Herceptin
`
`H.M. Shepard((cid:0)), P. Jin, D.J. Slamon, Z. Pirot, and D.C. Maneval
`
`1 Magic Bullets and Monoclonal Antibodies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
`2 The Discovery and Development of HER Therapeutics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
`2.1 Setting the Stage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
`2.2 The Development of a Preclinical Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
`2.3 A Perversion of Nature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
`2.4 Biologic Effects of Amplified HER2 Expression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
`2.5 Proof of Concept for Antagonists of p185HER2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
`2.6 In Vivo Proof of Concept with MuMAb4D5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
`3 Proof of Concept for Therapeutic Antibodies in Solid Cancer . . . . . . . . . . . . . . . . . . . . . . . . 194
`3.1 The Controversy About Making a Successful Monoclonal Antibody Therapeutic
`vs. Solid Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
`3.2 MuMAb4D5 Therapy for Breast Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
`3.3 Phase I Investigation with MuMAb4D5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
`4 Development of Herceptin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
`4.1 Humanized Versions of MuMAb4D5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
`4.2 Characterization of Herceptin In Vivo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
`4.3 Summary of the History of Clinical Trials with Herceptin . . . . . . . . . . . . . . . . . . . . . . . 201
`4.4 Current Status and Significance of Herceptin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
`5 Alternative Therapies Targeting p185HER2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
`5.1 Radiolabeled Herceptin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
`5.2 Active Specific Immunotherapy (ASI) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
`6 New Approaches to the Human EGFR Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
`6.1 Current Status of the Human EGFR Family . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
`6.2 Pan-HER Ligand Traps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
`6.3 Inhibition of Ligand-Induced Receptor Phosphorylation by Hermodulins . . . . . . . . . . 209
`6.4 Hermodulins Inhibit Tumor Cell Proliferation In Vivo . . . . . . . . . . . . . . . . . . . . . . . . . . 211
`7 Summary and Conclusions: What We Have Learned
`and What to Do Next . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
`References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
`
`H.M. Shepard
`Receptor BioLogix, Inc., 3350 W. Bayshore Rd., Suite 150, Palo Alto, CA 94303, USA
`e-mail: hms@rblx.com
`
`Y. Chernajovsky, A. Nissim (eds.) Therapeutic Antibodies. Handbook of
`Experimental Pharmacology 181.
`c(cid:2) Springer-Verlag Berlin Heidelberg 2008
`
`183
`
`
`
`Pfizer v. Genentech
`IPR2017-01488
`Genentech Exhibit 2028
`
`

`

`184
`
`H.M. Shepard et al.
`
`Abstract The biology of the human epidermal growth factor (EGF) receptor-2
`(HER2) has been reviewed numerous times and provides an excellent example for
`developing a targeted cancer therapeutic. Herceptin, the FDA-approved therapeutic
`monoclonal antibody against HER2, has been used to treat over 150,000 women
`with breast cancer. However, the developmental history of Herceptin, the key events
`within the program that created pivotal decision points, and the reasons why deci-
`sions were made to pursue the monoclonal antibody approach have never been ade-
`quately described. The history of Herceptin is reviewed in a way which allows the
`experience to be shared for the purposes of understanding the drug discovery and
`development process. It is the objective of this review to describe the pivotal events
`and explain why critical decisions were made that resulted in the first therapeutic to
`successfully target tyrosine kinases in cancer. New approaches and future prospects
`for therapeutics targeting the HER family are also discussed.
`
`1 Magic Bullets and Monoclonal Antibodies
`
`The specific targeting of disease-causing organisms, or diseased cells, was first artic-
`ulated by Ehrlich, who reasoned that because it is possible to differentially stain
`cancer and normal cells, it should be possible to specifically target cancer (perspec-
`tive by Witkop 1999). Based upon this work, many successful chemotherapeutics
`have been created. However, because diseased and normal cells share biochemical
`pathways, and are much more similar than they are different, targeting disease pro-
`cesses by interrupting cellular metabolism without toxicity to the host has remained
`a problem in drug discovery.
`Subtle differences between normal and tumor cells include a greater dependence
`of cancer cells on glucose metabolism instead of the citric acid cycle for the genera-
`tion of adenosine triphosphate (the “Warburg Effect”; Ashrafian 2006), and a greater
`use of uracil to support their growth (“uracil flux”), leading to the discovery of fluo-
`rouracil as a chemotherapeutic (Heidelberger et al. 1957). These targets for therapy
`share the inherent problem that the differences are a matter of degree. While the flu-
`orouracils are clearly effective, with a role to play in the treatment of many cancers
`(Dorr and Von Hoff 1994), their efficacy is more a result of the relative leakiness of
`blood vessels in tumors (leading to drug localization) than it is to the specific target-
`ing of cancer cell metabolism. In fact, thymidylate synthase, the enzyme inhibited by
`the fluorouracils, is predictably expressed to a higher degree in tumor cells than it is
`in normal cells. As a result, normal cells (with lower thymidylate synthase, like gut
`epithelium, skin fibroblasts, and hematopoeitic cells) are generally more sensitive to
`the cytotoxic effects of fluorouracils than are tumor cells, which have a higher intra-
`cellular concentration of the enzyme (Lackey et al. 2001; Li et al. 2001). The goal
`of the modern era of cancer treatment is to create drugs that preferentially damage
`tumor cells based upon their specific biochemical properties, and leave normal cells
`relatively free from injury: the realization of Ehrlich’s “magic bullet” hypothesis.
`
`

`

`Herceptin
`
`185
`
`The enablement of this goal in cancer treatment required several important
`advances in drug discovery. Key discoveries include identifying cancer-specific anti-
`gens (e.g., tyrosine kinases and other enzymes), understanding disease pathways,
`and developing a means for specifically targeting diseased cells. The discovery
`of viral oncogenes which encode tyrosine kinases, and subsequently the finding
`that mutations in some normal cellular tyrosine kinases can cause them to become
`oncogenic, resulting in cellular transformation (immortality, anchorage indepen-
`dent growth, and the ability to form tumors in immune-deficient mice), reviewed
`by Bishop (1989) and Varmus (1989), provided the basis for targeting the human
`epidermal growth factor receptor 2 (HER2) protooncogene with a monoclonal anti-
`body. Since the approval of Herceptin in 1998, the tyrosine kinases have become
`the archetypical example of a validated target in cancer, and monoclonal antibodies
`have become an accepted biopharmaceutical to target cell surface receptors.
`Especially relevant to Herceptin, the early enabling oncogene discovery was the
`finding that the v-erb-B oncogene, derived from chicken erythroblastosis virus,
`shared significant homology with the human epidermal growth factor receptor
`(EGFR or HER1), thereby giving rise to the hypothesis that under conditions of con-
`stitutive activation, EGFR might be implicated in human cancer (Kris et al. 1985).
`Further work proved this hypothesis and motivated the discovery of HER2, also
`known as human NEU, erb-B2, or NGL (Coussens et al. 1985; King et al. 1985;
`Schechter et al. 1985; Semba et al. 1985; Yarden and Ullrich 1988). The focus of
`this chapter is the discovery and pharmacological studies that led to the approval
`of Herceptin, a humanized monoclonal antibody targeting the receptor extracellu-
`lar domain encoded by HER2 (p185HER2). At the time when Investigational New
`Drug (IND)-enabling efforts began for Herceptin, only EGFR and HER2 had been
`described (Yarden and Ullrich 1988). The field has made tremendous advances since
`this time, powered in large part by the commercial success of Herceptin. Herceptin
`provided the first “magic bullet” targeted at tyrosine kinases to treat cancer. A dis-
`cussion later in the chapter will outline newer approaches to targeting the Human
`EGFR (HER) family.
`
`2 The Discovery and Development of HER Therapeutics
`
`2.1 Setting the Stage
`
`Direct causal relationships between oncogene amplification and/or overexpres-
`sion and certain types of cancer were less well defined in the 1980s (during the
`initial development efforts for Herceptin) than they are now. One of the most
`critical events in the research leading to Herceptin was reported by Weinberg
`and colleagues (Schechter et al. 1984). This involved the discovery of the first
`oncogenic receptor tyrosine kinase oncogene, NEU. It was discovered by gene
`
`

`

`186
`
`H.M. Shepard et al.
`
`transfection/transformation of fragmented DNA from a series of rat neuroblastomas
`into NIH 3T3 cells (the focus-forming assay; Shih et al. 1981).
`The product of the HER2 protooncogene (p185HER2) is a transmembrane Type
`1 receptor tyrosine kinase with extensive homology to the EGFR (Coussens et al.
`1985; Schechter et al. 1985; Yarden and Ullrich 1988) and now known to have simi-
`lar homology with HER3 and HER4 (Katoh et al. 1993; Zhou and Carpenter 2002).
`HER2 can be distinguished from HER1, 3, and 4 by differences in chromosomal
`location, transcript size, molecular mass, ligand activation of the associated tyro-
`sine kinase, and antigenicity, as determined by interaction with specific monoclonal
`antibodies (Citri and Yarden 2006; Kumar and Pegram 2006; Prenzel et al. 2001).
`We will review the science behind Herceptin development from a historical per-
`spective. It is our goal to provide a roadmap that can be generally applied to the
`assembly of the rationale for the development of other successful therapeutics. We
`will describe the progression of the science beginning with the discovery of the
`HER2 protooncogene through the demonstration that overexpression leads to cellu-
`lar transformation, tumor cell resistance to elements of the host immune system, and
`other characteristics that create a disease-specific signaling pathway in cancer cells.
`This pathway was the focus of efforts that resulted in the development of Herceptin,
`the first biopharmaceutical to demonstrate clinical proof of concept and the value of
`targeting tyrosine kinases in cancer.
`
`2.2 The Development of a Preclinical Rationale
`
`A striking convergence of basic science, clinical research, and translational medicine
`occurred within a short interval that resulted in the enablement of HER2 as a thera-
`peutic target (Fig. 1).
`
`2.3 A Perversion of Nature
`
`2.3.1 Growth Factor Activation of Tumor Cell Tyrosine Kinases Mediate
`Resistance to Immune Effector Molecules
`
`The discovery that activation of growth factor receptors can limit the ability of tumor
`necrosis factor-alpha (TNF-α) to inhibit tumor growth was a key finding in the his-
`tory of the development of Herceptin (Fig. 2).
`These results suggested that tumor cells may be able to secrete growth factors,
`not only to promote their own proliferation, as suggested by the autocrine growth
`factor model (Sporn and Todaro 1980), but also as a protective mechanism against
`host immune surveillance (Fig. 4).
`
`

`

`Herceptin
`
`187
`
`Fig. 1 Decision to develop Herceptin. Several of the most important events and data are summa-
`rized in this figure, with each step referenced by number. References: (1) Hudziak et al. (1987);
`(2) Hudziak et al. (1988, 1989); (3) Slamon et al. (1987, 1989); (4) Lewis et al. (1993); (5) Muller
`et al. (1988); (6) Maneval et al. (1991b)
`
`Fig. 2 Antagonism of rHuTNF-α-mediated growth inhibition by EGF or TGF-α on ME-180
`−1 TNF-α; triangles:
`cervical carcinoma cells. Open circles: growth factor alone; boxes: 50 u ml
`−1; closed circles: 5,000 u ml
`−1. The left axis (0) represents the effect of rHuTNF-α alone
`500 u ml
`(Sugarman et al. 1987)
`
`

`

`188
`
`H.M. Shepard et al.
`
`2.3.2 Activation of Receptor Tyrosine Kinases is Associated
`with Tumor Cell Resistance to Macrophages and TNF-α
`
`To further establish the link between tyrosine kinase activation in tumor cells and
`immune cell resistance, two approaches were taken. First, NIH 3T3 fibroblasts were
`transfected with expression plasmids encoding p185HER2 and cell lines were isolated
`as controls, or which express control plasmids and high levels of p185HER2. Second,
`spontaneous transformants of NIH 3T3 fibroblasts, which are often associated with
`amplification of the c-Met protooncogene (Giordano et al. 1989), were examined
`for associated resistance to TNF-α. These are two independent methods and very
`different examples of receptor tyrosine kinases.
`When cells selected for increased p185HER2 expression were tested for sensitiv-
`ity to macrophage-mediated cytotoxicity, it was found that high levels of p185HER2
`expression were associated with resistance to “effector” cells (Fig. 3a, Hudziak et al.
`1988). Similarly, when spontaneous transformants of NIH 3T3 fibroblasts were
`characterized for increased c-Met gene copy number, it was found that amplified
`c-Met was associated with increased resistance to TNF-α (Fig. 3b).
`These results support the concept that activation of tyrosine kinases by added
`growth factors, spontaneous gene amplification or gene transfection, are associated
`with resistance to immunosurveillance by macrophages/TNF-α.
`
`Fig. 3 Oncogene amplification and resistance to macrophage- or TNF-α-mediated cytotoxicity. (a)
`Macrophages are known to have a key role in eliminating incipient tumors (Urban and Schreiber
`1988), but the frequency of tumor cell resistance (∼10
`−4 for some cell lines; Lewis et al. 1987)
`means that escape from this single mechanism is common. Macrophage effector molecules, espe-
`cially TNF-α, can then act to stimulate tumor cell proliferation (Hudziak et al. 1988, Lewis et al.
`1987) and angiogenesis (Leibovich et al. 1987). (b) The c-Met protooncogene is often amplified in
`spontaneous transformants of NIH 3T3 fibroblasts. In this experiment, spontaneous transformants
`were subcloned, then tested for sensitivity to TNF-α. The results showed decreasing sensitivity
`correlates with increased c-Met gene copy number (Hudziak et al. 1990)
`
`

`

`Herceptin
`
`189
`
`Fig. 4 Tumor cells characterized by autocrine stimulation of receptor tyrosine kinase activity are
`selected via activated macrophages to form a clinical tumor
`
`2.3.3 Macrophage-mediated Antitumor Effects are Converted
`to Protumorigenic During In Vivo Tumor Progression
`
`While the macrophage may be instrumental in the initial detection and destruction of
`tumor cells (Urban et al. 1986), its effects, if unsuccessful in the first instance, may
`potentiate tumor cell growth and malignancy (Lewis et al. 1987, Fig. 4, see “Clinical
`Tumor”). Mechanisms that modulate tumor cell sensitivity to the host immune
`system play an important role in the growth of an incipient tumor, and the continued
`presence of activated macrophages within a resistant tumor may help to establish
`a protumorigenic environment. Our work established that antireceptor agents can
`reverse oncogene-associated immune resistance, and provided a critical element in
`building a rationale for the use of tyrosine kinase antagonists in the treatment of
`cancer.
`
`2.4 Biologic Effects of Amplified HER2 Expression
`
`Following the first molecular description of the sequence encoding p185HER2,
`its extensive homology with the rat NEU protooncogene was quickly established
`(Coussens et al. 1985; Yarden and Ullrich 1988). Further work distinguished NEU
`and HER2 by showing that a mutation in the transmembrane domain of p185NEU
`was sufficient to enable transforming activity (Weiner et al. 1989), while ampli-
`fied expression of the wild-type HER2 protooncogene was sufficient to transform
`
`

`

`190
`
`H.M. Shepard et al.
`
`fibroblasts in culture (Hudziak et al. 1987). Very rapidly, Slamon and colleagues at
`the University of California (Los Angeles) and Ullrich (Genentech, Inc.) initiated
`a collaboration that provided a critical link between amplified expression of HER2
`and aggressive breast and ovarian cancer (Slamon et al. 1989, 1987). In this work,
`it was demonstrated that threefold to fivefold overexpression of tumor-associated
`p185HER2 predicted a dramatically shortened survival in breast and ovarian cancer
`patients. A direct connection between HER2/NEU and breast cancer was further
`supported when transgenic mice overexpressing the NEU oncogene specifically
`developed mammary adenocarcinoma (Muller et al. 1988). Completing the cause
`and effect relationship between overexpression of p185HER2 and cellular malig-
`nancy, Hudziak et al. (1988, 1989) demonstrated that the earlier described resis-
`tance to macrophage killing that characterized growth factor activated tumor cells
`(Sugarman et al. 1987) also occurred in tumor cells which overexpress p185HER2
`(Fig. 3a) or which are characterized by amplification of the c-Met protooncogene.
`Many explanations have been offered for the connection between HER2 over-
`expression and disease progression. It is likely that HER2 overexpression leads
`to enhanced signaling with other members of the HER family (EGF formation;
`Pinkas-Kramarski et al. 1996), and potentially to coupling with other receptor tyro-
`sine kinases, like the IGF-1 receptor (Nahta et al. 2005). This strong coupling is sup-
`ported by in vitro proliferation data which show that cells overexpressing p185HER2
`can be much more sensitive to antibodies directed to the extracellular domain of
`p185HER2 (see Sect. 2.5.2).
`In summary, overexpression of p185HER2 transforms cells (enabling growth in
`soft agar and in nude mice) induces tumor cell resistance to the cytotoxic effects
`of macrophages, and is correlated with an aggressive form of breast and ovarian
`cancer.
`
`2.5 Proof of Concept for Antagonists of p185HER2
`
`2.5.1 In Vitro Proof of Concept
`
`No growth factor has yet been found which directly activates either p185NEU or
`p185HER2 (Hynes and Lane 2005). For this reason, p185HER2 is a noncanonical
`receptor and mechanisms which may be able to modulate it were not obvious.
`The most similar system is the human EGFR/HER1, for which Mendelsohn and
`colleagues had reported successful antibody-mediated inhibition (Gill et al. 1984).
`Similarly, the oncogenic version of rat p185NEU was found to be downregulated by
`monoclonal antibodies directed to its extracellular domain (Drebin et al. 1985).
`Fendly et al. (1990b) prepared a large array of monoclonal antibodies directed
`against the extracellular domain of p185HER2. These monoclonal antibodies were
`screened in proliferation assays against normal and tumor cells which expressed
`a spectrum of p185HER2 levels (Lewis et al. 1993; Park et al. 1992; Shepard
`et al. 1991). Some of these data are shown in Table 1. Multiple monoclonal
`
`

`

`Herceptin
`
`191
`
`Table 1 Effect of anti-p185HER2 monoclonal antibodies on the growth of human tumor cell linesa
`
`Cell Line
`
`Relative P185HER
`Relative Cell Proliferation (% of control)
`Expressionb
`6E9
`3H4
`2C4
`7F3
`7C2
`4D5
`1.0
`103
`114
`109
`116
`117
`116
`184
`0.3
`110
`110
`103
`106
`104
`129
`184A1
`0.8
`106
`107
`105
`108
`108
`108
`184B5
`1.0
`105
`102
`103
`96
`104
`104
`HBL-100
`1.2
`105
`113
`100
`111
`112
`101
`MCF7
`1.2
`013
`100
`93
`98
`104
`91
`MDA-MB-231
`3.3
`97
`105
`99
`97
`108
`102
`ZR-75-1
`3.3
`101
`91
`98
`93
`92
`97
`MDA-MB-436
`4.5
`96
`77
`29
`48
`87
`62
`MDA-MB-175
`16.7
`101
`65
`88
`80
`70
`61
`MDA-MB-453
`16.7
`99
`67
`64
`76
`105
`63
`MDA-MB-361
`25.0
`91
`29
`60
`21
`78
`27
`BT474
`33.0
`89
`40
`73
`51
`82
`33
`SK-BR-3
`16.7
`99
`85
`87
`91
`97
`77
`SK-OV-3
`16.7
`108
`102
`103
`111
`106
`99
`MKN7
`5.0
`99
`102
`101
`98
`107
`91
`KATO III
`8.3
`110
`132
`123
`125
`122
`107
`COLO201
`6.7
`96
`97
`99
`100
`98
`98
`SW1417
`a Cells were seeded in 96-well microtiter plates and allowed to adhere before the addition of differ-
`−1. Monolayers were
`ent anti-p185HER2 monoclonal antibodies at a final concentration of 10μgml
`stained with crystal violet dye after 5 days for determination of relative cell proliferation. Each
`group consisted of 8–16 replicates, with the coefficient of variation for each group always less than
`12%
`b Levels of anti-p185HER2 expression were measured by fluorescence-activated cell sorting, relative
`to the 184 mammary epithelial cell
`
`antibodies were active in monolayer proliferation assays. The monoclonals induced
`both inhibitory and stimulatory effects on cell proliferation, and response data
`were dependent on the cell line. For instance, the monoclonal 7F3 had the greatest
`antiproliferative effect of all antibodies tested vs. the HER2-overexpressing BT474
`breast tumor cell line in monolayer proliferation assays. However, 7F3 also stimu-
`lated growth of some tumor cell lines (e.g., COLO201). The cell lines shown in the
`box in Table 1 had high levels of expression of p185HER2, but were less sensitive
`than expected. The explanation for this difference of activity between cell lines
`and among antibodies is not known (see Sect. 6.1). Most of the antibodies with
`an antiproliferative effect on HER2-overexpressing tumor cells were found to stim-
`ulate growth of nontumorigenic fibroblasts (Table 1). Overall, the best correlation
`between p185HER2 expression and growth inhibitory activity in vitro (in these exper-
`iments, and in soft agar assays) was with the monoclonal 4D5 (muMAb4D5). Based
`upon this correlation, and animal xenograft studies (Park et al. 1992), future work,
`including humanization of antibody, focused on muMAb4D5. Follow-on preclinical
`studies have also been conducted with the 2C4 monoclonal antibody (muMAb2C4).
`The rationale for clinical testing of the humanized muMAb2C4 (Pertuzumab) is
`based upon its ability to bind to the dimerization domain of p185HER2 and prevent
`
`

`

`192
`
`H.M. Shepard et al.
`
`Fig. 5 Relationship between p185HER2 expression and growth inhibition mediated by muMAb4D5
`on human breast cell lines, normal and tumor (open circles), and other types of tumor cells over-
`expressing p185HER2 (filled circles, Lewis et al. 1993)
`
`receptor multimerization (Adams et al. 2005). Successful preliminary data with Per-
`tuzumab in combination with chemotherapy in a subset of ovarian cancer patients
`have recently been reported (Makhija et al. 2007).
`The data shown in Table 1 show that overall muMAb4D5 has the best activity vs.
`tumor cells which overexpress p185HER2.
`Figure 5 shows the nonlinear relationship between response to muMAb4D5 and
`p185HER2.
`There are two aspects of this nonlinear relationship which are particularly note-
`able: (1) the inflection point of the curve (3–5 × overexpression) is very similar
`to the level of expression that predicts more aggressive breast and ovarian cancer
`(Slamon et al. 1989, 1987); and (2) there are a number of cell types which fall
`above the best-fit regression line. These cells are “inherently resistant” to the growth
`inhibitory effects of muMAb4D5. Several theories have been advanced to help
`explain the inherent resistance. Probably the most common of these mechanisms
`is tumor cell co-expression of the human EGFR (HER) family members (Sergina
`et al. 2007).
`
`2.5.2 Overcoming Tyrosine Kinase Inhibition of Macrophage
`(TNF-α)-Mediated Tumor Cell Cytotoxicity
`
`The clear relationship between tyrosine kinase activation and resistance to TNF-α
`(Sect. 2.5.1) predicted that downregulation of tyrosine kinase activity could enhance
`the antitumor effect of TNF-α. To test this we treated breast tumor cells in culture
`with muMAb4D5 in the presence or absence of rHuTNF-α (Fig. 6).
`In most HER2-overexpressing cell types, the combination of rHuTNF-α and
`muMAb4D5 results in an additive antiproliferative effect, and in others the effect
`
`

`

`Herceptin
`
`193
`
`Fig. 6 MuMAb4D5 treatment sensitizes p185HER2 breast cancer cells to the cytotoxic effects of
`rHuTNF-α. Monoclonal antibody 4D5 sensitizes breast tumor cells to the cytotoxic effects of TNF-
`α. Cells were plated in 96-well microdilution plates (4× 104 cells per well) and allowed to adhere
`−1) or antihepatitis B surface antigen monoclonal antibody 40.1.H1
`for 2 h. MuMAb 4D5 (5μgml
`−1) was then added for a 4-h incubation prior to the addition of TNF-α to a final concen-
`(5μgml
`−1. After 72 h, the monolayers were washed twice with PBS and stained with
`tration of 104 units ml
`crystal violet dye for determination of relative cell proliferation. In addition, some cell monolayers
`were stained with crystal violet following adherence in order to determine the initial cell density
`for comparison with cell densities measured after 72 h (Hudziak et al. 1989)
`
`is more pronounced. In any case, it is likely that part of the in vivo activity of
`Herceptin stems from the renewed ability of macrophages to inhibit the growth of
`HER2-overexpressing cancers.
`
`2.6 In Vivo Proof of Concept with MuMAb4D5
`
`2.6.1 Tumor Xenograft Studies with MuMAb4D5
`
`Initial studies of the in vivo antitumor activity of muMAb4D5 were performed in
`nude mice bearing human tumor xenografts. Human breast or ovarian tumors that
`overexpressed p185HER2 were implanted into the subrenal capsule, and tumors were
`permitted to form. Mice were then treated intravenously with muMAb4D5 or an
`irrelevant isotype-matched control monoclonal antibody, muMAb5B6. Tumors were
`excised and weighed to evaluate effects on growth. Experimental results in models
`of ovarian (see Table 2) and breast cancer (see Park et al. 1992) demonstrated both
`target-specific and dose-dependent antitumor effects of muMAb4D5.
`
`

`

`194
`
`H.M. Shepard et al.
`
`Table 2 MuMAb4D5 inhibits growth of human ovarian tumor xenografts in mice
`−1)a
`Total dose (mg kg
`Tumor weight (mg)b
`Material
`1, 288± 865
`−
`PBS
`1,416± 483
`90.9
`Control IgG (MuMAb5B6)
`1,187± 825
`1.5
`MuMAb4D5
`1,026± 330
`3.0
`MuMAb4D5
`1,287± 919
`7.5
`MuMAb4D5
`812± 669
`15.0
`MuMAb4D5
`715± 529
`36.4
`MuMAb4D5
`698± 174
`90.9
`MuMAb4D5
`a Human ovarian tumors were implanted in the subrenal capsule of athymic mice. Monoclonal
`antibodies were administered as equally divided doses on days 12, 15, and 18 post tumor implan-
`tation
`b Tumors were excised on day 22 and weighed wet. Data are mean ± SD (n = 8/group; Maneval,
`Pegram, Shepard, and Slamon, unpublished data)
`
`The subrenal capsule nude mouse model was also used to characterize the biodis-
`tribution of muMAb4D5. 125I-labeled muMAb4D5 was injected intravenously to
`tumor-bearing mice, and mice were sacrificed at 5 min, 3 h, 24 h, 3 or 7 days. As
`a control, a separate cohort of tumor-bearing mice received 125I-muMAb5B6, and
`mice were sacrificed at 24 h or 7 days.
`Whole body sagittal sections were exposed to film together with 125I standards
`to generate autoradiograms (Fig. 7). Tumor accumulation was negligible 5 min after
`dosing of 125I-muMAb4D5, but continually increased over the initial 24-h interval
`(see arrows in Fig. 7). Tumor-to-blood ratios increased throughout the 7-day study.
`In contrast to 125I-muMAb4D5, tumor accumulation of 125I was not evident after
`administration of the radiolabeled control antibody. These results demonstrated the
`in vivo localization of muMAb4D5 to tumors overexpressing p185HER2. Further
`studies with the radiolabeled, humanized version of muMAb4D5 demonstrated that
`tremendous efficacy is achievable in these tumor models by combining the pref-
`erential localization and internalization by antibody with the cytotoxic effects of
`radioimmunotherapy.
`These initial experiments with muMAb4D5 led to more comprehensive studies
`demonstrating the antitumor effects of therapy directed at p185HER2.
`
`3 Proof of Concept for Therapeutic Antibodies
`in Solid Cancer
`
`3.1 The Controversy About Making a Successful Monoclonal
`Antibody Therapeutic vs. Solid Cancer
`
`By early 1990, a number of clinical studies had been performed with murine mon-
`oclonal antibodies (muMAb), yet only one agent was approved for clinical use
`
`

`

`Herceptin
`
`195
`
`Fig. 7 Tumor localization of muMAb4D5. Athymic mice bearing human tumors in the subrenal
`capsule received a single intravenous (IV) dose of 125I-muMAb4D5 (4D5) or control antibody
`125I-muMAb5B6 (5B6). Autoradiograms were generated by exposing sagittal 20 μm sections to
`film (nose at left in each panel). Radioactivity was evident in well-perfused tissues (e.g., heart,
`liver) at 5 min and 3 h post dosing. Tumor (T →) uptake of 125I was evident by 24 h and 7 days. In
`contrast, no specific localization of radioactivity was detected in the tumor of mice that received
`125I-muMAb5B6 control (Maneval et al. 1991b)
`
`(OKT-3). Despite more than a decade of intense research, no antibody-based drug
`had yet been approved for use in oncology. However, a wealth of information on
`the clinical use of murine antibodies, and the factors influencing effective antibody
`delivery to tumors was available (for perspective see Blumenthal et al. 1990; Gold-
`enberg 1991). The development of an anti-p185HER2 antibody (Herceptin) required
`an effective strategy to address key challenges including: (1) the neutralizing effect
`of human antimouse antibodies (HAMA) on the serum pharmacokinetics of active
`antibody; (2) the inefficient in vivo delivery of macromolecules to solid tumors; and
`(3) the potential for adverse events due to the specific binding of nontumor tissues
`in humans.
`
`3.2 MuMAb4D5 Therapy for Breast Cancer
`
`The initial in vitro and in vivo studies described above for muMAb4D5 indicated the
`potential therapeutic utility of this antibody for the treatment of breast cancer. More
`
`

`

`196
`
`H.M. Shepard et al.
`
`extensive preclinical investigation at UCLA and Genentech provided additional
`support for muMAb4D5 therapy. Engineered cell lines were generated with differ-
`ential expression of p185HER2 (Chazin et al. 1992). Subcutaneous xenografts with
`these cells provided an in vivo model to demonstrate the activity of muMAb4D5 and
`the potential synergy when added to chemotherapy (Pietras et al. 1994). Receptor-
`mediated uptake and concentration of radiolabeled muMAb4D5 supported the
`potential for radioimmunotherapy (DeSantes et al. 1992; Maneval et al. 1992)
`A series of preclinical studies was also completed in nontumor bearing animals to
`characterize the pharmacokinetics and biodistribution of muMAb4D5. These stud-
`ies provided a dosing rationale for subsequent investigation of efficacy and safety.
`Consistent with expectations for murine monoclonal antibodies, muMAb4D5 was
`cleared slowly from the blood of all species tested (mice, rats, rabbits, and cynomol-
`gus monkeys; Maneval et al. 1991c). Average terminal half-life ranged from 85 h in
`monkeys to 459 h in mice. Peak circulating concentrations of muMAb4D5 indicated
`an initial volume of distribution approximately equal to the plasma volume. A mon-
`key antimouse antibody (MAMA) response was detected within 3 weeks of dosing
`in a majority of monkeys treated with muMAb4D5 (Fig. 8). The MAMA response
`occurred as early as 12 days after single intravenous injection and corresponded to
`a rapid decline in measurable plasma concentrations of muMAb4D5.
`Efficacy studies with muMAb4D5 indicated the need for sustained concentra-
`tions of anti-p185HER2 antibody for therapeutic benefit. However, primate studies
`
`Fig. 8 Pharmacokinetics of muMAb4D5 in cynomolgus monkeys. Adult female cynomolgus mon-
`−1), and blood samples were
`keys received a single intravenous injection of muMAb4D5 (2.5mgkg
`collected over a 21-day interval. Plasma concentrations of muMAb4D5 were measured by ELISA,
`and data from individual monkeys (N = 3) were plotted vs. time (represented by squares, trian-
`gles, and circles). Monkey antimurine antibodies (MAMA) were detected in the plasma as early as
`day 12 (indicated by solid symbols). The MAMA response cor