization in the biopharmaceutical industry

24

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

[62] Quan S, Bardwell JC. Chaperone discovery. Bioessays 2012;34:973e81. [63] Sorokin AV, Kim ER, Ovchinnikov LP. Proteasome system of protein degradation and processing. Biochemistry (Mosc) 2009;74:1411e42. [64] Dunker AK, Lawson JD, Brown CJ, Williams RM, Romero P, Oh JS, et al. Intrinsically disordered protein. J Mol Graph Model 2001;19:26e59. [65] Dunker AK, Babu MM, Barbar E, Blackledge M, Bondos SE, Dosztanyi Z, et al. What’s in a name? Why these proteins are intrinsically disordered. Intrinsically Disord Proteins 2013;1:e24157. [66] Huang Y-M, Hu W, Rustandi E, Chang K, Yusuf-Makagiansar H, Ryll T. Maximizing productivity of CHO cellbased fed-batch culture using chemically defined media conditions and typical manufacturing equipment. Biotechnol Prog 2010;26:1400e10. [67] Britt KA, Schwartz DK, Wurth C, Mahler HC, Carpenter JF, Randolph TW. Excipient effects on humanized monoclonal antibody interactions with silicone oil emulsions. J Pharm Sci 2012;101:4419e32. [68] DiMasi JA, Grabowski HG, Hansen RW. Innovation in the pharmaceutical industry new estimates of R&D. J Health Econ 2016;47:20e32. [69] Smith LM, Kelleher NL. Proteoform: a single term describing protein complexity. Nat Methods 2013;10(3):186e7. [70] Ha E, Wang W, Wang YJ. Peroxide formation in polysorbate 80 and protein stability. J Pharm Sci 2002;91:2252e64. [71] Jiang Y, Nashed-Samuel Y, Li C, Liu W, Pollastrini J, Mallard D, et al. Tungsten-induced protein aggregation: solution behavior. J Pharm Sci 2009;98:4695e710. [72] Kerwin BA, Remmele Jr RL. Protect from light: photodegradation and protein biologics. J Pharm Sci 2007;96:1468e79. [73] Patel J, Kothari R, Tunga R, Ritter NM, Tunga BS. Stability considerations for biopharmaceuticals, Part I. Bioprocess Int 2011:20e31. [74] Wang W. Instability, stabilization, and formulation of liquid protein pharmaceuticals. Int J Pharm 1999;185:129e88. [75] Bossio RE, Marshall AG. Baseline resolution of isobaric phosphorylated and sulfated peptides and nucleotides by electrospray ionization FTICR ms: another step toward mass spectrometry-based proteomics. Anal Chem 2002;74:1674e9. [76] Geiger T, Clarke S. Deamidation, isomerization, and racemization at asparaginyl and aspartyl residues in peptides. Succinimide-linked reactions that contribute to protein degradation. J Biol Chem 1987;262:785e94. [77] Hambly DM, Banks DD, Scavezze JL, Siska CC, Gadgil HS. Detection and quantitation of IgG 1 hinge aspartate isomerization: a rapid degradation in stressed stability studies. Anal Chem 2009;81:7454e9. [78] Czajkowsky DM, Hu J, Shao Z, Pleass RJ. Fc-fusion proteins: new developments and future perspectives. EMBO Mol Med 2012;4(10):1015e28. [79] Beck A, Reichert JM. Therapeutic Fc-fusion proteins and peptides as successful alternatives to antibodies. mAbs 2011;3:415e6. [80] Philo JS, Arakawa T. Mechanisms of protein aggregation. Curr Pharmaceut Biotechnol 2009;10:348e51. [81] Chi EY, Krishnan S, Kendrick BS, Chang BS, Carpenter JF, Randolph TW. Roles of conformational stability and colloidal stability in the aggregation of recombinant human granulocyte colony-stimulating factor. Protein Sci 2003;12:903e13. [82] Mahler HC, Fischer S, Randolph TW, Carpenter JF. Protein aggregation and particle formation effects of formulation, interfaces, and drug product manufacturing operations. In: Wang W, Roberts CJ, editors. Aggregation of therapeutic proteins. John Wiley & Sons, Inc.; 2010. p. 301e5. [83] Ellis RJ, Minton AP. Protein aggregation in crowed environments. Biol Chem 2006;387:485e97. [84] Minton AP. Influence of macromolecular crowding upon the stability and state of association of proteins: predictions and observations. J Pharm Sci 2005;94:1668e75. [85] Shire SJ, Shahrokh Z, Liu J. Challenges in the development of high protein concentration formulations. J Pharm Sci 2004;93:1390e402. [86] Demeule B, Shire SJ, Liu J. A therapeutic antibody and its antigen form different complexes in serum than in phosphate-buffered saline: a study by analytical ultracentrifugation. Anal Biochem 2009;388:279e87. [87] Wright RT, Hayes DB, Stafford WF, Sherwood PJ, Correia JJ. Characterization of therapeutic antibodies in the presence of human serum proteins by AU-FDS analytical ultracentrifugation. Anal Biochem 2018;550:72e83.

I. Proteins and biophysical characterization in the biopharmaceutical industry

References

25

[88] Mathaes R, Koulov A, Joerg S, Mahler HC. Subcutaneous injection volume of biopharmaceuticalsdpushing the boundaries. J Pharm Sci 2016;105(8):2255e9. [89] Chaudhri A, Zarraga IE, Yadav S, Patapoff TW, Shire SJ, Voth GA. The role of amino acid sequence in the selfassociation of therapeutic monoclonal antibodies: insights from coarse-grained modeling. J Phys Chem 2013;117:1269e79. [90] Kayser V, Chennamsetty N, Voynov V, Helk B, Forrer K, Trout BL. Evaluation of a non-Arrhenius model for therapeutic monoclonal antibody aggregation. J Pharm Sci 2011;100:2526e42. [91] Li S, Xing D, Li J. Dynamic light scattering application to study protein interactions in electrolyte solutions. J Biol Phys 2004;30:313e24. [92] Scherer TM, Liu J, Shire SJ, Minton AP. Intermolecular interactions of IgG1 monoclonal antibodies at high concentrations characterized by light scattering. J Phys Chem 2010;114:12948e57. [93] Yadav S, Scherer TM, Shire SJ, Kalonia DS. Use of dynamic light scattering to determine second virial coefficient in a semidilute concentration regime. Anal Biochem 2011;411:292e6. [94] Bailon P, Won CY. PEG-modified biopharmaceuticals. Expert Opin Drug Deliv 2009;6:1e16. [95] Jevsevar S, Kunstelj M, Porekar VG. PEGylation of therapeutic proteins. Biotechnol J 2010;5:113e28. [96] Lambert JM, Berkenblit A. Antibody-drug conjugates for cancer treatment. Annu Rev Med 2018;69:191e207. [97] Podust VN, Balan S, Sim BC, Coyle MP, Ernst U, Peters RT, Schellenberger V. Extension of in vivo half-life of biologically active molecules by XTEN protein polymers. J Control Release 2016;240:52e66. [98] Ma Y, Nolte RJM, Cornelissen JJLM. Virus-based nanocarriers for drug delivery. Adv Drug Deliv Rev 2012;64:811e25. [99] Unverdorben F, Richter F, Hutt M, Seifert O, Malinge P, Fischer N, Kontermann RE. Pharmacokinetic properties of IgG and various Fc fusion proteins in mice. mAbs 2016;8(1):120e8. [100] Roopenian DC, Akilesh S. FcRn: the neonatal Fc receptor comes of age. Nat Rev Immunol 2007;7:715e25. [101] Roopenian DC, Christianson GJ, Sproule TJ, Brown AC, Akilesh S, Jung N, et al. The MHC class I-like IgG receptor controls perinatal IgG transport, IgG homeostasis, and fate of IgG-Fc-coupled drugs. J Immunol 2003;170:3528e33. [102] Peters RT, Low SC, Kamphaus GD, Dumont JA, Amari JV, Lu Q, et al. Prolonged activity of factor IX as a monomeric Fc fusion protein. Blood 2010;115:2057e64. [103] Shapiro AD, Ragni MV, Valentino LA, Key NS, Josephson NC, Powell JS, et al. Recombinant factor IX-Fc fusion protein (rFIXFc) demonstrates safety and prolonged activity in a phase 1/2a study in hemophilia B patients. Blood 2012;119:666e72. [104] Dumont JA, Liu T, Low SC, Zhang X, Kamphaus G, Sakorafas P, et al. Prolonged activity of a recombinant factor VIII-Fc fusion protein in hemophilia A mice and dogs. Blood 2012;119:3024e30. [105] Powell JS, Josephson NC, Quon D, Ragni MV, Cheng G, Li E, et al. Safety and prolonged activity of recombinant factor VIII Fc fusion protein in hemophilia A patients. Blood 2012;119:3031e7. [106] Houde D, Berkowitz SA. Conformational comparability of factor IX-Fc fusion protein, factor IX, and purified Fc fragment in the absence and presence of calcium. J Pharm Sci 2012;101:1688e700. [107] Li H, Bai S, Wei JY, Berkowitz SA, Brader ML. Calcium binding to a factor ix Fc fusion protein and effects on higher-order structure. J Pharm Sci 2011;100:4597e606. [108] Kulman J, et al. Structural comparability between recombinant FVIII-Fc and its isolated FVIII and Fc constituents, Abstract 1135 presented at the 54th Annual meeting and exposition in Altlanta Ga 8e11 December 2012. [109] Harris JM, Martin NE, Modi M. Pegylation: a novel process for modifying pharmacokinetics. Clin Pharmacokinet 2001;40:539e51. [110] Turecek PL, Bossard MJ, Schoetens F, Ivens IA. PEGylation of biopharmaceuticals: a review of chemistry and nonclinical safety information of approved drugs. J Pharm Sci 2016;105(2):460e75. [111] Ivens IA, Achanzar W, Baumann A, Brändli-Baiocco A, Cavagnaro J, et al. PEGylated biopharmaceuticals: current experience and considerations for nonclinical development. Toxicol Pathol 2015;43(7):959e83. [112] Adem YT, Schwarz KA, Duenas E, Patapoff TW, Galush WJ, Esue O. Auristatin antibody drug conjugate physical instability and the role of drug payload. Bioconjug Chem 2014;25(4):656e64. [113] Pan LY, Salas-Solano O, Valliere-Douglass JF. Conformation and dynamics of interchain cysteine-linked antibody-drug conjugates as revealed by hydrogen/deuterium exchange mass spectrometry. Anal Chem 2014;86(5):2657e64.

I. Proteins and biophysical characterization in the biopharmaceutical industry

26

1. The complexity of protein structure and the challenges it poses in developing biopharmaceuticals

[114] Berkowitz SA, Philo JS. Monitoring the homogeneity of adenovirus preparations (a gene therapy delivery system) using analytical ultracentrifugation. Anal Biochem 2007;362:16e37. [115] Watson JD, Crick FH. Molecular structure of nucleic acids; a structure for deoxyribose nucleic acid. Nature 1953;171:737e8. [116] Berkowitz SA. Analytical characterization: structural assessment of biosimilarity. In: Endrenyi L, Declerck P, Chow S-C, editors. Biosimilar drug product development, vol. 216. Bca Raton, FL: CRC Press; 2015. p. 15e82.

Further reading

[1] [2] [3] [4] [5] [6] [7]

For a comprehensive review on many of the topics covered in this chapter, the authors highly recommend the following useful books: Creighton TE. Proteins: Structures and Molecular Properties. 2nd ed. W.H. Freeman and Co., NY. Creighton TE. The Biophysical Chemistry of Nucleic Acids & Proteins, Helvetian Press. Cantor CR, Schimmel PR. Biophysical Chemistry Part I: The Conformation of Biological Macromolecules, W.H. Freeman & Company, New York. Dickerson RE, Geis I. The Structure and Action of Proteins, Harper & Row, New York. Kaltashov IA, Eyles SJ. Mass Spectrometry in Biophysics, Wiley-Interscience, Hoboken (NJ). Kessel A, Ben-Tal N. Introduction to Proteins: Structure, Function and Motion, CRC Press, Boca Raton (FL). Kyte J. Structure in Protein Chemistry, Garland Publishing, Inc., NY.

I. Proteins and biophysical characterization in the biopharmaceutical industry

C H A P T E R

2

Biophysical characterization and its role in the biopharmaceutical industry Damian J. Houdea, Steven A. Berkowitzb a

Biomolecular Discovery, Relay Therapeutics, Cambridge, MA, United States; bConsultant, Sudbury, MA, United States

2.1 Drug development process The process of developing a drug, whether it is a pharmaceutical (small molecule drug) or a biopharmaceutical (large biopolymer, usually a protein), can be stripped down to its most basic level of finding a molecular substance that will affect a relevant biological target, responsible for a disease state, to provide a favorable therapeutic benefit (e.g., slowing the progression of a disease or eliminating the disease entirely). This concept is illustrated by the dotted area enclosed in Fig. 2.1. As can be seen, this process simply proceeds via a research to commercialization interaction. However, given the risks and dangers associated with bringing a drug into the market place, the pharmaceutical and biopharmaceutical industries are highly regulated and controlled by governmental regulatory agencies to minimize the risks and dangers to the patient. The word “minimize” has been carefully chosen here to remind the reader that this process is not perfect or foolproof. All the risks and dangers cannot be removed. This situation exists because of the following: 1. Living organisms are complex entities, which in the case of humans consist of a hierarchy of systems, organs, tissues, cells and molecules that interact with each other, that we unfortunately have a limited understanding aboutdespecially at the molecular level. A quote from Dr. Bruce Alberts, former editor-in-chief of Science captured below from his editorial entitled “A Grand Challenge in Biology” written in the September 2, 2011,

Biophysical Characterization of Proteins in Developing Biopharmaceuticals, Second Edition https://doi.org/10.1016/B978-0-444-64173-1.00002-0

27

Copyright © 2020 Elsevier B.V. All rights reserved.

28

2. Biophysical characterization and its role in the biopharmaceutical industry

FIG. 2.1 A simplistic view of the overall biopharmaceutical development process.

issue of Science [1] best summarizes what we know we face in developing a biopharmaceutical (as well as pharmaceutical): "There are about 21,000 distinct proteins encoded by the human genome. At present, one can only guess the function on nearly half of these gene products. And even when we know the exact function and structure of a particular protein, embedding this protein in the cell often reveals a network of interactions so complex that the biological outcome of any perturbation, such as a drug treatment, is unpredictable."

2. Due to the underlining diversity of humansdespecially in their genetic makeup, responses to drugs can be surprisingly heterogeneous and unanticipated. A drug that may work for one person may not necessarily work for another and/or may give rise to heterogeneous adverse side effects that can be experienced by some, but not by others. 3. The inability to specifically deliver a drug to the disease site(s) without exposing normal patient tissues (etc.) to the drug. In the case of a protein biopharmaceutical, additional unique risks and dangers lurk due to the following: 1. The compounding issues concerning our inability to completely understand the full physical and chemical complexity of protein drugs, due to limitation in our analytical tools and the limited ability to reproducibly manufacture the identical microheterogeneity that characterizes them. 2. The unique problems of aggregation [2,3] and immunogenicity [4,5]. As a result, drugs go through lengthy, intense, and detailed development (or process development) phase, which is inserted between and overlaps with the areas of research and commercialization. Within this area is housed those activities for producing, purifying, formulating, and analytically testing and characterizing protein drugs for the purpose of supporting critical human (clinical) testing, drug filing, and drug approval. However, before one can file a biopharmaceutical for approval (i.e., which occurs via a biologics license application or BLA) several rounds of clinical trials must be undertaken that proceed through a rigorous series of steps that in its most basic form consists of the following three phases of testing: (1) assessing safety (phase I), (2) finding appropriate dosing (phase II), and (3) demonstrating efficacy (phase III), see Fig. 2.2. I. Proteins and biophysical characterization in the biopharmaceutical industry

2.1 Drug development process

29

FIG. 2.2 A detailed view of the overall biopharmaceutical development process showing the key contributing areas of biophysical characterization. Note: BLA ¼ biological license application, HOS ¼ higher order structure.

In the initial entry phase into development, significant overlap and integration with research is required to allow for an effective transition of a drug candidate. However, at the later stages of development, as the drug successfully passes through the various phases of development and clinical testing, different interactions are required. Here, interactions with the commercialization groups of the company become necessary in order to successfully bring the drug to the patient in a safe, timely, and economic manner (see Fig. 2.2). During the entire drug development process, information concerning the physical, chemical, and biological properties of the drug candidate is continually acquired in support of clinical and commercial activities. In the case of biophysical information, three general categories are needed for specific purposes. These are illustrated in Fig. 2.2 and include: (1) information on the biophysical properties of the drug, (2) characterization information that provides the biophysical fingerprint of the drug for assessing comparability, stability, and compatibility, and (3) a deeper understanding of the impact of the differences between its normal and variant forms (usually detected via biochemical analysis, e.g., mass spectrophotometry, MS) that are present in the drug product, in terms of their biophysical structure and properties. Details concerning these biophysical categories will be provided in subsequent sections within this chapter.

I. Proteins and biophysical characterization in the biopharmaceutical industry

30

2. Biophysical characterization and its role in the biopharmaceutical industry

2.2 Protein drugs (biopharmaceuticals) As discussed in Chapter 1, proteins are complicated molecules. This complexity is immediately evident if one considers their high molecular weight (MW), the diversity of their chemical composition, how they occupy three-dimensional space in a very dynamic manner (in comparison to the more static nature of pharmaceuticals), and the way they are made using natural biological factories, cells (which are far more complex than any man-made factory). As a result, when a biopharmaceutical company makes a protein drug, the end result is never a true homogeneous product where every molecule is identical to each other. Rather, protein biopharmaceuticals are produced as a heterogeneous drug product mixture. This is illustrated in Fig. 2.3, where capillary (free) zone electrophoresis (CZE) shows the microheterogeneity of a purified recombinant commercial biopharmaceutical called Interferon beta-1a, IFNb1a [6]. In addition, in manufacturing a protein drug this exact mixture of protein molecules unfortunately cannot be made in such a way that the mixture is identical every time the drug is manufactured. It will vary to some extent on a lot-to-lot basis due to the inherent difficulties and complexity of making these drugs in cells and to some extent because of the slight variability

FIG. 2.3 Capillary (free) zone electrophoresis analysis on an intact commercial (23kDa glycoprotein) biopharmaceutical called Interferon beta-1a which clearly reveals its complex microheterogeneity. Adopted from Fig. 2.2 in Berkowitz SA, Zhong H, Berardino M, Sosic Z, Siemiatkoski J, Krull IS, et al. Rapid quantitative capillary zone electrophoresis method for monitoring the micro-heterogeneity of an intact recombinant glycoprotein. J Chromatogr 2005;1079(1e2):254e65, reprinted with permission. Copyright © 2005 Elsevier.

I. Proteins and biophysical characterization in the biopharmaceutical industry

2.3 The role of biophysical characterization in biopharmaceutical drug development

31

in the purification process. Nevertheless, to successfully develop and obtain protein drug approval a company must demonstrate that every time the drug is made it will be “comparable” and/or “highly similar” to the previously made drug material on a lot-to-lot basis. This comparability or high similarity must be maintained over the entire shelf life of the drug (from its initial release until the end of its permitted expiry period). This is critical because the drug needs to provide clinical (or therapeutic) potency that consistently falls within a limited range of activity over its shelf life. Unfortunately, the concepts of comparability and high similarity are relative terms that are open to interpretation. This interpretation, however, is contained by using a battery of analytical methods that are capable of monitoring an appropriate number of critical quality attributes (CQAs) concerning the physical, chemical, and biological characteristics of the drug, which continually increases in detail as the drug moves through the various stages of its development. These analytical methods fall into two unique categories: (1) characterization methods and (2) release testing methods. Although both types of methods are important, the latter defines the necessary specification window agreed upon between the manufacturer and regulators that must be met before releasing any drug product lot into any clinical trial or commercial use. Release methods also tend to have characteristics of being easier to performer, more robust in their performance and require less expertise to run. These attributes make them very suitable for use in a routine-testing environment in comparison to characterization methods, which typically involve more specialized, intricate and timeconsuming methodology frequently requiring higher specialized training to properly execute and interpret data.

2.3 The role of biophysical characterization in biopharmaceutical drug development 2.3.1 Biophysical properties: the developability issue at the researchedevelopment interphase In the research phase of finding a protein biopharmaceutical, the initial focus is on target discovery. Once a potential target is identified, work on finding a drug that will interact with this target, which will elicit a desired biological effect that could lead to a favorable clinical outcome of reducing or eliminating a disease state, can begin. After one or more drug candidates have been identified, efforts may continue to optimize the drug’s interaction with its target; however, at some point thoughts and action need to turn toward assessing the developability of these possible drug candidates. In general, the developability of a protein drug in terms of its biophysical properties is concerned with the drug’s ability to withstand a range of stresses it will likely encounter during cell culture production, purification, compounding, vialing, long-term storage (stability), and accelerated stability testing conditions (see Chapter 17 for a more in depth discussion on biopharmaceutical developability). Some of the important biophysical properties of a protein drug that needs to be monitored at this stage are associated with understanding the following: 1. How high a drug candidate can be concentrated (under a range of conditions) before problems are encountered, e.g., aggregation, high viscosity, etc.

I. Proteins and biophysical characterization in the biopharmaceutical industry

32

2. Biophysical characterization and its role in the biopharmaceutical industry

2. The drug’s ability to withstand a range of stresses its will or may be exposed to, e.g., extremes in pH, ionic strength, temperature, light, etc. 3. The drug’s ability to withstand repeated freezeethawing cycles and agitation (e.g., in the presence of aireliquid interfaces, packaging-shipping-transportation, etc.). 4. The level of compatibility a drug has with different material surfaces that it will come in contact with, especially in terms of the final container closure used to store it, as well as the material and environment a drug comes in contact with during its delivery to the patient. This added focus concerning the developability of a protein drug, in terms of its biophysical properties, at the researchedevelopment interface arises because once the lead drug is chosen and transitioned to the process development team (for scale-up to initiate clinical studies) any attempt to change the drug’s primary structure to improve it will be very costly in terms of time, money, and resources [7e11]. This situation arises because the development process will most likely need to be recycled back to the beginning due to such a change. As a result, the deeper into the development process the initial drug candidate moves before the need for a change is realized, the greater the cost and the more likely the project will be eliminated. It should also be noted that there are additional indirect but negative impacts that may arise in making such changes after a drug candidate crosses the researchedevelopment interface. These impacts are primarily due to the disruptive effects of the change on other ongoing projects within a company, which is caused by the reshuffling of timelines, resources, and scheduled clinical trials. Thus, making good choices at this early point in a drug’s development can pay significant dividends in reducing the cost burden these changes have on the drug’s development process. Such cost reductions are important not only to the drug innovator, but also to the end user of the drug, the patient, as well as third-party payers, e.g., insurance companies. This can be easily understood when one realizes that for a drug company to stay in business, a successfully approved drug must not only cover its own cost, it must cover the cost of all drug failures plus provide a profit to the innovator company pursuing the commercialization of the drug [12e15]. If these failures are costly, the patient as well as health insurance providers will need to bear some of the burden of cost, via high drug price, in order for the drug company to stay in business to make future needed (and hopefully more cost-effective) drugs for other diseases afflicting humans. Consequently, having a good understanding of the developability of a protein drug candidate plays one important part in reducing drug failure and thus in reducing high attrition rate of these drugs which plays a key role leading to their high drug development cost [15e18]. Indeed, the ability of a drug company to effectively understand the developability of its protein drug candidates can be an influencing factor in the willingness of a drug company to find and develop a number of new drugs at the same time. This can be especially apparent in its avoidance of areas that require higher level of innovation and risk taking, even if the potential financial rewards and more importantly health needs are great, a situation characterized as risk aversion. Hence, to facilitate this important decision process, a collaborative overlap between research and development is necessary for generating the appropriate information concerning the biophysical properties of the different protein drug candidates they are pursuing to help in the process of fulling understanding their developability attributes (note: the same should also apply to the chemical and biological properties of the drug).

I. Proteins and biophysical characterization in the biopharmaceutical industry

2.3 The role of biophysical characterization in biopharmaceutical drug development

33

2.3.1.1 Limiting factors that influence the level of effort put into assessing a protein drug’s developability In assessing the protein drug selection process, several factors need to be considered before a drug candidate is moved through the researchedevelopment interface. These issues include the following: 1. The high failure rate encountered in drug development is typically not only due to the developability of the drug. Factors such as efficacy, pharmacokinetics, and adverse effects in human and animal toxicity also account for these failures [15,19] and tend to be thought as the key factors for drug failure. As a result of this situation, a strong driving force will appear at biopharmaceutical companies that will limit the amount of up-front work a company is willing to put into obtaining information on its drug candidates’ biophysical properties (as well as other types of studies) to assess a drug’s developability. The obvious reason for this is that if the high probability of drug failure is believed to be due to factors other than its developability, which can only be assessed in the clinic, then why commit a potential large amount of time and effort toward understanding and improving its biophysical properties to begin with! 2. The amount of drug material available to assess its developability in research and in the researchedevelopment interface is minimal. This situation puts considerable limitations on the range of experimentation that can be carried out. (It should be noted here that this situation has been a strong driving force in the search and development of an array of bioanalytical tools that can perform useful measurements in this area using very small amounts of material). 3. Optimum formulation and drug stability is minimally understood at this point, which could mislead the decision process. 4. Dosage and route of delivery is not clear or is unknown at this point, which could also influence the decision process. 5. The difficulty of modeling the impact of scale (e.g., in freezeethaw studies where the rapid freezing and thawing of microliters cannot model the freezing and thawing of liters, or in concentrating studies where the concentrating of a few milliliters cannot model the concentrating of hundreds of liters) in order to assess useful biophysical information. 6. The number of viable drug candidates that are being considered. 7. Material contact compatibility differences in a small analytical scale versus a large-scale commercial setting. 8. Sample handling differences in small analytical scale versus large-scale commercial setting. 9. The inherent and variable level of commitment each company is willing put into this phase of the drug development process. 2.3.1.2 Tools and approaches in assessing developability of a protein drug at the researchedevelopment interface While there are difficulties a biopharmaceutical company may face in assessing a drug’s developability, an optimized approach that maximizes the amount of information a company can obtain in a limited amount of time is of great importance. The key to this approach is to

I. Proteins and biophysical characterization in the biopharmaceutical industry

34

2. Biophysical characterization and its role in the biopharmaceutical industry

use simple and robust biophysical tools to routinely assess knowledge that is typically concerned with major biophysical structural changes (e.g., loss of drug material from solution and formation of soluble and insoluble (precipitates) aggregates). It is also desirable that these tools require only small amounts of drug sample (for replicate measurements) and are amendable to high-throughput screening formats. As a result, simple biophysical methods such as ultraviolet (UV) visible absorbance, fluorescence, light scattering (LS), and sizeexclusion chromatography (SEC) are found to be those most commonly employed at this junction in the drug development process. These relatively simple biophysical methods can, in general, meet these demands. In addition, the simplicity of these methods also helps make data interpretation easy and straightforward. Nevertheless, the authors would like to specifically point out that judicious use of some advanced biophysical techniques, even at this early stage in drug development, could be particularly useful. These techniques include velocity sedimentation analytical ultracentrifugation (SV-AUC, see Chapter 9), scanning differential calorimetry (DSC, see Chapter 11), and intact or global hydrogenedeuterium exchange with mass spectrometer (HDX-MS, see Chapter 12) detection. The two former biophysical tools typically require about a 0.5 mL of sample at a concentration of about 0.1e0.5 mg/mL per run while the latter requires approximately an order of magnitude less material. In the case of AUC, use of this technique at various points, especially in the researche development phase or the very earliest stages of development, could be significantly helpful in assessing the critical element of aggregation and concentration-dependent aggregation (self-association). This occurs because AUC requires minimal method development and is associated with minimal issues that could question the validity of the information it provides in comparison to SEC [20]. Hence, until an SEC method is in place, AUC can step in and cover this area, using only a single measurement per sample (due to material limitations). In addition, simple profiling of protein drugs at the start using DSC to detect protein candidates with low temperature domain melts, Tm values, can be very useful in comparing drug candidates to help find and avoid those drugs that have a less thermally stable structure (or conformational stability) than others [21]. In the case of global HDX-MS, the technique’s high sensitivity enables it to be used on very small amounts of sample. In addition, the ability to automate the whole process makes this approach very attractive operationally [22,23]. Use of this specific technique can be envisioned to proceed through a two-step process. In step one, only one time point (in addition to t ¼ 0) would be acquired in order to make a preliminary evaluation of drug candidates. This preliminary screening could then be followed with a second step screening process where a more complete time course study on those limited number of drug candidates that passed the initial screening process. Although this experimental HDX-MS format does not have the same resolving power as local or peptide-level HDX-MS (see Chapter 12), this approach would offer the biopharmaceutical scientist the ability to make use of the dynamic properties of a protein drug in assessing its developability. In this situation, it is probable, especially on a comparison basis, that the proteins showing the highest level of HDX are most likely to have the greatest instability in conformation when stressed, due to their increased dynamic properties. Although this thought paradigm may not be perfect, it could be utilized in much the same way that Tm values are used to assess the stability profile of proteins. In this latter case, proteins with the lowest Tm values are thought of as the least desirable to employ as drug candidate due their earlier onset of instability in terms of temperature. In addition, given the simple single point comparative nature of these head-to-head evaluations, global HDX-MS measurements could I. Proteins and biophysical characterization in the biopharmaceutical industry

2.3 The role of biophysical characterization in biopharmaceutical drug development

35

also be used to conduct meaningful protein drug comparisons under a wide range of identical experimental conditions that could also include pH and temperature, which are normally more complicated to deal when conducting HDX, due to their intrinsic effects on HDX itself (see Chapter 12 for more details on this). The ability to assess those protein drug candidates that have the poorest biophysical properties early on in the drug development process is extremely beneficial. This benefit will be realized given the link that those drug candidates with the poorest biophysical properties will likely have the worst limited range of tolerability to stresses. As a result, this may make their purification, formulation, and drug deliverability much more difficult and present significant stability challenges, in terms of the drug manufacturability that can result in lengthy and costly drug development or, worse, drug product failure.

2.3.2 Important early biophysical activities required after a protein drug transitions from research to development Once a protein drug makes the transition from research to development, two important activities need to be initiated and completed very early on, both of which require the support of biophysical measurements. These two activities are the establishment of a standard approach for assessing the protein drug’s concentration and the development of an appropriate formulation. The former is not only necessary in establishing a protein drug’s specific activity and for providing a known and constant drug dosage, but also required for generating meaningful biophysical comparability studies and for supporting a range of calculations that generate other key pieces of biophysical and non-biophysical information about the protein drug. In the case of developing a stable formulation, once a formulation is put into place it becomes very costly (for the same reasons as explained for developability in Section 2.3.1) to change it during the development process. Hence finding the most effective formulation early in development, preferably before the start of phase I clinical studies, would be optimal. In terms of the concentration, the presence of UV-absorbing chromophores in a protein (e.g., aromatic group-containing amino acids, typically at 280 nm, or if necessary and if the formulation matrix allows, the peptide bond, typically at 215 nm) generally allows a simple UV spectrophotometric absorbance measurement to be used to assess the total amino acid concentration of protein drug. The use of “total amino acid concentration” and not “total protein concentration” is important. The former is commonly employed due to the invariant nature of the amino acid sequence (although amino acid substitutions can occur during the production of a protein biopharmaceutical, this situation is usually carefully assessed during the development to avoid or minimize) relative to the more variant nature of total protein concentration which would also include post-translational modifications (PTMs, see Chapter 1 Section 1.4.1), e.g., glycosylation. To utilize UV spectroscopy, however, an appropriate extinction coefficient is required. In the past, this parameter was determined via several experimental procedures that tended to be very laborious and error prone [24e26]. Today, nearly all protein extinction coefficients are assessed using empirical equations that utilize the average molar extinction coefficients (determined from a collection of published literature data) for the three key chromophores that contribute to the UV absorbance of a protein at 280 nm (tyrosine, tryptophan, and the

I. Proteins and biophysical characterization in the biopharmaceutical industry

36

2. Biophysical characterization and its role in the biopharmaceutical industry

disulfide bond), in conjunction with the known amino acid sequence of the protein drug of interest [24,27]. Using the latter information, which provides the accurate number of moles for each of the above-mentioned chromophores per mole of the protein drug (based on its amino acid sequence), an extinction coefficient can be computed using an equation having the following form: Protein  molar extinction at 280 nm ¼ εtyr ðntyr Þ þ εtry ðntry Þ þ εdisulf ðndisulf Þ

(2.1)

where. εtyr ¼ molar extinction coefficient at 280 nm for tyrosine εtry ¼ molar extinction coefficient at 280 nm for tryptophan εdisulf ¼ molar extinction coefficient at 280 nm for disulfide bond ntyr ¼ number of moles of tyrosine per mole of protein X ntry ¼ number of moles of tryptophan per mole of protein X ndisulf ¼ number of moles of disulfide bonds per mole of protein X The accuracy in the extinction coefficient generated from this empirical approach has been assessed at about 5% but could at times be as great as 10% particularly if the protein has an unusual amino acid composition [27]. Nevertheless, there are two completely independent experimental approaches that are orthogonal to each other, easy to perform and reasonably accurate for obtaining a protein’s extinction coefficient. These can be compared and averaged with the empirical approaches outlined above to provide a more unbiased assessment of the accuracy of this key parameter. These two other approaches involve the use of the analytical ultracentrifuge [28,29] (see Chapter 9) and SEC with LS, UV, and refractometric detection [30] (see Chapter 7). It should be particularly noted that these two experimental methods are very useful in demonstrating the high similarity of this key parameter between an innovator drug and a biosimilar. In terms of formulation development, the same biophysical approaches and tools discussed in Section 2.3.1.2 that were used in the developability assessment are also typically employed in carrying out this task. Although at this stage in the development process the restrictions associated with the limited supply of purified protein material is generally no longer an issue, in its place is the enormous number of samples that are generated in conducting a search for the optimum formulation. Hence, the process of finding an appropriate formulation is still heavily dependent on the use of high-throughput methodologies to meet the high sample demands generated from the ensuing and extensive searches for the best formulation [31,32]. Nevertheless, once formulation candidates are narrowed to only a few, the use of more advance biophysical tools can be by employed to help support the final decision process as to which formulation would likely be the best choice.

2.3.3 Developing the biophysical higher order structure (HOS) fingerprint of a protein drug Due to the complexity of protein drugs, these molecules require a great deal of characterization work as they move through the development process. Once the protein drug moves into development, experiments are initiated to build a biophysical fingerprint about the drug, as indicated in Fig. 2.2. This fingerprint will house an important part of the knowledge base I. Proteins and biophysical characterization in the biopharmaceutical industry

2.3 The role of biophysical characterization in biopharmaceutical drug development

37

about the drug. In the case of biophysical characterization activities associated with developing this fingerprint, this task does not require a three-dimensional structure of the protein drug. Although generating such information is useful and informative, it is by no means a necessity. What is important, however, is the development of a sufficient body of information that can effectively proxy as a detailed fingerprint for the protein drug’s HOS to support the many activities involved in the drug’s development process. Some of the key activities that a biophysical fingerprint will support include the following: 1. The demonstration by the manufacturer of its ability to make a protein drug reproducibly, within a meaningful established window of acceptance. This amounts to ensuring that each protein lot is “comparable” or “highly similar” to previous lots [33e37]. 2. In establishing a meaningful battery of release and characterization tests that will adequately define the specification window to the satisfaction of the regulators. 3. Enable the drug innovator to make changes in the manufacturing process to improve its manufacturability, stability, and safety without the need to conduct clinical trials to specifically support these changes. 4. Support efforts associated with finding a stable drug formulation by adequately demonstrating the drug’s stability under a range of conditions that include various levels of stress. 5. Demonstrate to the regulators that the innovator understands the drug’s biophysical properties well enough to establish an adequate drug stability profile that assures its safety and performance over the lifetime of the drug, as long as the drug is stored, handled, delivered, and administered to the patient within a specified and defined protocol. 6. In the specific case of developing a biosimilar the demonstration of biosimilarity. It should be pointed out that this fingerprint is not completely generated at the very start of the development phase nor does it remain static. Rather, it is a process that continually expands in terms of obtaining more detail about this fingerprint as the probability increases that the drug will be eventually submitted for regulatory approval. For biophysical characterization, this leads to an increased effort to acquire more detailed information about the protein drug’s HOS. This enhanced data package will play an important role in making comparability studies more useful and powerful as an important analytical tool for convincing the innovator’s management staff, regulatory agencies, physicians, and patients that the innovator understands its drug molecule and is capable of making it safe, consistent, and efficacious, especially as/if changes are made during the protein drug’s development. 2.3.3.1 Developing the HOS fingerprint of a protein drug using biophysical tools that provide indirect rather than direct structural information As we have already mentioned, the nature and amount of biophysical analysis required to develop the HOS fingerprint of a protein drug is a challenging and evolving task that requires the use of both relatively simple and sophisticated biophysical methods. In general, methods used to investigate a protein’s HOS can essentially be divided into two categories: (1) those that can directly determine the HOS of a protein and (2) those that can provide indirect information about the HOS of a protein via measured biophysical properties. The main methods currently available for directly determining a protein's structure include X-ray

I. Proteins and biophysical characterization in the biopharmaceutical industry

38

2. Biophysical characterization and its role in the biopharmaceutical industry

crystallography [38], nuclear magnetic resonance (NMR) spectroscopy [39], cryo-electron microscopy (cryo-EM) [40,41], and small-angle X-ray scattering (SAXS) [42e44]. The first three tools are quite sensitive and capable of solving (modeling) the complete molecular (three-dimensional) structure of a protein. On the other hand, the latter tool delivers lower structural resolution of a protein’s HOS by providing information about the spatial arrangement of larger structural elements, such as domains or individual polypeptide chains in the quaternary structure of a protein or the protein subunit structure in more complex structures, e.g., a binding complex formed between a protein drug and its protein-binding target or possibly protein delivery system like a virus particle or virus like particle (VLP). Unfortunately, at this time the use of these techniques to provide direct structural information about a protein drug is not routine and/or typically used in a process development setting. This in large part is due to the complexity of the measurement process, the time needed to perform the analysis, the low sample throughput, the high expertise required to make these measurements, the very high cost of the necessary instrumentation and the large amount of time it can take to analyze the data gathered from only one sample to determine its structure (which can take weeks, months, or even years). In addition, some proteins are simply not amenable to direct structural determination by some of these methods (e.g., in the case of proteins that cannot be crystalized or are intrinsically disordered, see Chapter 1 Section 1.3, use of X-ray crystallography is not feasible while in the case of NMR for large protein and/or for proteins that cannot be brought to high enough concentration, adequate NMR signal processing makes the use of this technique impractically). However, it should be noted, there have been some reports in crystallography screening and x-ray crystallography techniques that are claiming that advances in the field may potentially make the use of this technique more amenable to some forms of routine use [45]. Given the strong focus of this book on the process development area, we will not be concerned with those methods used for direct structural determination, or at least used in that mode where the purpose is to determine structure. Nevertheless, it should be noted that in the case of NMR spectroscopy (Chapter 13) and SAXS (Chapter 8) opportunities exist where these two biophysical techniques could be applied to generate a very useful HOS fingerprint just for practical comparison purposesdthis will be discussed in Chapter 3. Rather, our focus will be on those biophysical methods that provide indirect information about a protein’s structure. Such methods are generally more practical, robust, and applicable to the routine day-to-day activities that are involved in protein drug development and therefore find their greatest use at this stage of the overall process. However, what will be critical in using these tools is the realization that the information that they do provide can proxy for the proteins HOS that direct biophysical measurements normally provide. Such techniques cover a range of biophysical areas, e.g., spectroscopy, calorimetry, hydrodynamics, electrophoresis and chromatography. For the most part, these methods tend not to be as sophisticated, are easier to implement, have fewer limitations, higher throughput (most can be fully automated), and can therefore be applied more routinely in the development process where their capability can be put to practical use. Using an array of different indirect biophysical methods, an indirect fingerprint picture of a protein’s structure is obtainable. Such a fingerprint picture can be generated in more formal and definable ways using data from several low-resolution techniques, e.g., circular dichroism (CD), DSC, Fourier transform infrared spectroscopy (FTIR), by combining them with sophisticated data analysis techniques. Such techniques are

I. Proteins and biophysical characterization in the biopharmaceutical industry

2.3 The role of biophysical characterization in biopharmaceutical drug development

39

illustrated by the work of Middaugh, Volkin, and coworkers at the University of Kansas in creating an “empirical phase diagram” [35,46e49]. The focus of these approaches is to highlight deviations between samples analyzed under many different conditions. The utility of such approaches for a fast assessment of HOS analysis can be applied in many applications, such as in assessing the impact of sample handing [50e54], contact surfaces [54e56], mechanical stress [54], light exposure [54], process changes [34,54], as well as for assessing other factors. 2.3.3.2 The widespread importance of biophysical characterization in biopharmaceutical comparability studies (stability, compatibility, biosimilarity, and variant assessments) The two hallmark areas in the development process of a biopharmaceutical protein drug that highlight the traditional concept and use of comparability studies are: (1) establishing the consistency to which a drug is manufactured and (2) establishing the absence of a change in a drug product when changes to the manufacturing process are made. In the last decade, however, a third (3) area has emerged where the concept of comparability plays an even more critical and central role. This is the development of biogenerics or, as is more accurately described, biosimilars [9,34,36,57e60]. The criticality of this additional form of comparability arises because the whole premises for developing biosimilar is not whether it will work better, but rather, will it work as well as the commercial drug already on the market. Or in regulatory language, will it be “highly similar” enough (have adequate “biosimilarity”) to the commercial innovator drug to justify its approval. In the first two areas mentioned above, comparability studies are an essential component for maintaining an effective linkage between the data acquired during all phases of a biopharmaceutical’s development. In the case of the third area (biosimilars), these studies are critical for establishing that one company can effectively make the same drug as another by maintaining an adequate level of drug “high similarity”, “comparability” or “biosimilarity”. In terms of the biophysical characterization part of this assessment, this will come down to demonstrating the comparability between the innovator drug and the biosimilar drug in an array of biophysical parameters and profiles, which corresponds to the overall biophysical fingerprint of the protein drug. One should also realize that in principle, the concept of comparability effectively plays a role in virtually all testing conducted on a protein biopharmaceutical. For all practical purposes, the concept of comparability is interwoven throughout all the testing activities in developing a drug, from stability to drug variant analysis to what is specifically called compatibility studies (studies concerned with determining potential physical and chemical interactions between drugs and excipients or contact surfaces that can affect the chemical, physical, therapeutic properties and stability of the drug). In all these activities, the concept of comparability is carried out via the requirement of a comparison to an actual reference standard or defined reference state that is characterized by a set of specifications. In all these testing scenarios the key element is to establish the absence of any significant change/ difference between the test drug sample and a reference material/state. The unique association of comparability with biopharmaceuticals stems from the fact that protein drugs are large complicated molecules (see Chapter 1). A further complexity of biopharmaceuticals that links them to comparability studies is that the production of protein drugs relies on the use of cells. As a result, control over their production is not as

I. Proteins and biophysical characterization in the biopharmaceutical industry

40

2. Biophysical characterization and its role in the biopharmaceutical industry

precise as the synthesis of a small molecule drug. During their in vivo synthesis and in vitro processing, biopharmaceuticals may undergo many chemical modifications (PTMs) that may proceed to various levels of completion [61,62]. In some cases these PTMs may be specifically introduced in vitro as part of the actual drug development, e.g., in making antibody drug conjugates (ADCs) and pegylated protein drugs. All these routes for introducing changes, which increase the chemical microheterogeneity of the biopharmaceutical, can lead to changes in its HOS and therefore function/behavior [63e65]. In vivo, the fingerprint pattern of PTMs can easily be altered (either knowingly or unknowingly) during biopharmaceutical manufacturing by using different cells lines, expression systems, and/or growth conditions (including different raw materials) [9,33]. This is illustrated in the case of the latter in Fig. 2.4, which shows differences in the CZE profile of the same glycoprotein grown under two different growth media conditions. In addition, further PTMs can occur in vitro through changes in purification strategies as well as in filling, vialing, and storage steps [8,9,52,56,66,67]. Even variations in material contact surfaces (including container closure)

FIG. 2.4 CZE analysis on the same intact glycoprotein (Interferon beta-1a) characterized in Fig. 2.3 showing the effect of two different growth media conditions A and B on the resulting microheterogeneity of the glycoprotein. Note the capillary used in this figure was significantly shorter than that used in Fig. 2.3. Adopted from Figure 10 in Berkowitz SA, Zhong H, Berardino M, Sosic Z, Siemiatkoski J, Krull IS, et al. Rapid quantitative capillary zone electrophoresis method for monitoring the micro-heterogeneity of an intact recombinant glycoprotein. J Chromatogr 2005;1079(1e2):254e65, reprinted with permission. Copyright © 2005 Elsevier.

I. Proteins and biophysical characterization in the biopharmaceutical industry

2.3 The role of biophysical characterization in biopharmaceutical drug development

41

[68,69] and site(s) [70] of manufacture can lead to changes in a biopharmaceutical and its HOS or alteration in the distribution of variants forms present due to preferential interactions, which can ultimately change the final biopharmaceutical product [8,56,71,72]. 2.3.3.3 The role of biophysical comparability in the biosimilar world In the above discussion on comparability we have been predominantly concerned with its traditional meaning and usage, concerning consistency of drug product manufacturing and the absence of changes in a protein drug with the introduction of some change in its manufacturing process. For the latter, the drug product made after a change needs to be comparable or highly similar to the drug product that was made before a change was implemented. Such changes frequently occur during development; nevertheless, changes can also occur post-approval. However, in recent years a newer and more unique meaning and usage of comparability, which was briefly discussed in the previous section, has appeared that has become associated with an entirely new and rapidly developing field called biosimilars. As a result, comparability studies can be divided into two modes. The first mode is associated with the classical activity of establishing consistency of manufacturing and the absence of changes in the biopharmaceutical when changes are introduced into the manufacturing process by any company attempting to make a biopharmaceutical (including biosimilars in terms of making a change in the biosimilar it is producing). Comparability studies made in these situations correspond to the original or traditional concept of comparability, which can be referred to as “internal comparability” [57]. The term “internal” is used here because all samples being compared were made internally within the same company. The reason for introducing this nomenclature is to distinguish it from the second and more recent mode of comparability, which is uniquely associated with companies attempting to make biosimilar versions of a commercial biopharmaceutical coming off patent protection. In this latter situation, an “additional” mode of comparability between the biosimilar drug and the commercial biopharmaceutical (which is frequently referred to as the “reference product”) will be required. However, in this form of comparability the material (the reference product and the biosimilar) being compared are not made by the same company. Given this difference we will refer to this additional mode of comparibility as an “external comparability” [57]. The term “external” is used here because the samples of the key comparison material, the reference product, were made externally (by another company) relative to the company who made the biosimilar samples and who is performing the comparison testing. Nevertheless, the authors would like to point out that in the literature, the distinction between the concepts of “internal” and “external” comparability made here is frequently reflected in the use of the terms “comparability” and “biosimilarity” respectively. However, the introduction of the latter term would seem to imply a significant difference between these two terms. In actuality there is no difference, other than what is being compared and the level of comparability that may be required. This additional area of comparability was born when the first wave of biopharmaceuticals blockbuster drugs started to come off patent protection [58,73,74]. This situation has given the opportunity, for the first time, to other companies to undertake the challenge of making copies or generic-like version of commercial biopharmaceutical drugs. This commercial drug area is effectively analogous to the pharmaceutical generics area that has existed for almost four decades within the pharmaceutical industry (that is concerned with production of small drug molecules), which is clearly defined and successfully regulated. In the United I. Proteins and biophysical characterization in the biopharmaceutical industry

42

2. Biophysical characterization and its role in the biopharmaceutical industry

States, the HatcheWaxman Act is the fundamental piece of legislation that empowered the US Food and Drug Administration (FDA) to establish the pharmaceutical generic industry, which has made a significant impact in reducing the cost of pharmaceutical drugs [75]. The key to the success of these generic pharmaceuticals is the ability to show “identity” of the generic to the innovator drug [76]. Once this is established, the clinical history of the innovator drug’s efficacy can be used to support approval of the generic drug. The implementation of this concept into the area of protein biopharmaceuticals, however, is much more complex and challenging (as the reader should realize) due to the complex nature of protein drugs and the way these drugs are made in comparison to a small pharmaceutical drug; see Chapter 1. The difficulty surrounding the idea of biopharmaceutical generics revolves around the common knowledge that no biopharmaceutical company can manufacture a protein biopharmaceutical that is “identical” to an innovator’s biopharmaceutical. In fact, it is well known that the innovator of a protein biopharmaceutical cannot manufacture its own protein biopharmaceutical such that every lot is “identical” or an “exact duplicate” copy. Rather, protein biopharmaceuticals can only be made so they are “highly similar” or “comparable” to each other on a lot-to-lot basis, as discussed in Section 2.3.3.2. Thus, the concept of biopharmaceutical generics, as spelled out for pharmaceutical generics, is not possible. This reality has given rise to a lower level of biopharmaceutical generics called biosimilars. As a result, biosimilar manufacturers need to show that their biosimilar drug is “highly similar” to the innovator’s drug in much the same way as the innovator demonstrated that its innovative drug showed high similarity or comparability to itself as it passed through all the stages of development and throughout its commercial history. For the Innovator the high similarity or comparability is established based on acceptable specifications that effectively define the drug’s physical, chemical, and biological fingerprint. In this latter situation, the level of similarity is based on an array of bioanalytical measurements that delineate the key attributes that adequately define the drug that were established by the innovator over the long history of the drug’s development. The measured bioanalytical values for these attributes must fall within acceptable specifications (within a specific design space for a range of release testing and characterization data) that were agreed upon by the innovator and the regulators. Unfortunately information concerning these specifications is not public knowledge. In addition, regulators cannot provide this information to the biosimilar manufacturer. Consequentially, the only approach a potential biosimilar manufacturer has to obtain these acceptable specifications is to purchase the innovator drug product (reference product), from a number of different lots on the open market in an attempt to characterize these different innovator samples to reestablish the innovator drug’s specifications. In all cases, in terms of chemical, physical, and biological testing, the higher the level of similarity (comparability) that can be established by the biosimilar manufacturer the greater the chance of the biosimilar manufacturer achieving drug approval. In the case of biophysical comparability, the basic tools (i.e., UV, fluorescence, CD, SEC, DSC, AUC, chromatography and electrophoresis), which are routinely used in the normal development of new innovative drugs, need to be applied to this process along with other useful pieces of biophysical information that can provide more detail and higher resolution information concerning the HOS comparability. It is the uniqueness of only needing to prove comparability with rigor and confidence once that makes the effort of bring to bear the biophysical tools capable of higher resolution, such as HDX-MS, SAXS and NMR spectroscopy very attractive. I. Proteins and biophysical characterization in the biopharmaceutical industry

2.3 The role of biophysical characterization in biopharmaceutical drug development

43

A key challenge in this process occurs when a manufacturer of a biosimilar places its biosimilar in a formulation that is different from the innovator drug product. In this situation, the biosimilar manufacturer must get both forms of the drug into the same matrix to initiate the comparisons, which is especially important in conducting biophysical comparison studies [57]. Given the need to undertake additional processing to achieve this task, a situation can arise where bias artifacts can be introduced due to differential sample handle and processing [77]. Unfortunately, there is no way to avoid the handling and processing of these samples; thus, one is left with a quandary of how to proceed to make a meaningful comparison. There are several ways to envision such a comparison, however all will require some form of buffer exchange, preferably via passive dialysis instead of centrifugation spin cartridges (unless the drugs are concentrated enough so they can be appropriately diluted and the remaining formulation components have no significant impact on the biophysical measurements), as outlined below [57]: 1. Place the biosimilar in the innovator’s formulation using passive dialysis. If there is a concern that some excipients may not partition well during buffer exchange and the biosimilar samples are concentrated enough, it may be useful to first perform an initial dilution of the biosimilar into the innovators formulation before dialysis. However, the dilution should not result in the final sample concentration being too low at the end of the buffer exchange process. This should help achieve a better matching of the final formulations between samples that will be compared. 2. Place both the biosimilar and innovator drugs into a new formulation using the same approached outlined in example “1”. This approach may be required because of the presence of excipients in both formulations that interfere with the biophysical measurements. 3. Carry out example “1” or “2”, but do so with a sample of the innovator material that is processed in the same way as the biosimilar sample to get it into the innovators formulation in the case of “1” or into a new formulation in the case of “2”. If high comparability is established using any of these conditions then there should be no concern. However, if noncomparability is encountered, the situation is more complex and the certainty of the outcome can be questioned. Example “3”, however, would appear to hold the highest ranking for detecting and observing a real difference since both innovator and biosimilar are being processed as closely as possible in the same way.

2.3.4 Understanding the impact of variant forms of a protein drug on its HOS and biophysical properties The third area of biophysical analysis activity, which is generally the focus of mid-to laterstage development in the drug development process (see Fig. 2.2), is concerned with acquiring a better understanding about the major variant forms within the protein drug product. At this stage, the probability that the drug will obtain approval has increased significantly. In addition, it is also a time where regulatory agencies will begin asking more detailed questions concerning the major variant forms contained within the protein drug. They will want to know and understand what the knowledge base of the manufacturer is concerning these variant forms of the drug. Therefore, a significant effort by the innovator is warranted for getting as much detailed understanding about these variant forms by I. Proteins and biophysical characterization in the biopharmaceutical industry

44

2. Biophysical characterization and its role in the biopharmaceutical industry

conducting more intense characterizations. It should be noted that, although the ability of today’s biopharmaceutical scientist to detect and quantitate the presence of various chemically altered forms of a protein drug (which corresponds to the biochemical characterization component of variant analysis) via MS is very good [78,79], the biopharmaceutical scientist knows very little about what impact these primary structural changes have on the HOS of the drug [63e65]. In addition, the presence of variant forms that are a result of noncovalent changes in the HOS, which we have referred to in Chapter 1 as “silent changes” [57] in the protein’s HOS (due to the inability to detect these variant forms by MS), will very likely go undetected. As a result, a more detailed biophysical assessment of a drug’s HOS will focus on attempting to detect the presence of these silent variant forms of the drug by constantly monitoring the consistency of the expanding biophysical fingerprint using higher resolution techniques such as HDX-MS or NMR spectroscopy. Although work gathered via biological assays and binding studies is very helpful in supporting these expanded biophysical studies, especially during structureeactivity relationship studies or structureefunction assessments, they have the following limitations in the level of information they can provide: 1. Accuracy limitations, e.g., biological assays typically have error limits of 20%. 2. Although they can provide information about the important binding region of the drug, they provide no information about potential changes in the HOS of a drug that are not part of the binding domain, which can still pose problems associated with immunogenicity. A key part of these expanded biophysical studies, concerning the characterization of these variant forms of the drug, is the need to isolate each variant drug form (which are also referred to as drug-related impurity) as a concentrated solution to improve the ability of biophysical measurements to detect whatever HOS differences may exist between the normal drug and the specific variant form isolated. In conducting this work, it is of great importance to consider the potential impact of the sample handling and processing that is necessary to obtain these enriched samples [8,77,80]. Consequently, it is also important that the “normal” form (i.e., nonvariant form) of the protein drug that is removed during the isolation of these variant forms be used and tested along with the variant forms for use as control material since it will have gone through the same sample handling and processing steps as the variant forms. By employing this control material in this manner, it will minimize or detect the introduction of handling and processing artifacts enabling a much better head-to-head comparison to be made to detect and assess the possibility of any differences in HOS that may exist between the variant forms and the “normal” form of the drug [57]. At this point, all the core biophysical tools will be put into use to make a more detailed and critical assessment. However, it is in this area that the more advanced and higher resolving biophysical tools like HDX-MS and NMR spectroscopy can find their greatest usage.

2.4 The challenges in conducting biophysical measurements to detect changes in a protein drug’s HOS The complex and fragile nature of proteins can present unique challenges in undertaking the biophysical characterization measurements on biopharmaceuticals. They can arise from I. Proteins and biophysical characterization in the biopharmaceutical industry

2.4 The challenges in conducting biophysical measurements to detect changes in a protein drug’s HOS

45

areas that, at first glance, would not be considered as factors capable of impacting the biopharmaceutical. Such areas include the measurements itself, the biopharmaceutical formulation matrix and the container closure systems. The list below illustrates some of these challenges.

2.4.1 Molecular size and molecular heterogeneity (PTMs) challenge Biophysical analysis on a protein biopharmaceutical typically becomes more difficult with the increase in the MW of the biopharmaceutical. This difficulty arises because with the increase in MW comes an increase in the number of different structural elements that will appear in the protein, which can increase the heterogeneity of the measured biophysical signal. Basically, a biophysical method may struggle to resolve the large number of individual signals and/or average them together, in which case a small difference arising from a change in one structural element can be buried in what might be called the large background of signals arising from all the remain structural elements in a drug. This difficulty is further amplified by the increased opportunity for further structural heterogeneity, due to PTMs, which can expand the number of different signals a biopharmaceutical can generate. Ramifications from this increase in the number of unique structural configurations for a fixed amount of drug sample are many. • In the case where each additional structural element yields another unique signal, the larger the MW, the weaker any given signal will tend to be for the same fixed amount of material. This reduction in signal response is simply due to the molar reduction in concentration of each (unique) structural element (relative to the whole sample). Hence, the protein drug concentration will need to be increased appropriately in order to detect an altered signal assuming the signal to noise (S/N) ratio and background signal does not change. • If the increased number of unique signals is too great, extensive overlap between these signals can occur making the detection of a specific signal more difficult to detect against a background of normal signals. • If the increased number of structural elements does not yield a unique signal (or the unique signals cannot be resolved from one another, due to instrumentation resolution limitations), then when a change occurs, which only shows up in terms of a change in intensity, the change will be more difficult to detect unless the S/N ratio of the measurement can be increased. This situation occurs because the small change in signal must now be detected against a larger background of normal signals. Basically, the weight fraction of the changed signal will be reduced (see Figure 3.1 and its figure legend). • An additional modification of the scenario above could arise where changes in some of these structural elements yield an increase in the normal signal, while other similar structural elements can yield a reduction in the normal signal. In this situation, the sum of signals that are recorded may not show any change due to the canceling effects from both changes. For example, in fluorescence measurements, a change in the conformation of the protein drug can cause some tryptophan’s to become exposed to solvent causing a decrease in their fluorescence signal intensity, while other tryptophan’s can become buried on the same protein drug molecule, causing their fluorescence signal intensity to increase. These two effects can cause a canceling effect leading to a significant reduction

I. Proteins and biophysical characterization in the biopharmaceutical industry

46

2. Biophysical characterization and its role in the biopharmaceutical industry

in the net change (increase or decrease) in fluorescence signal intensity (note: this is dependent on the fact that there is no major shift in the wavelength dependence of the fluorescence spectrum between the structural elements responsible for the enhanced vs. reduced fluorescence) making the resulting change difficult or impossible to detect given the normal S/N in the data. The above mentioned difficulties associated with increasing MW point to a key difficulty with many biophysical measurements on protein drugs: the weak ability to detect small differences between samples due to their increased complexity with MW, which is made even more difficult when the small difference is present in a small population of the drug molecules. In the end much of the work associated with the biophysical characterization of variant forms of a protein drug, frequently involves a significant effort of isolating a relatively large amount of each variant with high purity to adequately characterize and assess the biophysical differences between each variant.

2.4.2 Protein concentration challenge If the sensitivity of a biophysical method to measure a signal from a protein is low, higher sample concentrations will be required. However, bringing a protein to higher concentrations may not be achievable because some proteins will experience solubility issues or display concentration-dependent aggregation behavior. In addition, if the initial sample is too dilute, the sample processing steps required to concentrate it will need to be done carefully, reproducibly, and in a manner that itself does not risk changing the biophysical properties of the protein. Although many biopharmaceutical companies are moving toward the production of biopharmaceutical products that are being made at very high concentrations (50e150 mg/mL or even greater), it would appear that the issues mentioned above are not a significant problem. However, these high-concentration biopharmaceuticals have their own unique problems. Key among them is the difficulty of trying to understand the long-term impact on stability. Unfortunately, many biophysical techniques encounter physical problems that inhibit the measurements at high concentrations, e.g., in velocity sedimentation analysis, as a sample’s concentration is increased, the resulting concentration gradient generated can distort the recorded data, making it unusable for data analysis [81]. In addition, when measurements can be made at these high protein concentrations, the underlying science available to interpret the data is often too complex and/or incompletely developed to extract meaningful data [81,82], see Chapter 15 for a further discussion on this topic.

2.4.3 Challenges arising from a biopharmaceutical’s formulation One of the most bothersome problems in attempting many biophysical measurements on a biopharmaceutical is the interfering effects from its formulation. For example, investigating the spectroscopic properties of a protein in the presence of buffers or excipients can be problematic (e.g., the presence of amino acids, such as arginine or histidine, when one is interested in obtaining far-UV CD spectrum, see Chapter 6). In addition, the quality of some excipients can vary making attempts to blank out their effect difficult. This is often encountered with

I. Proteins and biophysical characterization in the biopharmaceutical industry

2.4 The challenges in conducting biophysical measurements to detect changes in a protein drug’s HOS

47

common additives, such as the nonionic detergent polysorbate [83]. This detergent can contain high molecular weight (HMW) impurities that absorb light at 280 nm (or lower wavelengths), which can appear to be HMW biopharmaceutical aggregates in SEC. This is typically encountered when the biopharmaceutical concentration is very dilute (e.g., less than a mg/mL) and the detergent concentration is relatively high (e.g., approximately 0.1% or greater) requiring the use of lower UV wavelength for improved detection (e.g., 215 nm), where the strong absorption of the peptide bond can be used to increase the sensitivity of the SEC method. In this situation, many impurities in excipient will adsorb with a much greater ability at 215 nm than 280 nm in comparison to proteins creating a higher probability of encountering interfering effects at these lower wavelengths that could vary from excipient lot to lot.

2.4.4 Container closure challenges In finding and developing a biopharmaceutical and dealing with the many challenges these drugs present to the scientist who are developing them it is sometime hard to image the need to consider the significant care needed in choosing the correct container closure or in considering the need to pay close attention to the nature of the different surfaces these biological macromolecules will come in contact with. Nevertheless, this all proves to be the case. This attribute between the biopharmaceutical and its container closure and the other surface the biopharmaceutical comes in contact to is captured by the term compatibility and can also play a role when performing some biophysical measurements. Some examples of this concern are listed below: • The use of silicon oil in prefilled syringes can contaminate protein drug samples to varying levels. Such interference is a particular problem for particle monitoring techniques. Silicon oil can also interact with the protein drug affecting its HOS profile [71]. Other possibilities could also involve the selective removal (via surface binding) of stabilizing recipients, e.g., detergents that stabilize a biopharmaceutical’s HOS preventing it from aggregating. • Extractable and leachable components from the drug container and container closure components, e.g., caps on vials, tip caps, plunger stoppers, and fix needles in syringes, can interfere with the biophysical measurement by generating an interfering signal or by interacting with the protein and affecting its HOS [52]. • Surface interactions between the protein drug and an array of surfaces the protein comes in contact with during processing, storage, and even during sample preparation for biophysical measurement. In addition, the nature of the surfaces of the various containers used in all types of measurement process including biophysical measurement itself can alter the conformation of the protein drug, e.g., cause aggregates [55].

2.4.5 Challenges of sample measurement time versus sample stability In some cases, the time required to make a biophysical measurement could be fairly long, e.g., sedimentation velocity runs can take a number of hours, while sedimentation equilibrium runs could actually take 1e3 days depending on the number of different speeds used

I. Proteins and biophysical characterization in the biopharmaceutical industry

48

2. Biophysical characterization and its role in the biopharmaceutical industry

during the experiment [20,84]. Another factor concerning time and stability of the protein is the issues of bacteria and enzyme contamination when handling, processing, and analyzing these samples. Many biopharmaceutical solution compositions can act as good growth media for microorganisms and because biophysical measurements are rarely conducted under sterile conditions, the possibility of contamination coupled with long measurements times can potentially give rise to altered and degraded protein drug material and therefore erroneous biophysical results [85e87]. Light exposure is yet another issue or concern, which can cause protein degradation (e.g., oxidation) [62,88]. Light can also interact with excipients, extractables, and leachables, especially metals [88e91], resulting in additional adverse effects on the protein drug.

2.5 Regulatory needs and considerations While there is no definitive regulatory guidance on exactly what methods or specifically when biophysical analysis should be performed, assessing the information about the HOS of a protein drug has over the years become more practical and accessible to those in the biopharmaceutical industry and is therefore recognized as a critical component throughout the life cycle of the biopharmaceutical. Indeed, improvements in obtaining this type of information along with improvements in biochemical and biological analysis has played an important role in eliminating the need to repeat lengthy clinical trials when implementing a change or changes in the production of a protein drug by generating this information to adequately show the absence of significant differences between drug material made pre- and post-change(s). The arrival of this capability has ushered in the world of “Well Characterized Biologics” [92], replacing the old-world concept of the “Process is the Product” [73,93]. Such an achievement has resulted in a significant reduction to the barriers of developing and improving these life-changing and life-saving drugs. Since biopharmaceuticals are not like small molecules, in that they cannot be manufactured exactly the same each time and given the importance of the concept of “structureefunction”, when it comes to biopharmaceuticals, HOS characterization is recognized as an important activity in the development of these drugs. Furthermore, it is also acknowledged that in developing a biopharmaceutical the process continuously evolves, necessitating the need for ongoing changes [9,11,33,94,95]. Since such changes to the biopharmaceutical manufacturing process (e.g., to raw materials, and cell line to mention just a few areas where changes are common [9,11]) are inevitable, to safeguard against the possibility that a manufacturing process change actually does not alter the HOS of a protein drug, biophysical characterization is a critical necessity (especially with high-resolution methods at key points in a drug's development). To provide better regulatory over site in the production and control of biopharmaceuticals to improve safety, in 2004 the Food and Drug Administration (FDA) introduced a new riskbased quality assessment system termed quality by design (QbD) [96]. The implementation of QbD represented a systemic approach for developing a biopharmaceutical based on product knowledge, manufacturing risks, and how best to mitigate these risks [96e102]. The basis of QbD begins with predefined objectives that emphasizes the need for product and processes understanding and process control. In creating QbD, the designing and developing of

I. Proteins and biophysical characterization in the biopharmaceutical industry

References

49

formulations and manufacturing processes to ensure a predefined product quality are key goals. Understanding the biophysical properties and HOS of the protein drug is a very critical part in achieving this goal. The implementation of QbD aims to enhance and modernize the regulation of biopharmaceutical manufacturing and product quality to enhance the biopharmaceutical’s manufacturability by reducing risks via an understanding as to what constitutes an acceptable deviation(s) in a process while maintaining a biopharmaceutical’s efficacy and safety profile. The achievement of this goal should also play an important role in reducing drug costs. The key to the success of this initiative requires an increase in product knowledge, which is only obtainable through increased bioanalytical measurements that include enhances biophysical characterization. Thus, the implementation of biophysical measurements that can provide meaningful information that expands our knowledge base about these complex drugs (protein biopharmaceuticals) should be considered at each step of biopharmaceutical’s development.

References [1] Alberts B. A grand challenge in biology. Science 2011;333(6047):1200. [2] Mahler H-C, Kiskoot W, editors. Analysis of aggregates and particles in protein pharmaceuticals. Hoboken (NJ): John Wiley & Sons, Inc.; 2012. [3] Wang W, Roberts CJ, editors. Aggregation of therapeutic proteins. Hoboken (NJ): John Wiley & Sons, Inc.; 2010. [4] Buttel IC, Chamberlain P, Chowers Y, Ehmann F, Greinacher A, Jefferis R, et al. Taking immunogenicity assessment of therapeutic proteins to the next level. Biologicals 2011;39(2):100e9. [5] De Groot AS, Scott DW. Immunogenicity of protein therapeutics. Trends Immunol 2007;28(11):482e90. [6] Berkowitz SA, Zhong H, Berardino M, Sosic Z, Siemiatkoski J, Krull IS, et al. Rapid quantitative capillary zone electrophoresis method for monitoring the micro-heterogeneity of an intact recombinant glycoprotein. J Chromatogr 2005;1079(1e2):254e65. [7] Dingermann T. Recombinant therapeutic proteins: production platforms and challenges. Biotechnol J 2008;3(1):90e7. [8] Kueltzo LA, Wang W, Randolph TW, Carpenter JF. Effects of solution conditions, processing parameters, and container materials on aggregation of a monoclonal antibody during freeze-thawing. J Pharm Sci 2008;97(5):1801e12. [9] Lubiniecki A, Volkin DB, Federici M, Bond MD, Nedved ML, Hendricks L, et al. Comparability assessments of process and product changes made during development of two different monoclonal antibodies. Biologicals 2010;39(1):9e22. [10] Narhi M, Nordstrom K. Manufacturing, regulatory and commercial challenges of biopharmaceuticals production: a Finnish perspective. Eur J Pharm Biopharm 2005;59(3):397e405. [11] Schenerman MA, Hope JN, Kletke C, Singh JK, Kimura R, Tsao EI, et al. Comparability testing of a humanized monoclonal antibody (Synagis) to support cell line stability, process validation, and scale-up for manufacturing. Biologicals 1999;27(3):203e15. [12] DiMasi JA, Hansen RW, Grabowski HG. The price of innovation: new estimates of drug development costs. J Health Econ 2003;22(2):151e85. [13] DiMasi JA, Hansen RW, Grabowski HG, Lasagna L. Research and development costs for new drugs by therapeutic category. A study of the US pharmaceutical industry. Pharmacoeconomics 1995;7(2):152e69. [14] DiMasi JA, Paquette C. The economics of follow-on drug research and development: trends in entry rates and the timing of development. Pharmacoeconomics 2004;22(2 Suppl. 2):1e14. [15] Samanen J. The structure and business of biopharmaceutical companies including the management of risk and resources. In: Ganellin CR, Jefferis R, Roberts S, editors. Introduction to biological and small molecule drugs research and development: theory and case studies. Elsevier; 2013. p. 225e53. [16] Bharate SS, Vishwakarma RA. Impact of preformulation on drug development. Expert Opin Drug Deliv 2013:1239e57.

I. Proteins and biophysical characterization in the biopharmaceutical industry

50

2. Biophysical characterization and its role in the biopharmaceutical industry

[17] Noiles K, Vender R. Biologic survival. J Drugs Dermatol JDD 2009;8(4):329e33. [18] Saxena V, Panicucci R, Joshi Y, Garad S. Developability assessment in pharmaceutical industry: an integrated group approach for selecting developable candidates. J Pharm Sci 2009;98(6):1962e79. [19] DiMasi JA, Feldman L, Seckler A, Wilson A. Trends in risks associated with new drug development: success rates for investigational drugs. Clin Pharmacol Ther 2010;87(3):272e7. [20] Berkowitz SA. Role of analytical ultracentrifugation in assessing the aggregation of protein biopharmaceuticals. AAPS J 2006;8(3):E590e605. [21] Garber E, Demarest SJ. A broad range of Fab stabilities within a host of therapeutic IgGs. Biochem Biophys Res Commun 2007;355(3):751e7. [22] Chalmers MJ, Busby SA, Pascal BD, He Y, Hendrickson CL, Marshall AG, et al. Probing protein ligand interactions by automated hydrogen/deuterium exchange mass spectrometry. Anal Chem 2006;78(4):1005e14. [23] Landgraf RR, Chalmers MJ, Griffin PR. Automated hydrogen/deuterium exchange electron transfer dissociation high resolution mass spectrometry measured at single-amide resolution. J Am Soc Mass Spectrom 2012;23(2):301e9. [24] Gill SC, von Hippel PH. Calculation of protein extinction coefficients from amino acid sequence data. Anal Biochem 1989;182(2):319e26. [25] Ander JC, Parten BF, Petrie GE, Marlowe RL, McEntire JE. Using amino acid analysis to determine absorptivity constants: a validation case study using bovine serum albumin. Biopharm Int 2003;16(2):30e7. [26] Lundbald RL, Price NC. Protein concentration determination: the achilles’ heel of cGMP. BioProcess Int 2004:38e47. [27] Pace CN, Vajdos F, Fee L, Grimsley G, Gray T. How to measure and predict the molar absorption coefficient of a protein. Protein Sci 1995;4:2411e23. [28] Gray RA, Stern A, Bewley T, Shire SJ. Rapid Determination of Spectrophotometric Absorptivity by Analytical Ultracentrifugation. Beckman Coulter Application Information. Beckman-Coulter Application Note Number A-1815-A. http://www.beckman.com/beckman/biorsrch/prodinfo/xla/pdf/a_1815a.pdf. [29] Voelker P. Measurement of the extinction coefficient of prostate specific antigen using interference and absorbance optics in the optima XL-A analytical ultracentrifuge. Prog Colloid Polym Sci 1995;99. [30] Wen J, Arakawa T, Philo JS. Size-exclusion chromatography with on-line light-scattering, absorbance, and refractive index detectors for studying proteins and their interactions. Anal Biochem 1996;240(2):155e66. [31] Chaudhuri R, Cheng Y, Middaugh CR, Volkin DB. High-throughput biophysical analysis of protein therapeutics to examine interrelationships between aggregate formation and conformational stability. AAPS J 2014;16(1):48e64. [32] Iyer V, Maddux N, Hu L, Cheng W, Youssef AK, Winter G, et al. Comparative signature diagrams to evaluate biophysical data for differences in protein structure across various formulations. J Pharm Sci 2013;102(1):43e51. [33] Chirino AJ, Mire-Sluis A. Characterizing biological products and assessing comparability following manufacturing changes. Nat Biotechnol 2004;22(11):1383e91. [34] FDA US Food and Drug Administration. FDA Guidance Concerning Demonstration of Comparability of Human Biological Products, Including Therapeutic Biotechnology-Derived Products; 1996. [35] Federici M, Lubiniecki A, Manikwar P, Volkin DB. Analytical lessons learned from selected therapeutic protein drug comparability studies. Biologicals 2013;41(3):131e47. [36] EMEA. Comparability of Biotechnological/Biological Products Subject to Changes in Their Manufacturing Process. Note for Guidance on Biotechnological/Biological Products Subject to Changes in their Manufacturing Process; 2006 (CPMP/ICH/5721/03) ICH 5QE. [37] Petricciani J. A global view of comparability concepts. Dev Biol 2002;109:9e13. [38] Keith J, Ilari A, Savino C. Protein structure determination by X-Ray crystallography. In: Bioinformatics. Humana Press; 2008. p. 63e87. [39] Wuthrich K, editor. Protein Structure Determination in Solution by NMR Spectroscopy; 1990. p. 22059e62. [40] Thuman-Commike PA. Single particle macromolecular structure determination via electron microscopy. FEBS Lett 2001;505(2):199e205. [41] Glaeser RM. Review: electron crystallography: present excitement, a nod to the past, anticipating the future. J Struct Biol 1999;128(1):3e14. [42] Luzzati V, Tardieu A. Recent developments in solution X-ray scattering. Annu Rev Biophys Bioeng 1980;9:1e29.

I. Proteins and biophysical characterization in the biopharmaceutical industry

References

51

[43] Mertens HD, Svergun DI. Structural characterization of proteins and complexes using small-angle X-ray solution scattering. J Struct Biol 2010;172(1):128e41. [44] Putnam CD, Hammel M, Hura GL, Tainer JA. X-ray solution scattering (SAXS) combined with crystallography and computation: defining accurate macromolecular structures, conformations and assemblies in solution. Q Rev Biophys 2007;40(3):191e285. [45] Maddux NR, Joshi SB, Volkin DB, Ralston JP, Middaugh CR. Multidimensional methods for the formulation of biopharmaceuticals and vaccines. J Pharm Sci 2011;100(10):4171e97. [46] Hu L, Olsen C, Maddux N, Joshi SB, Volkin DB, Middaugh CR. Investigation of protein conformational stability employing a multimodal spectrometer. Anal Chem 2011;83(24):9399e405. [47] Brader ML, Baker EN, Dunn MF, Laue TM, Carpenter JF. Using X-ray crystallography to simplify and accelerate biologics drug development. J Pharm Sci 2017;106(2):477e94. [48] Maddux NR, Rosen IT, Hu L, Olsen CM, Volkin DB, Middaugh CR. An improved methodology for multidimensional high-throughput preformulation characterization of protein conformational stability. J Pharm Sci 2012;101(6):2017e24. [49] Ramsey JD, Gill ML, Kamerzell TJ, Price ES, Joshi SB, Bishop SM, et al. Using empirical phase diagrams to understand the role of intramolecular dynamics in immunoglobulin G stability. J Pharm Sci 2009;98(7): 2432e47. [50] Evans DRH, Romero JK, Westoby M, Richard RB, Murray PD. Concentration of proteins and removal of solutes. In: Methods in Enzymology. Academic Press; 2009. p. 97e120 (Chapter 9). [51] Cordes AA, Carpenter JF, Randolph TW. Accelerated stability studies of abatacept formulations: comparison of freeze-thawing and agitation-induced stresses. J Pharm Sci 2012;101(7):2307e15. [52] Hammond M, Nunn H, Rogers G, Lee H, Marghitoiu AL, Perez L, et al. Identification of a leachable compound detrimental to cell growth in single-use bioprocess containers. PDA J Pharm Sci Technol/PDA 2013;67(2):123e34. [53] Zhang A, Singh SK, Shirts MR, Kumar S, Fernandez EJ. Distinct aggregation mechanisms of monoclonal antibody under thermal and freeze-thaw stresses revealed by hydrogen exchange. Pharm Res 2012;29(1):236e50. [54] Manning MC, Chou DK, Murphy BM, Payne RW, Katayama DS. Stability of protein pharmaceuticals: an update. Pharm Res 2010;27(4):544e75. [55] Majumdar S, Ford BM, Mar KD, Sullivan VJ, Ulrich RG, D’Souza AJ. Evaluation of the effect of syringe surfaces on protein formulations. J Pharm Sci 2011;100(7):2563e73. [56] Boven K, Stryker S, Knight J, Thomas A, van Regenmortel M, Kemeny DM, et al. The increased incidence of pure red cell aplasia with an Eprex formulation in uncoated rubber stopper syringes. Kidney Int 2005;67(6):2346e53. [57] Berkowitz SA, Engen JR, Mazzeo JR, Jones GB. Analytical tools for characterizing biopharmaceuticals and the implications for biosimilars. Nat Rev 2012;11(7):527e40. [58] Kozlowski S, Woodcock J, Midthun K, Sherman RB. Developing the nation’s biosimilars program. N Engl J Med 2011;365(5):385e8. [59] Rader RA. What is a generic biopharmaceutical? Biogeneric? Follow-on protetin? Biosimilar? Follow-on biologic? Bio Process Int 2007:28e38. [60] Woodcock J, Griffin J, Behrman R, Cherney B, Crescenzi T, Fraser B, et al. The FDA’s assessment of follow-on protein products: a historical perspective. Nat Rev 2007;6(6):437e42. [61] Walsh G, Jefferis R. Post-translational modifications in the context of therapeutic proteins. Nat Biotechnol 2006;24(10):1241e52. [62] Jenkins N. Modifications of therapeutic proteins: challenges and prospects. Cytotechnology 2007;53(1e3): 121e5. [63] Houde D, Peng Y, Berkowitz SA, Engen JR. Post-translational modifications differentially affect IgG1 conformation and receptor binding. Mol Cell Proteom 2010;9(8):1716e28. [64] Hu D, Qin Z, Xue B, Fink AL, Uversky VN. Effect of methionine oxidation on the structural properties, conformational stability, and aggregation of immunoglobulin light chain LEN. Biochemistry 2008;47(33):8665e77. [65] Liu D, Ren D, Huang H, Dankberg J, Rosenfeld R, Cocco MJ, et al. Structure and stability changes of human IgG1 Fc as a consequence of methionine oxidation. Biochemistry 2008;47(18):5088e100. [66] Dengl S, Wehmer M, Hesse F, Lipsmeier F, Popp O, Lang K. Aggregation and chemical modification of monoclonal antibodies under upstream processing conditions. Pharm Res 2013;30(5):1380e99.

I. Proteins and biophysical characterization in the biopharmaceutical industry

52

2. Biophysical characterization and its role in the biopharmaceutical industry

[67] Webb SD, Golledge SL, Cleland JL, Carpenter JF, Randolph TW. Surface adsorption of recombinant human interferon-g in lyophilized and spray-lyophilized formulations. J Pharm Sci 2002;91(6):1474e87. [68] Bee JS, Randolph TW, Carpenter JF, Bishop SM, Dimitrova MN. Effects of surfaces and leachables on the stability of biopharmaceuticals. J Pharm Sci 2011;100(10):4158e70. [69] Sharma B. Immunogenicity of therapeutic proteins. Part 2: impact of container closures. Biotechnol Adv 2007;25(3):318e24. [70] FDA Advisory Committee Open Session Information for Public Release, Alglucosidase Alfa Advisory Committee Briefing Package10. 2008. http://wwwfdagov/ohrms/dockets/ac/08/briefing/2008-4389b1-04Genzymepdf. [71] Thirumangalathu R, Krishnan S, Ricci MS, Brems DN, Randolph TW, Carpenter JF. Silicone oil- and agitationinduced aggregation of a monoclonal antibody in aqueous solution. J Pharm Sci 2009;98(9):3167e81. [72] Vermeer AWP, Giacomelli CE, Norde W. Adsorption of IgG onto hydrophobic teflon. Differences between the fab and fc domains. Biochim Biophys Acta Gen Subj 2001;1526(1):61e9. [73] Schellekens H. When biotech proteins go off-patent. Trends Biotechnol 2004;22(8):406e10. [74] Thayer AM. The new copycats. Chem Eng News 2013;91(40):15e23. [75] FDA US Food and Drug Administration. Drug Price Competition and Patent Term Restoration Act of 1984 (Hatch-Waxman Amendments). [76] Heavner GA, Arakawa T, Philo JS, Calmann MA, LaBrenz S. Protein isolated from biopharmaceutical formulations cannot be used for comparative studies: follow-up to “case study using Epoetin alfa from Epogen and EPREX”. J Pharm Sci 2007;96(12):3214e25. [77] Hirsch BR, Balu S, Schulman KA. The impact of specialty pharmaceuticals as drivers of healthcare cost. Health Aff 2014;33(10):1714e20. [78] Beck A, Sanglier-Cianferani S, Van Dorsselaer A. Biosimilar, biobetter, and next generation antibody characterization by mass spectrometry. Anal Chem 2012;84(11):4637e46. [79] Zhang Z, Pan H, Chen X. Mass spectrometry for structural characterization of therapeutic antibodies. Mass Spectrom Rev 2009;28(1):147e76. [80] Timms JF, Arslan-Low E, Gentry-Maharaj A, Luo Z, T’Jampens D, Podust VN, et al. Preanalytic influence of sample handling on SELDI-TOF serum protein profiles. Clin Chem 2007;53(4):645e56. [81] Mächtle W, Börger L, editors. Analytical Ultracentrifugation, Instrumentation. Analytical Ultracentrifugation of Polymers and Nanoparticles. Heidelberg: Springer Berlin; 2006. p. 17e46. [82] Shire SJ, Shahrokh Z, Liu J. Challenges in the development of high protein concentration formulations. J Pharm Sci 2004;93(6):1390e402. [83] Minton AP. Influence of macromolecular crowding upon the stability and state of association of proteins: predictions and observations. J Pharm Sci 2005;94(8):1668e75. [84] Singh SR, Zhang J, O’Dell C, Hsieh MC, Goldstein J, Liu J, et al. Effect of polysorbate 80 quality on photostability of a monoclonal antibody. AAPS Pharm Sci Tech 2012;13(2):422e30. [85] Liu J, Shire SJ. Analytical ultracentrifugation in the pharmaceutical industry. J Pharm Sci 1999;88(12):1237e41. [86] Armstrong SE, Mariano JA, Lundin DJ. The scope of mycoplasma contamination within the biopharmaceutical industry. Biologicals 2010;38(2):211e3. [87] Drexler HG, Uphoff CC. Mycoplasma contamination of cell cultures: incidence, sources, effects, detection, elimination, prevention. Cytotechnology 2002;39(2):75e90. [88] Robinson LB, Wichelhausen RH. Contamination of human cell cultures by pleuropneumonialike organisms. Science 1956;124(3232):1147e8. [89] Kerwin BA, Remmele Jr RL. Protect from light: photodegradation and protein biologics. J Pharm Sci 2007;96(6):1468e79. [90] Li S, Schoneich C, Borchardt RT. Chemical instability of protein pharmaceuticals: mechanisms of oxidation and strategies for stabilization. Biotechnol Bioeng 1995;48(5):490e500. [91] Stadtman ER, Oliver CN. Metal-catalyzed oxidation of proteins. Physiological consequences. J Biol Chem 1991;266(4):2005e8. [92] Agarkhed M, O’Dell C, Hsieh MC, Zhang J, Goldstein J, Srivastava A. Effect of polysorbate 80 concentration on thermal and photostability of a monoclonal antibody. AAPS Pharm Sci Tech 2013;14(1):1e9. [93] Henry C. FDA, reform, and the well-characterized biologic. Anal Chem 1996;68(21):674Ae7A.

I. Proteins and biophysical characterization in the biopharmaceutical industry

Further reading

53

[94] Bren L. The road to the biotech revolution: highlights of 100 years of biologics regulation. FDA Consum 2006;40(1):50e7. [95] Rosenberg AS, Verthelyi D, Cherney BW. Managing uncertainty: a perspective on risk pertaining to product quality attributes as they bear on immunogenicity of therapeutic proteins. J Pharm Sci 2012;101(10):3560e7. [96] Schiestl M, Stangler T, Torella C, Cepeljnik T, Toll H, Grau R. Acceptable changes in quality attributes of glycosylated biopharmaceuticals. Nat Biotechnol 2011;29(4):310e2. [97] Rathore AS, Winkle H. Quality by design for biopharmaceuticals. Nat Biotechnol 2009;27(1):26e34. [98] Rathore AS. Roadmap for implementation of quality by design (QbD) for biotechnology products. Trends Biotechnol 2009;27(9):546e53. [99] Yu L. Pharmaceutical quality by design: product and process development, understanding, and control. Pharm Res 2008;25(4):781e91. [100] FDA US Food and Drug Administration. Pharmaceutical cGMPs for the 21st Century: A Risk-Based Approach. [101] FDA US Food and Drug Administration. PAT Guidance for Industryda Framework for Innovative Pharmaceutical Development, Manufacturing and Quality Assurance. US Department of Health and Human Services, Food and Drug Administration, Center for Drug Evaluation and Research, Center for Veterinary Medicine, Office of Regulatory Affairs. [102] FDA US Food and Drug Administration. Guidance for Industry: Q9 Quality Risk Management. US Department of Health and Human Service; 2006.

Further reading [1] Shukla AA, Etzel MR, Gadam S, editors. Process Scale Bioseparations for the Biopharmaceutical Industry. Taylor & Francis Group; 2007. [2] Narhi LO, editor. Biophysics for Therapeutic Protein Development. New York (NY): Springer; 2013. [3] Lundbald RL, editor. Approaches to the Conformational Analysis of Biopharmaceuticals. Boca Raton (FL: CRC Press, Taylor and Francis Group; 2010. [4] Rathore AS, Mhatre R, editors. Quality by Design for Biopharmaceuticals: Principles and Case Studies. WileyInterscience; 2009 [Wiley Series in Biotechnology and Bioengineering]. [5] Ganellin CR, Jefferis R, Roberts S, editors. Introduction to Biological and Small Molecule Drugs Research and Development: Theory and Case Studies. Oxford (UK): Academic Press, Elsevier; 2013.

I. Proteins and biophysical characterization in the biopharmaceutical industry

C H A P T E R

3

Biopharmaceutical industry’s biophysical toolbox Damian J. Houdea, Steven A. Berkowitzb a

Biomolecular Discovery, Relay Therapeutics, Cambridge, MA, United States; bConsultant, Sudbury, MA, United States

3.1 Attributes of a single biophysical tool to characterize and detect changes in the higher order structure of a biopharmaceutical Although it is unlikely that any single biophysical tool will ever provide all the necessary information needed to characterize the higher order structure (HOS) of a protein drug, it is a worthy idea to contemplate the characteristics as to just what such a hypothetical biophysical tool might look like, in terms of its key attributes. Such characteristics are, in part, coarsely illustrated in Fig. 3.1. Preferably, the tool should be capable of providing a unique quantitative signal read out for each basic structural element that makes up the drug that has good sensitivity with high spatial resolution resulting no or minimum overlapping of each of these unique signals. These characteristics can be broken down into four essential parts. First is the ability to measure signals from as small a unit of the protein as possible (e.g., atomic resolution would be ideal, but amino acid level might be acceptable, see Fig. 3.2). Second is the ability to detect all such signals emitted from a single molecule; essentially, we want to detect and interrogate as much of the protein’s structure (hopefully the entire structure) as possible. Third is the ability to separate or resolve these signals spatially in a manner using some parameter space (e.g., relative atomic distances, wavelength, temperature, time, etc.) with minimal signal overlap. The fourth and final characteristic would be the ability to quantitatively record all these signals with the highest precision (and accuracy) possible. The highest resolution of a protein’s HOS would amount to knowing all the relative positions of the atoms in the entire protein molecule (again see Fig. 3.2). Having accurate data on every atom in a protein, in terms of their spatial coordinates, as well as their temporal behavior (i.e., dynamics) would correspond to the ultimate situation in detecting a difference in the smallest element of the protein drug. Such direct information (as mentioned in Chapter 2) can presently be approached by only two techniques, X-ray crystallography and nuclear

Biophysical Characterization of Proteins in Developing Biopharmaceuticals, Second Edition https://doi.org/10.1016/B978-0-444-64173-1.00003-2

55

Copyright © 2020 Elsevier B.V. All rights reserved.

56

3. Biopharmaceutical industry’s biophysical toolbox

FIG. 3.1 The impact of overlapping (A) versus nonoverlapping (B) signals from a biopharmaceutical on the ability of a biophysical measurement to detect changes in the HOS between different drug samples. In the example given in this figure, the biophysical measurement on a protein drug is derived from only six different structural elements on the drug molecule. Each structural elements outputs a signal, S(1)eS(6) in terms of this particular biophysical measurement. In this specific example, two drug samples (referred to as sample 1 and sample 2) are being compared in which the only difference that exists between them is due to a 20% reduction in the intensity of S(5) in sample 2. This difference is indicated by the same difference in the vertical spacing value between the two dotted black lines in Parts A and B (that corresponds to the intensity of level of S(5) in samples 1 and 2) and whose difference is represented in both parts by the symbol D5. However, in Part A, there is extensive overlapping of the six signals, which in the area where S(5) appears along the x-axis (of the data output), S(5) now only accounts for about 25% of the total output signal. Given this reduction in the contribution from S(5), the original difference (D5) in signal S(5) between samples 1 and 2 is also reduced to 0.25D5 as indicated by the blue (gray in print version) dotted lines (which now amounts to only a 5% reduction in the total signal where the S(5) appears in the data output). With the same 99% confidence limit, indicated by the error bars in Parts A and B, the biophysical measurement difference value of 0.25D5 in S(5) between samples 1 and 2 in Part A will go undetected in the case where the signal outputs overlap, due to the level of statistical noise in the data, while in the case of Part B where signal outputs are non-overlapping, the original D5 difference is maintained and can be detected.

I. Proteins and biophysical characterization in the biopharmaceutical industry

3.2 Studying the biophysical properties of a biopharmaceutical as an indirect approach for characterizing changes in its HOS

57

FIG. 3.2 A schematic representation of the hierarchy of spatial resolution of a biopharmaceutical’s HOS is shown. The different structural elements and their relationship to the three basic forms of chemical and physical changes (which include post-transtrantional modifications (PTMs) and fragmentation, non-covalent changes (silent changes) and aggregation) that can impact the HOS and biophysical properties of a biopharmaceutical are also illustrated.

magnetic resonance (NMR). However, the application of these tools are beyond the scope of this book and will not be discussed (also see Chapter 2, Section 2.3.3.1).

3.2 Studying the biophysical properties of a biopharmaceutical as an indirect approach for characterizing changes in its HOS In the last decade, the ability of more advanced biophysical tools to provide direct information on the HOS of proteins has greatly improved [1e3]. However, the impracticality or present limitations in utilizing these tools in the process development area of today’s biopharmaceutical industry still remains. As a result, many biophysical tools and their associated methods have evolved to support an alternative approach for extracting indirect information about a molecule’s HOS. This approach uses biophysical characterization studies to gather information on the “biophysical properties” of these molecules. The success of using biophysical properties to proxy for the direct determination of the HOS of a biopharmaceutical rests on the idea that the combination of a molecule’s chemical composition and its HOS give rise to unique biophysical properties (Fig. 3.3). Through this critical linkage, information

I. Proteins and biophysical characterization in the biopharmaceutical industry

58

3. Biopharmaceutical industry’s biophysical toolbox

FIG. 3.3 The impact of chemical and physical changes on the HOS (as indicated by the objects shape,

rectangle ¼ native HOS, hexagon ¼ altered HOS) and biophysical properties of the monomeric form of a biopharmaceutical after its primary structure is synthesized and folded to attain its normal HOS. (A) An example where a large specific chemical addition is made to a biopharmaceutical to give a new form of the drug that displays different biophysical properties and an altered HOS relative to the original biopharmaceutical. (B) The same as for “A” but the chemical modification does not alter the HOS of the original drug. However, the resulting drug still displays different biophysical properties because it is chemically modified. (C) An example where a small specific chemical addition (post-translational modifications, PTMs) occurs on a biopharmaceutical that does not alter the HOS, but may or may not alter biopharmaceutical’s biophysical properties. (D) The same example as in “C” but the original biopharmaceutical’s HOS and biophysical properties are altered. (E) A physical change that alters the HOS and biophysical properties of the biopharmaceutical. (F) The same as in “E” but the physical change does not alter the HOS or biophysical properties of the biopharmaceutical.

about changes in a biopharmaceutical’s biophysical properties should indirectly shed light on changes to its HOS. Although this information does not delineate the spatial coordinates of the protein drug’s atomic structure, it is hoped that by measuring a collection of simple parameters derived from the molecule’s various biophysical properties an indirect fingerprint of the biopharmaceutical’s HOS is obtained. Using such information, constraints on the variation of these biophysical properties can be put into place such that if the value for these parameters falls outside a predetermined, established, and agreed-upon range, they would imply a likely change in the biopharmaceutical’s HOS.

I. Proteins and biophysical characterization in the biopharmaceutical industry

3.2 Studying the biophysical properties of a biopharmaceutical as an indirect approach for characterizing changes in its HOS

59

In general, in the biopharmaceutical process development area, it is the simple biophysical tools that monitor the biophysical properties of a biopharmaceutical that forms the backbone of most of the biophysical studies conducted during the development of these drugs. In most cases, these simple biophysical tools often yield straightforward information, that is represented by a single numerical parameter (e.g., melting temperature, Tm, sedimentation coefficient, s) or by a two-dimensional graphical profile (e.g., in the case of circular dichroism, by a CD spectrum or in the case of chromatography by a chromatogram). Changes that can be introduced into a biopharmaceutical’s HOS during its production that can affect its efficacy, can be categorized into the following two areas: (1) those caused by chemical (covalent) changes, e.g., due to post-translational modifications, PTMs, and fragmentation, and (2) those caused by physical (noncovalent) changes, e.g., due to temperature and agitation. In most cases, these chemical and physical changes are typically confined to small areas of the biopharmaceutical molecule that may not lead to any significant change in the drug’s HOS (Fig. 3.3C and F). These changes are often so small that they can be very difficult to detect and therefore lead to no measurable change in the biophysical properties of the biopharmaceutical. On the other hand, there are cases where a small chemical or physical change could lead to a significant change in the HOS of the biopharmaceutical and therefore, lead to measurable changes in its biophysical properties [4e6] (Fig. 3.3D and E). Thus, there is significant variation from one protein to another in the ability of a protein to respond to the same chemical or physical change. Even within the same protein the same modification in one area of a biopharmaceutical can lead to no significant change in HOS, while the same type of modification in another part of the same protein molecule can lead to very significant HOS changes [7]. As a result, to minimize the inability of detecting a change in a protein drug’s biophysical properties (and therefore its HOS), we are left with the risk mitigating task of applying multiple biophysical tools that probe different biophysical properties of these drugs. In the case of chemical changes (PTMs), it should be noted, however, that it is also possible that a chemical change (especially on the surface of the biopharmaceutical) may lead to a very measurable change in the biophysical property of a biopharmaceutical but result in no measurable change in its HOS (Fig. 3.3B) [7]. The assessment of such situations in fact is an important development activity in making certain that novel classes of protein drugs that require direct chemical modification are not altered in terms of their HOS. Such protein drugs include those that involve the direct coupling of a second protein (in forming fusion proteins), a large chemical polymer (e.g., pegylation), or a small toxic compound to make a conjugated biopharmaceuticals, (antibody drug conjugate, ADC) [8e13]. While this indirect HOS approach may be more straightforward, there are still important considerations, such as “how well do the measured biophysical properties proxy for assessing information on a biopharmaceutical's HOS”? The answer to this question hinges on finding and using an adequate number of biophysical tools that can collectively provide information about the various parts of the protein’s HOS that regulate different physical properties of the protein drug. By measuring several different biophysical properties that are dependent on different attributes of the biopharmaceutical’s HOS, a better opportunity is created for generating a more effective fingerprint of the biopharmaceutical’s HOS. Hence, the success of this approach ultimately rests on understanding what attribute of a biopharmaceutical’s

I. Proteins and biophysical characterization in the biopharmaceutical industry

60

3. Biopharmaceutical industry’s biophysical toolbox

HOS is being probed, how sensitive the technique is at assessing a change, and are they important. As an example, consider the use of global hydrogen exchange (HX) with mass spectrometry detection for probing the conformational properties of backbone amide hydrogens on an intact biopharmaceutical (Chapter 12). When a biopharmaceutical is placed into a formulation buffer made with deuterated water instead normal water, many amide hydrogens present in a biopharmaceutical will start to exchange with deuterium in the deuterated water (H/DX). However, a change in the HOS of a biopharmaceutical can increase or decrease the H/DX of some amide hydrogens. In some cases, the result (change in the total number of exchanged hydrogens) may not show any change, indicating no change in HOS. Another example concerns intrinsic fluorescence measurements (Chapter 5). In this case, the signal recorded is from a limited number of reporter elements (i.e., aromatic amino acid residues only). If the change in a protein drug’s HOS is not located within a region where the reporter element is present (i.e., an aromatic residue), the change is likely to go undetected. Similarly, if there is a change in the drug’s conformation that contains more than one reporter element (e.g., two tryptophan amino acids), it is possible that the change in the HOS could cause one tryptophan to increase in fluorescence while the other may decrease in fluorescence. In the latter situation, unless the fluorescence signals from each tryptophan can be uniquely resolved (possibly via a large-enough wavelength shift), the overall fluorescence spectrum will look like that of the original unaltered protein drug. Clearly, having uniquely resolved readouts from as many parts of the protein drug as possible will increase our ability to better characterize and detect a change in the HOS of a biopharmaceutical. In addition, we will also see that, in some cases, the ability to have more than one biophysical technique to probe the same biophysical attribute using different physical principles (e.g., size exclusion chromatography (SEC) and analytical ultracentrifugation (AUC) in assessing protein aggregation) can also be beneficial in securing confidence in the data and the conclusions that are drawn [14,15].

3.3 General considerations in analyzing the biophysical properties of biopharmaceuticals Given the number of commercial biopharmaceuticals that are on the market where the active pharmaceutical ingredient (API) is a protein, the word “biopharmaceutical” has commonly been found to mean protein drug (or protein therapeutic). In 2012, the Food and Drug Administration (FDA) defined any polypeptide chain longer than 40 amino acids [16] as a protein. Although other chemical classes of biopharmaceuticals exist, where the API is not a protein (e.g., RNAi, DNA (in gene therapy) [17,18], peptides (polypeptide chain