Drug Discovery and Design
`
`433
`
`The role of therapeutic antibodies in drug
`discovery
`
`L.H. Stockwin and 5. Holmes 1
`Therapeutic Antibody Division, Oxford Glycosciences [UK) Ltd., The Forum, 86 Milton Park, Abingdon, Oxfordshire OX14 4RY, U.K.
`
`Abstract
`The last 5 years have seen a major upturn in the fortune of therapeutic monoclonal antibodies (mAbs),
`with nine mAbs approved for clinical use during this period and more than 70 now in clinical trials beyond
`phase II. Sales are expected to reach $4 billion per annum worldwide in 2002 and $15 billion by 2010.
`This success can be related to the engineering of mouse mAbs into mouse/human chimaeric antibodies or
`humanized antibodies, which have had a major effect on immunogenicity, effector function and half-life.
`The issue of repeated antibody dosing at high levels with limited toxicity was essential for successful clinical
`applications. Emerging technologies (phage display, human antibody-engineered mice) have created a
`vast range of novel, antibody-based therapeutics, which specifically target clinical biomarkers of disease.
`Modified recombinant antibodies have been designed to be more cytotoxic (toxin delivery), to enhance
`effector functions (bivalent mAbs) and to be fused with enzymes for prodrug therapy and cancer treatment.
`Antibody fragments have also been engineered to retain specificity and have increased the penetrability
`of solid tumours (single-chain variable fragments). Radiolabelling of antibodies has now been shown to be
`effective for cancer imaging and targeting. This article focuses on developments in the design and clinical
`use of recombinant antibodies for cancer therapy.
`
`Introduction
`Monoclonal antibodies (mAbs) have revolutionized bio(cid:173)
`logical research and clinical diagnostics. However, any
`suggestions that mAbs could ever realize potential as anti(cid:173)
`cancer therapeutics would, until recently, have been met
`with a degree of scepticism. The first (rodent) mAbs were
`potent immunogens provoking strong endogenous antibody
`responses. They were also unable to trigger effectorfunctions,
`had a short half-life (1-2days) and proved to be extremely
`expensive to manufacture. Consequently, several promising
`mAb-based cancer therapies were relegated to the realm of
`pre-clinical development.
`By contrast, the situation in 2002 sees mAbs undergoing
`what can best be described as a renaissance with sales of mAbs
`expected to reach $4 billion in 2002, and $15 billion by 2010.
`The major factors underpinning this shift are advances in
`both antibody engineering and discovery technologies, to
`the extent that it is now possible to rapidly derive high(cid:173)
`affinity humanized/human mAbs. To date, 12 mAbs have
`been approved by the regulatory authorities and of the > 450
`currently in clinical trials around 70 have progressed beyond
`phase II. In this review we discuss the design and clinical
`application of recombinant mAbs in cancer therapy with
`reference to several precedents.
`
`Key words: antibody, cancer, discovery, monoclonal, therapeutic
`Abbreviations used: EGF, epidermal growth factor; mAb, monoclonal antibody; scfv, single(cid:173)
`chain variable fragment
`'To whom correspondence should be addressed (e-mail Steve.Holmes@ogs.co.uk).
`
`Antibodies by design: reducing
`immunogenicity
`One of the fundamental problems with mouse mAbs is
`that administration of murine Ig induces a human anti(cid:173)
`mouse response in about 50% of patients after a single
`dose and >90% after repeated administrations. This response
`leads to rapid clearance, allergic reactions and complications
`relating to hypersensitivity [1]. Three general strategies
`have now been developed to reduce the immunogenicity
`of mAbs. Firstly, advances in molecular biology have made
`it possible to substitute human sequences for their rodent
`counterparts. Early attempts produced chimaeric mAbs with
`75% homology with human mAbs by retaining the rodent V
`genes linked to sequences encoding human constant regions.
`This homology was then improved to around 95% by
`Winter and colleagues by a process called V-region human(cid:173)
`ization where only the rodent complementarity-determining
`region is retained combined with human V-region frame(cid:173)
`works and constant regions of the heavy and light chain [2].
`Examples of chimaeric and humanized mAbs are Rituxan
`(anti-CD20 mAb; approved for the treatment of B-cell ma(cid:173)
`lignancies) and Synagis (approved for the treatment of Rous
`sarcoma virus infection), respectively. The next technology,
`phage display, involved selecting high-affinity recombi(cid:173)
`nant antibody fragments using libraries (> 109 members)
`of human antibody V regions presented on the surface of bac(cid:173)
`teriophage [3], thereby bypassing immunization strategies.
`Human V-region antibody fragments can be selected by
`several rounds of selection on antigen and then screened
`for activity, generating fully human single-chain variable
`fragment (scFv) or Fab [4,5]. A large number of mAbs
`
`~2003 Biochemical Society
`
`EX2107
`Eli Lilly & Co. v. Teva Pharms. Int'l GMBH
`IPR2018-01710
`
`1
`
`

`

`434
`
`Biochemical Society Transactions (2003) Volume 31, part 2
`
`currently in clinical development have been generated using
`this approach; one example is an anti-(transforming growth
`factor {32) mAb for the treatment of fibrotic disorders [6].
`Finally, transgenic animals have been generated [7,8] in which
`natural Ig genes arc deleted and replaced with the human
`loci necessary for the production of IgG!IgM. This in vivo
`approach has been used to generate several mAbs, including
`ABX-EGF (Abgcnix Corp.), a high-affinity anti-[epidermal
`growth factor (EGF) receptor] antibody for use in the
`treatment of EGF-responsive tumours [9]. In summary, these
`techniques have reduced the immunogcnicity of mAbs to
`an extent where repeated doses can be administered without
`impaired efficacy.
`
`Antibodies by design: the right tools
`for the job
`A sound understanding of the biochemical pathway being
`targeted is central to the 'wish list' for a mAb therapeutic. In
`many applications the recruitment of host effector functions
`through Fe receptors (including complement fixation and
`antibody-dependent cell-mediated cytotoxicity) is essential.
`In humans the IgG 1 isotypc is the preferred therapeutic
`choice for triggering effector cascades. An approved antibody
`with the potential to trigger effector function is Herccptin,
`an anti-HER2/neu antibody used in the treatment of breast
`cancer. In addition to the well-described ability to block
`Her2-dependcnt growth, some of the clinical benefit from
`Herccptin treatment is thought to arise from an ability
`to promote antibody-dependc'nt cell-mediated cytotoxicity
`[10]. Conversely, when target neutralization is the only goal,
`it may be preferable to usc the IgG4 isotypc that is incapable
`of triggering these cascades [11]. Irrespective of isotype there
`arc several additional benefits to the IgG format. Divalent IgG
`molecules have the potential to cross-link cell-surface antigen
`in the absence of macrophage/natural killer cell Fe receptors.
`For example, in some CD20-positivc cell lines cross-linking
`has been shown to induce apoptosis [12]. An alternative
`to using this is to usc antibody fragments such as Fab or
`scFv. Both of these formats arc unable to trigger effector
`function and because of a reduced size have an increased
`tumour penetration. These fragments arc especially suited to
`the delivery of 'payload' (see below). One obvious drawback
`of Fabs/scFvs is a reduced half-life (hours compared with 2-
`3 weeks with IgG). It is, however, possible to improve the
`serum longevity of these proteins by conjugating to inert
`polymers such as poly(cthylcnc)glycols [13]. Examples of
`cancer targets where neutralization is the primary goal include
`matrix metalloproteascs ('MMPs'), urokinase plasminogen
`activator ('uPA') and vascular endothelial growth factor
`('VEGF'), which arc all secreted proteins associated with
`tumour progression.
`
`Antibody conjugates: adding insult
`to injury
`I~ was Paul Erlich in 1900 who first coined the phrase 'magic
`bullet', accurately predicting the value of antibodies as smart
`
`"2003 Biochemical Society
`
`weapons in the delivery of destructive payload [14]. A
`number of approaches have been developed, which in the
`context of anti-cancer protocols provide measurable im(cid:173)
`provements
`in cell killing. The major categories arc
`radioisotopes, protein toxin-enzyme fusions and small(cid:173)
`molecule conjugates.
`Antibodies arc routinely used to concentrate doses of
`radiation in tissues for both therapeutic and diagnostic pur(cid:173)
`poses. Common isotopes used to this end include iodine-131 ,
`yttrium-90, indium-111 and tcchnicium-99. Tumour killing
`by unlabelled mAbs is limited by the degree of antigen
`density on the tumour cell and the ability to penetrate
`tumours adequately. Although radiolabellcd mAbs may be
`less restricted by antigen density in their efficacy they
`can gain an advantage by a 'bystander' effect in killing
`antigen-negative tumour cells. Conversely, this phenomenon
`would also be responsible for non-specific toxicity. The two
`most extensively studied radiolabelled mAbs are Zcvalin
`90Y-labclled anti-CD20) and Bexxar (131 I-labelled anti(cid:173)
`(
`CD20), the former receiving recent U.S. Food and Drug
`Administration (FDA) approval for the treatment of non(cid:173)
`Hodgkins lymphoma. For therapeutic purposes 90Y -labelled
`mAbs may be better debulking agents for larger tumours
`because of the increased path length of the emission compared
`with 131 I -labelled mAbs, which may be preferable for
`targeting post-therapy minimal disease [15].
`Protein toxins are a large group derived from a variety of
`microbial, plant and human sources. Above all other mAb
`conjugates, these toxins can be engineered directly into the
`antibody-constant regions, significantly reducing the manu(cid:173)
`facturing costs. Examples include Pseudomonas exotoxin,
`which when conjugated to anti-CD22 has been shown to
`dramatically increase cell killing [16]. For plant toxins, de(cid:173)
`glycosylated ricin a-chain has the longest and most successful
`history. In one recent report a ricin-CD19 conjugate w1s
`shown to confer a large increase in the ability of antibody to
`kill malignant B-cells [17]. The major foreseeable problem
`with protein toxins
`is
`their potential immunogenicity,
`especially when considering microbial toxins to which an
`individual may already have been sensitized. An alternatiYe
`would be to usc cytotoxic human proteins. Angiogcnin, a
`human RNase, has been shown to induce apoptosis when
`delivered into the cytoplasm. In one recent publication,
`bacterially expressed CD30L-angiogenin fusion protein was
`found to be capable of killing a wide range of CD30+
`Hodgkin-derived cell lines [18]. This toxin may represent
`a clever way to avoid host immune responses.
`The third payload group comprises toxic small molecules
`which arc usually DNA-complexing agents or inhibitors
`of the cell cycle. In this situation the antibody conjugate
`is internalized and the toxic drug liberated after cleav1gc
`of a pH- or enzyme-sensitive linker. As mAbs can target
`chemotherapy exclusively to cancer cells, more p tent
`chemotherapy can be used when attached to roAbs than
`when administered systemically, for example mayta.nsinc
`conjugates [19]. mall toxic drug molecules have potc.ntial
`advantages: in general they have a negligible immunogenicit)'
`
`2
`
`

`

`Drug Discovery and Design
`
`435
`
`Figure 1 I Mechanisms of action for therapeutic antibodies
`The therapeutic potential of mAbs resides in an ability to modulate
`several different pathways. (A) For a receptor involved in a growth(cid:173)
`promoting pathway, mAb binding may prevent ligand binding and signal
`transduction (e.g. mAbs against EGF receptors and Her2/c-erb-B2). In a
`similar fashion, mAbs may target cell-surface growth factor receptors
`tor degradation as opposed to a natural recycling process (Herceptin
`anti-Her2). (B) Cell-surface cross-linking may initiate a cascade that
`kills
`tumour cells
`(e.g. Rituxan-mediated cross-linking of CD20).
`(C) several mAb formats have the capacity to activate the classical (Clq)
`complement pathway. (D) Macrophages and natural killer cells express
`cell-surface Fe receptors which, when bound to immobilized mAb,
`become activated and secrete cytotoxic mediators. (E) Finally, mAbs
`can be engineered to deliver toxic payloads including radionuclides,
`protein toxins and small molecules (Bexxar, 131 1-labelled anti-CD20; and
`Mylotarg Zogamicin anti-CD33)
`
`y
`
`E
`
`and compared with radi onuclides arc rchtivcly easy to
`handle. Mylotarg is a humani zed anti-CDJJ linked to
`calichcamicin. This useful therapy for acute myelogenous
`lcukncmi'l ('AML') was the first toxin conju gnte ever to be
`nppmved showing that the s mall- molecule l?ayload group
`ha
`just as much validi ry as other regimes. Recruitment
`of effector cell s u sing bispecific antibodies (one pecificiry
`directed towards the tumour, the other to the effector cell) has
`nlso been rep orted wide! y [20]. T hese rcagems nrc expected to
`be difficult tO manufacture inlnrge quantLtics but hnve shown
`Promise in vitro, in animal models and now also in some
`
`phase I clinical trials. Mechanisms of action for conjugated
`and naked antibodies are illustrated in Figure 1.
`
`Problems, solutions and evolution
`We can infer from the previous discussion that the modern
`antibody investigator has the technological tool kit necessary
`to ensure the smooth transition of a project from antigen
`discovery to therapeutic mAb. What then is the major
`bottleneck that still exists in the discovery process? The
`completion of the human genome combined with advances
`in protcomics technologies have helped to enhance our
`understanding of the complex interplay between genetic,
`transcriptional and translational alterations in human cancers.
`Although bioinformaticians have made strides in identifying
`potentially interesting novel cancer targets there is a central
`bottleneck at the point when cancer biologists must invest(cid:173)
`igate each individual gene for therapeutic potential. These
`assays are time-consuming and labour-intensive. Looking at
`the current crop of approved antibody targets, they have
`all benefited from around 10-20 years of detailed academic
`research. The future goal will be to design high-throughput
`biology modes to move novel cancer targets quickly towards
`clinical development.
`Insight into the future of antibody therapeutics can
`probably be glimpsed through technologies such as ribosome
`display. Here, antibody-fragment libraries arc not displayed
`on bacteriophage but arc produced entirely in vitro and
`there is also the potential to introduce mutagenesis steps
`into the antibody-encoding RNA sequences. This system
`has the advantages of greater library size(> 10 12
`), potentially
`higher mAb affinity and improved speed compared with
`standard phage display [21]. If this technique tells us anything
`about the future of antibody discovery it suggests that soon
`the generation of humanized/human mAbs may become as
`routine as PCR is today: we await the release of the first
`human antibody-generation kit!
`
`Conclusions
`Kohler and Milsteins' dream of mAbs as exquisitely sensitive
`therapeutics has finally been realized. Two antibodies,
`Herceptin and Rituxan, have proved that this class of drugs
`can be effective (and highly lucrative) anti-cancer agents
`for those companies brave enough to enter this research
`'graveyard'. From a stalled start, mAbs now represent
`around 25% of the novel biological entities entering clinical
`trials, indicating that the biotechnology community at large
`has finally recognized the speed and efficiency of the
`antibody platform. It is also becoming clear (and it is ironic)
`that the classical monoclonal antibody may yet become
`the commercial saviour of the high-tech proteomics and
`gcnomics revolutions.
`
`References
`1 Shawler, D L, Bartholomew, R M, Smith, L M and Dillman, R 0 (1985)
`J lmmunol 135, 1530-1 535
`
`<1:;2003 Biochemical Society
`
`3
`
`

`

`436
`
`Biochemical Society Transactions (2003) Volume 31, part 2
`
`2 Queen, c., Schneider, W.P. and Selick, H. (1989) Proc. Natl. Acad.
`Sci. U.S.A. 86, 10029-10033
`3 Winter, G., Griffiths, A D., Hawkins, R E. and Hoogenboom, H.R. (1994)
`Annu. Rev. lmmunol 12, 433-455
`4 Marks, J.D., Hoogenboom, H.R., Bonnert, T.P., McCafferty, J., Griffiths, A.D.
`- and Winter, G. (1991) J. Mol. Bioi 222, 581-597
`5 Griffiths, A.D. (1993) Curr. Opin. rmmunol. 5, 263-267
`6 Thompson, J.E., Vaughan, T.J and Williams, A.J. (1999)
`J. lmmunol. Methods 227, 17-29
`7 Mendez, M.J., Green, L.L., Corvafan, J R., jia, x.c, Maynard-Currie, C.E .
`Yang, X.D., Gallo, M.L., Louie, D.M., Lee, D.V., Erickson, K.L. et al. (1997)
`Nat. Genet. 15, 146-156 ·
`8 Ishida, 1., Tomizuka, K., Yoshida, H., Tahara, T, Takahashi, N., Ohguma, A.,
`Tanaka, s., Umehashi, M., Maeda, H., Nozaki, c. et al. (2002) Cloning
`Stem Cells 4, 91-102
`9 Yang, X. D., Jia, X C., Corvalan, J.R., Wang, P. and Davis, C G. (2001)
`Crit. Rev. Oncol. Hematol. 38, 17-23
`10 Baselga, J and Albanell, J (2001) Ann. Oncol. suppl. 1,
`S35-S41
`11 Antoniw, P., Farnsworth, A.P., Turner, A., Haines, AM., Mountain, A.,
`Mackintosh, J., Shochat, D., Humm, J, Welt, s., Old, L.J et al (1996)
`Br. J. Cancer 74, 513-524
`
`12 Shan, D., Ledbetter, J.A and Press, O.W. (2000) cancer lmmunol.
`lmmunother. 48, 673-683
`13 Lee, L.S , Conover, C., Shi, C., Whitlow, M and Filpula, D. (1999)
`Bioconj. Chem. 10, 973-981
`14 Erlich, P. (1900) Proc R. Soc. London 66, 424
`15 Zelenetz, A D. (1999) Curr. Opin. Oncol. 11, 375-380
`16 Mansfield, E., Amiot, P., Pastan, I. and FitzGerald, D.J (1997) Blood 90
`2020-2026
`•
`17 Conry, R M, Khazaeli, M.B., Saleh, M.N., Ghetie, V, Vitetta, E.S., Liu, T
`and LoBuglio, A.F. (1995) J. lmmunother. Emphasis Tumor lmmunol
`18, 231-241
`18 Huhn, M., Sasse, s., Tur, M.K., Matthey, B., Schinkothe, T, Rybak, S.M,
`Barth, s. and Engert, A. (2001) Cancer Res. 61, 8737-8742
`19 Smith, s. (2001) Curr. Opin. Mol. Ther. 3, 198-203
`20 Peipp, M. and Valerius, T. (2002) Biochem. Soc. Trans. 30,
`507-511
`21 Hanes, J., Schaffitzel, C., Knappik, A. and Pluckthun, A (2000)
`Nat Biotechnol. 18, 1287-1292
`
`Received 20 November 2002
`
`C>2003 Biochemical Society
`
`4
`
`

`

`Honorary Editor
`john Wrigglesworth (London)
`
`- Subscription rates I Subscription rates for volume 31, parts 1-6
`(2003) paper and online:
`
`Publisher
`Pauline Starley
`
`Production Editor
`Michael Cunningham
`
`Staff Editor
`Mark Thorne
`
`,, General policy I Biochemical Society Transactions (ISSN 0300-
`5127 for print/online journal; ISSN 1470-8752 for online only
`journal) is published by Portland Press Ltd, the not-for-profit
`publishing subsidiary of the Biochemical Society, and captures the
`exciting science presented at the Biochemical Society meetings.
`The journal contains prestigious medal lectures presented in the
`form of authoritative reviews of the medallists' research areas, the
`published colloquia in the format of a series of 'mini-review'-type
`articles, and provides a timely snapshot of the topics chosen as
`the basis of various colloquia. Submitted primary research papers
`are not published and authors are reminded that detailed
`presentation of new experimental data may prejudice its
`subsequent publication in precisely the same format elsewhere.
`Articles are published subject to editorial acceptance and authors
`will be issued with instructions on the format of their article
`before the meeting.
`
`,, Biochemical Society Transactions Online I www.biochemsoctrans.org
`
`:o The Biochemical Society i For
`information about the
`advantages of membership of the Biochemical Society, contact the
`Executive Secretary, The Biochemical Society, 59 Portland Place,
`London W1 B 1 QW, U.K.
`(telephone : 020 7580 5530;
`e-mail: genadmin@biochemistry.org).
`
`' Editorial and publishing office 1 Biochemical Society
`Transactions, 59 Portland Place, London W1 B 1 QW, U.K .
`(telephone: 020 7637 5873;
`fax: 020 7323 1136;
`e-mail: editorial@portlandpress.com).
`
`Biochemical Society Transactions cj o Mercury Airfreight
`International Ltd, 365 Blair Road, Avenel, NJ 07001, U.S.A.
`
`' Design I Angie Moyes, www.fuz.co.uk
`
`I Display
`~' Advertising
`accepted.
`is
`advertising
`Contact PRN Media, Suite 8 and 9, Griggs, West Street,
`Coggeshall, Essex C06 1 NT
`(telephone: 01376 563811;
`fax : 01376 562453; e-mail: sales@prnmediasales.co.uk).
`
`' Periodicals I postage paid at Rahway, NJ.
`
`~1 Postmaster I send address corrections to: Biochemical Society
`Transactions, cj o Mercury Airfreight International Ltd, 365 Blair
`Road, Avenel, NJ 07001.
`
`Volume 31
`- - - -
`Paper and Online
`Institutional
`Online only
`Institutional
`Paper only
`Institutional
`Single issues
`
`North America Rest of world Euro Price
`
`$355.00
`
`£208.00
`
`€354 .00
`
`$327.00
`
`£191.00
`
`€326.00
`
`$344.00
`$90.00
`
`£203.00
`£55.00
`
`€332 .00
`€92.00
`
`VAT is due on some supplies to subscribers within the U.K. and to
`non-registered subscribers elsewhere in the EU . If in the EU,
`please state your VAT number otherwise VAT may be charged at
`the national rate. Subscribers to the Biochemical journal may take
`a joint subscription to that journal and to Biochemical Society
`Transactions, saving $30 (North America), €28 .50 (EU) or £17
`(rest of world), or
`to Biochemical Society Transactions,
`Biochemical
`journal and Biotechnology and Applied
`Biochemistry, saving $40 (North America), €39
`(EU) or £23 (rest
`of word) . Overseas rates include air freight delivery. Priority
`delivery is available at extra cost (please ask for details). Back
`issues are available.
`
`.:;; Electronic back archive I Free to all users with Internet access.
`
`•' Subscription department ! Portland Customer Services,
`P.O. Box 32, Commerce Way, Colchester C02 8HP, U.K.
`(telephone:
`01206
`796351;
`fa x:
`01206
`799331;
`e-mail: sales@portland-services.com).
`
`All rights reserved I Apart from any fair dealing for the
`purposes of research or private study, or criticism or review, as
`permitted under the Copyright, Designs and Patents Act, 1988,
`this publication may be reproduced, stored or transmitted, in any
`form or by any means, only with the prior permission granted by
`the copyright clearing houses such as the Copyright Licensing
`Agency (U.K.) or the Copyright Clearance Centre (U.S.A.). The CCC
`item-fee code for this publication is 0300-5127 j 2003/
`$30.00+0.00. Inquiries concerning reproduction outside those
`terms should be sent to the publishers at the London address.
`
`Although, at the time of going to press, the information contained
`in this publication is believed to be correct, neither the authors
`nor the publisher nor the editor assume any responsibility for any
`errors or omissions herein contained.
`
`'> Cover illustration I THP-1
`macrophages treated with
`acetylated low-density
`lipoprotein, stained with
`haematoxylin and eosin, and
`counter-stained with Oil Red 0 to
`detect accumulated lipid. Image
`kindly supplied by Marcel van
`Duin (Organon Research Laboratories).
`
`5
`
`

`

`_,.,./ Dctrans.org
`
`/
`
`HSL
`
`Genes in food- why the furore?
`
`Sulphotransferases in glycobiology
`
`Biological functions of sulphated
`glycoproteins
`
`·
`
`Proteasome interactions with viral and · · · ·
`cellular proteins
`
`·;t:ALTH SCIE~!CES LIBRAR ·.
`iJNIVERSITY OF V·/13 CONSit'\
`
`APR 2 2 L: ~1J3
`
`i305 Linden Drive
`~~adison. WI 5370R
`
`6
`
`