Seribantumab, an Anti-ERBB3 Antibody, Delays the Onset of Resistance and Restores Sensitivity to Letrozole in an Estrogen Recept

Published OnlineFirst August 26, 2015; DOI: 10.1158/1535-7163.MCT-15-0169

Cancer Biology and Signal Transduction Molecular Cancer Therapeutics Seribantumab, an Anti-ERBB3 Antibody, Delays the Onset of Resistance and Restores Sensitivity to Letrozole in an Estrogen Receptor–Positive Breast Cancer Model Michael D. Curley1, Gauri J. Sabnis2, Lucia Wille1, Bambang S. Adiwijaya1, Gabriela Garcia1, Victor Moyo1, Armina A. Kazi2, Angela Brodie2, and Gavin MacBeath1

Abstract

Heregulin-driven ERBB3 signaling has been implicated as a in letrozole-sensitive (MCF-7Ca) and -refractory (LTLT-Ca) mechanism of resistance to cytotoxic and antiendocrine ther- cells.Notably,inMCF-7Ca–derived xenograft tumors, cotreat- apies in preclinical breast cancer models. In this study, we ment with seribantumab and letrozole had increased antitu- evaluated the effects of seribantumab (MM-121), a heregulin- mor activity compared with letrozole alone, which was accom- blocking anti-ERBB3 monoclonal antibody, alone and in panied by downregulated PI3K/MTOR signaling both prior to combination with the aromatase inhibitor letrozole, on cell and after the development of resistance to letrozole. Moreover, signaling and tumor growth in a preclinical model of post- the addition of an MTOR inhibitor to this treatment regimen þ menopausal estrogen receptor–positive (ER ) breast cancer. did not improve antitumor activity and was not well tolerated. In vitro, heregulin treatment induced estrogen receptor phos- Our results demonstrate that heregulin-driven ERBB3 signal- phorylation in MCF-7Ca cells, and long-term letrozole-treated ing mediates resistance to letrozole in a preclinical model of þ þ (LTLT-Ca) cells had increased expression and activation levels ER breast cancer, suggesting that heregulin-expressing ER of EGFR, HER2, and ERBB3. Treatment with seribantumab, but breast cancer patients may beneﬁt from the addition of ser- not letrozole, inhibited basal and heregulin-mediated ERBB ibantumabtoantiendocrinetherapy.Mol Cancer Ther; 14(11); receptor phosphorylation and downstream effector activation 2642–52. 2015 AACR.

Introduction women, estrogen biosynthesis occurs predominantly in periph- eral tissues under the control of the aromatase enzyme. Hormone receptor–positive tumors, comprising both estro- þ Aromatase inhibitors (AI), such as letrozole, anastrozole, or gen and progesterone receptor–positive (ER/PR ) tumors, rep- exemestane, that block estrogen-mediated ER activation by resent the predominant subtype of breast cancer, accounting for suppressing estrogen production, have demonstrated clinically approximately 70% of breast cancer diagnoses (1, 2). The meaningful activity and have emerged as the standard therapy estrogen receptor mediates the effects of estrogen on normal þ for postmenopausal ER breast cancer (4). Despite the effec- and cancerous breast tissue. Binding of estrogen (17b-estradiol) tiveness of these agents, however, some patients are intrinsi- to the ER induces nuclear localization of this complex. Recruit- cally insensitive to AIs and most patients eventually develop ment to specific DNA-binding sites and subsequent activation resistance to these treatments. of transcription culminates in increased cell proliferation and þ Dysregulated expression of receptor tyrosine kinases (RTK) and survival (3). In the context of ER breast cancer, ER activation their growth factors (ligands) has been implicated in the devel- drives tumor growth and hence disrupting estrogen activity opment of resistance to therapy in breast cancer (5, 6). When represents the primary approach to therapy. In postmenopausal mutated or overexpressed, RTKs can act as primary oncogenic drivers of cancer growth, as is the case for HER2 amplification in breast cancer (7, 8). In addition, paracrine-, autocrine-, or juxta- 1 2 Merrimack Pharmaceuticals, Cambridge, Massachusetts. Depart- crine-derived growth factors in the tumor microenvironment bind ment of Pharmacology and Experimental Therapeutics, University of Maryland School of Medicine, Baltimore, Maryland. to RTKs, activating a variety of downstream signaling pathways, including the PI3K/AKT/MTOR prosurvival pathway (9). In this Note: Supplementary data for this article are available at Molecular Cancer Therapeutics Online (http://mct.aacrjournals.org/). way, cancer cells may become desensitized to treatment by exposure to growth factors that activate compensatory signaling path- M.D. Curley and G.J. Sabnis contributed equally to this article. ways (10). Corresponding Author: Gavin MacBeath, Merrimack Pharmaceuticals, One Bidirectional signaling crosstalk has been shown to exist Kendall Square, Suite 7201, Cambridge, MA 02139. Phone: 617-441-7461; Fax: between the ERBB family of RTKs and estrogen receptors 617-491-1386; E-mail: [email protected] (11, 12). The ERBB family of receptors comprises four related doi: 10.1158/1535-7163.MCT-15-0169 transmembrane proteins, EGFR, HER2, ERBB3, and ERBB4, that 2015 American Association for Cancer Research. form homo- or heterodimers with distinct binding affinities for

2642 Mol Cancer Ther; 14(11) November 2015

Seribantumab Delays Resistance, Restores Sensitivity to Aromatase Inhibitors

specific growth factors (13). Growth factor–induced receptor Cell culture dimerization stimulates intrinsic receptor kinase activity, result- MCF-7Ca (human breast cancer cells, which have been stably ing in activation of intracellular signaling cascades that regulate transfected with human aromatase gene) were provided by Dr. S and promote a variety of cellular functions, including cell Chen (City of Hope, Duarte, CA) in 2012. These cells were survival and proliferation (14, 15). Although no direct ligand maintained in DMEM base media (Life Technologies) supple- for HER2 has been identified (16), and although ERBB3 has mented with 5% charcoal–dextran–treated FBS (CDT-FBS), 1% very weak catalytic activity (17–19), the HER2::ERBB3 hetero- penicillin/streptomycin (P/S; Life Technologies) and 700 mg/mL dimer represents the most potent driver of PI3K/AKT signaling G418 (Life Technologies). Long-term letrozole-treated (LTLT-Ca) and cell proliferation (20, 21). The growth factor heregulin cells were developed from MCF-7Ca xenograft tumors that had (HRG), which binds directly to ERBB3, stimulates signaling been treated with letrozole (10 mg/day, i.p. injection) for 56 weeks through the HER2::ERBB3 heterodimer and can induce phos- in mice (28). These cells were routinely cultured in phenol red-free phorylation of the estrogen receptor independent of estrogen, Modified IMEM base media (Life Technologies) supplemented leading to ER activation (22). Moreover, HRG expression in with 5% CDT-FBS, 1% P/S, 750 mg/mL G418, and 1 mmoL/L breast cancer cells can contribute to cancer progression by letrozole. Both cell lines were maintained in a humidified atmo- activating HER2::ERBB3 signaling through autocrine mechan- sphere of 5% CO2 at 37 C. All cell lines used in this study were isms (23, 24). Thus, if HRG-driven HER2::ERBB3 signaling is authenticated with short tandem repeat (STR) profile analysis in þ active in ER breast cancer, blocking both estrogen-dependent August 2013 using CellCheck service provided by Idexx Radil. The and HRG-mediated, estrogen-independent signaling may be cell lines were found to be identical to the genetic profile reported necessary to fully block tumor progression and provide the for the MCF-7 cell line (ATCC# HTB-22). Frozen aliquots of cell maximum benefittopatients. lines were used within 6 months of culturing. Computational modeling and sensitivity analysis of the ERBB For in vitro assays, cells were seeded in phenol-red free IMEM signaling network previously identified ERBB3 as the key node containing 4% CDT-FBS, 1% P/S and 75 mg/mL G418. Twenty- driving AKT activation following growth factor stimulation four hours before treatment initiation, cells were washed in PBS (25). On the basis of these findings, seribantumab (MM-121) and media was replaced with low serum media (phenol-red free was developed as a fully human IgG2 antibody that recognizes IMEM containing 0.5% CDT-FBS, 1% P/S). Treatments were ERBB3 and blocks HRG binding to the receptor. In addition to administered as detailed in the Figure legends. preventing HRG-mediated ERBB3 activation, seribantumab also prevents heterodimerization of ERBB3 with other receptors in In vivo studies the ERBB family and induces ERBB3 receptor internalization All animal studies were performed according to the guide- and degradation. lines and approval of the Animal Care Committee of the HER2 and ERBB3 signaling have previously been implicated University of Maryland, Baltimore, MD. Female ovariectomized in the development of resistance to antiestrogens such as BALB/c athymic nude mice (4–6 weeks of age) were obtained tamoxifen, and inhibition of ERBB3 signaling was shown to from the NCI-Frederick Cancer Research and Development þ overcome HER2-mediated tamoxifen resistance in ER breast Center (Frederick, MD). Mice were housed in a pathogen-free cancer cells (26). More recently, studies using an in vivo mouse environment under controlled conditions and received food þ model of postmenopausal ER breast cancer indicated that the and water ad libitum. development of resistance to the nonsteroidal AI letrozole was Tumor xenografts were established by subcutaneous injection associated with upregulation of HER2 expression in xenograft of 100 mL of a cell suspension consisting of 2.5 106 MCF-7Ca tumors (27). Here, we used this same model of postmeno- cells, diluted 1:1 in Matrigel, into single sites on both flanks of þ pausal ER breast cancer to determine the effect of blocking each mouse. Mice were injected daily with D4A (100 mg/mouse), HRG-mediated ERBB3 signaling and/or estrogen-mediated ER a precursor for estrogen synthesis and substrate for aromatase. activation on tumor growth. We found that a combination of seribantumab and letrozole had increased antitumor activity in vivo compared with either agent alone, both before and after the development of resistance to letrozole. This effect was accompanied by decreased signaling through the AKT/ MTOR/S6 pathway, suggesting a mechanistic explanation for these observations. Overall, our results suggest that targeting þ ERBB3 signaling in ER breast cancer tumors that express HRG þ may extend the activity of AIs in postmenopausal ER breast cancer patients.

Materials and Methods Materials Letrozole and everolimus (RAD001) were purchased from Figure 1. HRG activates ER phosphorylation in MCF-7Ca cells. Serum-starved MCF-7Ca Selleck Chemicals, LLC. Androstenedione (D4A) was obtained cells were pretreated with seribantumab (Seri; 1 mmol/L) for 1 hour, followed from Sigma Aldrich. Matrigel Basement Membrane Matrix was by treatment with 10 nmol/L HRG for 10 minutes, where indicated (þ). Cell obtained from BD Biosciences. All antibodies were purchased lysates were analyzed by immunoblotting with antibodies for p-ERBB3 from Cell Signaling Technology, except for total ERBB3 (Abcam) (Y1289), p-AKT (S473), and p-ER (S305 and S167). Anti-b-actin antibody was and pER (EMD Millipore). used as a loading control.

www.aacrjournals.org Mol Cancer Ther; 14(11) November 2015 2643

Curley et al.

Tumor formation was monitored weekly and tumor volumes tumab was diluted in 0.9% NaCl solution, and everolimus was were calculated following caliper measurement according to the diluted in 5% glucose. Drug treatments were administered as formula (4/3 p(radius 1 radius 1 radius 2), where radius indicated in the Figure legends. Following the development of 1 < radius 2). Once the average measured tumor volume had resistance to letrozole, deﬁned as the time point at which the reached approximately 300 mm3, mice were randomized into average tumor volume in the letrozole group had doubled from groups and treatment was administered. Overall, the average the average starting tumor volume, mice in the letrozole-treated starting tumor volume per group was equivalent across all group were rerandomized to various treatment groups, main- groups. taining an equivalent average tumor volume per group. Mouse For injection, letrozole and D4A were prepared in 0.3% body weights were measured weekly over the course of treat- hydroxypropylcellulose (HPC) in 0.9% NaCl solution, seriban- ment and mice were euthanized if their total body weight loss

Figure 2. Seribantumab, but not letrozole, inhibits basal and HRG-induced activation of ERBB receptors and downstream effectors in MCF-7Ca and LTLT-Ca cells. MCF-7Ca and LTLT-Ca cells were serum-starved (0.5% serum) for 24 hours before treating with seribantumab (1 mmol/L), letrozole (1 mmol/L) or HRG (10 nmol/L) for 2 hours, alone and in combination, as indicated. Cell lysates were analyzed by immunoblotting with antibodies for phosphorylated or total EGFR, HER2, and ERBB3 (A), and phosphorylated or total AKT, ERK, MTOR, S6K1, and ER (B). Anti-b-actin antibody was used as a loading control. The blots show a single representative of two independent experiments.

2644 Mol Cancer Ther; 14(11) November 2015 Molecular Cancer Therapeutics

Seribantumab Delays Resistance, Restores Sensitivity to Aromatase Inhibitors

surpassed 15%. Tumors were flash-frozen in liquid nitrogen adding cell and tumor lysates. HRG1-b1wasdetectedwith following extraction for pharmacodynamic analyses, where biotinylated goat anti-human NRG1-b (R&D Systems; BAF377) indicated. for 2 hours at room temperature followed by streptavidin HRP (Jackson ImmunoResearch) for 30 minutes at room temperature. HRG1-b1 levels were quantified using purified Immunoblotting recombinant HRG1-b1 protein (R&D Systems, 377-HB) as a For in vitro studies, cells were treated as indicted for each standard. ELISA signal was visualized using One-step Ultra experiment, then media was removed, dishes were placed on ice, TMB solution (Thermo Scientific) and absorbance was mea- and cells were washed with ice-cold PBS. To generate cell lysates, sured at 450 nm. M-PER Mammalian Protein Extraction Buffer (Thermo Scientific) For detection of HRG1-b1 in media derived from cells in supplemented with protease and proteinase inhibitors (Roche) culture, the human NRG1-beta 1/HRG1-beta 1 DuoSet ELISA kit was added and cells were scraped and pipetted into fresh tubes. was used (R&D Systems, DY377). Signal was visualized as Cell debris was removed by centrifugation (14,000 rpm for 10 described for cell and tumor lysates. minutes). To generate tumor lysates, flash-frozen tumors were pulverized in a CryoPrep pulverizer (Covaris) and resuspended in Statistical analyses fi T-PER Tissue Protein Extraction Buffer (Thermo Scienti c), sup- Statistical analyses were used to evaluate potential differences plemented with protease and proteinase inhibitors (Roche). in tumor growth rates between the various groups of mice treated Following a 30-minute incubation on ice, cell debris was removed with different drugs and drug combinations. For in vivo models, by centrifugation. tumor growth rates Gi of a lesion i were estimated from the Total protein was measured using a BCA assay (Thermo Sci- obs measured volume V of each tumor lesion i at time t using the entific). Thirty micrograms of protein samples were analyzed by i;t following nonlinear mixed effect model: SDS-PAGE and bands were visualized using the Odyssey detection system (LI-COR). Quantitative analyses were performed using Git Vi;t ¼ Vi;0e ImageStudio (LI-COR), and measurements were corrected for h G ¼ G e 1 loading control. i TRT h Vi;0 ¼ e 2 obs V ¼ Vi;tðÞþ1 þ " " HRG ELISA i;t ÀÁ 1 ÀÁ2 b 2 2 For detection of HRG1-beta 1 (HRG1- 1) in tumor tissue "i ¼ N 0; Ei ; hi ¼ N 0; Wi lysates, Reacti-Bind plates (Thermo Scientific) were coated overnight with recombinant histidine-tagged ERBB3 capture where GTRT indicates the population average growth rates of antibody (29), then blocked with 1% BSA (w/v) in PBS before each treatment TRT, Gi indicates the individual growth rates for

Figure 3. Seribantumab and letrozole cotreatment delays the onset of resistance and restores sensitivity to letrozole in MCF-7Ca xenografts. MCF-7Ca xenograft tumors were generated in female, ovariectomized nude mice, which were randomized to receive vehicle [Control; 0.3% HPC in 0.9% NaCl, twice weekly (Q2W), i.p.; 15 mice/group], seribantumab (750 mg/ mouse, Q2W, i.p.; 15 mice/group), letrozole [10 mg/mouse/day 5 days/week (QD 5), subcutaneous injection (SQ); 60 mice/group], or letrozole in combination with seribantumab, dosed as indicated for the monotherapies (15 mice/group). Changes in mean tumor volume (SEM) were determined weekly by caliper measurement. Following the development of resistance to letrozole (week 14), mice in the letrozole-only group were rerandomized into 15 mice/group to receive: letrozole alone, seribantumab alone, or a combination of letrozole and seribantumab. All agents were administered as described above. Mice included for pharmacodynamic analyses were taken from each group at speciﬁed time points, and were not included in efﬁcacy study calculations.

www.aacrjournals.org Mol Cancer Ther; 14(11) November 2015 2645

Curley et al.

Figure 4. Seribantumab treatment blocks ERBB3 activation and downregulates MTOR pathway activation in combination with letrozole. A, MCF-7Ca xenograft tumors (n ¼ 4/treatment group) were harvested 24 hours after one dose of vehicle, letrozole, seribantumab, or letrozole þ seribantumab, administered as described in Fig. 3. Tumor lysates were generated and analyzed by immunoblotting with antibodies for phosphorylated or total ERBB3. Anti-b-actin antibody was used as a loading control. Quantiﬁed expression, normalized to b-actin, is shown in the dot plots. B, dot plots showing quantiﬁed expression (normalized to b-actin) of tumor- derived proteins analyzed by immunoblotting following 24 hour treatment as indicated. Antibodies were for t-MTOR, p-S6K1 (T389), p-S6 (S235/S236) and x p-S6 (S240/S244). , P < 0.05 vs. control and vs. letrozole; †, P < 0.05 vs. control; z, P < 0.05 vs. letrozole and vs. seribantumab; , P < 0.05 vs. all other groups.

2646 Mol Cancer Ther; 14(11) November 2015 Molecular Cancer Therapeutics

Seribantumab Delays Resistance, Restores Sensitivity to Aromatase Inhibitors

each tumor lesion (each mouse's left and right ﬂank of tumor cells (27). To further explore network adaptation to long-term obs lesions), Vi,t indicates the tumor volume at time t,andVi;t treatment with letrozole leading to resistance, we compared indicates the observed tumor volume at time t.Tumorvolumes the expression of ERBB receptors and downstream effector below the limit of detection were treated as censored observa- proteins in MCF-7Ca and LTLT-Ca cells following treatment tions. The resulting Gi estimates were compared across treat- with HRG, letrozole and seribantumab. The parental MCF-7Ca ments and lines of therapy using ANOVA, with the letrozole cell line displayed low levels and activation states of EGFR, alone treatment group used as the baseline comparator. The HER2, and ERBB3 receptors. In addition to the previously nonlinear mixed-effect modeling approach was implemented reported upregulation of HER2 in LTLT-Ca cells (27), we found in NONMEM version 7.3 and the ANOVA was implemented in that total and phosphorylated levels of EGFR and ERBB3 were JMP version 10.0.2. All P values < 0.05 were considered sta- also increased (Fig. 2A and Supplementary Table S1), poten- tistically signiﬁcant. tially implicating these receptors in contributing to letrozole resistance. Treatment with seribantumab alone decreased p- Results EGFR (Y1068), p-HER2 (Y1196), p-ERBB3 (Y1289), and total ERBB3 (t-ERBB3) levels in MCF-7Ca and LTLT-Ca cell lines in HRG stimulation of ERBB3 promotes ER phosphorylation the absence of exogenously added HRG, indicating that HRG- To explore the effect of inhibiting HRG-mediated ERBB3 sig- mediated ERBB3 signaling persists in these cells. As the only in naling on ER phosphorylation, we stimulated MCF-7Ca cells source of HRG in vitro is autocrine-derived, we measured HRG vitro with HRG, in the presence and absence of seribantumab, and levels in cell lysates and conditioned media by ELISA measured phosphorylated protein levels by immunoblotting. and found that HRG was present in both cell lines at low Treatment with HRG potently stimulated ERBB3 and AKT phos- levels (Supplementary Fig. S1). Both cell lines were highly phorylation, as well as ER phosphorylation at serine 167 and responsive to exogenously added HRG, as indicated by serine 305 (Fig. 1), indicating direct cross-talk between the ERBB3 increased levels of p-EGFR, p-HER2, and p-ERBB3. Whereas and ER signaling pathways in this cell line. Treatment with letrozole had no effect on HRG-induced ERBB3 receptor acti- seribantumab reduced HRG-stimulated phosphorylation of vation, seribantumab blocked HRG-induced HER2 and ERBB3 ERBB3, AKT, and ER in MCF-7Ca cells. receptor phosphorylation. Given the increase in ERBB receptor activation following the Seribantumab inhibits basal and HRG-induced receptor development of resistance to letrozole, we analyzed effector activation in MCF-7Ca and LTLT-Ca cells activation downstream of these receptors. LTLT-Ca cells dis- Upregulation of HER2 has been implicated as an adaptive played increased levels of p-AKT (S473), p-ERK (T202/Y204), mechanism of resistance to letrozole treatment in breast cancer p-MTOR (S2448), p-S6K1 (T389), and p-ER (S167 and S305),

Figure 5. Seribantumab has improved antitumor activity compared with everolimus when cotreated with letrozole but no increase in activity is observed for the combination of all three agents in MCF-7Ca xenografts. MCF-7Ca xenograft tumors were randomized to receive vehicle (Control; 0.3% HPC in 0.9% NaCl, Q2W, IP; 10 mice/group), letrozole (10 mg/mouse/day, QD 5, SQ; 40 mice/group), or letrozole in combination with seribantumab (750 mg/mouse, Q2W, i.p.; 10 mice/ group). Changes in mean tumor volume (SEM) were measured weekly. Following the development of resistance to letrozole (week 16), mice in the letrozole-only group were rerandomized into 10 mice/group to receive: letrozole; letrozole in combination with seribantumab; letrozole in combination with everolimus [2.5 mg/kg/d, orally (PO)]; or a triple combination of letrozole, seribantumab, and everolimus. All agents were administered as described above.

www.aacrjournals.org Mol Cancer Ther; 14(11) November 2015 2647

Curley et al.

Figure 6. þ Full inhibition of tumor cell proliferation and survival in HRG-expressing, ER tumors requires dual inhibition of HRG/ERBB3 and estrogen/ER signaling. A, ER-driven tumor growth can arise by estrogen-dependent and HRG-dependent, estrogen-independent mechanisms. Estrogen binding to the ER or phosphorylation and activation of ER via non–estrogen-mediated RTK signaling can induce gene transcription and ultimately promote tumor cell proliferation, growth, and survival. (Continued on the following page.)

2648 Mol Cancer Ther; 14(11) November 2015 Molecular Cancer Therapeutics

Seribantumab Delays Resistance, Restores Sensitivity to Aromatase Inhibitors

and reduced levels of t-ERK and t-ER compared with parental vival signaling pathway. HRG was detected in all tumor MCF-7Ca cells (Fig. 2B and Supplementary Table S1). Addition samples and no significant differences were observed between of exogenous HRG resulted in increased levels of p-AKT, p- the treatment groups harvested at various time points after ERK, p-S6K1, and p-ER, which were substantially reduced by treatment initiation (Supplementary Fig. S3A). In addition, cotreatment with seribantumab. As before, cotreatment with pERBB3 expression was elevated at the end of the study letrozole did not counteract HRG-mediated activation of these outlined in Fig. 3 following letrozole treatment (Supplemen- survival signaling-associated proteins. We conclude that, in the tary Fig. S3B). The elevated pERBB3 levels are decreased presence of HRG, blocking the action of estrogen alone is when seribantumab treatment is provided. On the basis of insufficient to block ER activation and survival signaling. these results, we attribute the response to seribantumab to an increased sensitivity to HRG rather than an increase in HRG Seribantumab delays the onset of resistance to letrozole expression. and restores sensitivity to letrozole in letrozole-resistant tumors The in vitro studies detailed above revealed that MCF-7Ca The antitumor activity of seribantumab and letrozole is not and LTLT-Ca cell lines are both responsive to HRG. We enhanced by an MTOR inhibitor therefore sought to determine whether HRG-induced signaling As cotreatment with seribantumab and letrozole substan- is active in tumors generated from these cell lines in vivo, tially affects MTOR signaling, we sought to determine (i) and if blocking this signaling pathway affects tumor growth. whether blocking this resistance pathway at a downstream To do so, we generated MCF-7Ca xenograft tumors in ovari- node (MTOR) would prove more or less effective than block- ectomized mice and tested the effects of treatment with ing it at the receptor level; and (ii) if dual pathway blockade seribantumab and/or letrozole on tumor growth in vivo.As at HRG/ERBB3 and MTOR would prove even more effective previously shown, the MCF-7Ca–derived xenograft tumors at restoring sensitivity to letrozole. Everolimus (RAD001), a initially responded to letrozole, but started to develop resis- specific inhibitor of MTOR, was selected for these studies as it tance after approximately 7 weeks of treatment (Fig. 3). When is currently approved by the Food and Drug Administration for mice were cotreated with letrozole and seribantumab, how- the treatment of postmenopausal women with advanced hor- ever, tumor growth was inhibited and resistance to letrozole mone receptor positive, HER2-negative breast cancer in com- substantially delayed (Fig. 3 and Supplementary Fig. S2A). bination with the aromatase inhibitor, exemestane based on a This suggests that HRG/ERBB3 signaling was either present randomized Phase III clinical trial (30). MCF-7Ca tumor cells at the start of the study or developed relatively quickly in were inoculated into ovariectomized nude mice and, following response to letrozole treatment. thedevelopmentoftumors,themicewererandomizedto Once resistance to letrozole was clearly established (week receive letrozole, alone or in combination with seribantumab. 14), mice in the letrozole-treated group were rerandomized As before, cotreatment with letrozole and seribantumab sig- to one of three cohorts: (i) continued letrozole monotherapy; nificantly inhibited tumor growth compared with letrozole (ii) seribantumab monotherapy; or (iii) seribantumab in alone (P < 0.05; Fig. 5 and Supplementary Fig. S4), confirming combination with letrozole (Fig. 3). Notably, the letrozole- our previous findings. Following the development of resistance resistant tumors displayed significantly decreased tumor to letrozole (week 16), mice in the letrozole-treated group were growth when cotreated with letrozole and seribantumab com- rerandomized to receive one of four treatment options: (1) pared with treatment with either drug alone (Fig. 3 and letrozole alone; (2) letrozole in combination with seribantu- Supplementary Fig. S2B). This is consistent with the hypoth- mab; (3) letrozole in combination with everolimus; and (4) esis that blocking both estrogen/ER- and HRG/ERBB3-driven letrozole in combination with both seribantumab and ever- signaling provides greater antitumor activity than blocking olimus. Interestingly, treatment with either seribantumab or either pathway alone. everolimus in combination with letrozole led to tumor regres- Xenograft tumors harvested 24 hours after initiating treat- sion. Seribantumab exhibited slightly more activity in combi- ment with letrozole and/or seribantumab were analyzed nation with letrozole than everolimus, but this effect was not to determine the effects of treatment on cellular signaling. significant (P > 0.05). Notably, cotreatment with letrozole and Treatment with seribantumab, alone or in combination with everolimus induced increased tumor activation of ERBB3 letrozole, resulted in a substantial decrease in total and phos- (Supplementary Fig. S5), suggesting that ERBB3 signaling may phorylated levels of ERBB3, consistent with the known mech- actasanescaperoutetoMTORinhibitioninthecontextof anism of action of seribantumab (Fig. 4A). Cotreatment antiendocrine therapy. In addition, adding everolimus to the with seribantumab and letrozole also led to decreased levels combination of letrozole and seribantumab did not increase of t-MTOR, p-S6K1, and p-S6 levels compared with the other antitumor activity and was more toxic, as evidenced by sig- treatmentgroups(Fig.4B),indicating an enhanced inhibitory nificantly reduced mouse body weightsat4weeksaftertreat- effect of combination treatment on this well-established sur- ment (P < 0.05; Supplementary Fig. S6).

(Continued.) B, treatment with AIs, such as letrozole, reduces the local availability of estrogen, thereby reducing signaling through the estrogen–ER axis. Compensatory upregulation of RTKs in response to letrozole treatment, however, sensitizes tumor cells to ligands, such as HRG, leading to activated intracellular signaling cascades. These cascades culminate in upregulated gene transcription, tumor cell proliferation, and survival. C, cotreatment of HRG- expressing tumors with letrozole and seribantumab can inhibit both estrogen-dependent and HRG-dependent, estrogen-independent mechanisms of ER- mediated gene transcription and can effectively inhibit tumor cell growth and proliferation. TF, transcription factor; CoAct, coactivator; E2-ER,estrogen– estrogen receptor complex; P, phosphorylated.

www.aacrjournals.org Mol Cancer Ther; 14(11) November 2015 2649

Curley et al.

Discussion than targeting MTOR directly. Notably, targeting both nodes in combination with letrozole (triple inhibition) was toxic in Innate or acquired drug resistance ultimately limits the mice (Supplementary Fig. S6). effectiveness of both cytotoxic and targeted therapies. Patient In our initial in vivo experiment, we observed increased outcomes could potentially be improved by identifying tumor growth near the very end of the studies following mechanisms of drug resistance, discovering biomarkers that long-term coadministration of seribantumab and letrozole are diagnostic for these mechanisms, and developing drugs that (Fig. 3). This suggests that alternative molecular pathways of combat these mechanisms. Coadministration of such drugs escape may also arise that drive tumor growth independent of with standard therapy could provide benefit either by enhanc- HRG and ER inhibition. This is consistent with the signaling ing or prolonging the effectiveness of existing therapies. redundancy that exists among RTKs and that allows for com- HRG-driven ERBB3 signaling has been proposed to mediate pensatory signaling upon targeted inhibition (10, 37, 38). resistance to a broad range of anticancer therapies by potently Identifying these compensatory signaling mechanisms remains activating the PI3K/AKT/MTOR prosurvival signaling pathway. the focus of future studies. In this study, we used a preclinical model of postmenopausal þ Our study, and those of others (39), implicate HRG-driven ER breast cancer to investigate the role of ERBB3 signaling ERBB3 signaling as an important mechanism of resistance in mediating resistance to the AI letrozole. Long-term exposure to conventional endocrine therapy. The overall model that of MCF-7Ca cells to letrozole induced resistance to letrozole and þ emerges is summarized in Fig. 6. In ER breast cancers in was accompanied by an increase in the total and phosphory- which HRG is expressed, ER stimulates tumor cell proliferation lated levels of EGFR, HER2, and ERBB3 (Fig. 2A). This is consistent in both an estrogen-dependent and HRG-dependent, estrogen- with a broad range of studies associating ERBB receptor activation independent fashion (Fig. 6A). In addition, HRG stimulates with acquired drug resistance (5, 31–34), and with clinical data prosurvival signaling through the PI3K/AKT/MTOR pathway. showing that expression levels of EGFR, HER2, and ERBB3 are Inhibition of estrogen-dependent ER signaling through associated with impaired overall survival in breast cancer patients the action of AIs like letrozole is insufficient to block HRG- (35, 36). Both parental and letrozole-resistant MCF-7Ca cells are mediated compensatory signaling (Fig. 6B). Dual blockade responsive to HRG in vitro, as evidenced by phosphorylation of of these pathways can be achieved through coadministration EGFR, HER2, ERBB3, and downstream proteins in the MAPK and of seribantumab and letrozole (Fig. 6C). This suggests that AKT signaling pathways (Fig. 2B). In addition, HRG induces þ patients with ER breast cancer whose tumors also express HRG activation of ER in an estrogen-independent manner (Fig. 1), and could potentially benefit from the addition of seribantumab to letrozole alone is unable to offset these effects (Fig. 2B). In their standard endocrine therapy. This hypothesiswasrecently contrast, the HRG-blocking anti-ERBB3 antibody seribantumab tested in a global, double-blinded, placebo-controlled Phase effectively blocks HRG-induced pathway activation. II study of postmenopausal women with locally advanced or When implanted in ovariectomized mice, MCF-7Ca–derived þ metastatic ER/PR , HER2-negative breast cancer (ClinicalTrials. xenografts acquire resistance to letrozole after 7 to 14 weeks. gov identifier: NCT01151046; ref. 40). Patients (N ¼ 115) were Resistance is delayed, however, when mice are cotreated with randomized to receive either seribantumab or placebo in com- seribantumab, and resistance is reversed when seribantumab is bination with the steroidal AI exemestane. Although the addi- coadministered with letrozole after tumors acquire resistance to tion of seribantumab to exemestane did not significantly pro- letrozole (Fig. 3). Consistent with our in vitro findings and with long progression-free survival (PFS) in the unselected popula- those of Linn and colleagues (9), molecular analyses of in vivo– tion, a subset of patients (45%) with high levels of tumor treated tumors implicate downregulation of the MTOR/S6 HRG mRNA was identified who benefited from the combina- signaling axis as the key mechanism underlying tumor growth tion of seribantumab and exemestane relative to placebo inhibition following dual inhibitor treatment (Fig. 4). and exemestane (hazard ratio of 0.26, P ¼ 0.003; ref. 41). Although seribantumab inhibited both AKT and ERK signaling These data are also in agreement with Phase II clinical trials in vitro, either alone or in combination with letrozole, the same in platinum-resistant ovarian cancer (ClinicalTrials.gov identi- inhibitory effects were not observed in our in vivo experiment. fier: NCT01447706) and EGFR wild-type non–small cell lung This discrepancy likely arises from a difference in timing; cancer (ClinicalTrials.gov identifier: NCT00994123), where tumors were acquired 24 hours after dosing mice in our in vivo tumor HRG mRNA levels were found to be the principal experiment. biomarker for seribantumab activity (42, 43). Consistent with the observed downregulation of the MTOR/ In conclusion, these data provide further evidence that HRG- S6 signaling axis by seribantumab in vivo, everolimus was also driven ERBB3 signaling mediates resistance to AI-based therapy able to reverse resistance to letrozole in MCF-7Ca–derived þ in postmenopausal ER breast cancer. Our in vivo data show xenografts (Fig. 5). The combination of seribantumab and that seribantumab can delay the onset of resistance and restore letrozole, however, appeared to have higher antitumor activity sensitivity to an aromatase inhibitor in an AI-sensitive model of than the combination of everolimus and letrozole. The dose of þ ER breast cancer. These and other results provide the molec- everolimus used in this study was the highest dose that dis- ular and biologic basis for the increased PFS observed in played antitumor activity in MCF-7Ca xenografts without patients with high HRG mRNA levels receiving seribantumab inducing weight loss when given as a monotherapy, based on in combination with exemestane compared with those receiv- adosefinding study in female ovariectomized BALB/c athymic ingexemestanealone.Wesubmitthatcontinuedeffortsto nude mice. Although the difference between seribantumab and identify mechanisms of resistance to therapy, along with their everolimus was not statistically significant, the observed trend biomarkers, will extend the effectiveness of existing therapies suggests that blocking HRG-mediated ERBB3 signaling at the and provide an armamentarium of treatment options for receptor level may result in more potent inhibition of resistance patient with breast cancer.

2650 Mol Cancer Ther; 14(11) November 2015 Molecular Cancer Therapeutics

Seribantumab Delays Resistance, Restores Sensitivity to Aromatase Inhibitors

Disclosure of Potential Conﬂicts of Interest Writing, review, and/or revision of the manuscript: M.D. Curley, G.J. Sabnis, G. MacBeath has ownership interest (including patents) in Merrimack A. Brodie, G. MacBeath Pharmaceuticals. No potential conﬂicts of interest were disclosed by the other Administrative, technical, or material support (i.e., reporting or organizing authors. data, constructing databases): G.J. Sabnis, A. Brodie Study supervision: M.D. Curley, G. Garcia, A. Brodie, G. MacBeath Authors' Contributions Conception and design: M.D. Curley, G.J. Sabnis, L. Wille, G. Garcia, V. Moyo, Grant Support A. Brodie, G. MacBeath This research was funded by Merrimack Pharmaceuticals. A. Brodie received Development of methodology: M.D. Curley, A. Brodie research support from Merrimack Pharmaceuticals. Acquisition of data (provided animals, acquired and managed patients, The costs of publication of this article were defrayed in part by the payment of advertisement provided facilities, etc.): M.D.Curley,G.J.Sabnis,L.Wille,A.A.Kazi, page charges. This article must therefore be hereby marked in A. Brodie accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M.D. Curley, G.J. Sabnis, B.S. Adiwijaya, G. Garcia, Received February 25, 2015; revised August 17, 2015; accepted August 20, A.A. Kazi, A. Brodie, G. MacBeath 2015; published OnlineFirst August 26, 2015.

References 1. Harvey JM, Clark GM, Osborne CK, Allred DC. Estrogen receptor status by shared coreceptor for multiple stroma-derived growth factors. Proc Natl immunohistochemistry is superior to the ligand-binding assay for predict- Acad Sci U S A 1999;96:4995–5000. ing response to adjuvant endocrine therapy in breast cancer. J Clin Oncol 17. Balko JM, Miller TW, Morrison MM, Hutchinson K, Young C, Rinehart C, 1999;17:1474–81. et al. The receptor tyrosine kinase ERBB3 maintains the balance between 2. Perou CM, Børresen-Dale A-L. Systems biology and genomics of breast luminal and basal breast epithelium. Proc Natl Acad Sci U S A 2012;109: cancer. Cold Spring Harb Perspect Biol 2011;3:1–17. 221–6. 3. Musgrove E, Sutherland RL. Biological determinants of endocrine resis- 18. Hynes NE, Lane HA. ERBB receptors and cancer: the complexity of targeted tance in breast cancer. Nat Rev Cancer 2009;9:631–43. inhibitors. Nat Rev Cancer 2005;5:341–54. 4. Winer EP, Hudis C, Burstein HJ, Wolff AC, Pritchard KI, Ingle JN, et al. 19. Tzahar E, Waterman H, Chen X, Levkowitz G, Karunagaran D, Lavi S, et al. A American Society of Clinical Oncology technology assessment on the use of hierarchical network of interreceptor interactions determines signal trans- aromatase inhibitors as adjuvant therapy for postmenopausal women with duction by Neu differentiation factor/neuregulin and epidermal growth hormone receptor-positive breast cancer: status report 2004. J Clin Oncol factor. Mol Cell Biol 1996;16:5276–87. 2005;23:619–29. 20.LaneHA,BeuvinkI,MotoyamaAB,DalyJohnM,NeveRM,Hynes 5. Olayioye MA, Neve RM, Lane HA, Hynes NE. The ErbB signaling network: NE. ErbB2 potentiates breast tumor proliferation through modula- receptor heterodimerization in development and cancer. EMBO J 2000; tion of p27 Kip1-Cdk2 complex formation: receptor overexpression 19:3159. does not determine growth dependency. Mol Cell Biol 2000;20: 6. Abd El-Rehim DM, Pinder SE, Paish CE, Bell J, Rampaul RS, Blamey RW, 3210. et al. Expression and co-expression of the members of the epidermal growth 21. Pinkas-Kramarski R, Soussan L, Waterman H, Levkowitz G, Alroy I, Klapper factor receptor (EGFR) family in invasive breast carcinoma. Br J Cancer L, et al. Diversification of Neu differentiation factor and epidermal growth 2004;91:1532–42. factor signaling by combinatorial receptor interactions. EMBO J 1996; 7. Slamon DJ, Clark GM, Wong S, Levin WJ, Ullrich A, McQuire W. Human 15:2452–67. breast cancer: correlation of relapse and survival with amplification of the 22. Pietras RJ, Arboleda J, Reese DM, Wongvipat N, Pegram MD, Ramos L, et al. HER-2/neu oncogene. Science 1987;235:177–82. HER-2 tyrosine kinase pathway targets estrogen receptor and promotes 8. Yarden Y. Biology of HER2 and its importance in breast cancer. Oncology hormone-independent growth in human breast cancer cells. Oncogene 2001;61:1–13. 1995;10:2435–46. 9. Beelen K, Hoefnagel LDC, Opdam M, Wesseling J, Sanders J, Vincent AD, 23. Li Q, Ahmed S, Loeb JA. Development of an autocrine neuregulin signaling et al. PI3K/AKT/MTOR pathway activation in primary and corresponding loop with malignant transformation of human breast epithelial cells. metastatic breast tumors after adjuvant endocrine therapy. Int J Cancer Cancer Res 2004;64:7078–85. 2014;135:1257. 24. Atlas E, Cardillo M, Mehmi I, Zahedkargaran H, Tang C. Heregulin is 10. Wilson TR, Fridlyand J, Yan Y, Penuel E, Burton L, Chan E, et al. Widespread sufficient for the promotion of tumorigenicity and metastasis of breast potential for growth-factor-driven resistance to anticancer kinase inhibi- cancer cells in vivo. Mol Cancer Res 2003;1:165. tors. Nature 2012;487:505–9. 25. Schoeberl B, Pace EA, Fitzgerald JB, Harms BD, Xu L, Nie L, et al. Thera- 11. Grunt TW, Saceda- M, Martin MB, Ruth LUPU, Krupitza G, Harant H, et al. peutically targeting ERBB3: a key node in ligand-induced activation of the Bidirectional interactions between the estrogen receptor and the c-erbB-2 ErbB receptor-PI3K axis. Sci Signal 2009;2:ra31. signaling pathway: heregulin inhibits estrogenic effects in breast cancer 26. Liu B, Ordonez-Ercan D, Fan Z, Edgerton SM, Yang X, Thor AD. Down- cells. Int J Cancer 1995;63:560–7. regulation of ERBB3 abrogates erbB2-mediated tamoxifen resistance in 12. Sainsbury RC, Farndon JR, Sherbet G V, Harris AL. Epidermal growth factor breast cancer cells. Int J Cancer 2007;120:1874–82. receptors and oestrogen receptors in human breast cancer. Lancet 1985;1: 27. Sabnis G, Schayowitz A, Goloubeva O, Macedo L, Brodie A. Trastuzumab 364–6. reverses letrozole resistance and amplifies the sensitivity of breast cancer 13. Yarden Y. The EGFR family and its ligands in human cancer. signalling cells to estrogen. Cancer Res 2009;69:1416–28. mechanisms and therapeutic opportunities. Eur J Cancer 2001;37 Suppl 4: 28. Jelovac D, Sabnis G, Long BJ, Macedo L, Goloubeva OG, Brodie AMH. S3–8. Activation of mitogen-activated protein kinase in xenografts and cells 14. Yarden Y, Sliwkowski MX. Untangling the ErbB signalling network. Nat Rev during prolonged treatment with aromatase inhibitor letrozole. Cancer Mol Cell Biol 2001;2:127–37. Res 2005;65:5380–9. 15. Holbro T, Beerli RR, Maurer F, Koziczak M, Barbas CF, Hynes NE. The 29. Horak E, Heitner T, Robinson MK, Simmons HH, Garrison J, Russeva M, ErbB2/ERBB3 heterodimer functions as an oncogenic unit: ErbB2 requires et al. Isolation of scFvs to in vitro produced extracellular domains of EGFR ERBB3 to drive breast tumor cell proliferation. Proc Natl Acad Sci U S A family members. Cancer Biother Radiopharm 2005;20:603–13. 2003;100:8933–8. 30. Baselga J, Campone M, Piccart M, Burris HA, Rugo HS, Sahmoud T, et al. 16. Klapper LN, Glathe S, Vaisman N, Hynes NE, Andrews GC, Sela M, et al. The Everolimus in postmenopausal hormone-receptor-positive advanced ErbB-2/HER2 oncoprotein of human carcinomas may function solely as a breast cancer. N Engl J Med 2012;366:520–9.

www.aacrjournals.org Mol Cancer Ther; 14(11) November 2015 2651

Curley et al.

31. Ma J, Lyu H, Huang J, Liu B. Targeting of ERBB3 receptor to overcome 38. Wagner JP, Wolf-Yadlin A, Sevecka M, Grenier JK, Root E, Lauffenburger resistance in cancer treatment. Mol Cancer 2014;13:105. DA, et al. Receptor tyrosine kinases fall into distinct classes based on their 32. McClelland RA, Barrow D, Madden T-A, Dutkowski CM, Pamment J, inferred signaling networks. Sci Signal 2013;6:ra58. Knowlden JM, et al. Enhanced epidermal growth factor receptor signaling 39. Sutherland RL. Endocrine resistance in breast cancer: new roles for ERBB3 in MCF7 breast cancer cells after long-term culture in the presence of the pure and ErbB4. Breast Cancer Res 2011;13:106. antiestrogen ICI 182,780 (Faslodex). Endocrinology 2001;142:2776–88. 40. Higgins MJ, Doyle C, Paepke S, Azaro A, Martin M, Semiglazov V, et al. A 33. Hutcheson IR, Knowlden JM, Madden T-A, Barrow D, Gee JMW, Wakeling randomized, double-blind phase II trial of exemestane plus MM-121 (a AE, et al. Oestrogen receptor-mediated modulation of the EGFR/MAPK monoclonal antibody targeting ERBB3) or placebo in postmenopausal pathway in tamoxifen-resistant MCF-7 cells. Breast Cancer Res Treat women with locally advanced or metastatic ERþ/PRþ, HER2-negative 2003;81:81–93. breast cancer. J Clin Oncol 32:15s, 2014 (suppl; abstr 587). 34. Knowlden JM, Hutcheson IR, Jones HE, Madden T-A, Gee JMW, Harper ME, 41. Macbeath G, Adiwijaya B, Liu J, Pujade-Lauraine E, Higgins M, Tabah-Fisch et al. Elevated levels of epidermal growth factor receptor/c-erbB2 hetero- I, et al. A meta-analysis of biomarkers in three randomized, phase 2 studies dimers mediate an autocrine growth regulatory pathway in tamoxifen- of MM-121, a ligand-blocking anti-ERBB3 antibody, in patients with resistant MCF-7 cells. Endocrinology 2003;144:1032–44. ovarian, lung and breast cancers. Ann Oncol 2014;25:iv58–iv84. 35. Tovey S, Dunne B, Witton CJ, Forsyth A, Cooke TG, Bartlett JMS. Can 42. Liu J, Ray-coquard IL, Selle F, Poveda A, Cibula D, Hirte HW, et al. A phase II molecular markers predict when to implement treatment with aromatase randomized open-label study of MM-121, a fully human monoclonal inhibitors in invasive breast cancer? Clin Cancer Res 2005;11:4835–42. antibody targeting ERBB3, in combination with weekly paclitaxel versus 36. Berghoff AS, Bartsch R, Preusser M, Ricken G, Steger GG, Bago-Horvath Z, weekly paclitaxel in patients with platinum-resistant/refractory ovarian et al. Co-overexpression of HER2/HER3 is a predictor of impaired survival cancers. J Clin Oncol 32:15s, 2014 (suppl; abstr 5519). in breast cancer patients. Breast 2014;23:637. 43. Sequist LV, Lopez-Chavez A, Doebele RC, Gray JE, Harb W, Modiano MR, 37. Schoeberl B, Faber AC, Li D, Liang M-C, Crosby K, Onsum M, et al. An et al. A randomized phase 2 trial of MM-121, a fully human monoclonal ERBB3 antibody, MM-121, is active in cancers with ligand-dependent antibody targeting ERBB3, in combination with erlotinib in EGFR wildtype activation. Cancer Res 2010;70:2485–94. MSCLC patients. J Clin Oncol 32:15s, 2014 (suppl; abstr 8051).

2652 Mol Cancer Ther; 14(11) November 2015 Molecular Cancer Therapeutics

Seribantumab, an Anti-ERBB3 Antibody, Delays the Onset of Resistance and Restores Sensitivity to Letrozole in an Estrogen Receptor−Positive Breast Cancer Model

Michael D. Curley, Gauri J. Sabnis, Lucia Wille, et al.

Mol Cancer Ther 2015;14:2642-2652. Published OnlineFirst August 26, 2015.

Updated version Access the most recent version of this article at: doi:10.1158/1535-7163.MCT-15-0169

Supplementary Access the most recent supplemental material at: Material http://mct.aacrjournals.org/content/suppl/2015/08/26/1535-7163.MCT-15-0169.DC1

Cited articles This article cites 43 articles, 17 of which you can access for free at: http://mct.aacrjournals.org/content/14/11/2642.full#ref-list-1

Citing articles This article has been cited by 2 HighWire-hosted articles. Access the articles at: http://mct.aacrjournals.org/content/14/11/2642.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mct.aacrjournals.org/content/14/11/2642. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.