Journal of Clinical Immunology, Vol. Jl, No. 3, 1991
`
`Special Article
`
`Monoclonal Antibody Therapy of Human Cancer: Taking
`the HER2 Protooncogene to the Clinic
`
`H. MICHAEL SHEPARD,1'4 GAIL D. LEWIS,1 JAY C. SARUP,1 BRIAN M. FENDLY,1 DANIEL
`MANEY AL, 1 JOYCE MORDENTI, 1 IRENE FIGARI, 1 CLAIRE E. KOTTS, 1
`MICHAEL A. PALLADINO, JR.,1 AXEL ULLRICH,2 and DENNIS SLAMON3
`
`Accepted: January 22, 1991
`
`the tumor site and can inhibit the growth of human tumor
`
`
`p185HER2• Modulation of
`
`xenografts which overexpress
`
`pl85HER2 activity by muMAb 405 can therefore reverse
`The HER2 protooncogene encodes a 185-kDa transmem­
`
`brane protein (pl85HER2) with extensive
`homology to the
`
`
`many of the properties associated with tumor progression
`
`epidermal growth factor (EGF) receptor. Clinical and
`mediated by this putative growth factor receptor. To­
`evidence supports a role for overexpression
`experimental
`
`
`gether with the demonstrated activity of muMAb 4D5 in
`of the HER2 protooncogene
`
`in the progression of human
`nude mouse models, these results support the clinical
`
`
`
`breast, ovarian, and non-sma11 cell lung carcinoma. These
`
`application of muMAb 4D5 for therapy of human cancers
`
`data also support the hypothesis that p1g5H£R2 present on
`
`characterized by the overexpression of pl85HER2.
`
`
`the surface of overexpressing tumor cells may be a good
`KEY WORDS: HER2; neu; TNF-o:; monoclonal antibody ther­
`
`
`
`target for receptor-targeted therapeutics. The anti­
`p l85HER2 murine monoclonal antibody (muMAb)
`apy.
`405 is
`
`one of over JOO monoclonals that was derived following
`
`immunization of mice with cells overexpressing
`BACKGROUND: THE HER2 PROTOONCOGENE
`p185l:lER2• The monoclonal antibody is directed at the
`AND HUMAN CANCER
`
`extracellular (ligand binding) domain of this receptor
`
`kinase and presumably has its effect as a result of
`tyrosine
`Ce!Jular protooocogenes encode proteins that are
`
`
`
`receptor function. In vitro assays have shown
`modulating
`
`
`thought to regulate normal ceJJular proliferation and
`
`
`that muMAb 4D5 can specifically inhibit the growth of
`
`
`
`differentiation. Alterations in their structure or am­
`the HER2 pro­
`tumor cells only when they overexpress
`
`
`plification of their expression lead to abnormal
`
`tooncogene. MuMAb 4D5 has also been shown to en­
`
`
`cellular growth and have been associated with car­
`
`hance the TNF-cx sensitivity of breast tumor cells that
`
`
`cinogenesis (1-4). Protooncogenes were first iden­
`
`
`overexpress this protooncogene. Relevant to its clinical
`
`tified by either of two approaches. First, molecular
`
`
`application, muMAb 405 may enhance the sensitivity of
`
`characterization of the genomes of transforming
`
`pl85HER2-overexpressing tumor cells to cisplatin,
`a
`
`
`retroviruses showed that the genes responsible for
`
`chemotherapeutic drug often used in the treatment of
`
`the transforming ability of the virus in many cases
`ovarian cancer. In vivo assays with a nude mouse model
`were altered versions of genes found in the genomes
`
`
`have shown that the monoclonal antibody can localize at
`
`of normal cells. The normal version is the protoon­
`cogene, which is altered by mutation to give rise to
`the oncogene. An example of such a gene pair is
`
`represented by the EGF receptor and the v-erbB
`
`gene product. The virally encoded v-erbB gene
`
`product has suffered truncation and other alter-
`
`1Department of Developmental Biology, Genentech, Inc., 460
`Point San Bruno Boulevard, South San Francisco, California
`94080.
`2Max Planck Institute for Biochemistry, Martinsreid, Germany.
`3Department of Hematology and Oncology, University of Cali­
`fornia, Los Angeles, California 90024.
`"To whom correspondence should be addressed.
`
`117
`
`0271·9142".11./0500-0117$06.50/0 0 1991 Plenum Publishing Corporation
`
`1 of 11
`
`BI Exhibit 1048
`
`

`

`118
`
`SHEPARD ET AL.
`
`direct role in the genesis or progression of disease
`
`ations that render it constitutively active and endow
`
`(15, l6). The data which support a role of HER2
`
`transformation it with the ability to induce cellular
`
`
`overexpression in the basic mechanisms of human
`(5).
`
`cancer are summarized below.
`
`
`The second method for detecting cellular trans­
`
`Amplified expression of p l85HER2 can lead to
`forming genes that behave in a dominant fashion
`
`
`cellular transformation as assessed by morphologi­
`
`
`
`involves transfection of cellular DNA from tumor
`
`cal alterations and growth of p185HER2-overex­
`
`celJs of various species into nontransformed target
`
`pressing celJs in soft agar and in nude mice (17, 18).
`
`
`cells of a heterologous species. Most often this was
`
`done by transfection of human, avian, or rat DNAs
`
`
`In addition, NIH 3T3 fibroblasts overexpressing
`p185HER2 have an increased
`
`resistance to cytotox­
`
`into the murine NIH 3T3 cell line (1-5). Following
`
`several cycles of genomic DNA isolation and re­
`
`
`icity mediated by activated macrophages or recom­
`
`transfection, the human or other species DNA was
`binant human TNF-cx (19), the cytokine that ap­
`pears to be mainly responsible
`
`molecularly cloned from the murine background
`for macrophage­
`
`and subsequently characterized. In some cases, the
`
`mediated tumor cell cytotoxicity (20). This
`
`observation extends also to breast tumor cells,
`
`
`same genes were isolated following transfe.ction and
`
`cloning as those identified by the direct character­
`
`
`which overexpress p l85HER2 (19), and suggests that
`
`high levels of pl85HER2 expression may be related
`
`ization of transforming viruses. In other cases,
`
`novel oncogenes were identifie.d. An example of a
`
`to tumor cell resistance to at least one component of
`
`
`novel oncogene identified by this transfection assay
`
`
`the host's antitumor surveillance armamentarium,
`
`
`is the neu oncogene. It was discovered by Weinberg
`
`
`This work has been the activated macrophage.
`
`
`
`and colJeagues in a transfection experiment in
`
`reviewed previously (21), and similar data have
`
`which the initial DNA was derived from a carcino­
`
`recently been reported for ovarian tumor cell lines
`p185HER2 (22). Further support
`which overexpress
`
`gen-induced rat neuroblastoma (6,7). Characteriza­
`
`tion of the neu oncogene revealed that it had the
`
`for a role of p185HER2 or the related neu oncogene­
`
`
`structure of a growth factor receptor tyrosine ki­
`
`encoded tyrosine kinase in tumorigenesis comes
`
`nase, had homology to the EGF receptor, and
`
`from work with transgenic mice that have been
`
`differed from its normal counterpart, the neu pro­
`
`
`manipulated to overexpress one or the other of
`
`
`tooncogene, by an activating mutation in its trans­
`
`these two related genes. Transgenic mice express­
`membrane domain (8).
`
`ing the activated form of the rat neu protooncogeoe,
`The association of the HER2 protooncogene with
`
`
`
`under the control of a steroid inducible promoter,
`
`
`cancer was established by yet a third approach, that
`uniformly develop mammary carcinoma (23). In
`
`is, its association with human breast cancer. The
`
`another transgenic mouse model the HER2 pro­
`
`
`HER2 protooncogene was first discovered in cDNA
`
`
`
`tooncogene product, ''activated'' by point mutation
`
`libraries by virtue of its homology with the EGF
`
`analogous to the rat neu oncogene product, or an
`
`unaltered form of the HER2 protooncogene,
`
`
`
`receptor, with which it shares structural similarities
`has
`
`been expressed in mice (24). The main malignancies
`
`
`throughout (5). When radioactive probes derived
`from the cDNA sequence encoding p185HER2 were
`induced in this model were either lung adenocarci­
`used to screen DNA samples derived from breast
`noma or lymphoma but not mammary carcinoma.
`
`While it is not known why the different transgenic
`
`
`cancer patients, amplification of the HER2 protoon­
`cogene was observed in about 30% of patient sam­
`
`
`mouse models give such distinct results, the latter
`ples (9). Further studies have confirmed this origi­
`
`
`model may be of particular significance given the
`
`recent report of an association between p185HER2
`
`
`nal observation and extended it to suggest an
`
`
`overexpression and poor prognosis in nonsmall celJ
`
`
`important correlation between HER2 protoonco­
`
`
`gene amplification and/or overexpression and wors­
`
`lung cancer (14). These differing results suggest
`
`
`
`some difference in the activity of activated neu and
`
`
`
`ened prognosis in ovarian cancer and non-small cell
`
`HERZ-encoded tyrosine kianses, although effects
`lung cancer (10-14).
`
`due to mouse strain differences cannot be excluded.
`
`The association of HER2 amplification/overex­
`
`
`The structural similarities between p l85HER2 and
`
`
`pression with aggressive malignancy, as described
`of p185HER2
`
`the EGF receptor suggest that function
`
`above, implies that it may have an important role in
`
`
`
`may be regulated similarly to the EGF receptor. In
`
`progression of human cancer; however, many tu­
`one expects that the tyrosine kinase
`particular,
`
`mor-related cell surface antigens have been de­
`
`
`
`activity associated with the cytoplasmic domain of
`scribed in the past, few of which appear to have a
`
`Journal of Clinical Immunology, Vol. 1I, No. 3, 1991
`
`2 of 11
`
`BI Exhibit 1048
`
`

`

`Fig. 1. Suppression of the TNF cytotoxic response by activation of p 1 g5HER2. The schematic shows both the TNF
`cytotoxic pathway (top) and the pl85HER2-stimulated cell proliferation/transformation pathway (bottom). Signaling
`from the TNF receptor following interaction with TNF has not been characterized. Binding of ligand to p185HER2
`is shown to activate the receptor-associated tyrosine kinase activity, resulting in stimulation of cellular
`proliferation and suppression of the tumor cell cytotoxic response to TNF.
`
`the receptor would be ligand activated. This pro­
`posal receives support from recent work describing
`a ligand for p185HER2 (25). These data lead to a
`model (Fig. 1) wherein antagonists that downregu­
`late the function of p185HER2 should have the effect
`of inhibiting growth of tumor ce!Js dependent upon
`pJ85HER2 function and of increasing the sensitivity
`of such tumor cells to TNF-o:. By analogy with
`previous work done with two related tyrosine ki­
`nases, the EGF receptor (26) and the activated neu
`protooncogene product (27), we hypothesized that
`monoclonal antibodies targeted to the extracellular
`domain of p185HER2 may have the desired proper­
`ties.
`
`DERIVATION OF muMAb 4D5
`
`A family of monoclonal antibodies focused
`against the extracellular domain of pl85HER2 were
`prepared (28). To do this, an NIH 3T3 fibroblast cell
`line that overexpresses p l85HER2 [NIH 3T3/HER2-
`3-400 (18)] was used to immunize BALB/c mice.
`The mice were subsequently boosted with NIH
`3T3/HER2-3-400 and, finally, with a preparation
`
`enriched for pl85HER2 by wheat germ agglutinin
`chromatography of membrane extracts of this cell
`line. Following splenocyte fusion with a mouse
`myeloma partner, the hybridomas were cultured in
`96-well microtiter plates. Hybridomas positive for
`anti-p 1s5HER2 activity, but with little or no anti­
`EGFR activity, were detected by ELISA (Fig. 2). A
`critical property of an anti-p185HER2 monoclonal
`antibody with potential for therapy would be its
`lack of cross-reactivjty with the closely related
`EGF receptor, which is expressed at elevated levels
`in multiple tissues (29). To select further monoclo­
`nal antibodies with this characteristic, a number of
`assays were performed, including immunoprecipita­
`tion assays utilizing in vivo labeled EGF receptor
`and p1s5HER2 (Fig. 3A) and FACS analysis of
`antibody binding to tumor cells overexpressing ei­
`ther p185HER2 or the EGFR (Fig. 3B). The screen­
`ing results are summarized in Table I. Based upon
`these results, nine of the p185HER2 monoclonal
`antibodies were chosen for further characte1ization,
`including a cell growth inhibition assay utilizing the
`SK-BR3 human breast adenocarcinoma cell line,
`which greatly overexpress p185HER2. The monoclo-
`
`.Journal of Clinical Immunology, Vol. II, No. 3, 1991
`
`3 of 11
`
`BI Exhibit 1048
`
`

`

`120
`
`SHEPARD ET AL.
`
`1.B
`
`1.6
`
`1.4
`
`1.2
`
`1.0
`
`O.B
`
`0.6
`
`0.4
`
`0.2
`
`0.0
`
`E
`c
`N
`0)
`""'
`
`0
`
`0
`
`2C4 7C2 7F3 2H11 3H4 3EB 405 SBB 6E9 703 586 40.1H1
`
`MAb (1 ug/ml)
`Fig. 2. ELISA screening of anti-pl85HERi monoclonal antibodies. Results shown measure the relative
`reactivities of the purified anti-pl85HER2 monoclonal antibodies (added to l µ,g/ml) with membrane extracts
`enriched in EGF receptor (open bars; from A43l squamous carcinoma cells) or enriched in pl85ffER2 (filled
`bars; from NlH 3T3/HER2-3-400).
`
`nal antibody, muMAb 405, was clearly the most
`effective of the group in this assay (Table II).
`The initial results characterizing the growth in­
`hibitory activity of these monoclonal antibodies
`were extended by comparing them for activity
`against a battery of human breast and ovarian tumor
`cell lines that expressed varying levels of p 185HER2.
`These results reveal that the monoclonal antibodies
`can be growth inhibitory, they may have no affect
`on cell proliferation, or they may stimulate the
`proliferation of breast tumor cells. Gwwth inhibi­
`tion appears to depend upon overexpression (Table
`III). This property, in particular, is shared by the
`monoclonal antibodies 405 and 3H4. These mono­
`clonal antibodies may exert their effects on cell
`growth by similar mechanisms since they compete
`for binding to the receptor (Tables I and Ill) (28)
`and, therefore, may recognize the same or overlap­
`ping epitopes. The other monoclonal antibodies
`vary in their ability to inhibit proliferation, but 7C2
`and 6E9 are consistently less active in this respect.
`
`The potent growth inhibitory activity of 2C4 for
`MDA-MB-175 breast tumor cells is not understood
`at present but may represent cross-reactivity with
`another receptor expressed on these cells. Simi­
`larly, the properties that distinguish 7C2 from the
`other antibodies with regard to its ability to stimu­
`late the proliferation of several of the tumor cell
`lines shown in Table III has not been determined.
`The 6£9 monoclonal antibody has been shown to
`bind to the extracellular domain of p l85HER2, al­
`though only to a subset of receptors on the surface
`of SK-BR-3 tumor cells (30). The functional signif­
`icance of this subset of receptors is unclear. In
`addition to its activity on breast tumor cells, which
`overexpress p185HER2, muMAb 405 is also clearly
`the most active of the monoclonal antibodies with
`respect to its ability to inhibit growth of SKOV-3, a
`human ovarian adenocarcinoma cell line that over­
`expresses p185HERi (Table Ill). Experiments are
`currently planned to try to understand in more
`detail how these monoclonal antibodies may exert
`
`Journal of Clinical Immunology, Vol. 11, No. 3, 1991
`
`4 of 11
`
`BI Exhibit 1048
`
`

`

`MONOCLONAL ANTIBODY THERAPY OF HUMAN CANCER
`
`121
`
`A.
`
`8.
`
`CVN
`
`HER2
`HER1
`.,....
`.,....
`.,....
`::i::
`::i::
`0
`�
`I/) co
`I/) co
`I/) co
`ci 0 0 ci 0 0 ci 0 0
`v v
`v v
`v v
`5
`6
`7 8
`9
`2
`1
`3 4
`
`-185
`-110
`
`-
`0 ....
`Q)
`.0
`E
`:::J
`z
`�
`�
`Q)
`CI:
`
`!
`
`\
`
`I � it
`/\
`\ ;
`!
`'
`'
`
`!
`
`I
`
`104
`103
`102
`10
`0.1
`Relative Fluorescence Intensity
`Fig. 3. MuMAb 4D5 does not cross-react with the EGFR. (A) Iromunoprecipitation of metabolically Jabelelled NIH 3T3 cells
`transfected by control plasmid (CVN), by a plasmid encoding the EGFR (HERi) or a plasmid encoding pl85HER2 (HER2).
`MuMAb 40. l.HI is directed against hepatitis B surface antigen (lanes I, 4, 7); muMAb 4D5 is directed against pl85H£Rz
`(lanes 2, 5, 8); muMAb 108 is directed against the EGFR (lanes 3, 6, 9). (B) Fluorescence-activated cell sorter histograms
`of muMAb 40.1.Hl (solid lines) or muMAb 4D5 (dotted lines) reacted with SK-BR-3 tumor cells (approx. 2 x 106 receptors
`per cell; upper panel) or the same antibodies reacted with the A431 squamous carcinoma cell line (approx. 2 x 106 EGFR
`per cell; lower panel).
`
`distinct effects on tumor cell proliferation. The in
`vitro results summarized in Table Ill clearly show
`that when the monoclonal antibodies are compared
`for efficacy, as measured by their abilities to inhibit
`growth of breast and ovarian tumor cells overex­
`pressing p185HER2, muMAb 4D5 is usually the most
`potent and is therefore a good candidate for further
`characterization in other models that may be pre­
`dictive of its efficacy in human clinical trials. Inter­
`estingly, the most dramatic activity of the antibody
`is seen in cell lines that overexpress greater than
`fivefold the level observed in MCF-7 breast tumor
`cell lines [a low expressor control cell line; Table III
`(19)]. Patients who overexpress greater than five­
`fold the normal level of p 1 g5HER2 have been shown
`to have a very poor prognosis (10). These results
`will aid in choosing patients who are most likely to
`respond in clinical trials.
`The model depicted in Fig. 1 predicts that down­
`regulation of p l85HER2 by a monoclonal antibody or
`
`other reagent should result in decreased ceUular
`proliferation, as shown in Table III, but also in­
`creased sensitivity to TNF-a. Results of experi­
`ments in which tumor cells overexpressing
`pl85HER2 were treated with muMAb 4D5 or a
`control monoclonal antibody, alone and in combi­
`nation with TNF-a, suggest the validity of this
`model (Fig. 4) (31). MuMAb 4D5 treatment of
`breast tumor cells overexpressing p 1 g5HER2 re­
`sulted in enhanced sensitivity of these cells to
`TNF-a. The growth and the TNF-a sensitivity of
`normal cells or tumor cells that do not overexpress
`the receptor were unaltered.
`In addition to the relationship between TNF-a
`resistance and p185HER2 overexpression, a possible
`relationship between protoncogene expression and
`resistance to the chemotherapeutic drug cisplatin
`has been investigated. A correlation between HER2
`protooncogene overexpression and resistance to
`chemotherapeutic drugs rests on the grounds that
`
`Journal of Clinical Immunology, Vol. 11, No. 3, 1991
`
`5 of 11
`
`BI Exhibit 1048
`
`

`

`122
`
`MAb
`
`4D5
`2C4
`2Hll
`3E8
`3H4
`SB8
`6E9
`7C2
`703
`7F3
`
`Isotype
`
`IgGl,k
`IgGl,k
`IgG2a,k
`Ig02a,k
`IgGl,k
`IgGl,k
`lgGl,k
`IgGJ,k
`IgGl,k
`JgGl,k
`
`EGFR
`
`EGFR
`
`Table 1. Summary Table of Monoclonal Antibodies Described
`RIP"
`ELISA0
`pJ85HER2
`++
`+++
`+
`+
`+
`+
`++
`++
`++
`+++
`
`SHEPARD ET AL.
`
`Epitopec
`
`FACSd
`
`I(p/c)
`F(p/c)
`H(p/c)
`H(p/c)
`l(p)
`nd(p)
`nd(p)
`G(p)
`F(p/c)
`G/F(p/c)
`
`+++
`+++
`++
`+++
`+
`+
`
`+++
`+++
`+++
`
`pl85HER2
`++
`++
`++
`+++
`+
`++
`+
`++
`++
`+++
`
`•summary of OD 492 nm:(-) <0.1; (+) 0.11-0.50; (++) 0.51-1.0; (+++) >1.0.
`bSummary of autoradiography from immunoprecipitations: (-) bands equal to negative control; ( +) weak bands but darker than negative
`control; ( + +) moderately exposed bands; ( + + +) strongly exposed bands.
`ci..etters were assigned to represent individual epitopes A through I (nd, not done). MAbs were considered to share an epitope if each
`blocked binding of the other by 50% or greater in comparison to an irrelevant MAb control. The epitope composition recognized by
`immunoprecipitations with each MAb from tunicamycin-treated cells is shown. The letters p, c, or p/c in parentheses indicate that the
`monoclonal antibody binds only to the polypeptide (p), the carbohydrate (c), or both (p/c) moieties in the extracellular domain of
`p!85HER2.
`dfluorescence staining of SK-BR-3 cells by the anti-pl85HER2 monoclonal antibodies:(-) MAbs equal to the negative control MAbs; ( +)
`1- to 9-fold higher than the negative controls;(++) 10· to 99-fold higher than the negative controls;(+++) >100-fold higher than the
`negative controls.
`
`patients exhibiting overexpression appear to have a
`worsened prognosis, especially in ovarian cancer
`(10, 13). In addition, recent work with the EGF
`receptor (32) has indicated that when the anti­
`EGFR monoclonal antibody 108.4 was added to­
`gether with cisplatin, the antitumor effect of the
`antibody was greatly enhanced. Because the 108.4
`monocJonal antibody and muMAb 4D5 appear to
`share the ability to inhibit soft agar growth of tumor
`
`Table II. Inhibition of SK-BR-3 Proliferation by Anti-pl85HBRZ
`Monoclonal Anlibodies0
`
`Monoclonal
`antibody
`
`4D5
`7C2
`2C4
`7D3
`3E8
`6E9
`7F3
`3H4
`2Hll
`40.1 Hie
`4F4
`
`Relative cell
`proliferationb
`
`44.2 ± 4.4
`79.3 ± 2.2
`79.5 ± 4.4
`83.8 ± 5.9
`66.2 ± 2.4
`98.9 ± 3.6
`62.l ± 1.4
`66.5 ± 3.9
`92.9 ± 4.8
`105.8 ± 3.8
`94.7 ± 2.8
`
`0SK-BR-3 breast tumor cells were plated at a density of 4 x Hf
`cells per well into 96-well microdilution plates, allowed to
`adhere, and then treated with monoclonal antibody (10 µg/ml).
`bRelative cell proliferation was determined by crystal violet
`staining of the monolayers after 72hr. Values are expressed as a
`percentage of results with untreated control cultures (100%).
`ccontrol monoclona.1 antibodies 40.lHI and 4F4 are directed
`against hepatitis B surface antigen and human interferon-"(,
`respectively (27).
`
`cells overexpressing their respective receptors, it
`seemed possible that such an interaction may also
`occur in the HER2 protooncogene system. The in
`vitro results (Fig. 5) show that treatment of SK­
`BR-3 breast tumor cells with muMAb 4D5 enhances
`their sensitivity to cisplatin.
`
`IN VIVO PRECLINICAL EFFICACY
`
`A critical part of the rationale supporting the
`application of muMAb 4D5 to human cancer ther­
`apy is its ability to inhibit the growth of tumor cells
`overexpressing p185HER2 in vivo. A human tumor
`xenograft model was used to test this property of
`muMAb 405 and to compare its activities with
`those of the other monoclonal antibodies in a rele­
`vant model of human disease. In this model, a
`human breast tumor, characterized with respect to
`HER2 protooncogene amplification and expression,
`was grafted into the subrenal capsules of nude mice.
`Therapy was initiated t week postimplantation. In
`order to be active in this model, the monoclonal
`antibody must be able to localize to the overex­
`pressing tumor cells in the lesion and subsequently
`exert a growth regulatory effect mediated through
`pl85HER2. Growth inhibition occurs only with tu­
`mors that overexpress the receptor. Heterotrans­
`plants (approximately 1 mg) of Murray breast tumor
`[a high expresser of the HER2 gene product (10)]
`were implanted into the subrenal capsule of 48
`
`Journal of Clinical Immunology, Vol. I 1, No. 3, 1991
`
`6 of 11
`
`BI Exhibit 1048
`
`

`

`MONOCLONAL ANTIBODY THERAPY OF HUMAN CANCER
`
`123
`
`Relative
`pJ85HER2
`expression°
`
`Table Ill. Inhibition of Human Breast and Ovarian Tumor Cell Growth by Monoclonal Antibodies Directed Against the Extracellular
`Domain of pJg5HER2
`Cell proliferation (% control)b
`2C4d
`7F3d
`4D5c
`3H4c
`Cell line
`101
`97
`94
`101
`MCF7
`l
`3
`ZR-75-1
`104
`100
`ll3
`106
`84
`61
`4
`MDA-MB-175
`24
`48
`7
`62
`91
`MDA-MB-453
`84
`68
`73
`17
`68
`60
`MDA-MB-361
`65
`20
`20
`53
`23
`BT474
`25
`56
`42
`33
`66
`SK-BR-3
`64
`77
`85
`87
`91
`17
`SK-OV-3
`"Based on FACS assay using muMAb 405 and fluorescence-labeled goat anti-murine lgGJ polyclonal antibody.
`bFive-day assay with 10 µg/ml of indicated monoclonal antibody (SE, -10%). Other methods as described in the footnotes to Table II.
`c405 and 3H4 define epitope "[."
`d2C4 and 7F3 will partially block one another, 2C4 is assigned epitope "F," and 7F3 is assigned epitope "F/G."
`'7C2 defines epitope "G" and will partially block 7F3 binding.
`f6E9 epitope determination not done.
`
`7C2'
`106
`149
`87
`78
`ll3
`74
`92
`97
`
`6E9'°
`110
`113
`103
`101
`J13
`94
`105
`99
`
`athymic mice on day 0. Groups of eight animals
`were injected intravenously with tissue culture­
`derived muMAb 4D5 (36.4 mg/kg), PBS, or control
`monoclonal antibody, muMAb 5B6 (directed
`against gpl20; 36.4 mg/kg), as single agents in
`equally divided doses on days 7, 10, and 13. Four
`mice from each group were sacrificed on day 20,
`and the remainder of the animals were sacrificed on
`day 34. Tumor sizes were measured using ocular
`micrometer and gravimetric techniques. A sum­
`mary of the tumor weights (mean ± SD) from
`animals sacrificed on days 20 and 34 is shown in
`Table IV. On day 20, average tumor weights of
`animals receiving muMAb 4D5 were significantly
`less than those receiving the same dose of the
`control antibody muMAb 5B6. Interactive effects
`between muMAb 4D5 and cisplatin have also been
`observed in this model (33). These studies in athy­
`mic mice bearing human breast tumor xenografts
`have demonstrated efficacy and suggested an en­
`hanced effect when muMAb 4D5 is given in combi­
`nation with cisplatin.
`
`a modulation of intracellular second messengers,
`including diacylglycerol. Diacylglycerol (DAG) is a
`product of phospholipase C breakdown of phos­
`phatidylinositol-4,5-bisphosphate. It is a cofactor
`
`A SK·BR-3
`
`B SK-BR-3 C. MDA-MB-175-Vll
`
`1.5
`
`1.0
`
`0.5
`
`....
`
`D. MDA·MB-231
`
`E. HBL-100
`
`F. T24
`
`c 0
`·� cu
`�
`0 ....
`a.. 0.0
`Q) 1.5
`{.)
`Cl> >
`·� 1.0
`
`Q)
`a:
`
`0.5
`
`0.0
`
`MECHANISM OF ACTION
`
`The results described above are consistent with
`muMAb 4D5 having receptor antagonist activity.
`Surprisingly, however, muMAb 4D5 treatment of
`SK-BR3 tumor cells stimulates receptor tyrosine
`kinase activity (Table V) (30, 34). In addition, it can
`mediate the phosphorylation of intracellular sub­
`strates by pl g5HER2 (34). Consistent with its ability
`to stimulate receptor activity, muMAb 4D5 treat­
`ment of SK-BR-3 or SK-OV-3 tumor cells results in
`
`Fig. 4. Monoclonal antibody 405 sensitizes breast tumor cells to
`the cytotoxic effects of TNF-a. Filled bars, cell number at
`initiation of the assay; dark cross-hatching, untreated control;
`dark stipples, TNF-a alone; light cross-hatching, MuMAb 405;
`open bars, MuMAb 405 combined with TNF-a. (B) Lack of
`growth inhibition of SK-BR3 tumor cells by nmMAb 40.l HI
`(anti-hepatitis B antigen; light stipples) and failure of the 40. l H
`l to enhance SK-BR-3 tumor cell sensitivity to TNF-a (broken
`cross-hatching). SK-BR-3 and MDA-MB-175-VII overexpress
`p185HER2 (see Table l l l). MDA-MB-231 and HBL-100 are
`breast cell lines which do not overexpress pl85HER2, and 1'24 is
`a nonoverexpressing human bladder carcinoma cell line. The
`assay was performed as described in Ref. 31.
`
`Journal of Clinical Immunology, Vol. 11, No. 3, 1991
`
`7 of 11
`
`BI Exhibit 1048
`
`

`

`124
`
`SHEPARD ET AL.
`
`8
`�
`
`140 A
`� 120
`:0
`co 100 Ol
`5 -
`�
`c
`80
`Q)
`�
`Q)
`0..
`Q)
`>
`".;:! ro
`Q)
`a:
`
`6;
`
`60
`
`40 �
`
`20
`
`0
`
`0
`
`.01
`
`10
`
`100
`
`0
`
`.01
`
`1
`.1
`10
`1
`.1
`Cisplatin (µg/ml)
`Cisplatin (µg/ml)
`Fig. 5. Treatment of SK-BR-3 breast tumor cells with muMAb 405 enhances sensitivity to
`cisplatin. MuMAb 405 (A) or muMAb 6E9 (control; B) and cisplatin were added at the indicated
`concentrations to SK-BR-3 breast tumor cells. The plate cultures were incubated for 3 days and
`relative cell proliferation was determined as described (31). No antibody (0); 0.156 µ.g/ml muMAb
`405 or muMAb 6E5' (•); 0.625 µ.g/ml muMAb 405 or muMAb 6E9 (t:.); 2.5 µ.g/ml muMAb 4D5
`or muMAb 6E9 (A).
`
`''"I
`100
`
`for activation of protein kinase C and has been
`closely associated with growth factor activity (:55).
`As may be predicted from its effect on cell prolifer­
`ation, muMAb 4D5· treatment of SK-BR-3 tumor
`cells results in downregulation of intracellular pools
`of DAG (Table VO (30). This result is consistent
`with overall antagonist activity, as is inhibition of
`tumor cell proliferation. Other data suggest that
`muMAb 405 may inhibit association of ligand with
`the receptor (25). Similar monoclonal antibodies
`have been reported for the EGFR system (26, 28).
`Further work is under way to characterize the
`ligand(s) that binds p185HER2 and the mechanism of
`action of muMAb 4D5.
`While the ability of muMAb 405 to stimulate
`phosphorylation of pl 85HER2 is consistent with an
`agonist of receptor function, it is important to note
`that our current data suggest that it does not behave
`as an agonist in our cell growth assays in vitro or in
`nude mice. The results of an experiment that com­
`pares the effects of muMAb 4D5 on the growth of
`
`Table IV. MuMAb 405 [nhlbits the Growth of a Human Breast
`Tumor (Murray) in Athymic Mice0
`
`Group (n = 4)
`
`Tumor weight (mg)"
`Day 34
`
`Day 20
`
`PBS
`Control .(gG (muMAb5B6)
`muMAb4D5
`
`6.79 :!: 9.79
`7.11 :!::5.48
`1.48 ::!:: 1.10
`0Administered as equally-divided intravenous doses on days 7,
`IO, and 13 post tumor implantation.
`hData are mean :t standard deviation (SD) (n = 4).
`
`36.0 ± 30.7
`88. l ± 91.4
`6.5 :!: 6.4
`
`MCF-7 and SK-BR-3 breast tumor cells are shown
`in Fig. 6. These data demonstrate that muMAb 4D5
`has no effect on the growth of nonoverexpressing
`tumor cells (MCF-7; Tables Ill, VII) at any of the
`doses tested between 0.7 pM and 67 nM. Also,
`whatever allows the muMAb 4D5 to have differen­
`tial effects on overexpressing tumor cells, this dif­
`ference does not lie in different receptor affinities
`for the monoclonal antibody. Table VII clearly
`shows that SK-BR-3 and SK-OV-3, both p185HER2
`overexpressors, which are growth inhibited by 405,
`and MCF-7, which is not, all have similar affinities
`for muMAb 4D5. The clearest difference between
`these cell lines is the number of binding sites per
`tumor cell. These data are consistent with other
`work that has been previously reported with tumor
`cells in monolayer culture or in soft agar (18, 19,
`25).
`A possible mechanistic explanation, which takes
`many of our experimental observations into ac­
`count, is a model in which muMAb 4D5 binds
`
`Table V. Effect of muMAb 405 on Phosphoamino Acid
`Content of p185t1ER2 in SK-BR-3 Cells
`Phospho-
`Phospbo-
`Phospho-
`threonine
`serine
`tyrosine
`cpm0 %b
`%
`cpm
`1.5
`11
`14.0
`827
`
`cpm
`
`564
`3,658
`
`%
`75
`62.0
`
`23.5
`176
`1,429 24.0
`
`Treatment
`
`None
`muMAb4D5
`
`0Phosphoamino acids as cpm are expressed following back­
`ground subtraction (17 cpm for none, 21 cpm for muMAb 405).
`6Percentage of total phosphoamino acids.
`
`J.ournal of Clinical Immunology, Vol. 11, No. 3, 1991
`
`8 of 11
`
`BI Exhibit 1048
`
`

`

`MONOCLONAL ANTIBODY THERAPY OF HUMAN CANCER
`
`125
`
`Table VJ. Effect of muMAb 4D5 or 6E9 Monoclonal
`
`
`Antibodies on sn-1,2-diacylglycerol levels in SK-BR-3 Cells
`
`Table Vll. Monoclonal Antibody Binding to Cultured Human
`
`Adenocarcinoma Cells
`
`Receptor no. Growth
`Cell line Antibody Kd(nM) (sites/cell)
`inhibitionn
`926,650 -*
`428,930 -*
`5,525
`
`SK-BR-3 muMAb4D5 6.0
`SK-OV-3 muMAb4D5 5.0
`MCF7 muMAb4D5 1.2
`
`Time
`
`5 min
`5 min
`5 min
`
`24 hr
`24 hr
`24 hr
`
`pmol sn-1,2-
`Treatment DAG/106 cells % change
`
`Vehicle
`muMAb4D5
`muMAb6E9
`
`111.0 ± 10
`133.2 ± 11.3
`133.3 ± 12.9
`
`Vehicle
`muMAb4D5
`muMAb6E9
`
`98.6 ± 9.6
`62.1 ± 7.4
`92.0 ± 12.7
`
`0
`+20
`+20
`
`0
`-37
`-7
`
`
`
`aGrowth inhibition was measured as described in the footnotes to
`
`
`
`Table II.
`*StatisticaJly significant inhibition of growth (P<0.05) as com­
`
`
`
`
`pared to an untreated control.
`
`with monoclonal antibody (33 nM) or
`nFollowing incubation
`
`
`
`
`vehicle (PBS) control, the reactions were terminated by aspi­
`
`rating the media and adding 1 ml of ice-cold 100% MeOH. Cells
`vations that the monoclonal antibody is not de­
`
`
`were scraped from the plates and transferred to 13 x 100-mm
`
`graded following internalization (30) and, in the
`
`glass tubes containing 1 ml I 00% chloroform. Plates were rinsed
`
`
`with an additional 1 ml of cold methanol, and the rinses
`
`nude mouse experiments, localizes to and remains
`
`
`combined and mixed thoroughly. Following phase separation at
`
`at the tumor site for more than 7 days following a
`for 30 min, 1 ml methanol and I ml NaCl were
`room temperature
`
`single administration of antibody (D.M., personal
`
`added, the samples were centrifuged at 3000 rpm for 5 min, and
`
`the top aqueous layer was aspirated. The remaining organic
`
`
`communication). Such downregulation could result
`
`phase was assayed for sn-1,2-diacylglycerol by standard proce­
`
`
`protein ki­from activation of the serine-threonine
`dures.
`
`nase C, which is known to downregulate the func­
`tion of other receptor tyrosine kinases (5). The
`to p I g5HER2, excludes ligand binding,
`stim­
`tightly
`mechanism of action of muMAb 4D5 remains a
`
`ulates receptor internalization, and downregulates
`
`subject for continuing work.
`
`
`receptor signaling pathways as a result of constitu­
`
`
`
`tive activation of tyrosine kinase activity that re­
`
`sults from nondissociation of the muMAB 4D5/
`p 185HER2 complex during receptor cycling.
`This
`
`
`hypothesis has additional support from our obser-
`110 i
`100 fFI
`90 i
`80'
`
`I
`
`70
`
`c
`0
`"+::
`('Cl
`.....
`�
`0 ....
`a.
`Qi 60
`()
`<l>
`>
`
`SUMMARY AND CONCLUSIONS
`
`
`A convincing body of clinical and experimental
`evidence supports the role of p l85HER2 in the pro­
`
`gression of human cancers