Nondinical Studies Addressing the Mechanism of
`Action of Trastuzumab (Herceptin)
`
`Mark X. Sliwkowski, Julie A. Lofgren, Gail D. Lewis, Timothy E. Hotaling, Brian M. Fendly, and Judith A. Fox
`
`HER2 is a ligand-Iess member of the human epidermal
`growth factor receptor or ErbB family of tyrosine Id(cid:173)
`nases. In normal biological systems, HER2 functions as
`a co-receptor for a multitude of epidermal growth fac(cid:173)
`tor-like ligands that bind and activate other HER famlly
`members. HER2 overexpression is observed in a num(cid:173)
`ber of human adenocarcinomas and resuits in consti(cid:173)
`tutive HER2 activation. Specific targeting of these tu(cid:173)
`mors can be accomplished with antibodies directed
`against the extracelluiar domain of the HER2 protein.
`One ofthese antibodies, 405, has been fully humanized
`and is termed trastuzumab (Herceptin¡ Genentech,
`San Francisco, CA). Treatment of HER2-overexpress(cid:173)
`ing breast cancer cell lines with trastuzumab results in
`induction of p27 K1P1 and the Rb·related protein, p 130,
`which in turn significantly reduces the number of cells
`undergoing S-phase. A number of other phenotypic
`changes are observed in vitro as a consequence of tras(cid:173)
`tuzumab blnding to HER2-overexpressing ce lis. These
`phenotypic changes inelude downmodulation 01 the
`HER2 receptor, inhibition of tumor cell growth, re(cid:173)
`versed cytol<ine resistance, resto red E-cadherin ex(cid:173)
`pression
`levels, and reduced vascular endothelial
`growth factor pl'oduction. Interaction of trastuzumab
`with the human immune system via its human immu(cid:173)
`noglobulin G 1 Fe domain may potentiate its antitumor
`activities. In vitro studies demonstrate that trastu(cid:173)
`zumab is very effective in mediating antibody-depen(cid:173)
`dent eell-mediated cytotoxicity against HER2-overex(cid:173)
`pressing tumor targets. Trastuzumab treatment of
`mouse xenograft models results in marked suppression
`of tumor growth. When given in combination with
`standard eytotoxic ehemotherapeutic agents, trastu(cid:173)
`zumab treatment generally results in statistically supe(cid:173)
`rior antitumor efficaey compared with either agent
`given alone. Tai<en together, these studies suggest that
`the meehanism of action of trastuzumab ineludes an(cid:173)
`tagonizing the constitutive growth-signaling properties
`of the HER2 system, enlisting immune cells to attack
`and Idll the tumor target, and augmenting ehemother(cid:173)
`apy-induced eytotoxicity.
`Semin Oncol 26 (suppl 12):60-70. Copyright © 1999 by
`W.B. Saunders Company.
`
`From Genenteeh, Ine, SOlith San Francisco, CA
`All authors are eml)loyees of Genentech, Ine.
`Address eorrespondence and reprint requests to Mari< .'\liwl<owski,
`PhD, Department ofMolec¡¡lar Oneology, Genentech, Ine, 1 DNA
`Way, Mail Stop 63. SOllth San Francisco, CA 94080.
`Copyright © 1999 by W.B. Sallnders Company
`0093-7754/99/2604-1208$10.00/0
`
`T HE EPIDERMAL growth factor (also known
`
`as HER orErhB) family of receptor tyrosine
`kinases are important mediators of cell growth,
`differentiation, and survival.1.2 As shown in Fig 1,
`the receptor family is composed of four distinct
`members, including epidermal growth factor re(cid:173)
`ceptor (EGFR; 1 IERl or ErbBl) , HER2 (neu or
`ErbB2), HER3 (ErbB3), and HER4 (tyro2 or
`ErhB4). Within a given tissue, these receptors are
`rarely, if ever, expressed alone, but are found in
`various combinations.3 At present there are nine
`ligands that bind direct1y to EGFR, HER3, or
`HER4. HER2 also hecomes activated as a result of
`ligand binding to other HER family members. Fur(cid:173)
`thermore, blocking the association of HER2 with
`these receptors inhibits ligand-mediated signal(cid:173)
`ing.4-8 Thus, HER2 is an essential component of
`these receptor complexes, although no ligand has
`been characterized at the molecular level that in(cid:173)
`te~acts with HER2 alone.
`Recently, the HER receptors have been studied
`either alone or in specific combinations. 9-11 Li(cid:173)
`gand-receptor complexes
`that
`include HER2
`appear to be more potent than other receptor
`combinations. Features of HER2-containing com(cid:173)
`plexes that may account for enhanced receptor
`signaling indude the conversion from a low-affin(cid:173)
`ity to a high-affinity growth factor binding site,l2
`which rcsults from a decrease in ligand dissocia(cid:173)
`tion from the complex.5 Second, the interna liza(cid:173)
`tion rate of HER2 receptor complex is nearly an
`order of magnitudc lower than that observed for
`EGFR.13 The higher-affinity binding state and the
`decreased internalization rate may result in a re(cid:173)
`ceptor-ligand complex that resides longer on the
`plasma membrane and prolongs the duration of
`receptor signaling. Third, HER2 is a very active
`tyros.ine kinase14 that also can be constitutively
`activated hy mutation or overexpression. In addi(cid:173)
`tion to transphosphorylating other HER receptors,
`activated HER2 phosphorylates PLCy,15 activates
`the Ras/Raf/MEK/MAPK pathway,16-18 and asso(cid:173)
`ciates with the cytoplasmic tyrosine kinases src 19- 21
`and CHK.22 Heterodimerization of HER recep(cid:173)
`tors23-25 appcars to increase the diversity of cellular
`
`60
`
`Seminars in Onc%gy. Vol 26. No 4. Suppl 12 (August). 1999: pp 60·70
`
`IMMUNOGEN 2049, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`THE MECHANISM OF ACTION OF TRASTUZUMAB
`
`61
`
`neu protein existing in a constitutively activated
`state. T o date, no analogous mutations have be en
`found in human tumors. 28,29,36,)7 Rather, the mo(cid:173)
`lecular alteration that occurs in human cancers is
`overexpression of the normal gene product. Over(cid:173)
`expression genera tes a constitulively activated re(cid:173)
`ceptor and
`leads
`ro
`transformaríon of both
`NIH3T3 cells and human breast cells, suggesting
`that this alteration may playa pathogenic role in
`tumorigenesis. 38-41 Additionally, female transgenic
`animals that are engineered to overexpress the
`nonmutated neu oncogcne in their hreast tissue
`dcvelop breast tumors after puberty with a high
`degree of penetrance.42 Taken together, these data
`suggest that HER2 overexpression has an active
`role in neoplastic transformation.
`
`TARGETING HER2-0VEREXPRESSING
`CANCERS
`
`A number of approaches have been used to
`therapeutically target HER2-overexpressing can(cid:173)
`cers,43 A common approach, based on similar stud(cid:173)
`ies with the dosely related EGFR, has been the
`generation of antibodics that inhibit the growth of
`cells which possess activated HER2/neu recep(cid:173)
`tors,44-49 Analysis of one of these panels of anti(cid:173)
`HER2 monodonal antibodies led to the identifi(cid:173)
`cation of the murine monoclonal, 4D5. 50 This
`antibody recognizes an extracellular epitope (ami(cid:173)
`no acids 529 to 627) in the cysteine-rich II domain
`that resides very close to the transmembrane re(cid:173)
`gion, Treatment of brcast cancer cells with 4D5
`partially blocks heregulin activation of HER2-
`HER3 complexes, as measured by receptor phos(cid:173)
`phorylation assays.6 To allow for chronic human
`administration, murine 4D5 was fully humanized
`to generate trastuzumab (Herceptin; Genentech,
`San Francisco, CA). Solution-phase binding stud(cid:173)
`ies detennined tbat trastuzumah binds to the re(cid:173)
`combinant HER2 extracellular domain with an
`affinity (Kd) of 0.1 nmol/L, which is threefold
`greater than 4D5.51
`
`IN VITRO EFFECTS OF ANTI-HER2
`MONOCLONAL ANTIBODIES
`
`All tbe antibodies generated against HER2 were
`assessed for their ahility to inhibit the anchorage(cid:173)
`dependent growth ofhuman cancer cellHnes. Sev(cid:173)
`eral of these antibodies exhibited antiproliferative
`activity toward cell Hnes that expressed higher(cid:173)
`than-normallevels ofBER2. This observation was
`
`Fig l. Schematic representatlon of the human epidermal
`growth factor receptor (HER) family and their ligands.5
`
`responses to a variety of proteins that serve as
`ligands for thesc receptors.9,1O
`Activation of nonmutated HER2 occurs
`through self-association and by transactivation
`with other related receptor family members. In
`breast cancer, gene amplification results in HER2
`protein levels in tumor cells that are 10- to 100-
`fold greater than that found in the adjacent nor(cid:173)
`mal breast epithelium. 26 Since HER2 is localized
`in the plasma membrane, its diffusion is Hmited to
`two dimensionsj thus, even modera te overexpres(cid:173)
`sion ofHER2 can lead to a constitutively activated
`receptor,27 Soon after the HER2 gene was doned,
`and well before its role as a common subunit was
`appreciated, compelling clinical data were gener(cid:173)
`ated demonstrating that HER2 gene amplification
`predicts a poor clinical outcome,28 Although these
`data were controversial at first, at present the
`general consensus is that HER2 overexpression is a
`indicator for breast can(cid:173)
`negative prognostic
`cer,29,30 Subsequent studies have shown that
`HER2 overexpression correlates with shorter dis(cid:173)
`ease-free interval and overa1l survivapl Moreover,
`comprehensive studies have shown that the pres(cid:173)
`ence of this molecular alteration is predictive of
`aggressive disease regardless of disease stage or
`node status.32-H
`A role for HER2 as an oncogene was first ap(cid:173)
`prcciated in rodent systems. The activatcd form of
`the rodent homolog of the BER2 gene, c-neu, was
`initially identified as an extremely potent trans(cid:173)
`forming oncogene in DNA isolated from rats
`treated with ethylnitrosourea.35 This version of
`the neu oncogene contained a point mutation in
`the transmembranc domain, which resulted in the
`
`IMMUNOGEN 2049, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`62
`
`SLlWKOWSKI ET AL
`
`a[80 extended to ínclude assays performed in soft
`agar or in an anchorage-independent format, and
`similar results were obtained. 52 The most patent
`growth-inhibitory antibody of this panel was 4D5.
`Cell cycle analysis following treatment of SK(cid:173)
`BR-} breast cancer cells, which express high levels
`of HER2, is shown in Fig 2. Trastuzumab Ol' 4D5
`significantly reduces the percentage of SK-BR-}
`cells undergoing S-phase and increases the per(cid:173)
`centage of cclls in 00/01. In contrast, HERI
`antibody treatmcnt does not affect the cell line
`MCF-7, which expresses normal amounts of
`
`HER2. The effcct of these antibodies on the cell
`cycle is dose-dependent, with maximal inhibitory
`activity occurririg at antibody concentfRtions
`greater than 1 fLg/mL.53 To address the molecular
`basis fo! growth arrest, SK-BR-3 cells were treated
`with trastuzumab¡ celllysates were then probed by
`immunoblot analysis wi th reagen ts specific for
`known inhibitors of the cell cycle. As shown in Fig
`3, treatment of SK-BR-3 cells with trastuzumab
`resulted in marked inductiün üf the CDK2 kinasl:
`inhibitor, pI 7K1P1 • Moreover, a similar induction
`of the retinoblastoma-related protein, p130, was
`
`MCF7
`
`A
`
`ControlMAb
`%S=23.88
`%0010.=67.17
`
`B
`
`H
`
`8
`Ñ
`
`!
`
`muMAb4D5
`%S=23.57
`%0.,t0,=67.53
`
`e
`
`H
`
`¡:;¡
`
`!
`
`rhuMAb HER2
`%S=23.76
`%0.,tO,=67.75
`
`SK-BR-3
`
`D
`
`E
`
`ControlMAb
`%S=20.88
`%0 010,=67.13
`
`N
`
`muMAb4DS
`%S=12.41
`%0010.=78.01
`
`F
`
`rhuMAbHER2
`%S=13.87
`%00/0,=76.99
`
`!
`
`o
`
`Fig 2. MCF-7 (A-C) or SK-BR-3 (O-F) cells were plated at a denslty o( 2 x 106/dish in 60 mm X 15 mm culture dishes and allowed
`to adhere overnight. Monolayers were washed with phosphate-buffered saline (PBS) and incubated with media containing trastu(cid:173)
`zumab or control antibodies at a concentration o( 10 fLg/mL. Following a 72-hour incubation, cells were trypsinized, washed with PBS,
`fixed in ice-cold methanol, and stored at -20·C. Fixed cells were then washed twice with PBS and allowed to incubate with RNAse 100
`fLg/mL (Worthington Biochemlcal, Lakewood, NJ) (or 30 minutes at 37°C. Following centrifugation, nuclei were subjected to
`propidium iodide 50 Ilg/mL (Molecular Probes, Inc, Eugene, OR) staining in PBS. Samples were analyzed by flow cytometry (Eplcs
`Elite; Coulter Corporation, Miaml, FL) using Modfit L T software (Verity Software House, Sunnyvale, CA).
`
`IMMUNOGEN 2049, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`THE MECHANISM OF ACTION OF TRASTUZUMAB
`
`63
`
`time
`(h)
`
`Western Blot: anti-p27
`
`o 8 24 30 kD
`·32.5
`- 25
`- 16.5
`
`082430
`
`time
`(h)
`
`Western Blot: anti-p130
`
`kD
`- 175
`
`-83
`- 62
`
`Induction of the cell cycle inhibitor p27 K1P1 and the
`Fig 3.
`retinoblastoma-related p 130 protein in SK-BR-3 ce lis treated
`with trastuzumab. Cells (1.5 x lOs cells/well) were plated in
`six-well dishes and grown to approximately 50% confluence in
`medium containing 10% fetal bovine serum. The medium was
`removed and treatments (trastuzumab, 16 f.tg/mL) were pre(cid:173)
`pared in fresh medium containing 10% fetal bovine serum and
`added to the culture dishes. Incubations were stopped by the
`addition of 0.5 mL sodium dodecyl sulfate sample buffer (5%
`sodium dodecyl sulfate, 2% p-mercaptoethanol, 10% glycerol,
`50 mmol/L Tris HCI, pH 6.8). Samples (20 f.tg of protein/lane)
`were electrophoresed on an 8% to 16% Tris-glyciné gel (Novex,
`San Diego, CA). Protein was transferred from the gel s to
`nitrocellulose. Western blot analysis and detection were done
`following the manufacturer's instructions (ECO; Amersham,
`Buckinghamshire, UK). Primary antibodies used for p27 K1P1
`and the retinoblastoma-related p 130 protein immunoblots
`were rabbit polyclonals (SC-317 and SC-528, respectively;
`Santa Cruz, Santa Cruz, CA) and were used at 0.1 f.tg/mL. Both
`proteins were detected with an anti-rabbit horseradish perox(cid:173)
`idase-conjugate secondary antibody used at a titer of 1: I 0,000
`(ALI 3404; Biosource International, Carmarillo, CA).
`
`also observed. These data are consistent with the
`observation that treatment of HER2-overexpress(cid:173)
`ing cells is antiproliferative and that cytostasis
`results from an inhibition of cell cycle progre ss ion.
`Cells that overexpress HER2 are intrinsically
`resistant to the cytotoxic effects of tumor necrosis
`factor-a.54 This observation suggests that HER2
`overexpression may potentiate tumorigenesis by
`inducing tumor cell resistance to host defense
`mechanisms. Furthermore, when HER2-overex(cid:173)
`pressing cells are treated with the murine mono(cid:173)
`clonal 4D5, they become sensitized to tumor ne(cid:173)
`crosis factor-a treatment. 49
`Molecules involved in cell adhesion are also
`thought to play a critical role in malignant pro(cid:173)
`gression. One of these molecules, E-cadherin, plays
`a central role in maintaining epithelial cell mor(cid:173)
`phology.55 A reduction in E-cadherin synthesis
`can be correlated with increased invasiveness.56
`Overexpression of HER2 in normal human breast
`
`epithelial cells reduces morphogenesis in vitro.57
`These HER2 transfectants express significantly
`lower levels of E-cadherin as well as the a2 inte(cid:173)
`grin subunit. Treatment of these HER2 transfec(cid:173)
`tants with 4D5 restores E-cadherin and a2 inte(cid:173)
`grin to normallevels.58 Taken together, these data
`suggest that HER2 overexpression affects adhesion
`molecule synthesis, which in turn may contribute
`to the metastatic process and malignant progres(cid:173)
`sion. T argeting HER2 overexpression with mono(cid:173)
`clonal antibodies such as 4D5 and trastuzumab
`may be effective in controlling metastatic poten(cid:173)
`tial.
`Angiogenesis is a critical survival function for
`solid tumors.59,60 Vascular endothelial growth fac(cid:173)
`tor is one of the more important mediators of
`tumor angiogenesis.61 Recently, it has been dem(cid:173)
`onstrated that treatment of HER2-overexpressing
`tumor cells with 4D5 de creases vascularendothe(cid:173)
`líal growth factor production.62 Suppression of an(cid:173)
`giogenesis may enhance the activity of anti-HER2
`monoclonal antibody therapy in vivo.
`
`AGONIST PROPERTIES OF TRASTUZUMAB
`
`Several reports have suggested that binding of
`4D5 to HER2 activates the receptor, as monitored
`by an increase in tyrosine autophosphorylation.63
`In addition, c-fos induction64 and increases in ino(cid:173)
`sitol polyphosphate and diacyl glycerol synthesis65
`following 4D5 treatment also have be en reported.
`These agonistic properties are difficult to reconcile
`the potent cytostatic activity of 4D5 or tras tu(cid:173)
`zumab on the growth of human tumor cells that
`overexpress HER2. 6,52
`With recent recognition of the role of HER2 as
`a shared signaling subunit with other HER recep(cid:173)
`tor family members, the lígand-like properties of
`trastuzumab have been re-examined. MCF-7 cells
`express normal/low levels of HER2 and are not
`growth
`inhibited by
`trastuzumab
`treatment.
`MCF-7 cells were treated with trastuzumab, 4D5,
`or recombinant heregulin (rHRG). Cell lysates
`were prepared using denaturing (sodium dodecyl
`,sulfate) 01' nondenaturing (nonionic detergent,
`Triton X-lOO) conditions. These latter conditions
`are frequently used for examination of receptor
`activation or co-immunoprecipitation of media(cid:173)
`tors involved in signal transduction. As shown in
`Fig 4, treatment of MCF-7 cells with rHRG results
`in HER2 tyrosine phosphorylation, regardless of
`the manner in which the cells are lysed. In con-
`
`IMMUNOGEN 2049, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`64
`
`SLlWKOWSKI ET AL
`
`Lysis Conditions:
`
`...
`
`Denaturing
`.o
`ro
`E
`::::l
`N
`::::l Q)
`....... a:
`en
`ro
`.....
`I
`'-
`+-'
`
`"O
`.....
`LO
`e
`o O
`"<t
`ü
`
`+-'
`
`Non-Denaturing
`.o
`ro
`E
`::::l
`"O
`....
`N
`.....
`::l Q)
`'-
`en a:
`e
`LO
`o O ~ I
`.....
`.....
`ü
`"<t
`
`kD
`
`- 175
`
`- 83
`- 62.5
`
`Fig 4. Phosphotyrosine analysis of HER2. MCF-7 cells were seeded in 24-well plates at 1.5 X 10' cells/well and allowed to grow for
`2 days. Before treatment, cells were transferred to medium without serum. Cells were treated for 30 minutes at room temperature
`with 4DS (16 ¡.tg/mL), trastuzumab (16 ¡.tg/mL), or rHRG (300 pmol/L). Cell Iysates were then prepared using denaturing or
`nondenaturing conditions and allowed to remain at room temperature for 3 hours. For nondenaturing conditions, Iysates were
`prepared in 0.5% Triton X-lOO, 50 mmol/L Hepes, 150 mmol/L NaCl, I mmol/L phenylmethylsulfonyl fluoride, 2 mmol/L NaVO., and
`5 KIU/mL aprotinin. Sodium dodecyl sulfate-polyacrylamide electrophoresis buffer was used for denaturing conditions. Samples were
`prepared for electrophoresis by adding an equivalent volume of 5% sodium dodecyl sulfate and 2% /3-mercaptoethanol to each sample
`and then heated at I OO·C for 5 minutes. Following sodium dodecyl sulfate-polyacrylamide electrophoresis, proteins were transferred
`to nitrocellulose membranes and probed with anti-phosphotyrosine monoclonal antibody (#E 120H; Transduction Labs, San Diego,
`CA) linked to horseradish peroxidase used at a final concentration of 0.1 25 ng/mL.
`
`trast, 4D5 or trastuzumab results in HER2 tyrosine
`phosphorylation only when lysates are prepared in
`nonionic detergent. Although biochemically in(cid:173)
`teresting, this phosphorylation event does not ap(cid:173)
`pear to be physiologic. A similar artifact has been
`reported for anti-EGFR monoclonal antibodies. 66
`When similar experiments are performed with SK(cid:173)
`BR-3 cells, which overexpress HER2 and are
`growth inhibited by trastuzumab treatment, the
`results are confounded by a high degree of basal
`tyrosine phosphorylation that exists on HER2.
`Taking this high baseline phosphotyrosine level
`into account, trastuzumab treatment of SK-BR-3
`cells results in a modest but reproducible increase
`in tyrosine phosphorylation content, which is in
`agreement with previous published studies. 63
`T o assess whether trastuzumab treatment causes
`further activation of the HER2 signal transduction
`pathway, the common adapter molecule, shc, was
`examined. MCF-7 or SK-BR-3 cells were treated
`with trastuzumab, 4D5, or rHRG, and celllysates
`were then prepared usingdenaturing conditions.
`Immunoprecipitates of shc were prepared and im(cid:173)
`munoblots were then probed with anti-phospho(cid:173)
`tyrosine antibody to determine the level of shc
`tyrosine phosphorylation. As shown in Fig 5,
`rHRG treatment of either MCF-7 or SK-BR-3
`cells results in a rapid increase inshc phosphory(cid:173)
`lation that is sustained for approximately 120 min-
`
`utes. In contrast, neither trastuzumab nor 4D5
`treatment ~auses a significant increase in shc phos(cid:173)
`phorylation. Mitogen-activated protein kinases are
`common mediators of growth factor receptor sig(cid:173)
`nal transduction pathways. T o further investigate
`potential differences in downstream signaling, mi(cid:173)
`togen-activated protein kinase activation was ex(cid:173)
`amined after exposure to trastuzumab or rHRG. As
`shown in Fig 6, strong mitogen-activated protein
`ldnase activation occurs in either MCF-7 (Fig 6,
`top) or SK-BR-3 cells (Fig 6, bottom) only after
`rHRG treatment. These data suggest that while
`trastuzumab or 4D5 may cause autophosphoryla(cid:173)
`tion under some conditions, downstream signaling
`pathways are not significantly activated.
`Other detailed investigations have been under(cid:173)
`taken with well-characterized panels of anti-HER2
`monodonal antibodies in an attempt to correlate
`agonistic activities of these antibody panels with
`the desired biological outcome,
`ie, cytosta(cid:173)
`sis.47.48•67.68 To date, no strict correlation exists
`between any'Üne measured biochemical parameter
`and antiproliferative activity. Rather, a number of
`attributes emerge as components in assessing anti(cid:173)
`HER2 cytostatic activity. These indude antibody
`affinity, bivalency, pH dependence of binding,
`downmodulation, internalization, and the ability
`to block HER2 association with other HER family
`members. lt appears that the more potent antipro-
`
`IMMUNOGEN 2049, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`THE MECHANISM OF ACTION OF TRASTUZUMAB
`
`65
`
`MCF ... 7
`
`shc .......
`
`time
`(min.)
`
`1 2 3 4 5 6 7 8 9 10 11 12
`
`10
`
`40
`
`120
`
`240
`
`SK-BR-3
`
`shc .......
`
`control
`1
`trastuzumab
`2
`rHRG
`3
`control
`4
`trastuzumab
`5
`rHRG
`6
`7 control
`trastuzumab
`8
`9
`rHRG
`10 control
`11 trastuzumab
`12 rHRG
`
`- 83
`- 62
`- 47.5
`
`- 83
`- 62
`- 47.5
`
`time
`(min.)
`
`1 2 3 4 5 6 7 8 9 10 1112
`10
`40
`120
`240
`
`Fig 5. Effects oftrastuzumab or rHRG treatment on she phosphorylation in MCF-7 or 5K-BR-3 eells. Cells were grown in six-well
`dishes. They were serum-starved 24 hours before treatment and then Iysed under denaturing eonditions. Cell Iysates were immu(cid:173)
`noprecipltated with anti-she pAb (#514630; Transduction Labs) and immunoblotted with anti-phosphotyroslne monoclonal antibodies
`as deseribed in Fig 3.
`
`liferative HER2 antibodies share a number of these
`characteristics, which together diminish the con(cid:173)
`tribution of HER2 to ligand-dependent and inde(cid:173)
`pendent signaling.
`
`TRASTUZUMAB-MEDIATED RECEPTOR
`DOWNMODULATION
`
`Downmodulation of receptor-ligand complexes
`is thought to be a major attenuation mechanism
`for receptor-induced signaling. Recently, it has
`be en shown that HER2, HER3, and HER4 are
`impaired relative to EOFR with regard to their
`ability to undergo ligand-mediated endocytosis. 13
`Although the rate and precise mechanism for
`downmodulation of HER2 are vastly slower than
`observed for EOFR, signincant removal of HER2
`from the plasma membrane occurs with 4D5 treat(cid:173)
`ment.49,65,69
`In summary, these data are consistent with the
`notion that trastuzumab or 4D5 induces obligate
`formation of HER2 homodimers, particularly in
`cells that overexpress HER2, and that these ho(cid:173)
`modimers are capable of receptor transphosphoryla(cid:173)
`tion, However, antibody-induced receptor activation
`does not appear to induce effective downstream sig-
`
`treatment leads to
`trastuzumab
`naling. Instead,
`downmodulation of HER2 from the tumor cell sur(cid:173)
`face. The removal of HER2 from the plasma mem(cid:173)
`brane presumably results in less receptor available for
`dimerization with itself or other HER family mem(cid:173)
`bers, which in turn diminishes the HER2-initiated
`constitutive growth signal.
`
`COMPLEMENT -DEPENDENT
`CYTOTOXICITY
`
`Because trastuzumab contains a human immu(cid:173)
`noglobulin 01 (lgO 1) Fc region, it may be capable
`of activating the human complement cascade. Im(cid:173)
`munochemical and functional analysis demon(cid:173)
`strated that trastuzumab activated complement af(cid:173)
`ter binding to its target antigen HER2 on nve of
`six breast adenocarcinoma cell lines (M. Kir(cid:173)
`schfink, University of Heidelberg, personal com(cid:173)
`munication, July 1996). However, trastuzumab(cid:173)
`induced complement-mediated tumor cell lysis
`was not observed, which is likely due to the pres(cid:173)
`ence of membrane-associated complement regula(cid:173)
`tory proteins such as CD35 (complement receptor
`1), CD55 (decay accelerating factor), or CD46
`(membrane cofactor protein),7° Complement-ac-
`
`IMMUNOGEN 2049, pg. 6
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`66
`
`SLlWKOWSKI ET AL
`
`kD
`
`83
`62
`47.5- i
`
`treatment :t trastuzumab
`
`treatment :t rHRG
`
`47.5-i. - + -
`
`time
`(min.)
`
`~~
`
`time
`(min.)
`
`.. +-+-+-+-+-+
`10 20 40 80 120 240
`
`-..
`
`+-+-+-+ -+
`10 20 40 80 120 240
`
`Fig 6. Detection of MAP Idnase activation in MCF-7 or SK-BR-3 cells. Cells (1.2 X 105 cells/well) plated in 12-well dishes were
`changed to low serum-containing medium (0.1 % fetal bovine serum) 24 hours after adhering. After an additional 48 hours, the
`medium was changed again to serum-free for 1.5 hours to minimize any mitogen activated protein kinase activation in the untreated
`cells. Cells were treated for various times as indicated at 37·C with trastuzumab (10 ¡.tg/mL) or rHRG (10 nmoIlL). Incubations were
`stopped by the addition of 0.4 mL sodium dodecyl sulfate sample buffer with reducing agent. Lysates were heated to 100·C for S
`minutes and then 25 ¡.tL was electrophoresed on a 4% to 20% Tris-glycine gel (Novex, San Diego, CA). Western blots were probed
`with a polyclonal antibody recognizing the dually phosphorylated form (#V667A; Promega, Madison, WI) or antibodies that recognize
`the unactivated forms of ERK I and ERK2, each used at a dilution of 1:2,000 (SC-094-G and SC-I S4-G; Santa Cruz).
`
`tivating ability of the antibody was further sup(cid:173)
`ported by results from solid-phase binding assays.
`On binding of trastuzumab to the immobilized
`purified extracellular domain of HER2, the com(cid:173)
`plement cascade reaction could be demonstrated
`to occur up to the binding of the CSb-9 (mem(cid:173)
`brane attack complex) (M. Kirschfink, University
`of Heidelberg, personal communication, July
`1996).
`
`ANTIBODY-DEPENDENT CELL-MEDIATED
`CYTOTOXICITY
`
`Antibody-dependent cell-mediated cytotoxicity
`(ADCC) is triggered when antibody-coated (op(cid:173)
`sonize-d) target cells interact with Fc receptors that
`are present on effector cells. Activated effector
`cells destroy their targets by release of cytoplasmic
`granules that contain perforin and granzymes. Pre(cid:173)
`viously, it was reported that trastuzumab-mediated
`ADCC was very effective against tumor cells that
`overexpress HER2.S1.S2 In these studies, human
`peripheral blood mononuclear cells were used as
`effector cells and were pretreated with interleu(cid:173)
`kin-2. Since trastuzumab contains an IgG1 Fc,
`trastuzumab-dependent ADCC was expected to be
`mediated by Fey receptors , which are present on
`peripheral blood mononuclear cells. T o assess the
`type ofFey receptors thatare primarily responsible
`
`for m~diating trastuzumab-dependent ADCC
`without prior interleukin-2 treatment, cytotoxicity
`assays were performed in the presence of antibod(cid:173)
`ies known to block the interaction of human IgG 1
`Fc with specific Fc)' receptors. Data from these
`experiments demonstrated that trastuzumab-de(cid:173)
`pendent ADCC occurred primarily through inter(cid:173)
`actions with the type 1II receptor, or CD16.
`Fc)'RIlI is expressed primarily on natural killer
`cells, some monocytes, and activated T cells. As
`shown in Fig 7, specific types of effector cells were
`isolated and tested in SICr release cytotoxicity
`
`100 - . . - NK cells + trastuzumab
`--e--- NK cells - trastuzumab
`____ Monocytes + trastulumab
`----Er- Monocytes - tr'asluzumab
`
`~ 80
`~
`~ 60
`'(3
`'x o 40
`....
`~ 20
`o
`
`0.1
`
`10
`1
`Effector to Target (E:T) Ratio
`
`100
`
`Fig 7. Trastuzumab-mediated ADCC using purified natural
`Idller (NK) cells or monocytes as effector cells and SI Cr-Iabeled
`SK-BR-3 cells as targets.
`
`IMMUNOGEN 2049, pg. 7
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`THE MECHANISM OF ACTION OF TRASTUZUMAB
`
`67
`
`assays using the HER2-overexpressing breast can(cid:173)
`cel' cellline, SK-BR-3, as a target. By varying rhe
`c[fector to target rarios fro111 0.1 to 100, the effec(cid:173)
`tor to target ratio where half-maxima¡'cytotoxicity
`occurred could be determined by four-parameter fit
`analysis. F.or natural killer cells and monocytes,
`these values were determined to be 1 and 3, re(cid:173)
`spectively, suggesting that these effectol' cells are
`cxtremely potent in killing trastuzumab-coated
`target cells.
`The precise role of ADCC in vivo is difficult to
`assess in cancer patíents undergoing immunothcr(cid:173)
`apy or in animal models. Arming of natural killer
`cells by tumor-specific a,ntibody has be en postu(cid:173)
`lated as a mechanism for directing these effector
`ce lis to tumors. 71 Since the binding of human
`IgOl with FcyRIII is a relatively weak interaction
`and the concentration of 19O is higb in serum,
`effector cell arming witb trastuzumab is unlikely to
`occur in vivo. In contrast, the binding between
`trastuzumab and HER2 is governed by a very high(cid:173)
`affinity interaction. To determine whether trastu(cid:173)
`zumab-mediated ADCC was affected by physio(cid:173)
`logic concentratíons of 19O, cytotoxicity assays
`were performed in the presencc 01' absencc of 10
`mg/mL human IgO. The results of rhese experi(cid:173)
`ments dcmonstrated that serum leve! concentra(cid:173)
`tions of IgO did not affect either the half-maximal
`effector to target ratio or the degree of cytotoxic(cid:173)
`ity. Moreover, since trasruzumab is bivalent and
`HER2 protein density is very high in these tumor
`cells, an avidity component also tikely contributes
`to the near-irreversible binding of trastuzumab to
`cells that overexpress HER2. Furthermore, over(cid:173)
`expressing tumor cell tines were more susceptible
`to cytotoxic damage than intermediate- or low(cid:173)
`expressing tumor celllines. Taken togetber, these
`data suggest that the affinity and avidity (lf the
`tl'astuzumab-HER2 interaction may also contrib(cid:173)
`ute to the safety profile of trastuzumab in patients,
`since HER2-overexpressing tumor cells would
`likely be preferential1y targeted for ADCC rather
`than tissues tbat Express normallevels of HER2.
`
`EFFICACY IN HUMAN BREAST TUMOR
`XENOGRAFTS
`
`T o determine the relative potency of trastu(cid:173)
`zumab, hoth MCF-7 -HER2 and BT-474 xenograft
`models have been used. In these experiments, an(cid:173)
`imals bearing HER2-overexpressing MCF-7 -HER2
`xenografts were treated with trastuzumab at total
`
`doses of 3, 10,30, and 100 mg/kg. As shown in Fig
`'S, dose-dependent antitllmor activity was observed
`with 3 to 100 mg/kg trastuzumab treatment com(cid:173)
`pared with mice tl'cated with control antibody.
`N ude
`111 ice bearing BT -474 xenografts were
`treated with trastuzllmab doses ranging fram 0.1 to
`30 mgfkg. In this case, dose-dependent antitumor
`activity was obscrved with 0.1, 0.3, and 1 tngfkg
`trastuzllmab
`treatment compared with mice
`treated with control antibody. Antitllmor activity
`appeared to plateau at doses aboye 1 mg/kg. 7.1 The
`BT-474 xenograft model is thllS more sensitive to
`trastuzumab treatment than the MCF-7 -HER2
`xenograft modeJ.53
`
`COMBINATION EFFICACY STUDIES WITH
`CYTOTOXIC CHEMOTHERAPEUTIC
`AGENTS
`
`Treatment ofhuman cancer typically involves a
`multidisciptinary approach that frequently
`in(cid:173)
`eludes surgery, radiation, chemotherapy, and/or
`hormonal manipulation. T o determine how to use
`trastuzumab both alone and in combination with
`established therapeutic agents, a series of in vitra
`'and in vivo studies was conducted. These studies
`were based on previously pubtished reports of en(cid:173)
`hanced activity of cisplatin when used in combi(cid:173)
`nation with antibodies directed against
`the
`EOFR,73 Combination studies with cisplatin re(cid:173)
`vealed tbat trastuzllmab, 4D5, 01' other anti-HER2
`monoclonal antibodies potentiate cytotoxicity by
`decreasing DNA repair foUowing c