FEATURE
`
`Monoclonal antibody successes in the clinic
`
`Janice M Reichert, Clark J Rosensweig, Laura B Faden & Matthew C Dewitz
`
`Most monoclonal antibodies in clinical trials are owned by small biotech companies. But with blockbuster-sized
`revenues and approval rates higher than those for small-molecule drugs, that all may be set to change.
`
`From somewhat inauspicious beginnings in
`the 1980s, therapeutic monoclonal antibod(cid:173)
`ies (mAbs} have developed into a beneficial
`and profitable group of products. Monoclonal
`antibodies now comprise the majority of
`recombinant proteins currently in the clinic,
`with more than 150 products in studies
`spoqsored by companies located worldwide.
`Starting with ideas generated in academia,
`biotech companies pioneered the technologies
`and techniques to produce therapeutic mAbs,
`persevered despite failures and are now being
`rewarded. Major pharmaceutical firms were
`initially reluctant to adopt the new technolo(cid:173)
`gies, but most now have one or more mAbs
`in clinical study. As a group, genetically engi(cid:173)
`neered mAbs generally have higher probabili(cid:173)
`ties of approval success than small-molecule
`drugs, and so are useful for diversification of
`the therapeutics pipeline.
`We previously reported on trends in the
`development of therapeutic mAbs sponsored by
`US-based companies1•2 and trends for products
`from all companies as of 2003 (ref. 3). Because
`of the dynamic nature of the clinical develop(cid:173)
`ment process, we have continued to accumulate
`data (for example, initiation of clinical study
`of new mAbs, changes in clinical-phase status,
`and terminations) for these products (Box I).
`Analysis of the current data set indicates trends
`toward the study of human mAbs and mAb
`fragments. In addition, results presented here
`verify our previous findings that approval suc(cid:173)
`cess rates for chimeric and humanized mAbs
`are consistently in the 18-29% range.
`
`Janice M. Reichert and Laura B. Faden are at
`Tufts University Center for the Study of Drug
`Development (CSDD), 192 South Street, Suite
`550, Boston, Massachusetts 02111, USA. Clark
`J. Rosensweig and Matthew C. Dewitz were
`summer interns at CSDD.
`e-mail: janice.reichert@tufts.edu
`
`Hybridoma production methods, shown here, are increasingl,y being replaced by recombinant
`production in mammalian cells as more humanized and human mAbs enter the clinic.
`
`Monoclonal technology evolution
`The original method for production of mAbs
`involved fusing mouse lymphocyte and
`myeloma cells, yielding murine hybridomas
`(see Box2 for glossary of terms). The first report
`describing the preparation of hybridomas
`came from the UK's Medical Research Council
`Laboratory of Molecular Biology ( Cambridge)
`in 1975 (ref.4). However, the impact of the new
`method was not immediate5. The initial proce(cid:173)
`dure was unreliable; as protocols were improved
`and materials were shared among laboratories,
`the technology filtered into the scientific com(cid:173)
`munity. Notably, the hybridoma technique was
`never patented (seep. 1047).
`Therapeutic murine mAbs entered clini(cid:173)
`cal study in the early 1980s, but problems
`with lack of efficacy and rapid clearance due
`to patients' production of human anti-mouse
`
`antibodies (HAMA) became apparent. These
`issues became driving forces for the evolution
`of mAb production technology: the poten(cid:173)
`tial utility of therapeutic mAbs was obvious,
`but the initial execution was unsatisfactory.
`Attempts were made to increase efficacy and
`decrease immunogenicity through two paral(cid:173)
`lel paths: production of mAb chimeras derived
`from both human and mouse DNA, and pro(cid:173)
`duction of fully human mAbs. In the case of
`chimeric mAbs, genetic engineering techniques
`were used to replace the murine Fe region with
`one of human sequence6•7. The second avenue
`involved applying immortalization methods
`and hybridoma techniques to human cells8•9.
`Surprisingly, the chimeric mAbs have
`proved to be superior to the early human
`mAbs-replacement of the murine Fe region
`was sufficient for improving efficacy and
`
`NATURE BIOTECHNOLOGY VOLUME 23 NUMBER 9 SEPTEMBER 2005
`
`1073
`
`

`

`FEATURE
`
`Box 1 Anal
`
`1s criteria
`
`Since it was founded in 1976, Tufts CSDD has collected data
`on the development and approval of therapeutics and vaccines.
`Data for mAbs were collected by surveys of pharmaceutical
`and biotechnology firms, on company web sites and from
`public documents (such as press releases and annual reports).
`Commercially available databases (11Ddb3, I MS R&D Focus, and
`Pharma Projects) were accessed to verify the status of products
`currently in clinical study. Products in phase 1/2 were assigned
`to phase 2 and products in phase 2/3 were assigned to phase 3.
`Data were updated with all changes noted through June 2005.
`The mAb data set contained 355 therapeutic products in
`clinical study sponsored by more than 100 commercial firms
`located worldwide. Of the 355 products, 152 mAbs are currently
`in clinical study, with 63 in phase 1, 74 in phase 2 and 15
`in phase 3. Products developed in-house or licensed from any
`source were included. The years in which clinical studies were
`initiated were not available for nine (2.5%) of the products. These
`products, and an additional eight mAbs that entered clinical
`study in 2005 as of June, were not included in the data used for
`
`Figure 1. Bispecific and primatized mAbs, as well as products of
`unknown mAb type (4.5% of the data set), were excluded from
`probability-of-success calculations. The remaining four types of
`mAbs (murine, chimeric, humanized and human) comprised 90%
`of the data set.
`Approval success calculations were based on products
`with a known fate (US approval or discontinuation). Percent
`completion was defined as the percentage of products with
`a known fate in a given cohort. Phase transition percentages
`(Fig. 2) were calculated as follows: the number of products that
`completed a given phase and entered the next was divided by
`the difference between the number of products that entered
`the phase and those that were still in the phase at the time
`of the calculation. Transitions occurring between phases of
`clinical studies conducted worldwide were included. The
`immunological therapeutic category includes treatments for
`diseases characterized by defects in the immune system that are
`not currently accepted as having an infectious cause (such as
`rheumatoid arthritis, Crohn disease or multiple sclerosis).
`
`redticing the HAMA response for at least some
`pr~ducts. In fact, five chimeric mAbs are now
`marketed in numerous countries as treatments
`for a variety of diseases (Table I). The fault
`of the early human mAbs lay not necessarily
`
`in themselves, but in the production method.
`Establishment of human hybridomas and the
`production of sufficient amounts of human
`mAbs from human cell lines proved to be
`difficult8•10. Nevertheless, one human anti-
`
`Antibody. Complex protein-based molecules produced by 8-lymphocytes that bind to and
`help eliminate foreign and infectious agents in the body. Antibodies are Y-shaped, having
`two sets of branches attached to a single stem. The arms of the Y (Fab) are the so-called
`variable regions, the tips of the arms contain antigen-binding regions (complementarity(cid:173)
`determining regions or CDRs) and the stem (Fe) is a constant region. The constant regions
`trigger effector functions (phagocytosis, cytolysis by cytotoxic lymphocytes or initiation
`of complement cascade followed by cell lysis) by linking the complex to other cells of the
`immune system.
`Monoclonal antibody (mAb). Originally, mAbs were antibodies produced from a single
`8-lymphocyte. Genetic manipulation now allows genes from multiple sources of 8-
`lymphocytes (for example, mouse and human) to be combined. mAbs of a defined peptide
`sequence have identical antigen-binding regions and bind to the same site (the epitope) of
`an antigen.
`Murine (mouse) mAb. A mAb derived entirely from mice, specifically murine hybridomas
`generated from the fusion of mouse myeloma and mouse 8-lymphocyte cells.
`Chimeric mAb. A mAb constructed from variable regions derived from a murine source and
`constant regions derived from a human source.
`Humanized mAb. A mAb constructed with only antigen-binding regions (also called
`complementarity-determining regions or CDRs) derived from a mouse, and the remainder of
`the variable regions, and constant regions, derived from a human source.
`Primatized mAb. A mAb constructed from variable regions derived from cynomolgus
`macaques and constant regions derived from a human source. Primatized is a registered
`US trademark of Cambridge, Massachusetts-based Biogen ,Idec.
`Human mAb. A mAb derived entirely from a human source, currently transgenic mice or
`phage display. Human mAbs can also be produced from human hybridomas or human
`8-lymphocyte cell lines immortalized by Epstein-Barr virus. However, these cell lines are
`unstable and produce only small amounts of mAbs.
`
`endotoxin IgM mAb, nebacumab ( Centoxin;
`Centocor, now a wholly owned subsidiary of
`Johnson & Johnson, New Brunswick, New
`Jersey) was briefly marketed for septic shock
`and Gram-negative bacteremia outside the
`United States in the early 1990s, but was with(cid:173)
`drawn for safety reasons likely specific to the
`indication and mechanism of action rather
`than the type of mAb 11 .
`Chimeras were not the only possible result
`of mAb genetic engineering, though. Soon
`after chimeric mAbs were described, work on
`humanized mAbs was also reported12. The
`perception of risk associated with the pres(cid:173)
`ence of murine protein sequence inexorably
`drove the technology from the murine prod(cid:173)
`ucts toward humanized ones, bypassing the
`chimeras, though the ultimate goal was an
`efficient production method for fully human
`mAbs. The development of phage display tech(cid:173)
`nology and transgenic mice in the early 1990s
`provided the desired methods, but realization
`of the goal was delayed until the end of the
`decade because of disputes over patents on the
`technologies (seep. 1079).
`Advancements in genetic engineering
`techniques have finally provided the long(cid:173)
`sought human mAbs, and have also opened
`the door to the study of varied mAb frag(cid:173)
`ments, including single-chain variable frag(cid:173)
`ments and antigen-binding fragments (see
`p. 1126). However, even mAb fragments are
`complex molecules, and a variety of factors
`must be considered (for example, specificity
`or avidity) before one is selected as a clinical
`candidate. Increased control over the design
`of these molecules might increase the rate of
`success in the future.
`
`1074
`
`VOLUME 23 NUMBER 9 SEPTEMBER 2005 NATURE BIOTECHNOLOGY
`
`

`

`FEATURE
`
`Generic
`
`Company/location
`
`Trade
`
`Description
`
`Therapeutic category
`
`Approval date
`
`M uromonab-CD3
`
`Johnson & Johnson
`New Brunswick, New Jersey
`
`Orthoclone OKT3 Murine, lgG2a,
`anti-CD3
`
`Abciximab
`
`Centocor
`
`Rituximab
`
`Genentech
`
`Daclizumab
`
`Basiliximab
`
`Palivizumab
`
`Hoffmann-La Roche
`Basel
`
`Novartis
`Basel
`
`Med Immune
`Gaithersburg, Maryland
`
`ReoPro
`
`Rituxan
`
`Zenapax
`
`Simulect
`
`Synagis
`
`lnfliximab
`
`Centocor
`
`Remicade
`
`Trastuzumab
`
`Genentech
`
`Gemtuzumab ozogamicin Wyeth
`Madison, New Jersey
`
`Herceptin
`
`Mylotarg
`
`Chimeric, lgGl,
`anti-GPllb/llla; Fab
`Chimeric, lgG 1 K,
`anti-CD20
`
`Humanized, lgGl K,
`anti-CD25
`
`Chimeric, lgGlK,
`anti-CD25
`
`Humanized, lgGlK,
`anti-respiratory
`syncytial virus
`
`Chimeric, lgGlK,
`anti-tumor necrosis
`factor (TNfo)
`
`Humanized, lgGlK,
`anti-HER2
`
`Humanized, lgG4K,
`anti-CD33; immunotoxin
`
`,Immunological
`
`06/19/86 (US)
`
`Hemostasis
`
`12/22/94 (US)
`
`Oncological
`
`Immunological
`
`Immunological
`
`Anti-infective
`
`Immunological
`
`Oncological
`
`Oncological
`
`11/26/97 (US)
`06/02/98 (EU)
`
`12/10/97 (US)
`02/26/99 (EU)
`
`05/12/98 (US)
`10/09/98 (EU)
`
`06/19/98 (US)
`08/13/99 (EU)
`
`08/24/98 (US)
`08/13/99 (EU)
`
`09/25/98 (US)
`08/28/00 (EU)
`
`05/17/00 (US)
`
`Alemtuzumab
`
`Genzyme
`Cambridge, Massachusetts
`
`Campath-1 H
`
`Humanized, lgG 1 K,
`anti-CD52
`
`Oncological
`
`lbritumomab tiuxetan
`
`Biogen Idec
`
`Zevalin
`
`Murine, lgGlK, anti-CD20; Oncological
`radiolabeled (Yttrium 90)
`
`Humira
`
`Human, lgGlK, anti-TNfo
`
`Immunological
`
`Adalimumab
`
`Abbott
`Deerfield Park, Illinois
`
`Omalizumab
`
`Tositumomab-1131
`
`Efalizumab
`
`Cetuximab
`
`Genentech
`Corixa
`Seattle
`
`Genentech
`
`I mclone Systems
`New York
`
`Bevacizumab
`
`Genentech
`
`Natalizumaba
`
`-----
`
`Biogen Idec
`
`Xolair
`Bexxar
`
`Raptiva
`
`Erbitux
`
`Avastin
`
`. Tysabri
`
`05/07/01 (US)
`07/06/01 (EU)
`
`02/19/02 (US)
`01/16/04 (EU)
`
`12/31/02 (US)
`09/1/03 (EU)
`
`06/20/03 (US)
`
`06/27/03 (US)
`
`10/27/03 (US)
`09/20/04 (EU)
`
`02/12/04 (US)
`06/29/04 (EU)
`
`02/26/04 (US)
`01/12/05 (EU)
`
`Humanized, lgGlK, anti-lgE
`
`Immunological
`
`Murine, lgG2a1.., anti-CD20; Oncological
`radiolabeled (Iodine 131)
`
`Immunological
`
`Oncological
`
`Oncological
`
`Humanized, lgGlK,
`anti-CD! la
`
`Chimeric, lgGlK,
`anti-Epidermal growth
`factor receptor
`
`Humanized, lgGl,
`anti-vascular endothel ial
`growth factor
`
`Humanized, lgG4K,
`anti-a4-integrin
`
`Immunological
`
`11/23/04 (US)
`
`See Box 1 for methodology.
`"Voluntary suspension of natalizumab marketing announced February 28, 2005.
`
`Entering the clinic
`The therapeutic potential of mAbs was rec(cid:173)
`ognized early on, but the number of products
`ready for clinical study was limited in the early
`1980s (Fig. l).A surge in the clinical initiations
`of mAb products started in 1984, but reached a
`local maximum in 1987. Most of the products
`( 89%) entering clinical study between 1980 and
`1987 were murine mAbs that ultimately failed
`in the clinic. Clinical development of murine
`products declined dramatically between 1987
`and the mid-1990s, then gradually dropped to
`zero in 2003.
`A plateau in clinical initiations occurred
`while less immunogenic mAbs were developed.
`Starting in 1992, the loss of the murine mAbs
`was offset by a rise in the number of human-
`
`ized products entering clinical study. Of the
`four types-murine, chimeric, humanized and
`human-human mAbs constituted the largest
`number entering clinical study per year by 2001.
`In contrast, the number of chimeric mAbs
`entering clinical study per year has been con(cid:173)
`sistently low since the mid- l 980s, when these
`products first began to enter clinical trials.
`
`Moving through the phases
`We calculated probabilities for making transi(cid:173)
`tions between clinical phases for the chimeric
`and humanized mAbs in the data set (Fig. 2).
`Results for human mAbs were not included
`because most of these products are still in clini(cid:173)
`cal study, and therefore few have known fates.
`The probability of transition from US Food and
`
`DrugAdministration (FDA) review to approval
`was 100% for the cohorts presented, so results
`for the transition from phase 3 to regulatory
`review were combined with those for transi(cid:173)
`tion from review to approval. For any cohort,
`the arithmetic product of the phase transitions
`exactly equals the overall approval success rate
`only when the fates of all products are known
`(that is, when the percent completion is 100%).
`Thus, calculated results should be considered
`estimates when the percent completion is below
`100%.
`Overall, the chimeric mAbs tended to have
`lower probabilities of transitions from phases
`1 to 2 and phases 2 to 3 than the humanized
`products. However, the probability of receiving
`FDA approval once the product was in phase 3
`
`NATU RE BIOTECHNOLOGY VOLUME 23 NUMBER 9 SEPTEMBER 2005
`
`1075
`
`

`

`FEATURE
`
`32
`28
`24
`20
`16
`12
`8
`
`(cid:127) Murine
`(cid:127) Chimeric
`(cid:127) Human
`
`Humanized
`
`All
`
`Figure 1 Number of therapeutic mAbs entering clinical study per year
`(1980-2004). Data are presented as two-year moving averages.
`
`was higher-actually 100%-for the chimeric
`mAbs. Taken together, these results suggest that
`stringent: selection criteria for advancement
`were applied to the chimeric clinical candidates
`after phases 1 and 2, thus ensuring that only
`successful products entered critical, and expen(cid:173)
`sive, phase 3 studies. It should be noted that sev(cid:173)
`eral' of the chimeric mAb cohorts included only
`a ,small number of products.
`
`. ·Approval success
`The FDA has approved a total of 18 thera(cid:173)
`peutic mAbs (Table 1 ), although one, natali(cid:173)
`zumab (Tysabri; Biogen Idec, Cambridge,
`Massachusetts) has since been voluntarily
`withdrawn. Most of the approvals have come in
`the past decade; only two products ( 11 % ) were
`approved before 1997. Most mAbs have been
`studied as treatments for either oncological or
`immunological indications. FDA-approved
`mAbs in these two therapeutic categories
`comprise 89% of the total. The chimeric and
`humanized products comprise 14 (78%) of the
`approved mAbs, and one human mAb, adali(cid:173)
`mumab (Humira; Abbott), was approved in
`2002.
`Only three (17%) approved mAbs are
`murine products. The first mAb approved,
`muromonab-CD3 ( Orthoclone; Johnson &
`Johnson) was a murine anti-CD3 product used
`as a treatment to reverse transplant rejection in
`immunosuppressed patients. More recently, two
`radiolabeled murine mAbs, ibritumomab tiux(cid:173)
`etan (Zevalin; Biogen Idec) and tositumomab(cid:173)
`I-131 ( Bexxar; Corixa, Seattle, Washington, now
`a wholly owned subsidiary of GlaxoSmithKline,
`Brentford, UK), were approved for non(cid:173)
`Hodgkin lymphoma. The potency of a radio(cid:173)
`label allows administration of very small
`amounts of these products, so immunogenicity
`from the murine protein is less of a concern.
`Because of their limited utility, however, clini(cid:173)
`cal study of most of the murine mAbs has been
`discontinued. Overall, the approval success rate
`calculated for murine products was 3%. Few
`
`0
`
`2
`
`3
`
`Figure 2 Clinical phase transition percentages for therapeutic mAbs,
`according to FDA data. (See Box 1 for methodology.) 1, chimeric mAbs, all
`products (n = 39); 2, oncological chimeric mAbs (n = 21); 3, immunological
`chimeric mAbs (n = 9); 4, chimeric mAbs, 1987-1997 (n = 20); 5,
`humanized mAbs, all products (n = 102); 6, oncological humanized mAbs
`(n = 46); 7, immunological humanized mAbs (n = 34) ; 8, humanized mAbs,
`1988-1997 (n = 46). See Table 2 for completion rates for these.
`
`1076
`
`VOLUME 23 NUMBER 9 SEPTEMBER 2005 NATURE BIOTECHNOLOGY
`
`The Tufts University Center for the Study of
`Drug Development ( CSDD) reported approval
`success rates for therapeutic mAbs sponsored
`by only US-based companies in 2001 (ref. 1).
`The chimeric and humanized mAbs from the
`US-focused data set had approval success rates
`of 24% and 25%, respectively. The current
`data set includes approximately twice the total
`number of products (355 versus 186 mAbs) and
`nearly twice the number of approved products
`(18 versus 10 mAbs) compared with the data
`set analyzed in the 2001 study. The success rates
`reported here for chimeric and humanized
`mAbs (21 % and 18%, respectively) are slightly
`lower, but this is not surprising because 'suc(cid:173)
`cess' was defined here as FDA approval only,
`although products from companies located
`worldwide were included in the data set.
`Several
`therapeutic mAbs have been
`approved only outside the US. One murine
`mAb, edrecolomab (Panorex; Centocor), and
`one human mAb, nebacumab ( Centoxin;
`Centocor), were approved in the 1990s for
`treatment of colorectal cancer and septic
`shock/Gram-negative bacteremia, respectively.
`However, the products were subsequently with(cid:173)
`drawn from their non-US markets. In addition,
`three therapeutic mAbs were recently approved
`in Asia. China's State Drug Administration
`(Beijing) has approved two oncology mAbs:
`in 2003, a chimeric, I131 -radiolabeled mAb
`that targets histone Hl-DNA complexes in
`necrotizing tumor cells from MediPharm
`Biotech (Shanghai, China); and in 2005,
`nimotuzumab, a humanized mAb from YM
`Bioscience (Mississauga, Ontario, Canada).
`Japan's Ministry of Health, Labor and Welfare
`approved tocilizumab, a humanized anti-inter(cid:173)
`leukin 6 receptor mAb from Chugai (Tokyo) as
`a treatment for Castleman disease in 2005.
`
`(cid:127) Phase 1 to 2
`(cid:127) Phase 2 to 3
`(cid:127) Phase 3 to FDA
`approval
`
`5
`
`6
`
`7
`
`8
`
`4
`mAbs
`
`1/)
`
`100
`c ao
`C: g 60
`'iii
`C:
`_£l 40
`
`Q)
`1/)
`CCI
`-&_ 20
`
`murine mAbs remain
`in clinical study; the
`success rate for the
`group would not dra(cid:173)
`matically
`improve,
`even if some were to
`be approved.
`In contrast to the
`murine products, the
`chimeric and human(cid:173)
`ized mAbs have been
`much more success(cid:173)
`ful-these products
`comprise 28% and
`50%,
`respectively,
`of FDA-approved
`mAbs. Approval success rates for the chimeric
`and humanized mAb therapeutics included in
`the data set were 21 % and 18%, respectively
`(Table 2). Only slight variations in the success
`rates were observed when the products were
`stratified by therapeutic category. Chimeric
`mAbs for immunological indications had a
`somewhat higher approval success rate com(cid:173)
`pared with the oncology products (22% versus
`18%, respectively), but this order was reversed
`for the humanized products ( 19% versus
`24%, respectively).
`Success rates were also calculated for prod(cid:173)
`ucts that entered clinical study before 1998.
`The periods selected correspond to the inter(cid:173)
`vals between the year clinical study was initiated
`for the first of each mAb type through 1998.
`On average, the clinical development phase
`for the FDA-approved therapeutic mAbs is 6.5
`years. Thus, a reasonable amount of time has
`passed (a minimum of 7.5 years from 1998 to
`mid-2005) for determination of the fates of the
`products. Approval success rates for the .chime(cid:173)
`ric and humanized mAb therapeutics stratified
`by time were 29% and
`25%,
`respectively.
`These values repre(cid:173)
`sent the most accu(cid:173)
`rate overall approval
`success rates for chi(cid:173)
`meric and human(cid:173)
`ized mAbs we have
`calculated to date. It
`remains to be seen
`whether the success
`rates will be the same
`for mAbs currently in
`the clinical pipeline.
`Again, it should be
`noted that the chime(cid:173)
`ric mAb cohorts are
`much smaller when
`compared with simi(cid:173)
`lar groups of human(cid:173)
`ized mAbs.
`
`(cid:127)
`

`

`FEATURE
`
`able 2 Approval success rates for mAbs
`
`mAb type a•nd area of applicationa
`
`~ meric~Abs, all products
`
`Onc_ological chimeric mAbs
`
`lm~
`
`olog~ al chimeric mAbs
`
`Chimeric mAbs, 1987-1997
`
`Humanized mAbs, all products
`
`Oncologic~ humanized mAbs
`
`lmmu ~ logical humanized mA~
`
`Humanized mAbs, 1988-1997
`
`Total number of
`mAbs
`
`Number
`discontinued
`
`Number FDA
`approved
`
`Completionb (%)
`
`Approval successc
`(%)
`
`39
`
`21
`
`9
`
`20
`
`102
`
`46
`
`34
`
`46
`
`19
`
`9
`
`7
`
`12
`
`41
`
`13
`
`17
`
`24
`
`5
`
`2
`
`2
`
`5
`
`9
`
`4
`
`4
`
`9
`
`62
`
`52
`
`100
`
`85
`
`49
`
`37
`
`62
`
`72
`
`21
`
`18
`22 - - - -
`29
`
`18
`
`24
`
`19
`
`27
`
`8See Box 1 for inclusion criteria. bpercent completion, percentage of products with a known fate in a given cohort. cphase 1 to approval in the United States.
`
`Upcoming monoclonals
`The current pipeline of therapeutic mAbs in
`clinical study contains more than 150 prod(cid:173)
`ucts. Of the total in the clinic, 63 ( 41 % ) are in
`phase 1. This group of products is the least well
`defined of the data set because often only lim(cid:173)
`ited information is available at this early stage
`of thr'process. For example, of the entire data
`set, _mAb type is unknown for only 4.5%, but
`two--thirds of these are found in phase 1. Taking
`into account the uncertainty in the data, we can
`say that humanized and human mAbs comprise
`a minimum of 24% and 43%, respectively, of
`the products in phase 1. In contrast, humanized
`products constitute the greatest portion of the
`mAbs in phase 2 and phase 3 (39% and 47%,
`respectively), though the human mAbs run
`a close second ( comprising 31 % and 33% of
`those in phase 2 and phase 3, respectively).
`Most therapeutic mAbs are studied as treat(cid:173)
`ments for diseases in only three therapeutic
`categories: oncological, immunological and
`anti-infective (Fig. 3). The relative propor(cid:173)
`tion of each category has even remained fairly
`consistent. One notable exception is the reduc(cid:173)
`tion in the number of immunological mAbs
`in phase 1. This decrease likely is due to com(cid:173)
`petition from the approved immunological
`mAb products.
`The group of products in phase 3 (Table 3)
`highlights the trend toward the study of frag(cid:173)
`ments, rather than full-size mAbs. Only one
`FDA-approved product, abciximab (ReoPro;
`Centocor ), is a mAb fragment. In contrast, seven
`of the mAbs currently in phase 3 are variations
`on the mAb theme. Some targets are inacces(cid:173)
`sible to full-sized mAbs, so decreasing size while
`maintaining functionality and specificity might
`increase the potential utility of the products.
`
`Where to from here?
`Therapeutic mAbs, and the technology to pro(cid:173)
`duce them, have evolved over the past quarter
`century and will continue to do so in the future.
`Products derived from early technology have
`
`Immunogenicity is not a fatal flaw now, but
`rather a problem that can be managed.
`Therapeutic mAbs are now big business. Six
`FDA-approved products each had global sales
`over $500 million in 2004. This income helps
`keep firms in business and spurs mAb devel(cid:173)
`opment at other firms. Analysis of the average
`wholesale price13, dosing information and indi(cid:173)
`cations suggests that FDA-approved mAbs are
`differentially priced by therapeutic category.
`Oncology mAbs, with smaller markets, are
`priced much higher than mAbs for immuno(cid:173)
`logical diseases that tend to have larger markets
`(for example, rheumatoid arthritis, Crohn dis(cid:173)
`ease, psoriasis and asthma). The cost of pro(cid:173)
`tein therapeutics to the US health care system,
`which is already considered overburdened, has
`raised the question of approval for biosimilars
`(generic versions of biologics, also known as
`follow-on biologics). However, three factors
`suggest mAb biosimilars will not be seen any
`time soon: the complexity of mAb production,
`current lack of a regulatory pathway to approval
`and patent protection covering all the recently
`approved products.
`According to our analysis of the therapeutic
`protein pipeline, study of human mAbs and of
`mAb fragments is the wave of the future. On
`
`(cid:127) Oncological
`(cid:127) Immunological
`(cid:127) Anti-infective
`(cid:127) Misc.
`
`All
`
`60
`
`50
`& 40
`~ CD 30
`~
`8?_ 20
`
`10
`
`0
`
`Phase 1 Phase 2 Phase 3 Approved
`mAbs
`
`mostly fallen by the wayside, though some
`chimeras are still in active development. Small
`biotech companies sponsor the majority of
`chimeric mAbs now in clinical study. The chi(cid:173)
`meric products might be attractive for several
`reasons: use of the older technology may allow
`firms to avoid technology licensing or royalty
`fees levied for production of humanized and
`human mAbs, and, despite the purported risk
`of HAMA, careful selection of chimeric clinical
`candidates can yield blockbusters. For example,
`several chimeric mAbs are truly billion dollar
`molecules-infliximab (Remicade; Centocor)
`and rituximab (Rituxan; Genentech, S. San
`Francisco, California) each had over $2 billion
`in global sales during 2004.
`Still, a trend toward clinical study of human
`products is clear. This result follows from the
`attempts to reduce mAb immunogenicity by
`reducing the quantity of murine sequence in
`the products. Even so, the immunogenicity
`problem does not disappear with the murine
`component. The human immune system can
`potentially produce antibodies to any pro(cid:173)
`tein therapeutic. Labels for most of the FDA(cid:173)
`approved mAbs report that at least some
`patients developed detectable antibodies to the
`products. The labels also note that the identi(cid:173)
`fication of anti-mAb
`antibodies is highly
`dependent on
`the
`sensitivity and speci(cid:173)
`ficity of the assay, thus
`making comparisons
`between immunoge(cid:173)
`nicity rates difficult.
`Nevertheless, on the
`basis of labeling infor(cid:173)
`mation, three each
`of the humanized
`and chimeric mAbs,
`and the one human
`mAb, have
`immu(cid:173)
`nogenicity rates in
`the range of 5-10%.
`
`NATURE BIOTECHNOLOGY VOLUME 23 NUMBER 9 SEPTEMBER 2005
`
`1077
`
`Figure 3 Therapeutic categories for mAbs in clinical study. Misc.,
`miscellaneous category including ophthalmic, neuropharmacologic and
`cardiovascular indications.
`
`

`

`FEATURE
`
`Company/location
`
`Abgenix
`Fremont, California
`
`Generic/code name
`
`Brand name
`
`Description
`
`Panitumumab
`
`Human, lgG2K, anti-EGF receptor
`
`Amgen
`Thousand Oaks, California
`
`AMG-162
`
`Human , lgG2, anti-RANKL
`
`Genmab
`Copenhagen
`Human, lgGl K, anti-CTLA 4
`MDX-010
`Medarex
`Princeton, New Jerse_y ______________________________________________ _
`
`Humax-CD4
`
`Human, lgG, anti-CD4 receptor
`
`Zanolimumab
`
`Neutec
`Manchester, UK
`
`Alexion Pharmaceuticals
`Cheshire, Connecticut
`
`Alexion Pharmaceuticals
`
`Genentech
`
`GlaxoSmithKline
`
`Anti-MRSA mAb
`
`Aurograb
`
`Human, anti-MRSA; single chain
`
`Pexelizumab
`
`Eculizumab
`
`Ranibizumab
`
`Mepolizumab
`
`Humanized, anti-complement C5; single-chain variable fragment
`
`Humanized, lgG, anti-C5; single chain
`
`Lucentis
`
`Humanized, lgGl, anti-VEGF; Fab fragment
`
`Humanized, lgGl, anti-lL5
`
`Humanized, lgGl, anti-CD22
`Lymphocide
`Epratuzumab
`lmmunomedics
`Morris Plains, New Jersey ________________________ _
`
`Med Immune
`
`UCB
`Brussels
`
`Abbott
`
`Trion
`Martinsried, Germany
`
`Anti-RSV mAb
`
`Certolizumab Pegol
`
`Numax
`
`Cimzia
`
`Humanized, anti-RSV
`
`Humanized, lgG, anti-TNFa, pegylated Fab' fragment
`
`Afelimo_m_a_b ______ Seg~
`Removab
`
`Catumaxomab
`
`Murine, lgG3K, anti-TNFa; F(ab) 2
`Bispecific, anti-CD3/Epcam; trifunctional
`
`· Wilex
`Munich, Germany
`CTLA, cytotoxic T-lymphocyte associated protein; EGF, epidermal growth factor; MRSA, methicillin-resistant Staphylococcus aureus; RAN KL, receptor for activation of nuclear
`factor-KB ligand; RSV, respiratory syncytial virus; TNF, tumor necrosis factor; VEGF, vascular endothelial growth factor.
`
`Rencarex
`
`Chimeric, lgGl, anti-carbonic anhydrase IX
`
`WX-G250
`
`the basis of the historical data set, approval
`success rates in the 18-29% range might be
`expected; success rates for the newer prod(cid:173)
`ucts will be estimates for the foreseeable
`future, because most of the products are still
`in clinical study. The overall success rates for
`new chemical entities (NCEs) as a whole and
`NCEs in oncology were recently reported to
`be 11 % and 5%, respectively14• Our results 1- 3
`have repeatedly shown that chimeric and
`humanized mAbs have higher success rates
`than NCEs, notably in the oncology category.
`Results for the immunological category cannot
`be directly compared because of differences in
`the definition of the NCE and mAb cohorts.
`The combination of the efficient production
`of mAbs specifically designed to be safe and
`efficacious and the approval and marketing
`
`success of an increasing number of mAbs will
`continue to draw interest from biotech and
`pharmaceutical firms alike.
`
`ACKNOWLEDGMENTS
`The authors gratefully acknowledge the assistance
`of the companies that provided survey data. We also
`thank H. Pu jar for reviewing the manuscript and for
`providing helpful comments and suggestions.
`
`1. Reichert, J.M. Monoclonal antibodies in the clinic. Nat.
`Biotechnol. 19, 819-822 (2001).
`2. Reichert, J.M . Therapeutic monoclonal antibodies:
`trends in development and approval in the US. Curr.
`Opin. Mot. Ther. 4, 110-118 (2002) .
`3. Reichert, J. & Pavolu, A. Monoclonal antibodies market.
`Nat. Rev. Drug Discov. 3, 383-384 (2004).
`4. Kohler, G. & Milstein, C. Continuous cultures of fused
`cells secreting antibody of predefined specificity. Nature
`256, 495-497 (197 5),
`5. Milstein, C. The hybridoma revolution : an offshoot of
`basic research. Bioessays 21, 966-973 (1999).
`6. Morrison, S.L. et al. Chimeric human antibody
`
`molecules: Mouse antigen-binding domains with human
`constant region domains. Proc. Natl. Acad. Sci. USA
`81, 6851- 6855 (1984).
`7. Boulianne, G.L. et al. Production of functional chi(cid:173)
`maeric mouse/human antibody. Nature 312, 643-646
`(1984).
`8. Cole, S.P., Campling, B.G., Atlaw, T., Kozbor, D. &
`Roder, J.C. Human monoclonal antibodies. Mo/. Cell.
`Biochem. 62, 109-120 (1984).
`9. Carson, D.A. et al. Human lymphocyte hybridomas and
`monoclonal antibodies. Adv. lmmunol. 38, 275-311
`(1986).
`10. Pasqualini, R. & Arap, W