4/21 /18 PlanetDepos -Tricia Rosate, ROR. CRR. CSR No. 10891
`
`Page 1
`
`
`
`1 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Applications and Engineering of
`Monoclonal Antibodies
`
`HOSPIRA EX. 1 029
`Page 2
`
`
`
`2 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Applications and Engineering of
`Monoclonal Antibodies
`
`DAVID J. KING
`Celltech Therapeutics
`Slough, UK
`
`HOSPIRA EX. 1 029
`Page 3
`
`
`
`3 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Taylor & Francis Ltd, l Gunpowder Square, London EC4A 3DE
`UK
`USA Taylor & Francis Inc., 325 Chestnut Street, 8th Floor, Philadelphia, PA 19lOG
`Copyright © D.J. King 1998
`Eve1y effort has heen made to ensure that the advice and information in this book is trne
`British Library Cataloguing-in-Publication
`
`All rights reserved. No part of this publication may be reproduced, stored in a retrieval
`
`
`
`
`
`sys/em, or transmiued, in any form or by any means, electronic, electrostatic, magnetic
`
`
`
`
`
`tape, mechanical, photocopying, recording or otherwise. without the prior permission of"
`
`the copyright owner.
`
`and accurate at tlze time of going lo press. However, neither the publisher nor the author
`
`
`
`can accept any legal responsibility or liability for any errors or omissions that may he
`
`made. In the case of drug administration, any medical procedure or the use of 1ech11ical
`equipment mentioned within this book. you are strongly advised lo consult the
`
`manufacturer's guidelines.
`
`Data
`A catalogue record for this book is available from the British Librar1
`ISBN 0-7484-0422-8 (hb)
`0-7484-0423-6 (pb)
`
`...
`
`
`
`Lib1·ary of Congress Cataloging Publication Data arc available
`
`Cover design by Youngs Design in Production
`
`Typeset in Times 10/l2pt by Graphicraft Limited, Hong Kong
`
`Printed by T.J. International Ltd, Padstow, UK
`
`HOSPIRA EX. 1 029
`Page 4
`
`
`
`4 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Acknowledgements
`
`I would like to thank my many colleagues at Celltech for their helpful suggestions and
`advice, and particularly Alastair Lawson and Martyn Robinson for their comments on the
`manuscript. Thanks also to Tina Jones for assistance with the figures.
`And lastly, sincere thanks to my wife, Jane, for her support and patience through the
`long hours and weekends spent writing this book.
`
`HOSPIRA EX. 1 029
`Page 5
`
`
`
`5 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Contents
`
`1 !Preparation, siruciure and function of monoclonal antibodies
`LS Antigen-binding affinity and avidity
`
`1. 1
`Introduction
`1.2 The role of antibodies in the immune response
`1.3 Structure and function of antibodies
`1 .4 The organisation of antibody genes
`
`1.5 . 1 Affinity
`1.5.2 Avidity
`1 . 6 Generation of monoclonal antibodies
`1.6. 1 Hybridoma technology
`1.6.2 Human monoclonal antibodies
`1.6.3 Human MAbs from transgenic mice
`1.6.4
`Isolation of antibodies by phage display
`
`2 Antibody engineering: design for specific applications
`
`2 . 1
`Introduction
`2.2
`Isolation of variable region genes
`2.3 Overcoming irnmunogenicity
`2.3. 1 Chimeric and humanised antibodies
`2.3.2 Antibody fragments to reduce immunogenicity
`2.3.3 Chemical modification to reduce immunogenicity
`2.3.4 Immunosuppressive therapy
`2.4 Antibody fragments
`2.4. 1 Antibody fragments from proteolysis of IgG
`2.4.2 Recombinant antibody fragments
`Fab-based fragments
`Fv-based fragments
`Multivalent antibody fragments
`2.5 Antibodies with multiple specificities
`2.6 Engineering effector functions
`
`page 1
`I
`13 I.,
`• .J
`14 14 14
`
`2
`2
`11
`
`18
`20
`21
`
`27
`27
`28
`29
`29
`37
`38
`40
`40
`40
`42
`43
`45
`48
`
`50 54
`
`HOSPIRA EX. 1 029
`Page 6
`
`
`
`6 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Vlll
`
`2.6. l Engineering natural effector functions
`2.6.2 Attachment of diagnostic or therapeutic agents
`Chemical conjugates
`Site-specific attachment
`Fusion proteins
`2.7 Engineering phamrncokinetics and biodistribution
`2.7. l Pharmacokinetics of IgG
`2.7.2 Pharmacokinetics of antibody fragments
`2.7.3 Clearance
`2.7.4 Chemical modification
`
`2.7.5 Fe region to extend half-life
`3 Monoclonal antibodies in research and diagnostic applications
`
`3.1
`3.2
`
`Introduction
`Immunoassays in diagnostics and research
`3.2.l Radioimmunoassay
`3.2.2
`Immunoradiometric assay
`3.2.3 Non-isotopic immunoassays
`3.2.4
`Improving sensitivity
`3.2.5 Assay formats
`3.2.6 Advantages of monoclonal antibodies in immunoassay
`Immunosensors
`3.3.1 Mass-detecting immunosensors
`3.3.2 Electrochemical immunosensors
`3.3.3 Optical immunoscnsors
`3.4
`Immunocytochemistry
`3.5 Flow cytometry and cell sorting (FACS)
`3.6 Western blotting (immunoblotting)
`3. 7
`Immunopurification
`3.8 Antibodies in structural biology
`3.9 In vivo diagnostics
`3.9.1 Radioimrnunodetection of human tumours
`Tumour-associated antigens
`Form of antibody
`Radioisotopes
`Two- and three-step targeting approaches
`3.9.2 Radioimmunoguided surgery
`3.9.3 Non-tumour radioimmunodetection
`
`3.3
`
`4 Monoclonal antibodies in therapeutic applications
`
`4.1
`Introduction
`4.2 Cancer
`4.2. l Cancer therapy with unmodified (naked) antibodies
`4.2 2 Anti-idiotype antibodies
`4.2.3 Bispecific antibody-mediated effector cell targeting
`4.2.4 Other approaches to recruit the immune system using MAbs
`4.2.5 Radioimmunotherapy
`Form of antibody
`Radioisotopes
`
`Contents
`58 58 63
`
`54
`
`65
`67
`68
`71
`74
`75
`75
`
`77
`77
`77
`78
`79
`83
`85
`89
`91
`92
`93
`94
`95
`97
`99
`100
`102
`106
`106
`107
`107
`109
`110
`114
`116
`116
`
`119
`119
`120
`121
`123
`125
`127
`128
`129
`132
`
`HOSPIRA EX. 1 029
`Page 7
`
`
`
`7 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Clinical results with RIT
`
`Clearance mechanisms and two-step targeting
`
`Contents
`
`4.3
`
`4.2.6
`Immunotoxins
`4.2.7 Drug conjugates
`4.2.8 Antibody-directed enzyme prodrug therapy (ADEPT)
`4.2.9 Vascular targeting
`Infectious disease
`4.3.1 Antiviral antibodies
`4.3.2 Bacterial sepsis
`4.4 Cardiovascular disease
`4.4.1
`Inhibition of platelet aggregation
`4.4.2 Thrombolysis
`4.5 Disorders of the immune system/inflammatory diseases
`4.5. l The inflanunatory response
`4.5.2 Blocking inflammatory mediators
`Anti-TNF antibodies in rheumatoid arthritis and
`inflammatory bowel disease
`Anti-CS
`4.5.3 Blocking adhesive interactions
`4.5.4 Antibodies which directly inhibit T cell activation and
`proliferation
`4.5.5 Antibody treatment of allergy
`
`5 Production of monoclonal antibodies
`5.3 Expression in Escherichia coli
`Intracellular expression of antibody fragments in E. coli
`5.3.2 Secretion of antibody fragments from E. coli
`
`5.1
`Introduction
`5.2 Expression of antibodies in mammalian cells
`5.2.1 Transient expression systems
`5.2.2 Stable expression systems
`5.2.3 Expression of antibody fragments in mammalian cells
`
`5.3. l
`
`5.4 Expression in other microbial systems
`5.5 Expression in plants
`5.6 Production in transgenic animals
`5.7 Expression in insect cells
`5.8 Production of monoclonal antibodies - cell culture
`5.9 Purification of monoclonal antibodies
`5.9.1 Purification of IgG
`5.9.2 Purification of IgM
`5.9.3 Purification of monoclonal antibody fragments
`5.9.4 Purification for therapeutic use
`
`6 Prospects for engineered antibodies in biotechnology
`
`6.1 Gene therapy
`6.1.1
`Intracellular antibodies
`6.1.2 Other applications of MAbs in gene therapy
`6.2 Applications of antibodies in plants
`6.3 Catalytic antibodies
`
`ix
`
`136
`137
`139
`142
`144
`146
`147
`147
`148
`149
`149
`150
`151
`151
`153
`
`153
`154
`154
`
`155
`158
`
`161
`161
`161
`162
`163
`167
`167
`168
`170
`172
`173
`174
`175
`175
`176
`176
`181
`181
`185
`
`187
`187
`187
`188
`189
`189
`
`HOSPIRA EX. 1029
`Page 8
`
`
`
`8 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`x
`6.6 Summary and prospects
`7 References
`
`6.4 Towards drug design
`6.5
`Improving affinity
`
`Index
`
`Contents
`
`190
`191
`
`192
`
`195
`
`241
`
`HOSPIRA EX. 1029
`Page 9
`
`
`
`9 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`1
`
`Preparation, Structure and function
`
`
`of Monoclonal Antibodies
`
`1.1
`
`introduction
`
`Antibodies are proteins produced by an individual in response to the presence of a foreign
`molecule in the body. These foreign molecules are known as antigens, and they usually
`result from invading organisms such as bacteria, fungi or viruses. Antibodies bind to
`antigens and elicit a range of effector mechanisms to destroy the invading organism.
`Therefore, the generation of an antibody response is a key step in the immune system
`which has evolved to protect individuals from invading pathogenic organisms. However,
`antibodies are not restricted in specificity to pathogens, but can be formed to a huge
`variety of antigens including proteins, carbohydrates and organic compounds, including
`totally novel structures.
`In 1 975 Kohler and Milstein described a method for the 'production of antibodies of
`predefined specificity'. This technical breakthrough allowed, for the first time, the pro­
`duction of antibody molecules of a single specificity which could be characterised and
`defined. Such monoclonal antibodies immediately became valuable research tools, and
`applications in the diagnosis and therapy of human disease began to be widely investig­
`ated. Monoclonal antibodies (MAbs) have become increasingly important as diagnostic
`agents allowing precise molecular structures to be mapped and analysed. However, initial
`enthusiasm for their development as therapeutic agents was premature and many prob­
`lems limited their use in humans (see Chapter 4). A second technical revolution has now
`arrived in the ability to manipulate antibody genes and to design and produce anti­
`body molecules tailor-made for their application. Such redesigned antibody molecules
`are now rapidly becoming valuable reagents for therapy of human diseases as well as
`improved diagnostics and research reagents. The scope of these applications, and how
`
`antibody molecules are redesigned, or engineered, in the most suitable form for a particular
`
`application is the subject of this book. To understand the nature of this process requires
`an understanding of antibody structure and function and of how antibody specificities are
`generated.
`
`HOSPIRA EX. 1 029
`Page 1 0
`
`
`
`10 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Applications and Engineering oj"Monoc!u11a! Antibodies
`2
`1.2 The role of antibodies
`The immune system can be considered in two paiis. Innate immunity is mediated by a
`Two major types of lymphocytes are involved in the adaptive immune response, T
`
`illll the immune [l'esporise
`
`variety of physical and biochemical barriers and by cells responding non-specifically to
`the foreign organism or molecule. Innate immunity is characterised by a similar reponse
`on re-exposure to the same foreign agent. Adaptive immunity, the second part of the im­
`mune system, is mediated by cells tenned lymphocytes and is characterised by improved
`efficiency on re-exposure to the same foreign agent. It is in this system that antibodies
`play a major role.
`
`lymphocytes and B lymphocytes. T lymphocytes can be subdivided into cytotoxic T
`lymphocytes and 'helper' T lymphocytes. Cytotoxic T lymphocytes bind to foreign or
`infected cells through a surface antigen receptor, the T cell receptor, and lysc them.
`Helper T lymphocytes play a regulatory role in controlling the response of both T and B
`lymphocytes. B lymphocytes exert their effect through producing antibody molecules
`which bind to the foreign agent and invoke specific mechanisms for its elimination.
`Antigen is recognised by B lymphocytes through the use of antibody molecules on the
`surface of the cell. Each B lymphocyte carries antibody molecules on its surface with a
`single specificity as a consequence of the rearrangement of immunoglobulin genes by the
`individual cell during its development in a random, antigen-independent process. Each
`individual will have many millions ofB cells at one time thus comprising a 'polyclonal'
`population. When an antigen enters the body it will be recognised by any B cell which has
`antibody molecules able to bind to antigenic determinants (epitopes) present on the anti­
`gen. This recognition leads to activation of the cell leading to proliferation and differ­
`
`entiation (Figure 1. 1 ).
`
`In the case of protein antigens, bound antigen molecules are internalised into the B cell
`and degraded into peptides. Some of these antigen peptides are then bound to major
`histocompatibility complex (MHC) class II molecules to fonn a complex which is trans­
`ported to the cell surface and 'presented' by the B cell. Although B cells are relatively
`weak at antigen presentation compared to other cell types such as dcndritic cells, these
`complexes can be recognised by receptors present on T lymphocytes, the T cell receptors.
`If appropriate recognition takes place then the T cell may deliver 'help' in the form of
`signals back to the B cell to stimulate proliferation and differentiation. This process
`therefore gives rise to a 'T-cell-dependent B-ccll response'.
`Proliferation results in a clone of identical cells which can differentiate to fom1 either
`plasma cells capable of secreting large amounts of soluble antibody of the same specificity
`as the original activated B cell, or memory cells which mount an accelerated immune
`response on re-exposure to the original antigen. The memory cell-mediated 'secondary'
`response results in the production of higher affinity antibodies due to hypermutation in
`the immunoglobulin gene loci followed by antigen driven selection, a process known as
`affinity maturation. This process of B cell clonal selection and the generation of antibody
`responses is covered in more detail in an excellent review (Raj cw sky, 1996).
`
`1.3 Structure and function of antibodies
`
`Antibody molecules have two principal functions, firstly to bind to antigen and secondly
`to trigger its elimination from the body. Antibodies are therefore 'adaptor' molecules
`which have both the ability to bind to the antigen molecule and the ability to bind and
`
`HOSPIRA EX. 1 029
`Page 1 1
`
`
`
`11 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Preparation, structure and jimction of monoclonal antibodies
`
`i3 eel!
`(\
`)
`(
`
`Cfo1Drn1B
`§elledh11n
`
`Antigen
`driven
`proliferntion
`
`0 M•mocy ""' 0
`
`Figure 1 .1 Differentiation of B cel ls to antibody-secreting plasma ce lls
`
`bring into action molecules of the effector system which can then remove the foreign
`material. For effective defence systems to operate it is essential that an individual is able
`to recognise a wide variety of foreign material and thus antibody molecules of many
`different binding specificities are required. However, in each case the antibody must also
`
`HOSPIRA EX. 1 029
`Page 1 2
`
`
`
`12 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`4
`
`Applications and Engineering of Monoclonal Anribodif:'s
`
`Cm
`Heavy chain
`Cm
`Figure 1.2 Representation of the organisation of protein chains of an lgG molecule
`
`be able to trigger the same effector systems. To achieve this the antibody molecule has
`evolved variable regions which vary in protein sequence and structure to accommodate
`the development of binding specificities to a wide range of different antigens, and con­
`stant regions which are largely the same in each antibody. The constant regions main­
`tain a common structure of the molecule and allow interaction with the effector systems
`such as complement binding and binding to Fe receptors on macrophages to activate
`phagocytosis.
`Antibody molecules are also known as immunoglobulins. The term 'immunoglobulin'
`applies to the antibody protein whether or not the binding specificity of the molecule is
`characterised, whereas an antibody is an antigen-specific immunoglobulin. In practice the
`two terms are usually used interchangeably.
`Higher mammals have five classes (isotypes) of immunoglobulin, termed IgG, IgM,
`IgA, IgE and IgD. The most abundant of these, which is used in most applications of anti­
`bodies, is the IgG class, the main class of antibody generated by the secondary immune
`response. The IgG molecule consists of four polypeptide chains, two heavy chains of
`approximately 50 kDa and two light chains of approximately 25 kDa (Figure 1.2). Each
`of these is divided into discretely folded structural domains of approximately 110 amino
`acids stabilised by an internal disulphide bond. These are linked together by short regions
`of comparatively flexible protein chain which allow movement of the domains relative to
`one another. Each light chain comprises a variable domain (VL) and a constant domain
`(CL), and each heavy chain a variable domain (VH) and three constant domains (CH!, CH2
`and Cm). The heavy and light chain variable domains associate to form the antigen­
`binding site. The IgG molecule thus has two antigen-binding sites and is capable of
`
`HOSPIRA EX. 1 029
`Page 1 3
`
`
`
`13 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`5
`
`Preparation, structure andjimction of mouoclonal antibodies
`
`F(ab')2
`
`figure 1.3 Domain structure of lgG, demonstrating the ant igen-binding F(ab')2 region and
`the Fe region with carbohydrate attached at the CH2 domain
`
`binding antigens divalently to allow high binding avidity (see Section 1.5). The heavy
`and light chains are linked by a disulphide bond between the CL and Cm domains. A
`flexible region, known as the hinge, links the CHI and CH2 domains of each heavy chain
`and it is at this point that the two heavy chains are linked by disulphide bonds. The CH2
`domain is normally glycosylated and the Cm domains of each heavy chain associate with
`each other by non-covalent interactions.
`Much early work on the structure and function of antibodies made extensive use of
`proteolysis of the antibody protein and this has led to commonly used terminology for the
`different regions of the antibody (Figure 1.3). The area around the antibody hinge is more
`susceptible to proteolysis than the tightly folded domains and thus this is the point at
`which proteolytic cleavage usually takes place. Proteolysis above the disulphide bonds in
`the hinge region results in monovalent Fab (fragment antigen-binding) fragments which
`comprise light chain together with the N-terminal two domains of the heavy chain. The
`Fab fragment thus contains one functional antigen binding site. Proteolysis immediately
`below one or more of the hinge disulphide bonds results in the divalent F(ab')2 fragment.
`The CH2 and Cm domains together make up the Fe fragment (fragment crystalline)
`which contains the sites for binding effector molecules. These fragments were the starting
`points for crystallographic determinations of antibody structure, as the flexibility of the
`intact IgG molecule prevented crystallisation until fairly recently. Thus while there are
`more than 50 structures for Fab fragments in the literature and several for hinge deleted
`immunoglobulins (reviewed by Padlan, 1994) there is only one fully defined structure of
`
`an intact IgG (Harris et al., 1997).
`
`Crystallographic structure determinations of many Fab fragments and several Fe frag­
`ments of antibodies have revealed that the folded domains have similar overall structures.
`This structure comprises two stacked �-sheets twisted into a characteristic fold, termed
`the immunoglobulin fold, and stabilised by a disulphide bond (Poljak et al., 1973). In the
`constant domains one sheet has three and the other four antiparallel beta strands, while in
`the variable domains there are nine strands.
`The primary sequences of several thousand antibody molecules are known, represent­
`ing the largest number of known sequences in one protein family. Analysis of these
`sequences has revealed that the variability of antibody variable domains is largely re­
`stricted to three 'hypervariable' regions in each of the heavy and light chain variable
`domains, with comparatively little variation in the intervening 'framework' regions (Kabat
`
`HOSPIRA EX. 1029
`Page 14
`
`
`
`14 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`6
`
`Applications and Engineering o/ Monoclonal Antibodies
`
`el al., 199 l ). These hypervariable regions, also known as complementarity determining
`regions, or CD Rs, form loops at the tip of the Fab structure making up an antigen-binding
`
`surface of approximately 2800 A". Structures of several Fab:antigen complexes have
`
`confirmed that antigen binding takes place at this surface; the hypervariability of these
`loops is therefore responsible for the different binding specificities of different antibody
`molecules. The CDR loops vary in both amino acid sequence and length between anti­
`bodies. This results in the variety of antigen-binding surfaces, some of which contain
`grooves or clefts in the surface to accommodate antigen binding whereas others are
`comparatively smooth. In some cases three-dimensional structures have been determined
`for antibody fragments with and without bound antigen which show that the antigen­
`binding site is not always completely rigid and may move to allow tight antigen binding
`
`may also lead to some confom1ational changes at sites in the antibody further away from
`
`There are four different subclasses (or isotypes) of IgG, called IgG 1, IgG2, IgG3 and
`IgG4 in humans and IgGl, IgG2a, IgG2b and IgG3 in mice. Although all within the IgG
`class, these subclasses vary to some extent in their structure and function. The subclasses
`have different heavy chains termed yl, y2, y3 and y4 in humans and each of these can use
`
`in an 'induced fit' mechanism (Bhat et al., 1990; Rini et al., I 992). In some cases this
`the antigen-binding region (Guddat et al., 1994).
`light chains from one of the two light chain isotypes, K or A. The structures of the
`
`subclasses vary in their pattern of disulphide bonding (Figure 1.4). The number of
`disulphide bonds between the heavy chains varies from two in human IgG 1 and IgG4 to
`15 in human IgG3. Also, the position of the disulphide bond between the heavy and light
`chains varies from between the CL and Cm region to between the CL and VH/CHl
`interdomain region.
`The antibody effector functions arc mediated through the constant regions. The effec­
`tor functions can be mediated through complement activation or by cellular interactions
`through specific Fe receptors expressed on a range of cell types, which can result in the
`process of antibody-dependent cellular cytotoxicity (A DCC). Complement activation via
`the classical complement cascade is initiated through binding of the IgG to the comple­
`ment component Clq. This is subject to conformational restraints which are only partially
`understood, and requires two molecules of IgG bound to an antigenic surface to bind C 1 q
`efficiently. The binding site for Clq on IgG has been known for some time to be localised
`to the Cm domain, with evidence that the N-terminal region of Cm is important for
`
`binding C 1 q in the case of human lgG I (Morgan et al. , 1995) and also that the C-tenninal
`region of Cm is required for efficient complement lysis (Tao et al., 1993; Greenwood
`et al., 1993). Mutagenesis studies have suggested that three amino acid residues on the
`are also required (Tao et al., I 993).
`
`surface of the Cm domain of murine IgG2b, numbers 318, 320 and 322, are involved in
`the binding interaction (Duncan and Winter, 1988), although these residues arc also
`present on antibodies which cannot activate complement, suggesting that these residues
`are necessary but not sufficient to activate complement. Mutagenesis studies with human
`IgG I have also suggested that C 1 q binding alone is not sufficient for complement activa­
`tion but that interactions of the IgG molecule with other steps in the complement pathway
`
`Fe receptors have been identified for all the classes of immunoglobulin, IgG, lgM,
`Ig D, IgE and lgA. Of these the best characterised are those for IgG and IgE. Three types
`of Fe receptors have been identified for IgG in humans; FcyRl (also known as CD64),
`FcyRlI (also known as CDw32), and FcyRlII (also known as CD16). These receptors are
`structurally related and distributed on various blood cell types. FcyRl is a high affinity
`receptor, capable of binding monomeric IgG, which plays a key role in ADCC and is
`
`HOSPIRA EX. 1 029
`Page 1 5
`
`
`
`15 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Human faG
`
`Preparation, structure and function of monoclonal antibodies
`
`7
`
`IgGl
`
`Mui-ine faG
`
`IgG2
`
`IgG3
`
`IgG4
`
`figure 1.4 Variation in the d isulphide bonds between heavy chains in lgG subclasses
`
`lgG2b
`
`l!gG3
`
`IgGl
`
`IgG2a
`
`found on macrophages, monocytes and neutrophils. FcyRII and FcyRIIl are low affinity
`receptors which bind aggregated IgG. FcyRII is found on most leukocytes including
`monocytes, macrophages and neutrophils and FcyRIII is found on neutrophils, macrophages
`and NK cells. These low affinity receptors are also capable of eliciting ADCC and
`phagocytosis. The binding of a single antibody to Fe receptor is reversible and does not
`elicit a response. If several antibodies are clustered together Fe receptor clustering will
`take place and a cellular response will be elicited (Figure 1 .5). All three Fey receptors
`recognise sites on the lower hinge/Cm domain of IgG but not in an identical fashion. In
`particular, the sequence of heavy chain residues at positions 234-237 is involved in the
`FcyRl receptor binding site (Duncan et al., 1 988; Sannay et al., 1 992).
`
`The ability to elicit effector functions varies between isotypes of IgG (Table 1 . 1 ), and
`
`this may reflect important differences in the functions of the individual isotypes. For
`example, human IgG 1 and IgG3 are highly active isotypes with respect to complement
`activation and elicitation of ADCC responses whereas IgG2 and IgG4 are relatively
`inactive, being only poorly able to activate complement through C 1 q binding with little
`binding to Fe receptors.
`The presence of carbohydrate attached to the CH2 domain has been shown to be
`required for both Fe receptor binding and complement activation (Tao and Monison,
`
`;\
`
`HOSPIRA EX. 1 029
`Page 1 6
`
`
`
`16 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`8
`
`Applications and Engineering of Monoclonal Antibodies
`
`Fe receptor bearing
`
`effector cell
`
`\. _____ ___,...)
`i Signalling
`-.
`Fe receptor clustering
`�
`figure 1.5 Fe receptor c lustering due to mu ltiple antibody molecules bind ing to antigen
`
`CeUufa1r response§
`
`(phagocytosis, exocytosis)
`
`leading to cellu lar responses
`
`Table 1 .1
`
`Effector fu nctions of hu man and mouse JgG subclasses
`
`Human IgG
`
`IgGI
`
`+++
`+++
`++
`++
`
`lgG2
`
`+
`+
`
`IgG4
`
`+
`
`lgG3
`
`++
`+++
`++
`++
`
`Complement
`activation
`
`Human
`FqRl (CD64)
`
`Human
`FcyRll (CDw32)
`
`Human
`FcyRlII (CD 16)
`
`Mouse IgG
`
`IgGI
`
`+
`+
`
`IgG2a
`
`++
`++
`+
`+
`
`IgG2b
`
`+++
`+
`
`T
`
`IgG3
`
`++
`++
`++
`
`HOSPIRA EX. 1 029
`Page 1 7
`
`
`
`17 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Table 1.2 Human immunoglobul ins
`
`Preparation, structure and fimction of monoclonal antibodies
`
`Antibody
`
`Approximate
`molecular weight
`(kDa)
`
`Heavy chain
`
`IgGl
`IgG2
`IgG3
`IgG4
`IgM
`IgAl
`IgA2
`IgD
`IgE
`
`146
`146
`170
`1 46
`970
`1 60
`160
`200
`200
`
`yl
`y2
`y3
`y4
`µ
`
`al
`a2
`
`8
`€
`
`9
`
`in normal human serum
`
`Approximate concentration
`
`(mg/ml)
`
`9
`
`0.5
`1 .5
`
`0.5
`0.03
`0.000 1
`
`1989). The role of the carbohydrate is not fully understood but it is believed to be
`important in maintaining the tertiary structure and disposition of the Cm domains. IgG
`without carbohydrate has been shown to have an altered conformation in the lower hinge
`region and may result in disruption of the interaction sites (Lund et al., 1990).
`IgG represents approximately 70-75% of the total immunoglobulin in human serum,
`the remainder being molecules of the other classes mentioned above, IgM, IgA, IgD and
`IgE. Although the IgG structure is a good general model for antibody structure, these
`immunoglobulin classes differ in their structure, reflecting differences in their normal
`
`functions. The different classes of immunoglobulin use light chains of the same type (K
`or A.) and thus differ predominantly in their heavy chains (Table 1.2).
`
`IgM is the predominant antibody raised in primary responses to many antigens and
`represents approximately 10% of the immunoglobulin in human serum. IgM antibodies
`have a pentameric structure of the basic four chain unit such that a total of ten antigen­
`
`binding sites per molecule are present (Figure 1.6). The µ heavy chain does not have a
`than the y heavy chain, with sugar attached at five glycosylation sites. Although it is
`
`hinge region but has an extra constant region domain, which is inserted between the Cm
`and Cm domains in the analogous IgG structure, such that the CH4 domain of IgM is
`analogous to the Cm domain of IgG. There is also an extra tail of 19 amino acids on the
`heavy chain involved in polymer assembly. The heavy chains are held together by
`disulphide bonds between Cm domains and there is also an extra polypeptide chain, the
`J chain, whose role is not fully understood although it may assist the process of assembly
`
`of the pentameric IgM molecule. The µ heavy chain is also more heavily glycosylated
`
`presented as a pentameric structure in Figure 1.6, this is not always the case and the
`occurrence of IgM molecules with different numbers of binding units is now well estab­
`lished. The production of hexamers and tetramers has been reported (Eskeland and
`Christensen, 1975) and in some cases hexamers may be a predominant structure (Cattaneo
`and Neuberger, 1987; Davis et al., 1988). This multivalent structure provides the IgM
`molecule with a high binding avidity, to help compensate for the relatively low affinity of
`each binding site. Another consequence of the multivalent structure is the ability of one
`molecule of IgM to activate complement, which is thought to occur via binding to the
`IgM Cm domain. Therefore IgM molecules are well adapted for their role in the primary
`immune response.
`
`HOSPIRA EX. 1 029
`Page 1 8
`
`
`
`18 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`Applications and Engineering of Monoclonal Antibodies
`
`10
`
`IgM
`
`IgG
`
`Serum IgA
`
`IgE
`
`IgD
`
`Secretory IgA
`
`figure 1.6 Organisation of prote in chains in the immunoglobu lin classes
`
`There are two isotypes of IgA present in humans (IgAI and IgA2) which together
`make up 15-20% of the total serum immunoglobulin. In addition, a form of IgA, termed
`secretory IgA or sigA, is the predominant immunoglobulin present in external secretions
`such as saliva, milk, tears, genitourinary secretions and tracheobronchial secretions (Tomasi,
`1992). The mucosa! surfaces bathed in these secretions are a major site of exposure to the
`environment and thus secretory IgA is an important defence mechanism against invading
`organisms. In terms of quantity, sigA is the major type of lg produced, with estimates of
`2g per day being produced in humans, more than other Ig forms. The IgA present in
`serum is largely monomeric and consists of two heavy chains and two light chains
`assembled in a similar manner to IgG. IgAl is the predominant isotype, making up
`approximately 90% of the serum IgA. IgA heavy chains have three constant domains
`with IgA 1 having a longer hinge region than IgA2. In addition, IgA heavy chains also
`
`HOSPIRA EX. 1 029
`Page 1 9
`
`
`
`19 of 254
`
`Celltrion, Inc. 1195
`Celltrion v. Genentech
`IPR2017-01374
`
`

`

`11
`have a similar tail to fgM which allows polymerisation, and binding of the J chain to the
`a: chain, through a C-tenninal cysteine residue in a similar manner to IgM. sigA exists
`
`Preparation, structure andf!!nction of monoclonal antibodies
`
`predominantly in dimeric form with a molecular weight of approximately 385,000. slgA
`also contains a secretory component which is an extra polypeptide chain of approxim­
`ately 70 kDa, synthesised by epithelial cells and not the plasma cells. The secretory
`component is part of the receptor involved in transport of IgA into mucosa! secretions
`which is cleaved during transport across endothelial surfaces to release the 70 kDa sec­
`retory component. This then becomes bound to the lgA by disulphide bonds (Fallgreen­
`Gebauer et al., 1993). slgA is more stable to proteolytic attack than the