Search

This abstract was withdrawn by the authors. Citation Format: Metzger Filho O, Janiszewska M, Guo H, Yardley D, Mayer I, Spring L, Arteaga C, Wrabel E, DeMeo M, Freedman R, Tolaney S, Waks A, Bardia A, Parsons H, Partridge A, Mayer E, King T, Polyak K, Viale G, Winer E, Krop I. Withdrawn [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P1-15-01.

Background: Ribociclib (ribo) + letrozole (LET) demonstrated efficacy and safety in first-line treatment of patients (pts) with hormone receptor–positive (HR+), HER2− advanced breast cancer (ABC). Most endocrine therapy (ET)–naive pts with de novo ABC respond to ET, but first-line monotherapy with ET or fulvestrant elicited only moderate efficacy in the FALCON study of pts with visceral metastases (mets). Visceral mets are associated with poor prognosis and may warrant novel treatment options in addition to ET. Here, we present clinical outcomes of treatment with ribo in pts with de novo ABC and visceral mets. Methods: In the MONALEESA-2 study, postmenopausal women with HR+, HER2− ABC (N=668) were randomized (1:1) to receive either ribo 600 mg/d or placebo (3 wk on/1 wk off) + LET 2.5 mg/d (continuous). The primary endpoint was median progression-free survival (mPFS). Results: All pts in this analysis (n=227) had de novo ABC; 108 pts had visceral mets (ribo, n=53; placebo, n=55). Baseline characteristics were similar between the ribo and placebo groups. The median follow-up time was 26.4 months. In pts with visceral mets, mPFS was significantly higher in the ribo group (30.3 months) than in the placebo group (14.1 months; hazard ratio, 0.52; 95% CI, 0.295–0.905; P=0.019). Rates of PFS in these pts in the ribo vs placebo group, respectively, were 76% vs 60% at 12 months and 60% vs 36% at 24 months. Among the pts with visceral mets, clinical benefit rate (CBR) was 79% vs 71% in the ribo vs placebo group, and overall response rate (ORR) in pts with measurable disease at baseline was 55% vs 46%, respectively. In pts without visceral mets (ribo, n=61; placebo, n=58), PFS rates in the ribo vs placebo group, respectively, were 87% vs 72% at 12 months and 66% vs 51% at 24 months. Safety results were consistent with the overall population, and responses were durable with ribo + LET irrespective of whether the pts had visceral or nonvisceral mets. Conclusions: Clinical outcomes from this analysis of the MONALEESA-2 study are consistent with those of the overall population, and ribo + LET prolonged the mPFS and improved ORR and CBR in pts with de novo ABC with visceral mets vs placebo + LET. Citation Format: O'Shaughnessy JA, Arteaga CL, Hart LL, Lindquist DL, Beck JT, Purkayastha DD, Caria N, Hortobagyi GN. Treatment benefit of ribociclib + letrozole in patients with de novo disease and visceral metastases from the MONALEESA-2 study [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P6-18-01.

Approximately 20% of patients with early ER+ breast cancer (BC) treated with adjuvant antiestrogen therapy relapse with metastatic disease. Previously, we identified 3 amplicons (11q11.3, 8p11.23, and 17q21-23) associated with endocrine-resistance (Giltnane et al. Sci Transl Med 2017). The 17q21-23 amplicon has been associated with highly proliferative luminal B tumors and cancers with high genomic instability. A causal role of this region in endocrine resistance is unclear. We performed whole transcriptome analysis on RNA extracted from 58 ER+ breast cancers of patients treated with letrozole for 5.4-9.2 months (median 7.2 months). PRR11 (Proline rich 11), located in 17q21-23, was significantly upregulated in non-responding tumors as defined by cancer relapse after a median follow up of 5 years and/or a preoperative endocrine prognostic index (PEPI) ≥4. Differential gene expression analysis between tumors expressing low vs high PRR11 mRNA showed that BC signatures associated with proliferation, cell cycle, IGF-1 and PI3K signaling were enriched in tumors with high PRR11 expression. In the Metastatic Breast Cancer Project and TCGA, PRR11 amplification was higher in metastatic vs. primary BCs (16.5% and 8.5%, respectively; Fisher's p=0.0088). Gene Set Enrichment Analysis of mRNA expression in METABRIC and TCGA revealed significant enrichment of hallmark gene sets associated with proliferation in PRR11 amplified ER+ BCs. Genome-scale RNAi screening in Project Achilles showed that among all genes in the 17q21-23 amplicon, PRR11 knockdown results in the 4th strongest anti-proliferative effect in MCF7 cells. PRR11 knockdown with siRNA inhibited proliferation, cell cycle progression, and RB phosphorylation in HCC1428 LTED (long-term estrogen deprived), MCF7 LTED, and fulvestrant-resistant MCF7 cells. Using a PCR array with 84-cell cycle genes, we identified SKP2, CDKN1A, CCNB2, CCNA2, CKS2 and CCNB1 as genes downregulated by PRR11 knockdown. Except for CDKN1A, expression of all those genes was elevated ER+ BCs with PRR11 gain or amplification in TCGA. PRR11 associates with the p85 regulatory subunit of PI3K via its SH3 domain. We speculated this association would suppress p85 homodimers, thus permitting binding of PI3Kα (p110α)-p85 dimers to IRS1 and, hence, activating PI3K/AKT. To test this, we co-transfected HEK293T cells with HA-p85 and FLAG-p85. Transfection of PRR11 into these HEK293T cells reduced HA-p85 and FLAG-p85 homodimers as shown by HA and FLAG pulldowns followed by FLAG and HA immunoblots, respectively. Finally, PRR11 knockdown resulted in a reduction of p110a and S473 P-AKT levels and inhibition of IGF-1/2 stimulated P-AKT. Not inconsistent with these data, PRR11 amplification and PIK3CA mutations in METABRIC and TCGA are exclusive of each other, suggesting these alterations are functionally linked with the same signaling pathway. These data support a role of PRR11 in PI3K/AKT activation that may be causal to resistance to estrogen deprivation. We propose PRR11, located in the 17q21-23 amplicon, is a potential mediator of resistance to antiestrogen therapy by amplifying PI3K/AKT signaling, suggesting PI3K may be a therapeutic target in ER+ BCs harboring PRR11 amplification. Citation Format: Lee K-M, Guerrero-Zotano A, Formisano L, Jansen V, Gonzalez Ericsson P, Arteaga C. Neoadjuvant trial with letrozole identifies PRR11 in 17q21-23 amplicon as a resistance mechanism to endocrine therapy in ER-positive breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr PD7-10.

Background: Neratinib is a potent, irreversible pan-HER tyrosine kinase inhibitor. The phase III trial ExteNET showed improved disease-free survival in patients (pts) with HER2+ breast cancer treated with neratinib vs placebo after trastuzumab-based adjuvant therapy. The benefit from neratinib appeared to be greater in pts with ER+ tumors. Thus, we sought to elucidate mechanisms that may explain the benefit from extended adjuvant therapy with neratinib in pts with ER+/HER2+ breast cancer using a human-in-mouse model that simulates the clinical outcomes seen in ExteNET. Results: Mice with established ER+/HER2 amplified MDA-361 tumors were treated with trastuzumab (tz) + paclitaxel (pac) for 4 weeks, and then randomized to fulvestrant (fulv) ± neratinib for 4 weeks. All MDA-361 tumors exhibited a prompt and marked reduction in volume after tz/pac treatment; 10 mice achieved a complete response (CR) before receiving 'extended adjuvant' therapy with fulv (n=5) or neratinib/fulv (n=5). A CR was maintained with neratinib/fulv following tz/pac. However, mice treated with fulv alone, relapsed rapidly (p Conclusions: Neratinib/fulv but not fulv alone maintained complete responses of ER+/HER+ tumors following treatment with tz/pac or pertuzumab/tz/pac, reminiscent of the results in ExteNET. Neratinib treatment promoted ER transcriptional activity whereas ER downregulation with fulv was associated with increased HER2 signaling. In ER+/HER2+ breast cancer cells and tumors, neratinib/fulv synergistically inhibited growth, cyclin D1 expression, and AKT and MAPK activation, thus providing a plausible mechanism to explain the results in the ExteNET trial. Citation Format: Sudhan DR, Schwarz LJ, Guerrero-Zotano AL, Nixon M, Formisano L, Croessmann S, Gonzalez Ericsson PI, Sanders ME, Balko JM, Avogadri-Connors F, Cutler RE, Lalani AS, Bryce R, Auerbach A, Arteaga CL. Extended adjuvant neratinib/fulvestrant blocks ER/HER2 crosstalk and maintains complete responses of ER+/HER2+ tumors following treatment with chemotherapy and anti-HER2 therapy [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P1-13-08.

Background: BELLE-2 (NCT01610284) met its primary endpoint of a statistically significant improvement in progression-free survival (PFS) with the pan-phosphoinositide 3-kinase (PI3K) inhibitor BUP+FUL compared with PBO+FUL in postmenopausal pts with HR+, HER2– ABC (Baselga et al. Lancet Oncol 2017). Here, we report the final results for OS. Methods: Postmenopausal women with HR+, HER2– ABC refractory to aromatase inhibitors were enrolled in this randomized, double-blind, placebo-controlled, Phase III study. Pts were randomized (1:1) to receive either oral BUP (100 mg/day) or PBO plus intramuscular FUL (500 mg per standard of care). Randomization was stratified by PI3K pathway status in archival tumor tissue (activated [PIK3CA mutation or loss of phosphatase and tensin homolog expression], non-activated or unknown) and visceral disease status (present or absent). Baseline PIK3CA mutation status in circulating tumor (ct) DNA was assessed in a subset of pts. The key secondary endpoint was OS, defined as time from randomization to date of death due to any cause, in the full population (PI3K pathway activated, non-activated and unknown), main study cohort (PI3K pathway activated and non-activated) and PI3K pathway-activated subpopulation. Other secondary endpoints included safety. OS by PIK3CA status in ctDNA was also investigated. Results: 1147 pts received BUP+FUL or PBO+FUL. In the full population and main cohort, OS results trended in favor of BUP+FUL vs PBO+FUL, but were not statistically significant (Table). Post-treatment use of antineoplastic medications was balanced between both arms. In pts with PI3K pathway-activated status, OS improved by approximately 6 months (Table). In pts with mutant PIK3CA ctDNA status, there was a slight trend in favor of BUP+FUL vs PBO+FUL (Table). Median OS, monthsHR (95% CI)P valueFull population N=1147BUP+FUL n=57633.20.87 (0.74–1.02)0.045* PBO+FUL n=57130.4 Main cohort n=851BUP+FUL n=42730.90.91 (0.75–1.09)0.144* PBO+FUL n=42428.9 PI3K pathway activated n=372BUP+FUL n=18833.60.81 (0.61–1.08)0.075** PBO+FUL n=18427.5 ctDNA PIK3CA mutant n=200BUP+FUL n=8726.00.81 (0.56–1.17)0.127** PBO+FUL n=11324.8 ctDNA PIK3CA non-mutant n=387BUP+FUL n=19931.41.12 (0.83–1.50)0.774** PBO+FUL n=18836.8 *One-sided P value. **Nominal P value. The safety profile of BUP+FUL was consistent with that previously reported, with no new safety concerns. Grade 3/4 adverse events with at least 5% difference between BUP+FUL vs PBO+FUL, respectively, were increased alanine aminotransferase (25.7% vs 1.1%), increased aspartate aminotransferase (18.0% vs 2.8%), hyperglycemia (15.4% vs 0.2%) and rash (8.0% vs 0.0%). Conclusions: The BELLE-2 OS results showed a trend in favor of BUP+FUL, but did not meet statistical significance. More selective PI3K inhibition may improve treatment benefit and safety compared with pan-PI3K inhibition. Further assessment of the predictive benefit of ctDNA-confirmed PIK3CA mutations in this setting is required. Citation Format: Campone M, Im S-A, Iwata H, Clemons M, Ito Y, Awada A, Chia S, Jagiełło-Gruszfeld A, Pistilli B, Tseng L-M, Hurvitz S, Masuda N, Cortés J, De Laurentiis M, Arteaga CL, Jiang Z, Jonat W, Sellami D, El-Hashimy M, Le Mouhaër S, Sankaran B, Bourdeau L, Baselga J. Buparlisib (BUP) or placebo (PBO) plus fulvestrant (FUL) in postmenopausal patients (pts) with hormone receptor-positive (HR+), human epidermal growth factor receptor 2-negative (HER2–) advanced breast cancer (ABC): Overall survival (OS) results from BELLE-2, a randomized, phase III study [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr PD5-12.

Background: FGFR1 amplification (amp) occurs in ˜15% of breast cancers (BC) and associates with poor prognosis and resistance to endocrine therapy. CCND1 regulates cell cycle progression and is amplified in 15-20% of BC. Co-amp of FGFR1 occurs in 30-40% of CCND1amp tumors, suggesting the possibility of oncogene cooperativity. CDK4/6 inhibitors, which block the action of cyclin D1/CDK4 complexes at the G1-to-S transition, are approved for treatment of ER+ BC and FGFR inhibitors are in early phase clinical trials. Results: Between 11/2013 and 03/2017, 191 BCs from 188 patients with metastatic (M) or refractory locoregional recurrent (RLRR) BC at Vanderbilt (VICC) were profiled by targeted next gen sequencing (Foundation OneTM). These are included within the 2131 publicly available BC sequencing results in GENIE (Foundation OneTM and MSK-IMPACT). Among the GENIE cohort, rates were: FGFR1amp 7% (n=156), CCND1amp 12% (n=261) and CCND1/FGFR1co-amp 3% (n=58). Additional cases showed FGFR1 missense mutations (n=16) and deep deletions (n=5). When the analysis was limited to the VICC cohort allowing restriction to ER+ BC, FGFR1amp (16%) and CCND1amp (23%) rates are similar to rates in primary BC in TCGA (13% FGFR1 [p = 0.44] and 19% CCND1 [p = 0.24]). In GENIE, the most frequent co-mutations in FGFR1amp tumors were TP53 (31%), PIK3CA (21%), GATA3 (13%), CDH1 (11%) and MAP3K1 (10%). However, TP53 and PIK3CA mutations were less common among FGFR1amp tumors than FGFR1non-amp cases (p 0.016). Histopathologic correlation on tumors from our institution show a majority of FGFR1 and/or CCND1 amp BC (64%) were ER+/HER2–; 33% of ER+/FGFR1amp tumors were PR–. Distinctive histologic features associated with FGFR1 and/or CCND1 amp were lobular histology (17%) and neuroendocrine differentiation (14%), 0-10%TILs (94%) and high proliferative rate (46%). Conclusion: FGFR1amp and CCND1amp rates in TCGA are similar to those seen in MBC/LRRBC (GENIE) suggesting FGFR1 can function as both a driver mutation and de novo mechanism of endocrine resistance early in tumorigenesis. Frequent co-amp with CCND1 and lower rates of TP53 and PIK3CAmut also support a driver role for FGFR1amp and FGFR1/CCND1co-amp. The observation of neuroendocrine features in a subset of these tumors suggests lineage plasticity. This may be a consequence of genomic alterations promoting anti-estrogen resistance and is consistent with recently published BC outcome data associating neuroendocrine differentiation with higher grade ER+ tumors, frequent 8p amp, which includes FGFR1, and worse disease-free and overall survival. The frequency of FGFR1amp suggests genotype specific trials with FGFR inhibitors would be highly feasible. Whether FGFR1/CCND1 co-amplified tumors are candidates for treatment with a combination of FGFR and CDK4/6 inhibitors requires further investigation. Citation Format: Gonzalez-Ericsson PI, Estrada MV, Formisano L, Jansen VM, Mayer IA, Rexer BN, Abramson VG, Levy M, Balko JM, Stricker TP, Arteaga CL, Sanders ME. Genenomic landscape of breast cancers with FGFR1 amplification and FGFR1/CCND1 co-amplification revealed by targeted capture next generation sequencing [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr PD4-07.

ERBB2, the gene encoding HER2, is mutated in 2-4% of breast cancers. The HER2 tyrosine kinase inhibitor neratinib has shown clinical activity against breast cancers harboring HER2 activating mutations, suggesting these tumors depend on HER2 signaling. Co-occurring HER2 and HER3 (ERBB3) mutations have been reported in patients who respond to neratinib (Hanker et al., Cancer Discov. 2017) suggesting the possibility of cooperativity of both oncogenes. Co-expression of the mutant intracellular domains of HER2 and HER3 in HEK293 cells enhanced phosphorylation of HER3 and ERK compared to expression of either mutant alone, which was blocked by 100 nM neratinib. Interrogation of TCGA, METABRIC, Project GENIE, and Foundation Medicine datasets revealed that gain-of-function mutations in ERBB2 and ERBB3 co-occur with a statistically significant frequency. For example, in GENIE, ERBB2 mutations co-occur with mutations in ERBB3 (8.3% of ERBB2-mutant vs 2.3% of ERBB2 WT; q=1.37x10-10). We hypothesized that co-occurring mutations in HER2 and HER3 cooperate to enhance HER2 signaling and dependence and breast cancer progression. Thirty-four unique breast cancers were found to harbor co-occurring mutations in HER2 and HER3, the most common of which were ERBB2L755S/ERBB3E928G (n=10), ERBB2V777L/ERBB3E928G(n=6), and ERBB2L869R/Q/ERBB3E928G (n=4). Using co-immunoprecipitation assays with HER2 and HER3 antibodies in transfected HEK293 cells, we found that co-expression of HER3E928G with wild type (WT) HER2, or co-expression of HER2L755S or HER2L869R with HER3WT, slightly increased HER2-HER3 dimerization. However, binding was strongest between double mutants. This was accompanied by the highest levels of Y1289 p-HER3 in cells expressing both HER3E928G and each HER2L755S, HER2V777L, or HER2L869R compared to cells expressing each HER2 or HER3 mutant with a respective WT heterodimer partner. Structural modeling of the HER2L869R/HER3E928G double-mutant predicted that the HER3 mutation, located at the dimer interface, may enhance heterodimerization of the kinase domains through decreased bulk and electrostatic repulsion. We also noted that the HER2L755S mutation is predicted to be in close proximity to HER3E928G ( MCF7 “knock-in” cells incorporating HER2L755S, HER2V777L, or HER2L869R (or HER2WT) were stably transduced with HER3E928G or HER3WT. Co-expression of double mutants strongly enhanced estrogen-independent growth in 3D Matrigel over cells expressing either mutant alone. We are currently testing inhibitors of HER2/HER3 signaling, including neratinib ± trastuzumab, trastuzumab + pertuzumab, and the ERBB1-3 antibody mixture Sym013, to determine therapeutic strategies to block the cooperative growth induced by co-occurring HER2 and HER2 mutations. Conclusions: Co-expression of mutant HER2 and mutant HER3 promotes HER2/HER binding, HER3 phosphorylation, and breast tumor cell proliferation. We aim to identify therapeutic vulnerabilities for patients with co-occurring HER2 and HER3 mutations. Citation Format: Hanker AB, Koch JP, Ye D, Sliwoski G, Sheehan J, Kinch LN, Red Brewer M, He J, Miller VA, Lalani AS, Cutler, Jr. RE, Croessmann S, Zabransky DJ, Meiler J, Arteaga CL. Co-occurring gain-of-function mutations in HER2 and HER3 cooperate to enhance HER2/HER3 binding, HER-dependent signaling, and breast cancer growth [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr PD3-05.

Background: Around 15,000 US patients per year are diagnosed with de novo ABC. Due to the absence of prior systemic treatment for breast cancer, tumors of patients with de novo ABC may exhibit a different disease biology, which could result in different tumor responses compared with patients who have relapsed breast cancer. Ribociclib is an orally bioavailable cyclin-dependent kinase (CDK) 4/6 inhibitor. Results from MONALEESA-2, a double-blind, placebo-controlled, randomized Phase 3 trial (NCT01958021), demonstrated that first-line therapy with ribociclib + letrozole significantly improved progression-free survival (PFS) vs placebo + letrozole in patients with hormone receptor-positive (HR+), human epidermal growth factor receptor 2-negative (HER2–) ABC. Many patients with de novo ABC receive endocrine therapy in the first line and in subsequent lines; here we present results from the MONALEESA-2 study in a subpopulation of patients with de novo ABC. Methods: Postmenopausal women (N=668) with HR+, HER2– ABC who had no prior systemic therapy for ABC were randomized 1:1 (stratified by liver and/or lung metastases) to receive ribociclib (600 mg/day; 3-weeks-on/1-week-off) + letrozole (2.5 mg/day; continuous) or placebo + letrozole. Patients with de novo ABC were eligible. Additional eligibility criteria included measurable disease or ≥1 predominantly lytic bone lesion, Eastern Cooperative Oncology Group performance status ≤1, and adequate bone marrow/organ function. Prior CDK4/6 inhibitors or systemic therapy for ABC were prohibited. Patients may have received ≤14 days of letrozole or anastrozole for ABC. The primary endpoint was locally assessed PFS; a predefined subgroup analysis evaluated PFS in patients with de novo ABC. Results: In total, 227 patients with de novo ABC were enrolled. Patients with de novo ABC were equally distributed with 114 (34%) and 113 (34%) in the ribociclib + letrozole and placebo + letrozole arms, respectively. Median duration of exposure to study treatment in the ribociclib + letrozole vs placebo + letrozole arms was 14.1 vs 12.8 months. Treatment was discontinued in 84 (37%) patients with de novo ABC (ribociclib + letrozole vs placebo + letrozole, n [%]; 34 [30%] vs 50 [44%]). Reasons for treatment discontinuation (ribociclib + letrozole vs placebo + letrozole, n [%]) included disease progression (21 [18%] vs 41 [36%]), patient/physician decision (5 [4%] vs 6 [5%]), and adverse events (6 [5%] vs 3 [3%]). PFS was increased in patients with de novo ABC who received ribociclib + letrozole vs placebo + letrozole (hazard ratio=0.448 [95% confidence interval: 0.267–0.750]). The 12-month PFS event-free probability in patients with de novo ABC was 82% in the ribociclib + letrozole arm vs 66% in the placebo + letrozole arm. Conclusions: The combination of ribociclib + letrozole significantly improved PFS compared with placebo + letrozole in postmenopausal women with HR+, HER2– de novo ABC at diagnosis and therefore may become an important treatment option in the de novo ABC setting. Keywords: Advanced breast cancer; CDK4/6 inhibitor; Letrozole; Ribociclib Citation Format: O'Shaughnessy J, Petrakova K, Sonke GS, André F, Conte P, Arteaga CL, Cameron DA, Hart LL, Villanueva C, Jakobsen EH, Lindquist D, Souami F, Li X, Germa C, Hirawat S, Hortobagyi GN. First-line ribociclib plus letrozole in patients with de novo HR+, HER2– advanced breast cancer (ABC): A subgroup analysis of the MONALEESA-2 trial [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P4-22-05.

Background: Somatic mutations in ERBB2 are a new class of oncogenic drivers in HER2–non amplified MBC. Neratinib is an irreversible pan-HER tyrosine kinase inhibitor that inhibits the growth of ERBB2-mutant breast tumors in preclinical models and has encouraging single-agent clinical activity in patients (pts) with ERBB2-mutant, HER2–non amplified MBC. Bi-directional signaling between HER2 and ER may limit the effectiveness of endocrine and HER2 directed therapy, if each is given alone, in ER+ MBC with ERBB2 amplifications/mutations. Preclinical data suggest that dual blockade of ER and HER2 signaling results in enhanced anti-tumor activity in ER+ HER2+ MBC. SUMMIT, a multicenter multi-histology phase II 'basket' trial, is investigating the efficacy of neratinib monotherapy (in ER+ and ER– pts) and neratinib + fulvestrant (ER+ pts only) in ERBB2-mutant MBC. Methods: MBC pts with ERBB2 mutations documented by local testing were eligible and received oral neratinib 240 mg qd. Pts with ER+ MBC received fulvestrant 500 mg, a selective ER degrader, in addition to neratinib on d1 & 15 of month 1 then on d1 q4w. Patients received high dose loperamide prophylaxis during cycle 1. Primary endpoint is objective response rate (ORR) at 8w, defined using RECIST 1.1 and/or modified PERCIST assessments. Secondary endpoints include ORR, clinical benefit rate (CBR), progression free survival (PFS), and safety. Mutation profiling and central confirmation of ERBB2 mutation(s) from available fresh or archival tumor tissues and plasma DNA were performed retrospectively by next-generation sequencing (MSK-IMPACT). Clinicaltrials.gov: NCT01953926. Results: As of 23 Sep 2016, 35 efficacy-evaluable ERBB2-mutant MBC pts received neratinib, either as monotherapy (n=24) or in combination with fulvestrant (n=11). Efficacy findings are shown in the table. The overall safety profile of neratinib + fulvestrant was similar to that previously reported with neratinib monotherapy. Grade 3 diarrhea rate was 24% with neratinib monotherapy and 18% with neratinib + fulvestrant. Neratinib monotherapyNeratinib + fulvestrant(n=24)(n=11)Best Overall Response (confirmed and unconfirmed), n (%)8 (33.3)6 (54.5)[95% CI][15.6–55.3][23.4–83.3]CR3 (12.5)2 (18.2)PR5 (20.8)4 (36.4)aORR at 8 weeks, n (%)8 (33.3)5 (45.5)[95% CI][15.6–55.3][16.7–76.6]CR2 (8.3)2 (18.2)PR6 (25.0)3 (27.3)ORR confirmed, n (%)5 (20.8)2 (18.2)b[95% CI][7.1–42.2][2.3–51.8]CR3 (12.5)1 (9.1)PR2 (8.3)1 (9.1)CBR, n (%)10 (41.7)6 (54.5)[95% CI][22.1–63.4][23.4–83.3]CR, complete response; PR, partial response.aThere was 1 pt with PR at week 16; bAt time of data cut-off 4 pts are still on treatment. Conclusions: Encouraging clinical activity has been observed with neratinib + fulvestrant in heavily pretreated pts with ERBB2-mutant, ER+ MBC. Clinical efficacy in the ER+ MBC cohort met pre-specified efficacy requirements; a confirmatory trial of neratinib + fulvestrant for targeting ERBB2 mutations in ER+ MBC is warranted. The safety profile of neratinib was acceptable and diarrhea was manageable with loperamide prophylaxis. Citation Format: Hyman D, Piha-Paul S, Saura C, Arteaga C, Mayer I, Shapiro G, Loi S, Lalani A, Xu F, Cutler R, Butturini A, Bryce R, Meric-Bernstam F, Baselga J, Solit D. Neratinib + fulvestrant in ERBB2-mutant, HER2–non-amplified, estrogen receptor (ER)-positive, metastatic breast cancer (MBC): Preliminary analysis from the phase II SUMMIT trial [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr PD2-08.

The Human Epidermal Growth Factor Receptor 2 (HER2) oncogene is amplified in one-fifth of breast cancers (BC). However, development of resistance against standard anti-HER2 therapies poses a major clinical challenge. Anti-tumor efficacy of HER2-targeting agents depends on inhibition of the downstream phosphatidylinositol-3 kinase (PI3K) signaling cascade. Gain-of-function somatic mutations in the gene encoding the PI3K catalytic subunit p110alpha (PIK3CA), co-expressed in about 40% of HER2+ BC, have been implicated in conferring resistance to HER2 monoclonal antibody herceptin. The single amino acid alteration H1047R within the kinase domain of PIK3CA is one of three hot spot mutations prevalent in BC. Previously, we demonstrated that introduction of H1047R mutation in HER2-overexpressing MCF10A mammary epithelial cells enhances cellular transformation and decreases herceptin sensitivity by inducing secretion of endogenous ErbB ligand heregulin. However, genetic ablation of HER3, the major co-receptor for HER2 and the solitary receptor for heregulin, was insufficient for complete inhibition of cell growth, indicating the existence of additional mechanism/s responsible for the heightened aggressiveness and decreased drug sensitivity of HER2/H1047RPI3K cells. In the current study, we looked further into the molecular changes within these cells that might be responsible for these phenomena. When compared with the HER2/WTPI3K cells, the HER2/H1047RPI3K cells revealed a significant increase in CD44high/CD24low/negative populations, common markers of BC stem cells, as well as molecular and phenotypic changes associated with epithelial-to-mesenchymal transition. These observations are in agreement with previously published report on mouse model of HER2/H1047RPI3K BC. Further analyses demonstrated additional stem cell-associated characteristics in HER2/H1047RPI3K cells, such as expression of angiogenic and inflammatory cytokines, ability to induce chemotaxis and invasion, activation of TGFb and NFKb signaling pathways. Connectivity map (CMap) analysis of the gene expression signatures from HER2/H1047RPI3K cells revealed a negative association with those from BC cells treated with 17AAG, an inhibitor of the heat shock protein 90 (HSP90). In line with this, HER2/H1047RPI3K-expressing cells are found to be more sensitive to HSP90 inhibition compared to the pan-ErbB inhibitor lapatinib. Cancer stem cells are implicated in drug resistance and tumor recurrence. Enrichment of cell population expressing high levels of stem cell markers and stem cell-related features could be one of major mechanisms by which BC cells co-expressing HER2 and H1047RPI3K adapt to anti-HER2 therapeutic agents. Acute dependence on molecular chaperone HSP90 provides a unique, yet practical opportunity to effectively inhibit tumors harboring both molecular alterations, since HSP90 inhibitors have already shown encouraging clinical activity in herceptin-resistant setting. Citation Format: Surendran S, Bhola N, Arteaga CL, Chakraborty K, Chakrabarty A. Introduction of H1047R oncogenic mutation of PI3K p110alpha subunit in HER2-overexpressing mammary epithelial cells confers a "stem-like" phenotype and acute sensitivity to HSP90 inhibition [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-08-10.

BACKGROUND: Lucitanib is a potent, oral antiangiogenic tyrosine kinase inhibitor of Vascular Endothelial Growth Factor Receptors 1-3 (VEGFR1-3), Platelet-Derived Growth Factor Receptors alpha and beta (PDGFRα/β), and Fibroblast Growth Factor Receptors 1-3 (FGFR1-3). FGF aberrancies (amplification of FGFR1,or 11q[amplicon containing FGF ligands 3, 4, and 19]), are genomic alterations observed in over 20% of breast cancer pts and promote cancer proliferation and survival. METHODS: MBC pts who had received at least 1 metastatic line of therapy were randomized 1:1 to 10 or 15 mg QD of lucitanib. Stratification was based on local assessment of FGF aberrancy; pts with both FGFR1 and 11q-amplified tumors were stratified as FGFR1 amplified. Central confirmation of FGFR1 or 11q amplification was done using Abbott FISH probes (FGFR1 or 11q copy number ≥ 6 and a ratio of FGFR1 or 11q to centromere ≥ 2). Investigator-assessed progression-free survival (PFS) was the primary endpoint. Secondary endpoints included objective response rate (ORR) per RECIST 1.1, disease control rate (DCR), duration of response (DR), and incidence of treatment-emergent adverse events (TEAE). RESULTS: Enrollment completed in 3/2016; 178 pts that received at least 1 dose of lucitanib are included in this analysis (baseline characteristics in Table 1). Due to grade 3 hypertension in the 15 mg group (46% vs 37% in 10 mg group), enrollment to the 15 mg group was halted. Overall, most pts (97%) experienced at least 1 TEAE, with the most frequently (≥ 30%) occurring events being hypertension (73%), fatigue (48%), nausea (43%), hypothyroidism (40%), and headache (33%). Grade ≥ 3 TEAEs occurred in 66% of pts, with hypertension as the most frequent event (40%) followed by proteinuria and hyponatremia (both 6%). AEs were manageable with dose interruption or reduction, with approximately 8% of pts ending treatment due to an AE. Current median PFS is 3.5 mos (95% CI 2.8-4.6; range 0.62-12.95) and 2.6 mos (95% CI 1.8-2.9; range 0.82-18.87) respectively for the 10 mg and 15 mg treatment groups. No differences in clinical activity were observed by treatment group, FGF aberrancy, hormone receptor or HER2 status. Of the 168 evaluable pts, confirmed ORR was 3%; overall DCR was 27% (32% for pts in the 10 mg group compared to 20% for the 15 mg group); overall mean (standard deviation) DR of 3.3 (1.8) mos. Baseline Characteristics 10 mg QD15 mg QD N=109N=69Age (years)Median5653Range27-8227-80SexFemale109 (100%)67 (97%)Male02 (3%)ECOG PSmissing5 (5%)2 (3%)051 (47%)30 (43%)153 (49%)37 (54%)Number of prior anticancer therapies in the metastatic setting> 332 (29%)21 (30%)3-648 (44%)32 (46%)> 629 (27%)16 (23%)Endocrine/HER2 statusmissing7 (6%)1 (1%)ER+ or PR+74 (68%)50 (73%)HER2+12 (11%)7 (10%)TNBC16 (15%)11 (16%)FGFR aberrancyFGFR1 amplified54 (49%)29 (42%)11q amplified31 (28%)24 (35%)FGFR1 and 11q amplified13 (12%)9 (13%)FGFR1 and 11q non-amplified11 (10%)7 (10%) CONCLUSION: At 10 mg QD, lucitanib has modest activity with manageable toxicity in this heavily pretreated pt population. Future clinical development for lucitanib may focus on alternative biomarkers to identify sensitive tumors and rational combinations with other anti-cancer drugs. Citation Format: Mayer IA, Arteaga CL, Nanda R, Miller KD, Jhaveri K, Brufsky AM, Rugo H, Yardley DA, Vahdat LT, Sadeghi S, Audeh MW, Rolfe L, Litten J, Knox A, Raponi M, Tankersley C, Isaacson J, Wride K, Morganstern DE, Vogel C, Connolly RM, Gradishar WJ, Patel R, Pusztai L, Abu-Khalaf M. A phase 2 open-label study of lucitanib in patients (pts) with FGF aberrant metastatic breast cancer (MBC) [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-11-03.

Background: Amplification/overexpression of ERBB receptors and/or ligands has been associated with resistance to anti-HER2 therapies. Pan-HER is a mixture of six antibodies targeting each of the ERBB receptors, EGFR, HER2 and HER3, with synergistic pairs of antibodies. Each pair of antibodies simultaneously blocks ligand binding and/or induces target degradation, thus preventing compensatory mechanisms to anti-ERBB therapies. We examined the antitumor activity of Pan-HER against drug-sensitive and -resistant HER2+ breast cancer cells and xenografts. Results: Pan-HER exhibited potent growth inhibitory activity against a panel of HER2+ breast cancer cells (BT474, MDA-453, MDA-361, SUM190, HCC1954, UACC893 and SKBR3). Growth inhibition was associated with internalization and degradation of EGFR, HER2 and HER3. Pan-HER was superior to the combination of trastuzumab/pertuzumab (TP) against HER2+/PIK3CA mutant MDA-361, HCC1954, UACC893 and MDA-453 cells. We next compared the effect of Pan-HER against BT474, HCC1954 and MDA-361 xenografts established in nude mice to that of trastuzumab/lapatinib (TL), TP and T-DM1. All treatments were effective across the panel of xenografts. In mice with MDA-361 tumors, Pan-HER and TP were superior to TL. Immunoblot analysis showed significant downregulation of EGFR, HER2 and HER3 only in tumors treated with Pan-HER. After a complete response, treatment was discontinued. Among mice with BT474 xenografts treated with TP, TL and T-DM1, 25-50% of mice exhibited a tumor recurrence within 50 weeks of follow-up, while no recurrences were registered in mice treated with Pan-HER. Tumors recurring after TP and T-DM1 expressed significantly higher HER3 and P-HER3 protein levels and NRG1 mRNA levels. HCC1954 xenografts recurring after T-DM1 also overexpressed NRG1 mRNA compared to tumors before therapy. We next examined the effect of Pan-HER against trastuzumab-resistant HR6 (BT474) cells (Ritter et al. CCR 2007) and HCC1954 and UACC893 cells with acquired resistance to T-DM1 (TDR; IC50 >5-, >6- and 600-fold in HR6, UACC893-TDR and HCC1954-TDR cells, respectively, vs. parental cells). All T-DM1-resistant cells expressed significantly higher HER3 and P-HER3 protein levels and NRG1 mRNA and protein levels. Treatment with the HER3 neutralizing antibody LJM716 resensitized HR6 and HCC1954-TDR cells to T-DM1, suggesting a causal association between the NRG1-HER3 axis and drug resistance. Mice with HR6 tumors were treated with Pan-HER, TL, TP and T-DM1. Only Pan-HER arrested HR6 tumor growth and downregulated EGFR, HER2, HER3, P-HER3 and P-AKT. Finally, HCC1954-TDR tumors rapidly grew in vivo despite treatment with T-DM1. Administration of Pan-HER to mice bearing HCC1954-TDR xenografts growing in the presence of T-DM1, induced rapid tumor regressions. Conclusions: These data suggest that multitarget therapeutic interventions, such as Pan-HER, which simultaneously remove and/or block all ERBB receptors and ligands, are a feasible and effective approach against HER2-overexpressing cancers both sensitive and resistant to anti-HER2 therapies. Citation Format: Schwarz LJ, Hutchinson KE, Estrada MV, Sanders ME, Dugger TC, Formisano L, Guerrero AL, Red-Brewer M, Young CD, Lantto J, Pedersen MW, Kragh M, Horak ID, Arteaga CL. Pan-HER, an antibody mixture with antitumor activity against drug-resistant HER2-overexpressing breast cancers with high ERBB ligand expression [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-12-09.

Background: Approximately 20% of patients with early ER+ breast cancer (BC) treated with adjuvant antiestrogen therapy eventually relapse with endocrine-resistant metastatic disease. We hypothesized that profiling newly diagnosed ER+ BC that persist following prolonged estradiol deprivation with letrozole would identify genomic alterations associated with endocrine resistance. Methods: We treated 57 postmenopausal women (median 77 years; range 60-86) with ER+/HER2– BC with neoadjuvant letrozole (median 7.5 months; range 3-36) followed by surgery and adjuvant endocrine therapy. Patients were followed with serial ultrasounds and defined as non-responders if they developed recurrent locally or metastatic disease, or had a preoperative endocrine prognostic index (PEPI) ≥4 (composite score of post-treatment ER, Ki67, T and N status). Post-treatment specimens were profiled by RNA-seq and targeted capture NGS of >300 cancer-related genes. We screened for variants with a high probability of disrupting protein function (GERP score >4) and excluded likely germline variants by filtering out every alteration not present in COSMIC, if the variant had an allele frequency >0.1% as per the ExAC dataset. Results: Ten patients (17.5%) had a PEPI 0 score, 31 (54%) were PEPI 1-3, and 16 (28%) were PEPI ≥4. After a median follow-up of 50 months (12-100), 9 patients (15.7%) had recurred with metastatic disease (4 with PEPI 1-3, 5 with PEPI ≥4). We identified 294 variants with a median coverage >250x (206 nonsynonymous, 21 nonsense, 58 indels, 8 splice site). Recurrent mutations included PIK3CA (38%), KMT2C (28%), CDH1 (15%), NF1 (12 %), TP53 (10%), MAP3K1 (7%), ERBB2 (7%) and ESR1 (5%). Recurrent amplifications were identified in MCL1 (31%), GNAS (19%), CCND1 (16%), FYN (14%), AURKA (12%), and ERBB2 (10%), while recurrent deletions were found in DUSP4 (12%), NCOR1 (8%) and NF1 (6%). Compared to alterations reported in untreated ER+ breast cancers in TCGA, we observed a significant increase in KMT2C, NF1, MCL1 and FYN alterations (FDR Conclusions: Genomic profiling of residual ER+ breast cancers treated with prolonged neoadjuvant letrozole revealed a different mutational landscape than primary untreated ER+ BC. These alterations may be associated with poor response to estrogen deprivation in early breast cancer and deserve further study. Citation Format: Guerrero AL, Stricker T, Hutchinson KE, Formisano L, Giltnane J, Fidalgo A, Schwarz LJ, Gavila J, Guillen V, Lluch A, Ruiz A, Arteaga CL. Genomic profiling of residual ER+ breast cancers treated with prolonged neoadjuvant letrozole reveals novel alterations in clinically resistant tumors [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-10-01.

Background: Human epidermal growth factor receptor 2 (ERBB2 or commonly known as HER2) missense mutations have been reported in 2-4% of breast cancers and occur primarily in the absence of HER2 gene amplification. Based on TCGA, approximately 60% of these tumors are hormone-dependent and express estrogen receptor (ER) a. Among ER+ breast cancers with HER2 missense mutations, more than 80% are in the HER2 kinase domain. We examined herein whether ER+/HER2 mutant breast cancer cells are resistant to anti-estrogen therapies and, thus, whether they should be treated with combined ER and HER2 inhibitors. Methods: Three common HER2 activating mutations (G309A, L755S, V777L) and wild type (WT) HER2 were incorporated into ER+ MCF7 cells using AAV-mediated homologous recombination. The isogenic incorporation of a heterozygous mutation more accurately represents primary human tumors as compared to transfection and overexpression of exogenous vectors. We examined cell viability and ER transcriptional activity, using an ERE-luciferase reporter, in response to estrogen deprivation and treatment with fulvestrant (a selective ER downregulator) and neratinib (an irreversible, pan-HER tyrosine kinase inhibitor), either alone or in combination. Signaling downstream mutant HER2 was examined by immunoblot analysis. In vivo anti-tumor efficacy of fulvestrant ± neratinib is currently being assessed in ovariectomized athymic mice bearing MCF7/HER2V777L xenografts. Results: MCF7 cells containing HER2 kinase missense mutations (L755S and V777L), but not cells with HER2WT or an extracellular domain mutation (G309A), were able to proliferate exponentially in estrogen-free medium. MCF7/HER2L755S and MCF7/HER2V777L were also resistant to 1 mM fulvestrant, despite fulvestrant's ability to downregulate ER in these cells. Additionally, MCF7/HER2L755S and MCF7/HER2V777L showed increased levels of pERK and p70S6K. Treatment with 200 nM neratinib potently inhibited growth of MCF7/HER2L755S and MCF7/HER2V777L in estrogen-free conditions and resensitized them to fulvestrant while partially downregulating HER2 levels. Addition of 1 nM estradiol markedly rescued all three HER2 mutant cells from neratinib-induced cell death suggesting that the inhibition of both ER and mutant HER2 is required for tumor cell apoptosis. Using ERE-luc reporter assays, neratinib did not inhibit basal or estrogen-induced ER transcriptional activity or ERα Ser118 phosphorylation, thus not supporting HER2 mutation-to-ER crosstalk in these genetically engineered cells. This result also suggests that the ER and HER2 mutant pathways can operate independently and it is the dual pathway inhibition that results in synthetic lethality. Conclusions: These data suggest that, in ER+ breast cancers and similar to HER2 gene amplification, HER2 kinase domain mutations induce resistance to antiestrogen therapies. Therefore, we propose simultaneous therapeutic targeting of both ER and HER2 signaling pathways is required for maximal inhibition of ER+ breast cancers also harboring HER2 activating mutations, as is currently being investigated in the phase II SUMMIT trial (NCT01953926). Citation Format: Croessmann S, Zabransky DJ, Cutler, Jr. RE, Lalani AS, Park BH, Arteaga CL. Inhibition of mutant HER2 results in synthetic lethality when combined with ER antagonists in ER+/HER2 mutant human breast cancer cells [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr PD2-05.

Background: The addition of first-line ribociclib (RIB; cyclin-dependent kinase 4/6 inhibitor) to letrozole (LET) significantly improved progression-free survival (PFS) compared with placebo (PBO) + LET in patients (pts) with hormone receptor-positive (HR+), human epidermal growth factor receptor 2-negative (HER2–) advanced breast cancer (ABC) in the Phase III MONALEESA-2 study. Identifying biomarkers that predict response to treatment remains a key challenge in pts with HR+ ABC. Here we analyze results from MONALEESA-2 by molecular alterations detected in circulating tumor DNA (ctDNA) at baseline, including PIK3CA mutations and other alterations considered to be important in HR+ ABC. Methods: Postmenopausal women (N=668) with HR+, HER2– ABC who had not received any prior therapy for ABC were randomized 1:1 to RIB (600 mg/day; 3-weeks-on/1-week-off) + LET (2.5 mg/day; continuous) or PBO + LET. The primary endpoint was PFS. Biomarker analysis of the ctDNA mutation profile was an exploratory endpoint. Plasma samples for ctDNA analysis were collected at baseline and end of treatment. ctDNA was analyzed using next-generation sequencing with a targeted panel of ˜550 genes. Results: Baseline ctDNA was successfully sequenced in 494 pts (RIB + LET: n=212; PBO + LET: n=215); 67 (14%) of 494 pts were removed from the analysis due to limited tumor DNA in circulation. 427 (86%) pts had ≥1 alteration, including 1,573 mutations, 513 short insertions/deletions, 166 amplifications, and 8 translocations. Alterations (frequency) were commonly observed in the following genes: PIK3CA (33%), TP53 (12%), ZNF703/FGFR1 (5%), and ESR1 (4%), and in genes involved in receptor tyrosine kinase (RTK) signaling (12%). RIB + LET treatment benefit was consistent in pts with wild-type (WT) and altered PIK3CA, and in pts with WT and altered TP53 (Table). RIB + LET improved PFS regardless of RTK or ZNF703/FGFR1 alterations. However, there was a weak trend for increased benefit in pts with WT vs altered RTK genes and in pts with WT vs altered ZNF703/FGFR1 genes. These results should be interpreted with caution due to the small number of pts with these alterations. There were too few ESR1 alterations for firm conclusions to be drawn. Events, n/NMedian PFS, months Gene(s)RIB + LETPBO + LETRIB + LETPBO + LETHazard ratio (95% confidence interval)PIK3CAWT54/14393/14229.614.70.44 (0.31–0.62)Altered40/6955/7319.212.70.53 (0.35–0.81)TP53WT72/180129/19427.614.70.44 (0.33–0.59)Altered22/3219/2110.25.50.43 (0.23–0.83)ZNF703/FGFR1WT88/202139/20524.814.60.47 (0.36–0.62)Altered6/109/1010.611.40.73 (0.23–2.29)RTKWT81/189128/18724.814.40.46 (0.35–0.61)Altered13/2320/2821.311.40.72 (0.34–1.53) Conclusions: Consistent RIB + LET treatment benefit was observed compared with PBO + LET, irrespective of the status of baseline ctDNA biomarkers. Citation Format: Hortobagyi GN, Stemmer S, Campone M, Sonke GS, Arteaga CL, Paluch-Shimon S, Petrakova K, Villanueva C, Nusch A, Grischke E-M, Chan A, Jakobsen E, Marschner N, Hart LL, Alba E, Ohnstand HO, Blau S, Yardley DA, Solovieff N, Su F, Germa C, Yap Y-S. First-line ribociclib + letrozole in hormone receptor-positive, HER2-negative advanced breast cancer: Efficacy by baseline circulating tumor DNA alterations in MONALEESA-2 [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr PD4-06.

Background: Activation of the phosphoinositide-3-kinase (PI3K) pathway has been associated with resistance to endocrine therapy in estrogen receptor-positive (ER+) breast cancer. Recently, we reported a Stand Up 2 Cancer (SU2C) Phase Ib trial of buparlisib, an oral, reversible, pan-PI3K inhibitor in combination with the aromatase inhibitor letrozole in patients with metastatic ER+/HER2– breast cancer (n=51) who had previously progressed on endocrine therapy. In this study, 31% of patients demonstrated clinical benefit (CR, PR and SD ≥6 months; Mayer et al. JCO 2014). Clinical activity was observed in patients with PIK3CA-wild type (PIK3CA-WT) and PIK3CA-mutant (PIK3CA-MUT) tumors. We performed targeted next-generation sequencing (tNGS) to identify somatic alterations associated with clinical benefit to the combination therapy. Methods: tNGS was performed using Memorial Sloan Kettering-Integrated Mutation Profiling of Actionable Cancer Targets (MSK-IMPACT) in DNA extracted from tumors of 33 study patients (22 samples from archived primary/untreated and 11 from metastatic biopsies). Detected alterations were tested for association with clinical benefit (Fisher's exact test) and progression-free survival (PFS; log-rank test). For PFS analysis, patients were censored if they discontinued buparlisib for toxicity. Results: The most commonly detected alterations occurred in PIK3CA (36%), TP53 (30%), MAP3K1 (27%), GATA3 (24%), CCND1 (24%), CDH1 (21%) and PTEN (21%). Additional alterations of note included FGFR1 amplification (15%), MYC amplification (12%), ESR1 mutations (6%) and ERBB2 mutations (6%). Probable inactivating mutations occurring in MAP3K1 (MAP3K1-MUT) were significantly associated with improved clinical benefit, regardless of other mutations (6/9 patients, 67%, P=0.044). PIK3CA-MUT tumors trended toward greater clinical benefit (7/12, 58%, P=0.067). Patients with coexisting PIK3CA-MUT and MAP3K1-MUT tumors derived the largest clinical benefit (5/7, 70% P=0.07) compared to patients with only PIK3CA-MUT (2/5; 40%, P=1.0) or only MAP3K1-MUT tumors (1/2; 50%, P=1.0). Only 2/19 (11%) patients with PIK3CA-WT/MAP3K1-WT cancers benefitted from treatment. Both MAP3K1 and PIK3CA alterations were each also associated with increased PFS (p=0.03 and p=0.009, respectively). Three of 5 (60%) patients with tumors with FGFR1 amplification achieved clinical benefit (including a MAP3K1-MUT tumor and a PIK3CA-MUT/MAP3K1-MUT tumor), suggesting that FGFR1 may preferentially signal via PI3K and/or FGFR1 amplifications are not associated with resistance to the combination of aromatase inhibitors and PI3K inhibitors. Conclusions: Both MAP3K1 and PIK3CA are mutated at higher frequencies in luminal A breast cancer, suggesting that this alteration pattern (MAP3K1-MUT + PIK3CA-MUT) is a surrogate for low grade ER+ breast cancers with PI3K dependence. It is also possible that MAP3K1 mutations may predispose tumor cells to sensitivity to PI3K inhibition, but this speculation requires further investigation. Finally, patients with ER+/FGFR1-amplified cancers appeared to derive clinical benefit from combined therapy with letrozole and buparlisib. Citation Format: Balko JM, Hicks M, Berger MF, Solit DB, Bouvier N, Sanders ME, Estrada MV, Won H, Cantley LC, Mayer IA, Arteaga CL. Genomic alterations indicative of a luminal A subtype associate with clinical benefit to buparlisib and letrozole in endocrine-resistant ER+/HER2– metastatic breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-41.

Background: Dysregulation in cell cycle checkpoints is common in cancer. Small molecule inhibitors that target the CDK4/6/cyclinD1 pathway of the cell cycle are in clinical development. Recently the combination of the CDK4/6 inhibitor palbociclib and the aromatase inhibitor letrozole was approved for the treatment of post-menopausal women with ER+/HER2- advanced breast cancer. However, not all patients benefit from CDK4/6 inhibitors and a significant fraction of them eventually progress on these agents, underscoring the need to develop potent therapeutic strategies to circumvent drug resistance. Methods: We performed a high-throughput RNA interference (RNAi) kinome screen targeting 720 kinases to identify targetable molecules whose inhibition, in combination with the CDK4/6 inhibitor LEE011 (ribociclib), induced synthetic lethality in MCF7 ER+ breast cancer cells. PDK1 RNAi oligonucleotides and the PDK1 inhibitor GSK2334470 in combination with each of the CDK4/6 inhibitors, palbociclib and LEE011, were tested against ER+ breast cancer cells. In vivo anti-tumor efficacy of LEE011 and GSK2334470 was assessed in ovariectomized athymic nude mice bearing MCF7 xenografts. Results: A siRNA kinome screen identified PDK1 as the top RNA whose downregulation sensitized MCF7 cells to CDK4/6 inhibitors. This was confirmed with independent siRNAs in ER+ MCF7, T47D, HCC1428 and HCC1500 breast cancer cells. Pharmacological inhibition of PDK1 with the ATP-competitive, small molecule inhibitor GSK2334470 in combination with each of the CDK4/6 inhibitors, LEE011 and palbociclib, synergistically inhibited proliferation and increased apoptosis of MCF7 and T47D cells (combination index 0.19-0.89). LEE011-resistant MCF7 and T47D cells were generated by chronic treatment with doses of LEE011 up to 1 µM. Drug-resistant cells displayed increased levels of PDK1, phosphorylated Rb, and phosphorylated S6 ribosomal protein (pS6), an effector of the PDK1 substrate p70S6K, compared to parental drug-sensitive cells. Inhibition of PDK1 with siRNA or GSK2334470 re-sensitized the LEE011-resistant cells to the CDK4/6 inhibitors. Genetic (RNAi) and pharmacological inhibition of PDK1 (with GSK2334470) abrogated pS6 levels whereas inhibition of AKT with the small molecule inhibitor MK2206 did not affect pS6 levels, suggesting PDK1 can induce resistance to CDK4/6 inhibitors via p70S6K/pS6 signaling in an AKT-independent manner. The effects observed in cell lines in culture were recapitulated in vivo using MCF7 xenografts established in ovariectomized nude mice in the absence of estrogen supplementation. Treatment with GSK2334470 and LEE011 induced tumor regressions (8/8 tumors by RECIST criteria) more potently than either drug alone. Conclusions: These data support a critical role of PDK1 in mediating acquired resistance to CDK4/6 inhibitors in ER+ breast cancer cells. Co-targeting of the PDK1 and CDK4/6 pathways may overcome resistance to CDK4/6 inhibitors and is worthy of further translational and clinical investigation in patients with ER+ breast cancer. Citation Format: Jansen VM, Bhola NE, Bauer JA, Formisano L, Moore P, Koch J, Arteaga CL. Inhibition of 3-phosphoinositide dependent protein kinase 1 (PDK1) synergizes with CDK4/6 inhibitors against ER-positive breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr PD2-06.

Background: Neoadjuvant chemotherapy (NAC) is frequently used in triple negative breast cancers (TNBC), and patients who achieve pathological complete response (pCR) following NAC have improved outcomes over those with residual disease (non-pCR). Unfortunately only 30% of TNBC patients achieve pCR, with no good treatment options for other 70% with non-pCR. Tumor infiltrating lymphocytes (TILs) in TNBC are predictive of pCR to NAC, and TILs found in the residual disease further prognosticate patients with residual disease into an improved prognosis subset. These data suggest there is an immune component to TNBC that might be affected by chemotherapy. Given the advent of immunotherapy trials in breast cancer, specifically in combination with NAC, there is an unmet need to gain a better understanding of the immune microenvironment in TNBC. Objective: We investigated the role of NAC on the immune microenvironment by examining breast cancer samples before (diagnostic biopsy) and after (residual disease) NAC, as these data are pertinent to the design and analysis of ongoing clinical trials in breast cancer combining immunotherapy with concurrent chemotherapy. Methods: RNA was extracted from 51 paired (pre- and post-NAC) breast cancer specimens with extensive residual disease, and analyzed by nCounter analysis for expression of >750 immune-related genes. Functional immune signature scores were generated, compared between matched pre- and post-NAC, and stratified by breast cancer molecular subtype (luminal (n=4), Her2-enriched (n=5), basal-like (n=17)). DNA extracted from a subset of these samples was utilized for T cell receptor sequencing to explore changes in T cell clonality following NAC in each subtype. Results: Across all subtypes of breast cancer, immune scores decreased after NAC consistent with a broad decrease in TILs observed histologically. When samples were stratified by molecular subtype using the PAM50 analysis, luminal A/B and basal-like patients demonstrated a decrease in immune signatures after NAC, while Her2-enriched patients exhibited a global increase in immune scores. Importantly, basal-like patients had the greatest immune signature changes, including decreases in T cell functions and CD8+ T cell, T helper and T regulatory signatures (p>0.05). Conversely, these signatures were increased after NAC in the Her2-enriched molecular subtype. Analyses of T-cell clonality are ongoing, but should yield insight into the effect of NAC on the landscape of effector T-cells in the micro-environment. Conclusions: Our work suggests that NAC decreases immune infiltrate signatures, specifically related to effector T cells in patients who do not achieve pCR. However, we observed an increase after NAC in the same signatures in Her2-enriched patients, suggesting that these patients' tumors may respond differently to chemotherapy on the immune-molecular level. As clinical trials progress in TNBC with the combination of chemotherapy and immune therapy, more work is needed to understand who might benefit from these therapy regimens. Citation Format: Hicks MJ, Estrada MV, Sanders ME, Salgado R, Cook RS, Arteaga CL, Balko JM. Differential effect of chemotherapy on immune gene expression signatures based on molecular subtype of breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-04-03.

Background: mTOR inhibitors can activate p73, a pro-apoptotic member of the p53 family, and enhance the sensitivity of breast cancer cells to cisplatin and paclitaxel. Thus, we hypothesized that combined use of the mTOR inhibitor everolimus, cisplatin, and paclitaxel would have synergistic anti-tumor effects in TNBC. Methods: Patients with clinical stage II/III TNBC were assigned (2:1) to weekly cisplatin 25 mg/m2 + paclitaxel 80 mg/m2 ± daily everolimus 5 mg for 12 weeks, until definitive surgery. Biopsy specimens were obtained in 100% of patients at baseline, at day 5 of cycle 1 and at surgery. Primary endpoint was pathological complete response (pCR). The study design provided 90% power to detect a difference in pCR rate of 35% vs. 20% with a two-sided significance level equal to 0.1 (type I error) for each arm. Results: A total of 145 patients were accrued between 2009 and 2013. To date, 14 patients have not yet completed surgery, and 11 patients were not evaluable (study discontinuation due to disease progression, toxicity or withdrawal). Baseline characteristics between arms were similar and well balanced: median age was 52 (28 – 81), median breast tumor size was 2 cm (0.1 – 7.6), 72% of tumors were histologic grade III, and 70% of patients had clinical stage III disease. Clinical outcomes are summarized in Table 1. Clinical outcomesEvaluable patientsEverolimusN = 82PlaceboN = 38OverallN = 120Pathological responseN%N%N%pCR (pT0N0)293516424538Near pCR (pT1aN0)17215132218Residual disease364417455344Clinical ResponseN%N%N%CR465721556759PR232814373629SD1113381511PD220-21 Despite similar rates of pCR and clinical response in both arms, the combination of cisplatin/paclitaxel provided comparable pCR rates to anthracycline/cyclophosphamide/taxane containing regimens administered for longer periods of time. Most common adverse events are summarized in Table 2. Adverse events(%)Grade 1 and 2Grade 3 and 4 EvePlacEvePlacNeutropenia26272611Thrombocytopenia409 Anemia5975 2Rash4927 2Fatigue61753 Nausea6064 2Diarrhea30292 Dyspepsia3035 Mucositis3920 Hyperglycemia5140 2Transaminase elevation60183 Pneumonitis1 TNBC subtyping (Lehmann et al. JCI 2009), DNA mutations and alterations, as well as markers of proliferation, apoptosis, PI3K/mTOR and DNA damage response signaling will be presented. Preliminary analysis of Ki67 in a subset of tumors suggest that a reduction in Ki67 (day 5 biopsy) is associated with increased pCR rate. Tumors with androgen receptor expression exhibited a very low pCR rate. Conclusion: To our knowledge, this is the largest randomized neoadjuvant study in TNBC with a PI3K/mTOR pathway inhibitor. Results suggest that the paclitaxel/cisplatin combination is well tolerated and active in TNBC. The addition of Everolimus was associated with more adverse events and did not improve pCR or clinical response rates. A molecular signature or biomarker predictive of benefit from the paclitaxel/cisplatin combination is currently under investigation, and will be presented at the time of the meeting. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr PD1-6.

Introduction: The lack of estrogen receptor (ER) expression is the primary cause of de novo resistance of breast cancers to endocrine therapy. In contrast, in most cases of acquired endocrine resistance, ER is expressed and other mechanisms of resistance have been proposed, such as ER mutations. Pre-clinical studies demonstrated a small number of specific point mutations that can enhance ER function. However, the studies on clinical samples performed in the 1990's were limited by small sample size, lack of detailed clinical correlation and lacked the sensitivity of next-generation sequencing (NGS). Therefore, in this study we sought to comprehensively investigate the frequency and functional significance of ER mutations throughout the progression of breast cancer from primary disease to advanced metastatic disease using targeted NGS. Methods: In this retrospective study, a total of 249 tumor specimens were analyzed. The specimens include 134 ER positive and, as controls, 115 estrogen receptor negative tumors. The estrogen receptor positive samples consist of 58 primary breast cancers and 76 metastatic sample. All tumors were sequenced with high coverage using NGS targeting the coding sequence of ER and an additional 181 cancer-related genes. Results: Recurring somatic mutations at codons 537 and 538 within the ligand-binding domain of the estrogen receptor were detected in ER positive metastatic tumors. Overall, the frequency of these mutations was 12% (95% CI 6%-21%) in metastatic patients compared with none in the primary cases. In total there were 9 recurring somatic mutations; Y537C (11%), Y537N (33%), Y537S (22%) and D538G (33%). In addition in a small number of paired primary and metastatic samples from the same patient, these mutations were found only in the metastatic specimens. In a subset of heavily pre-treated patients the frequency was 20% (5/25, 95% CI 7%-41%). ER activating mutations were not detected in any stage of ER negative disease. ER alterations were not mutually exclusive with any of the other commonly altered genes and of the most frequently altered genes, all but ER alterations displayed similar frequencies across primary and metastatic specimens. Functional studies in cell line models demonstrated that these ER mutations render ER constitutively active and confer resistance to hormone deprivation, tamoxifen and fulvestrant. Conclusions: Herein, we reveal functional ER mutations as potential drivers of endocrine resistance during the progression of ER positive breast cancer. The absence of detectable mutations in the primary tumors suggests clonal evolution as the mechanism of resistance. Thus, these mutations have the potential to be an important genetic biomarker of endocrine resistance in ER positive metastatic breast cancer and could assist in clinical decision making as disease progresses. Our findings also underscore the value of repeated biopsies of metastatic lesions. Lastly, since the frequencies of these mutations are substantial when sensitive testing methods are used in the correct clinical context, pre-clinical and clinical studies to identify novel therapeutics that can overcome this resistance are warranted. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr S3-06.

A mass spectrometry-based proteomic screen in ER+ breast cancer cells revealed that levels of P-REX1 are decreased upon PTEN loss, and increased upon PI3K inhibition. P-REX1 integrates signaling inputs from receptor tyrosine kinases (RTKs)/PI3K (via the PI3K phospholipid product PIP3) and G protein-coupled receptors (via Gbg subunits) to drive guanine exchange factor (GEF) activity on Rac1. RNAi knockdown of PREX1 and overexpression of exogenous PREX1 in ER+ breast cancer cells respectively decreased and increased activation of insulin-like growth factor receptor-1 (IGF-1R)/insulin receptor (InsR), PI3K/AKT/SGK3, and MEK/Erk under steady-state and growth factor (IGF-1, Heregulin)-stimulated conditions. While the P-REX1 homologue P-REX2a was previously shown to inhibit PTEN phosphatase activity to activate the PI3K/AKT pathway, we did not detect an effect of P-REX1 on PTEN activity. PREX1 knockdown suppressed PI3K/AKT signaling in PTEN-null breast cancer cells; therefore, P-REX1 and P-REX2a may not be functionally redundant. Knockdown and overexpression of PREX1 respectively increased and decreased doxorubicin-induced apoptosis in ER+ breast cancer cells, linking PREX1 with a pro-survival phenotype. Inhibition of signaling nodes downstream of PI3K (AKT, mTOR) derepresses feedback to activate RTKs and PI3K; knockdown of PREX1 abrogated the PI3K activation induced by inhibition of mTORC1/mTORC2. Structural analysis of P-REX1 revealed that the DH domain (which binds Gβγ and is required for GEF activity) is dispensable for P-REX1 effects on PI3K signaling, while the PH domain [which binds PIP3 and PI(3,4)P2] is required. These data place P-REX1 in a positive feedback loop, whereby PI3K generates PIP3 and PI(3,4)P2, and P-REX1 binds these phospholipids at the plasma membrane. In turn, P-REX1 promotes RTK activation, and RTKs activate PI3K/AKT and MEK/Erk signaling. Gene expression profiling of diverse types of solid tumors (n = 2,009) and cancer cell lines (n = 807) revealed that PREX1 mRNA is most abundant in ER+ breast tumors compared to other subtypes. Reverse-phase protein array (RPPA) analysis of lysates from 712 breast tumors showed that P-REX1 levels are inversely correlated with markers of PI3K/AKT/mTOR pathway activation. Further, P-REX1 levels are higher in ER+ tumors than ER- tumors. In another series of 1,293 carcinomas, PREX1 was genetically amplified or mutated in 6.2% of cases, and in 5% of breast cancers. We then tested whether PREX1 lesions co-exist with other PI3K pathway-activating lesions. Among genes encoding proteins implicated in RTK/PI3K signaling and phosphatidylinositol metabolism, we found a significant enrichment for PREX1 mutation/amplification in 54/79 (68%) genes across 1,523 carcinomas. We tested the effects of 7 PREX1 mutants found in breast tumors on PI3K signaling in vitro. A G344R mutation in the PREX1 PH domain conferred increased affinity for PIP3 and PI(3,4)P2, and increased the levels of phospho-AKT. These findings suggest that P-REX1 is an ER+ breast tumor-specific oncogene, and PREX1 mutations increase its oncogenic effects in breast cancer. We propose that neutralizing P-REX1 function is a novel therapeutic approach to selectively abrogate oncogenic signaling in ER+ breast cancers while sparing normal tissues. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr PD1-9.

Background: RO4929097 is a potent and oral selective inhibitor of γ-secretase (GSI), producing inhibitory activity of Notch 1–4 signaling in tumor cells. Preclinical studies have demonstrated synergism of GSIs in combination with endocrine therapy, in ERα-positive xenografts. This suggests that optimal treatment of ER+ BC would benefit from an antiestrogen and GSI combination. Materials and Methods: To explore the safety, tolerability and preliminary antitumor activity of Exemestane (25 mg/d) with R04929097 combination, we conducted a Phase Ib trial in patients with ER+/ HER2− MBC, whose disease relapsed during treatment with or within 6 months of discontinuation of an adjuvant nonsteroidal aromatase inhibitor or Tamoxifen, or whose MBC progressed during treatment with 1st or 2nd line endocrine therapy. Goserelin (3.6 mg q28 days Sub-Q) was given to all premenopausal patients. R04929097 was administered orally daily for 3 consecutive days, followed by 4 days off on a 21 day cycle. Doses of R04929097 were escalated on a standard 3+3 design over 5 doses (20, 30, 45, 90, and 140 mg). Patients were fully evaluated for dose limiting toxicity (DLT) for 3 weeks. Patients were treated until disease progression or occurrence of unacceptable toxicity. Results: A total of 15 patients were enrolled. Median age was 56; average number of chemotherapies received in the adjuvant or metastatic setting was 3.3. Eleven patients had received adjuvant endocrine therapy. One patient received this as 1st line endocrine therapy, eleven patients received this study as 2nd line endocrine therapy; and three patients received this study as 3rd line endocrine therapy in the metastatic setting. Of the 14 patients evaluable, 1 had a partial response, 6 had stable disease, and 7 had progressive disease. All patients have discontinued the study at this time, 11 due to progression, 2 due to toxicity, and 1 withdrew after beginning protocol therapy. Twenty percent of patients had stable disease for six months or greater. Of the patients with stable disease for six months, one patient progressed at 9 months and 2 patients progressed at 10 months. The most common toxicities were nausea (73.3%), anorexia (60%), hyperglycemia (53.3%), hypophosphatemia (46.7%), fatigue (66.7%), and cough (33.0%). Grade 3 toxicities were uncommon and included hypophosphatemia (13.0%) and rash (6.3%). There were no grade 4 toxicities. Rash was the only DLT and was seen at 140 mg. Conclusion: The combination of R04929097 and Exemestane was overall well tolerated, but MTD was not determined as access to drug ended. Further studies with GSIs in combination with endocrine therapy are warranted on the basis of a favorable safety profile and preliminary evidence of stable disease seen in patients with ER+ MBC refractory to previous endocrine therapies. PK, PD and biomarker analysis is ongoing. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P2-14-04.

Background: Using a mouse model of HER3 ablation in the mammary epithelium, we previously showed that HER3 is required for the genesis of transgenic HER2-overexpressing breast cancers. However, increasing evidence suggests that transgenic models of HER2 overexpression result in luminal B-like mammary tumors, which resemble human breast tumors that express estrogen receptor (ER) but are often not responsive to endocrine therapy. Furthermore, we found that HER3 is required in the untransformed mouse mammary epithelium for differentiation and survival of luminal cells, but not basal cells. Herein we report that HER3 drives growth and survival of luminal breast cancers in a manner that parallels its role in maintaining the untransformed luminal epithelium in the breast. Results: Expression profiling of a large human breast cancer dataset demonstrated that HER3 expression was highest in Luminal breast cancers as compared to different molecular subtypes of breast cancers, including HER2-enriched, claudin-low, and basal-like breast cancers. In luminal breast cancer cell lines MCF7, ZR75-1, and MDA-MB-361, HER3 targeting using siRNA or a monoclonal antibody against HER3 decreased Akt, S6, and Erk1/2 phosphorylation, cell growth and cell survival under conditions of estrogen deprivation. Luminal breast tumor xenografts treated with U3-1287, a humanized monoclonal anti-ErbB3 antibody, or EZN-3920, an anti-sense oligonucleotide against ErbB3, decreased tumor growth and phosphatidyl inositol-3 kinase (PI3K)/mTOR signaling in vivo. Interestingly, luminal breast tumor xenografts responded to the ER inhibitor fulvestrant with increased HER3 expression and phosphorylation, and with increased PI3K/mTOR signaling as shown by enhanced S6 phosphorylation. Fulvestrant-mediated upregulation of HER3 was also seen in matched clinical breast tumor specimens harvested before treatment and after 21 days of fulvestrant treatment. AMG-888 blocked fulvestrant-induced HER3 and phospho-S6 upregulation in vivo, resulting in regression of luminal breast tumor xenografts. Conclusions: These data suggest that HER3 targeting may improve the outcome of luminal breast cancers treated with endocrine therapy, and are consistent with our findings that a HER3-dependent expression signature positively correlates with relapse in tamoxifen-treated patients. Taken together, these results define a novel role for HER3 in luminal breast cancer cells, and support future investigations into therapeutic strategies to target HER3 in luminal breast cancers. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P5-19-01.

Background: Mutations in PIK3CA, the gene encoding the p110alpha subunit of PI3K, have been associated with antiestrogen resistance in ER+ BC. In general, antiestrogen-resistant cancers retain ER and responsiveness to estradiol, suggesting that treatment of ER+/PI3K mutant BC should include PI3K inhibitors plus endocrine therapy. Methods: We are conducting a phase Ib open-label trial of letrozole (2.5 mg/d) with the dual PI3K/ mTOR inhibitor BEZ235 (400 mg BID, sachet formulation) in post-menopausal patients with ER+/HER2 negative MBC. Upon toxicity, BEZ235 was reduced to 400 mg in AM/ 200 mg in PM (dose level 1), and 200 mg BID (dose level 2), in a standard 3+3 de-escalation design. Treatment continues until unacceptable toxicity or disease progression. Tumor assessments are performed every 2 months. All tumor biopsies are being analyzed for PI3K pathway alterations. Results: To date, 9 patients have been accrued; 6 on dose level 1 (400 mg BID) and 3 on dose level 1 (400 mg in AM/ 200 mg in PM). Of note, the latter 3 patients have been on treatment for less than 2 weeks at the time of abstract submission. Median age was 55 years (range 40–65); all patients had disease progression on previous endocrine therapy, and 100% had bone metastases and 30% had visceral metastases. Preliminary toxicities for dose level 1 cohort (400 mg BID) are summarized in the table below. The DLT was grade 3 mucositis, which happened less than 3 weeks of study initiation, in a total of 3 out of 6 patients. All grade 3 mucositis cases improved after BEZ235 interruption for 7 days and subsequent dose reduction. These 3 events triggered activation of the dose level 1 cohort (400 mg AM/ 200 mg PM). The first 3 patients on dose level 1 cohort (2 of which have PIK3CA mutated tumors) had their tumor assessment scans at 8 weeks post treatment initiation; all patients were found to have stable disease. Discussion: The combination of letrozole/BEZ235 is still being evaluated in patients with ER+/HER2MBC. Safety, tolerability and preliminary antitumor activity data will be updated at the meeting. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P6-10-05.

The ErbB receptor tyrosine kinase family regulates breast development and cancer. ErbB2/HER2 expression correlates with HER2-enriched breast cancers, while epidermal growth factor receptor (EGFR) expression often correlates with triple negative breast cancers. Less is known regarding ErbB3, which harbors only weak kinase activity, but strongly activates phosphatidylinositol-3 kinase (PI3K)/Akt signaling upon heterodimerization with EGFR or ErbB2. We report herein that ERBB3 mRNA expression strongly correlates with Luminal A/B breast cancers, which express lower levels of EGFR and ErbB2. Mammary-specific loss of ErbB3 in mice reduced Akt phosphorylation and caused cell death in the luminal epithelium. A decreased luminal population in ErbB3-deficient epithelium correlated with expansion of the mammary stem/progenitor fraction. In normal breast samples, ERBB3 mRNA expression was highest in mature/progenitor luminal populations, and lowest in the stem/basal population. Loss of ErbB3 shifted gene expression in mammary epithelial cells to resemble a mammary stem cell signature. The genes most greatly impacted by ErbB3 loss produced a signature that correlated with decreased overall survival in tamoxifen-treated patients. Knock-down of ErbB3 in human luminal breast cancer cells caused decreased cell growth under conditions of estrogen deprivation, decreased expression of the gene encoding aromatase, and impaired expression of estrogen receptor-induced genes, including those that encode progesterone receptor and TFF1. Treatment of MCF7 luminal breast cancer xenografts with AMG-888, a humanized monoclonal antibody directed against ErbB3, decreased tumor growth. Combination of AMG-888 with the estrogen receptor inhibitor fulvestrant caused MCF7 tumor regression. Taken together, these results suggest that ErbB3 is required to sustain normal and transformed luminal epithelial cells of the breast, and suggest that targeting ErbB3 may improve the clinical outcome of breast cancers treated with endocrine therapy. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr S2-6.

Estrogen receptor-positive (ER+) breast cancers are treated with strategies aimed at inducing estrogen deprivation such as aromatase inhibitors (AIs). However, many ER+ tumors adapt to hormone deprivation and acquire resistance to AIs. The purpose of this study was to identify kinases essential for the growth of endocrine-resistant ER+ breast cancer cells. To model acquired resistance to AIs, we maintained four ER+, estrogen-dependent human breast cancer cell lines (MCF-7, MDA-361, HCC-1428, and ZR75- 1) under hormone-depleted conditions for several months until hormone-independent populations emerged (termed long-term estrogen-deprived, LTED). We then performed an RNAi screen targeting 779 protein kinases to identify molecules required for growth of MCF-7/LTED cells in hormonefree conditions. Individual knockdown of 42 kinases significantly inhibited MCF-7/LTED cell growth ≥33% (p≥0.05) in at least 3/4 independent experiments. Of these 42 hits, knockdown of the insulin receptor (InsR) inhibited MCF-7/LTED growth (n=4 experiments) compared to control siRNA. Knockdown of InsR expression using independent siRNA oligonucleotides significantly inhibited growth of MCF-7/LTED, ZR75-1/LTED, and HCC-1428/LTED cells. The InsR can heterodimerize with the highly homologous insulin-like growth factor-I receptor (IGF-IR) to potently activate the phosphatidylinositol-3 kinase (PI3K)/AKT signaling pathway. Immunoblot analysis revealed that siRNA knockdown of InsR or IGF-IR caused a compensatory upregulation of IGF-IR and InsR expression, respectively. Knockdown of InsR or IGF-IR inhibited growth of 3/4 LTED lines, but knockdown of both InsR and IGF-IR was required for maximal inhibition of PI3K/AKT. Inhibition of these receptors by treatment with the InsR/IGF-IR tyrosine kinase inhibitor (TKI) OSI-906 decreased PI3K signaling and disrupted the basal association between p85 and the IGF-IR/InsR effectors IRS-1 and IRS-2 in MCF-7/LTED cells. Further, OSI-906 inhibited the hormone-independent growth of all four LTED lines and prevented the emergence of hormone-independent ER+ parental cells following estrogen deprivation. Treatment with the TKI or the neutralizing IGF-IR monoclonal antibody MK-0646 induced feedback upregulation and phosphorylation of the ErbB3/HER3 receptor tyrosine kinase (RTK). Induction of P-HER3 was blocked by the EGFR/HER2 TKI lapatinib. Combined treatment of lapatinib and OSI-906 prevented the emergence of hormone-independent (HER2+/ER+) MDA-361 cells more effectively than either drug alone. Finally, treatment of ovariectomized nude female mice with OSI-906 suppressed the growth of established MCF-7 xenografts. In addition, treatment with OSI-906 decreased PI3K signaling as measured by immunoblot analysis of tumor extracts using P-AKT antibodies. These results suggest that 1) InsR/IGF-IR signaling is required for PI3K/AKT signaling and hormone-independent growth of ER+ breast cancer cells; 2) these cells compensate for the inhibition of InsR/IGF-IR by upregulating alternative RTK pathways such as HER3; 3) combined targeting of InsR/IGF-IR and ER signaling may be effective for the treatment of endocrine-resistant breast cancer. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr S3-8.

Recombinant methodologies have allowed the production of monoclonal antibodies (MAbs) with higher percentages of human content to decrease antibody immunogenicity and reduce or eliminate the anti-antibody response. Trastuzumab (Tz) is a humanized MAb approved for treatment of HER2-overexpressing breast cancer. We detected anti-Tz antibodies in the serum of patients treated with Tz. We initially studied the effects of these antibodies on the response to Tz in a transgenic mouse model of breast cancer. Twenty-two wild-type FVB mice were transplanted with mammary tumors arising in FVBMMTV/HER2 transgenic mice. Mice with established tumors were treated with Tz (n=12) or PBS (n=10) for 4 weeks. Five of 12 mice responded to treatment while 7/12 mice exhibited progressive disease (PD). All 5 responders had elevated anti-Tz antibodies, whereas in mice that had tumor progression anti-Tz levels were low or undetectable (p=0.002). These findings were confirmed in a second mouse cohort (n=16), in which anti-Tz antibodies were detected in responder mice as early as one week after starting treatment. MAbs produced by hybridomas stemming from spleen cells derived from responder mice bound Tz. These findings argue against the current understanding that an anti-xenoglobulin response to Tz would produce blocking antibodies. Therefore, we addressed if detectable anti-Tz antibodies would predict benefit from antibody therapy in patients with HER2+ breast cancer. Samples from 21 patients with stage IV HER2 positive breast cancer experiencing clinical response or disease stabilization upon treatment with Tz that had been enrolled in an anti-HER2 vaccine phase I trial (J Clin Oncol 2009;4685) were retrospectively analyzed in a blinded fashion. With a median follow-up of 36 months, 14 patients had had disease progression while on Tz plus the vaccine. The presence of anti-Tz F(ab')2 antibodies was quantified among the enrolled patients. A univariate Cox regression analysis of the progression free survival (PFS) was performed. Analysis revealed that higher anti-Tz F(ab')2 antibodies were significantly associated with lower risk of disease progression (p=0.023). Fine specificity studies confirm the presence of a human polyclonal response to Tz, directed to epitopes present in both its Fc and Fab regions. In conclusion, in mice, response to Tz was significantly associated with the presence of anti-Tz antibodies in serum, while in patients with stage IV breast cancer treated with Tz, the presence of anti-Tz F(ab')2 antibodies was associated with longer PFS. The fine specificity of anti-Tz antibodies reveals epitopes that could be potentially relevant for vaccination strategies. Information about cognate epitopes on Tz may be also useful in the design of new therapeutic antibodies. These data also support prospective evaluation of patients undergoing therapy with other chimeric or humanized antibodies to address if detectable anti-therapeutic antibodies would predict benefit from long antibody therapy regimens, especially in the adjuvant setting, in which clinical benefit can only be measured by survival rates. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P2-09-04.

Background: CDK4/6 inhibitors have been approved in combination with endocrine therapy for treatment of ER+ metastatic breast cancer. The goal of this study was to discover mechanisms of resistance to ER antagonists alone and in combination with CDK4/6 inhibitors. Results: To achieve this goal, we used lentiviral vectors to individually express 559 human kinase open reading frames (ORFs) in ER+ MCF7 human breast cancer cells treated with fulvestrant ± the CDK4/6 inhibitor ribociclib. In MCF7 cells treated with fulvestrant alone or with ribociclib, we identified 15 and 17 kinases associated with resistance, respectively. Ten of these kinases overlapped in both groups. In a secondary screen, MCF7 cells were stably transduced with V5-tagged lentiviruses expressing the positive 'hits' for treatment with fulvestrant/ribociclib. Five of 17 kinases (FGFR1, FRK, HCK, FGR, CRKL) were confirmed to induce resistance to fulvestrant/palbociclib and fulvestrant/ribociblib. Survey of TCGA for copy number alterations and/or expression of these 5 genes showed only FGFR1 to be amplified/overexpressed in ˜15% of ER+ breast cancers. Experiments in vitro showed that ER+/FGFR1-amplified (amp) MDA-134, CAMA-1 and HCC1500 human breast cancer cells and MCF7 cells stably transduced with FGFR1 were relatively resistant to estrogen deprivation, fulvestrant and fulvestrant/palbociclib compared to non-FGFR1 amp MCF7 cells. This resistance was abrogated by treatment with the FGFR tyrosine kinase inhibitor (TKI) lucitanib. Treatment with fulvestrant or palbociclib alone modestly delayed growth of ER+/FGFR1-amp breast cancer patient-derived xenografts (PDX) established in nude mice. However, addition of the FGFR TKI erdafitinib to fulvestrant/palbociclib resulted in marked PDX regression in all mice without associated toxicity and a complete cell cycle arrest measured by Ki67. Treatment of FGFR-amp cells with FGF-2 strongly induced CCND1 (cyclin D1) expression. Downregulation of CCND1 with CCND1 RNAi oligonucleotides restored sensitivity of FGFR1-amp cells to fulvestrant/palbociclib, thus phenocopying the effect of FGFR TKIs. Conversely, overexpression of CCND1 in MCF7 cells induced resistance to estrogen deprivation and to fulvestrant ± palbociclib. Finally, we examined next gen sequencing of cell free tumor DNA by Guardant360 in 34 patients before and after progression on CDK4/6 inhibitor. In 10/34 (29%) post-progression specimens, we detected alterations in the FGFR pathway: FGFR1 amplification (n=7), FGFR1 N546K (n=1), FGFR2 N549K (n=1), and FGFR2 V395D (n=1) activating mutations. Conclusions: These data suggest aberrant FGFR signaling is a mechanism of resistance to anti-ER therapies ± CDK4/6 inhibitors. We posit overexpression of cyclin D1 induced by both FGFR signaling and ER transcription plays a role in drug resistance. Based on these findings we propose ER+/FGFR1 amplified breast cancers are endocrine resistant and should be candidates for treatment with combinations of ER and FGFR antagonists. Accordingly, we have initiated a phase Ib trial of fulvestrant, palbociclib and erdafitinib in patients with antiestrogen resistant ER+/HER2-negative breast cancer with FGFR1-4 amplification. Citation Format: Formisano L, Lu Y, Jansen VM, Bauer JA, Hanker A, Gonzalez Ericsson P, Lee K-M, Nixon MJ, Guerrero-Zotano AL, Schwarz LJ, Sanders M, Sudhan D, Dugger TC, Cruz MR, Behdad A, Cristofanilli M, Bardia A, O'Shaughnessy J, Mayer IA, Arteaga CL. Gain-of-function kinase library screen identifies FGFR1 amplification as a mechanism of resistance to antiestrogens and CDK4/6 inhibitors in ER+ breast cancer [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr GS6-05.

This abstract was not presented at the symposium.

Background: CDK4/6 inhibitors in combination with antiestrogens are approved for the treatment of ER+ advanced breast cancer. However, not all patients benefit from CDK4/6 inhibitors, underscoring the need to develop therapeutic strategies to circumvent de novo and acquired drug resistance. Methods: ER+ breast cancer cells (MCF-7, T47D, HCC1428, and HCC1500) were made resistant to increasing doses to the CDK4/6 inhibitor ribociclib (LEE011; Novartis). LEE011-resistant cells were characterized by 2D/3D growth, cell cycle, and immunoblot analyses. GSK2334470 (PDK1 inhibitor) and dinaciclib (CDK2 inhibitor) were used to modify resistance to ribociclib. PDK1 and pS6 immunohistochemistry (IHC) were performed on primary human tumor explants treated ex vivo with palbociclib. Results: Resistant cell lines (MCF-7/LR, T47D/LR, HCC1428/LR, and HCC1500/LR) exhibited an IC50 at least 20-fold higher than that of their parental cells. They displayed cross-resistance to the CDK4/6 inhibitors palbociclib and abemaciclib. Immunoblot analysis of ribociclib-resistant cells showed increased levels of 3-phosphoinositide dependent protein kinase 1 (PDK1), S227 pRSK2 (target of PDK1), T308 pAKT (target of PDK1), and pS6 (downstream effector of the PDK1 target p70S6K), compared to parental drug sensitive cells. PDK1 is a master kinase that functions downstream of phosphoinositide 3-kinase (PI3K) and is crucial for the activation of AKT and many other AGC kinases including PKC, S6K, SGK, and RSK. Primary tumor explants treated ex vivo with palbociclib for 96 h also exhibited upregulation of PDK1 and pS6 by IHC. Cell cycle analysis revealed that CDK4/6 inhibition failed to induce G1 arrest, a reduction in S phase, and senescence in MCF-7/LR and T47D/LR compared to parental cells. Progression into S phase in the presence of ribociclib suggested upregulation of S-phase cyclins/CDKs. Indeed, the resistant cells exhibited significantly higher levels of pCDK2, cyclin A, cyclin E and S477/T479 pAKT, a CDK2-dependent phosphorylation of AKT required for full kinase activity and limited to the S-phase of the cell cycle. Pharmacological inhibition of PDK1 (with GSK2334470) or CDK2 (with dinaciclib) re-sensitized the ribociclib-resistant cells to CDK4/6 inhibitors. However, ribociclib/GSK2334470 inhibited MCF-7/LR and T47D/LR cell proliferation better than ribociclib/dinaciclib. Further, ribociclib/GSK2334470 but not ribociclib/dinaciclib completely abrogated pRb, pS6, pRSK2, pCDK2, cyclin A, and cyclin E, suggesting the PI3K/PDK1 pathway mediates acquired resistance to CDK4/6 inhibitors through dysregulation of the cell cycle. Consistent with these data, ribociclib/GSK2334470 inhibited growth of established MCF-7 xenografts in nude mice, significantly more potently than each drug alone. Conclusions: These data support a critical role for PI3K/PDK1 in acquired resistance to CDK4/6 inhibitors in ER+ breast cancer cells. Co-targeting of PI3K/PDK1 and CDK4/6 may overcome resistance to CDK4/6 inhibitors and is worthy of further translational and clinical investigation in patients with ER+ breast cancer. Citation Format: Jansen VM, Formisano L, Witkiewicz A, Estrada MV, Sanchez V, Dugger TC, Knudsen ES, Arteaga CL. PI3K/PDK1 mediates resistance to CDK4/6 inhibitors through dysregulation of S-phase cyclins/cyclin dependent kinases (CDKs) [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-03-05.

ERBB2, the gene encoding HER2, is mutated in 2-4% of breast cancers. The HER2 irreversible tyrosine kinase inhibitor (TKI) neratinib has shown clinical activity against breast cancer cells harboring HER2 activating mutations. Here, we report for the first time an acquired gatekeeper HER2T798I mutation in a patient with HER2-mutant breast cancer after an initial exceptional response to neratinib. A patient with ER+/PR+/HER2-negative invasive lobular breast cancer progressing on standard therapy was found to harbor a L869R kinase domain mutation in HER2. HER2L869R is homologous to the known activating mutation EGFRL861R/Q. MCF10A breast epithelial cells expressing HER2L869R displayed enhanced HER2-mediated signaling and were resistant to lapatinib and trastuzumab but sensitive to neratinib. The patient was enrolled in the phase II SUMMIT trial (NCT01953926) and treated with neratinib, achieving a partial response lasting 16 months before developing progression. Next gen sequencing of DNA from both a new skin metastasis and plasma cell-free DNA (cfDNA) identified HER2L869R (8.7% cfDNA), whereas a novel HER2T798I mutation was detected only in plasma at 1.3%. Deep sequencing of pre-therapy tumor tissue and plasma did not detect HER2T798I, suggesting that this mutation arose upon resistance. HER2T798I has not been reported in TCGA, COSMIC, or among plasma samples from 17,345 cancer patients subjected to digital DNA sequencing using the Guardant360 assay. HER2T798I is homologous to the EGFRT790M, KITT670I and BCR-ABLT315I gatekeeper mutations known to mediate resistance to erlotinib/gefitinib and imatinib. To examine if HER2T798I mediates resistance to neratinib, we employed biochemical and biological assays and molecular modeling of wild-type (WT) HER2 and HER2T798I. Structural modeling showed the increased bulk of the isoleucine at position 798 would result in a steric clash with neratinib, thus reducing drug binding. We stably expressed HER2WT, HER2T798I, HER2L869R and HER2L869R/T798I in MCF10A cells and NR6 mouse fibroblasts. Neratinib (10-100 nM) blocked HER2-mediated signaling in cells expressing HER2WT or HER2L869R but did not in cells expressing HER2T798I. The EGFR irreversible TKI osimertinib (100 nM), which isselective for mutant EGFR (including EGFRT790M) and approved for treatment of NSCLC expressing EGFRT790M, failed to inhibit HER2WT, HER2L869R or HER2T798I. In contrast, either the EGFR/HER2 irreversible TKI afatinib or AZ5104, a metabolite of osimertinib, strongly blocked signaling induced by HER2WT, HER2L869R or HER2T798I. Cells expressing HER2T798M displayed a significantly higher IC50 to neratinib than cells expressing HER2WT, whereas afatinib or AZ5014 were very active against all cells (IC50 Conclusions: The acquisition of a T798I gatekeeper mutation in HER2 upon development of clinical resistance to neratinib in a breast cancer with an initial activating mutation in HER2 strongly suggests that HER2L869R is a driver mutation. We speculate that HER2T798I may arise as a secondary mutation following response to effective HER2 TKIs in other cancers with HER2 activating mutations. Certain irreversible EGFR inhibitors may be effective in patients with HER2-driven breast cancer resistant to neratinib. Citation Format: Hanker AB, Red Brewer M, Sheehan JH, Koch JP, Lanman R, Hyman DM, Cutler, Jr. RE, Lalani AS, Cross D, Lovly CM, Meiler J, Arteaga CL. An acquired HER2 T798I gatekeeper mutation induces resistance to neratinib in a patient with HER2 mutant-driven breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-03-03.

Background: Estrogen receptor (ER)-positive breast cancers (BC) initially respond to antiestrogens but eventually become hormone-independent and recur. FGFR1 is amplified in ∼10% of ER+ BC and is associated with early recurrence on antiestrogen therapy. Notably, one third of FGFR1-amplified tumors have simultaneous amplification of CCND1, FGF3, FGF4 and FGF19 on chromosome 11q12-14. Herein, we investigated the mechanisms by which FGFR1 amplification confers resistance to antiestrogen therapy in ER+ BC cells. Results: We performed whole exome sequencing in tumor biopsies from 130 patients with an operable ER+/HER2- BC who had received letrozole for 10-21 days prior to surgery. Tumors were categorized by the natural log (ln) of post-letrozole Ki67 as sensitive (ln ≤1 or ≤2.7% Ki67+ cells; n=68) or resistant (ln ≥2 or ≥7.4%; n=18). We found amplifications in FGFR1 and/or 11q12-14 in 6/11 (55%) resistant tumors compared with 5/34 (15%) in sensitive tumors (p=0.006); all cases were confirmed by FGFR1-fluorescence in situ hydridization (FISH). Resistant tumors with FGFR1 and/or 11q12-14-amplification showed a marked increase in nuclear FGFR1 with letrozole. ER+/FGFR1-amplified CAMA1 and MDA134 cell lines also exhibited co-localization of ER and FGFR1 in the nucleus. Cell proliferation was partially reduced by estrogen deprivation, and FGFR1 siRNA further reduced cell growth in hormone-depleted medium. We generated CAMA1 and MDA134 cells resistant to long-term estrogen deprivation (LTED). These cells exhibited overexpression of FGF3/4/19 and ERα with a concomitant increase in ligand-independent ER transcriptional activity and growth. An ER-FGFR1 interaction was observed in the nucleus and cytosol of CAMA1 parental cells with enhanced interaction in CAMA1 LTED cells. Genetic (with siRNA) and pharmacologic (with lucitinib) inhibition of FGFR1 reduced a) nuclear localization of FGFR1; b) ER transcriptional activity; and c) cell proliferation. Nuclear localization and ER-FGFR1 interaction were disrupted by a kinase-deficient FGFR1. Conversely, addition of FGF3 ligand stimulated ER-FGFR1 interaction and ER transcriptional activity, suggesting FGFR activation can regulate ER function. Inhibition of FGF receptor-specific substrate (FRS2), a principal mediator of FGFR1 signal transduction to the MAPK and PI3K pathways, with siRNA or pharmacologic inhibition of PI3K with buparlisib or MEK with GSK1120212 did not reduce ER transcriptional activity suggesting that, in ER+/FGFR1-amplified cancer cells, ER function is not modulated by FGFR signal transducers. Finally, using chromatin immunoprecipitation (ChIP) we showed that FGFR1 binds directly to estrogen response elements (ERE). This association was reduced with lucitanib. We are currently investigating genes modulated by ER/FGFR1 in ER+ BC and the in vivo anti-tumor efficacy of dual inhibition of FGFR1 and ER in ER+/FGFR1-amplified patient-derived breast cancer xenografts. Conclusions: These data support a critical role of ER and FGFR1 interaction in endocrine resistance in ER+/FGFR1-amplified breast cancer. Targeting of FGFR1 in combination with antiestrogens may abrogate resistance to endocrine therapy in these tumors and is worthy of clinical investigation. Citation Format: Formisano L, Young CD, Bhola NE, Bulen B, Estrada VM, Wagle N, Van Allen E, Red Brewer ML, Jansen VM, Guerrero AL, Giltnane JM, Strcker T, Arteaga CL. Nuclear FGFR1 interaction with estrogen receptor (ER) α is associated with resistance to endocrine therapy in ER+/FGFR1-amplified breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr S3-03.

Background: PI3K pathway activation is a hallmark of hormone receptor-positive (HR+) BC cells resistant to endocrine therapy (ET). Preclinical and early clinical data suggest that combining the pan-PI3K inhibitor BUP (BKM120) with ET may provide clinical benefits in this setting. BELLE-2 (NCT01610284) is the first randomized Phase III trial to assess the efficacy and safety of a PI3K inhibitor combined with FULV in HR+ advanced BC, including a prospective analysis of whether PI3K pathway activation status measured in archival tumor tissue and ctDNA is predictive of clinical benefit. Methods: Postmenopausal women with HR+/HER2– locally advanced or metastatic BC refractory to aromatase inhibitor therapy were enrolled. After a 14-day run-in with FULV (500 mg), patients (pts) were randomized (1:1) to receive oral BUP (100 mg/day) or placebo (PBO) with FULV (500 mg per standard of care). Randomization of all pts was stratified by PI3K pathway status measured in archival tumor tissue (PIK3CA mutation or PTEN loss; activated, non-activated, unknown) and visceral disease status (present, absent). Baseline PIK3CA mutation status in ctDNA was assessed in a subset of 587 pts at trial entry. The primary endpoint was locally-assessed progression-free survival (PFS; RECIST v1.1) in the full population and PI3K pathway-activated group. Secondary endpoints included overall survival, safety, overall response rate (ORR), and clinical benefit rate (CBR). Results: 1147 pts received FULV with BUP or PBO, with 187 (16%) ongoing at data cut-off. Baseline characteristics were well balanced between the two arms: median age was 62 years, 98% had ECOG performance status of 0/1, 61% had visceral disease, 69% were sensitive to prior ET, 28% had received prior chemotherapy for metastatic BC. BELLE-2 met its primary endpoint in the full population (Table). In pts with PI3K pathway-activated tumor tissue, PFS increase for BUP+FULV vs PBO+FULV did not meet the planned endpoint. Among pts with PIK3CA status measured in ctDNA, median PFS, ORR, and CBR were significantly improved for BUP+FULV vs PBO+FULV in pts with PIK3CA-mutant tumors but not pts without. Median PFS, monthsHR95% CI; P valueORRCBR BUP+FULVPBO+FULV Full population N=11476.95.00.780.67–0.89;P The most common Grade 3/4 adverse events (AEs; ≥5% of pts; BUP+FULV vs PBO+FULV) in all pts were increased alanine aminotransferase (26 vs 1%), increased aspartate aminotransferase (18 vs 3%), hyperglycemia (15 vs 0.2%), and rash (8 vs 0%). The most common reasons for treatment discontinuation were disease progression (54 vs 73%) and AEs (13 vs 2%). Conclusions: BELLE-2 met its primary endpoint in the full population. Prespecified analyses showed that characterizing PIK3CA mutation in ctDNA at trial entry identifies pts with endocrine- resistant HR+/HER2–advanced BC, for whom BUP+FULV results in meaningful clinical benefits. Citation Format: Baselga J, Im S-A, Iwata H, Clemons M, Ito Y, Awada A, Chia S, Jagiello-Gruszfeld A, Pistilli B, Tseng L-M, Hurvitz S, Masuda N, Cortés J, De Laurentiis M, Arteaga CL, Jiang Z, Jonat W, Hachemi S, Le Mouhaër S, Di Tomaso E, Urban P, Massacesi C, Campone M. PIK3CA status in circulating tumor DNA (ctDNA) predicts efficacy of buparlisib (BUP) plus fulvestrant (FULV) in postmenopausal women with endocrine-resistant HR+/HER2– advanced breast cancer (BC): First results from the randomized, phase III BELLE-2 trial. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr S6-01.

Background: We have previously identified the MAPK phosphatase dual specificity phosphatase 4 (Dusp4) as frequently down-regulated in triple-negative and basal-like breast cancer (TNBC and BLBC, respectively). DUSP4 is deregulated in BLBCs by heterozygous and homozygous loss, as well as by promoter CpG methylation. In TNBC and BLBC, DUSP4 deregulation frequently co-occurs with MYC amplification and mutations in TP53. In addition, loss of DUSP4 promotes activation of ETS-1 and cJUN, transcription factors downstream of Ras/MAPK and JNK, while imparting resistance to chemotherapy and promoting tumor-initiating cells, suggesting DUSP4 is a tumor suppressor. Thus, we hypothesized that loss of Dusp4 in the mammary epithelium, together with MYC amplification and mutations in Trp53, would promote tumorigenesis. Methods: We generated transgenic mice with a conditional floxed Dusp4 allele (Dusp4FLOX) and isolated mammary epithelial progenitor cells. The resulting cells were transduced with adenovirus to deliver either Cre recombinase to excise the Dusp4 locus (Dusp4NULL) or GFP control. CRISPR-mediated mutagenesis (or non-targeting CRISPR control) was used to disrupt exon 7 or 8 of Trp53 causing relevant p53 mutations and MYC was overexpressed (or LACZ control). Isogenic cell lines were assessed for tumorigenicity, ploidy, sensitivity to chemotherapy, proliferation, and cell signaling. Results: The isolated parental (Dusp4FLOX) mammary cell line maintained proliferation through >50 passages without senescing and was highly cytokeratin (Ck) 5/14+ and moderately Ck 8/18+, suggestive of a bipotent progenitor-like lineage. As expected, p53 disruption, regardless of Dusp4 status, resulted in tetraploidy and loss of p21 expression. With concurrent Dusp4, p53 and MYC alterations, cells became increasingly resistant to chemotherapy and targeted agents, as measured with proliferation assays. In order to determine whether cells formed tumors in vivo, we implanted isogenic cell lines subcutaneously into nude mice. We found that only Dusp4NULL cells, together with p53 disruption and MYC overexpression (DPM) generated aggressive tumors (forming and ulcerating within 1-2 weeks), while the Dusp4FLOX counterparts (p53 disruption and MYC overexpression; PM) did not. Syngeneic transplantation revealed that DPM cells could also form tumors in an immune competent setting. Cells with Dusp4 loss and p53 mutation were also able to develop tumors without MYC amplification, but at a much slower rate. Tumors derived from the DPM cells were primarily metaplastic or basal-like in nature, including intra-tumor heterogeneity (pockets of Ck5 and pockets of Ck8, with regions of Ck8+ keratin pearls). Conclusions: We have developed a genetically relevant mouse model to study BLBC in immune competent mice for unique opportunities to understand the pathogenesis and therapy response in BLBC, including immunotherapy. These results support the hypothesis that DUSP4 loss promotes tumorigenesis in BLBC. An understanding of the mechanism whereby Dusp4 loss contributes to this process in vivo should provide a rational for novel therapeutic strategies to treat BLBCs with DUSP4 loss. Citation Format: Hicks M, Owens P, Moore P, Giltnane J, Estrada M, Sanders M, Cook R, Arteaga C, Balko J. Cooperation of clinically-relevant genomic alterations transforms mammary progenitor cells to metaplastic breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr PD6-01.

Background: The PI3K/AKT/mTOR pathway has been reported altered in up to 40% of hormone receptor-positive (HR+) breast cancers (BC). Furthermore, pathway activation has been implicated in resistance to endocrine therapy, including aromatase inhibitors (AI). BKM120 is an oral inhibitor of all 4 isoforms of class I PI3K (α, β, γ, δ). In combination with fulvestrant, BKM120 has demonstrated pro-apoptotic effects in vitro, and achieved near-complete tumor regression in HR+ PIK3CA-mutated estrogen-independent xenografts. Preliminary clinical activity was observed with BKM120 in patients with BC, both as a single agent and in combination with letrozole (Mayer, et al. ASCO 2012: #510) or trastuzumab (Saura, et al. SABCS 2011: #PD09-03). Study design: This is a Phase III randomized, double-blind, placebo-controlled trial of BKM120 or placebo in combination with fulvestrant in postmenopausal women with HR+/HER2– locally advanced or metastatic BC that progressed on or after AI treatment (BELLE-2; CBKM120F2302; NCT01610284). After a 14-day fulvestrant run-in phase to allow determination of tumor PI3K pathway activation status, patients are randomized (1:1) to receive continuous BKM120 (100 mg/d) or placebo, and fulvestrant 500 mg at C1D15, and D1 of all cycles thereafter; 1 cycle=28 days. Randomization is stratified according to PI3K pathway activation (PIK3CA mut and/or PTEN mut and/or loss of PTEN protein expression by IHC), and visceral disease status. Study treatment is given until disease progression or until study discontinuation for any reason. Eligibility criteria include: postmenopausal women with HR+/HER2– locally advanced or metastatic BC refractory to AI (progression while on AI, or within 4 wks [metastatic] or 12 mos [adjuvant] of last AI dose); no more than 1 previous line of chemotherapy for advanced disease; and available archival tumor tissue for analysis of PI3K-related biomarkers. Co-primary endpoints: PFS in the full and PI3K pathway-activated populations based on local investigator assessment (RECIST 1.1). Key secondary endpoints: overall survival (OS) in both full and PI3K pathway-activated populations. Other secondary endpoints: PFS and OS (PI3K non-activated/unknown population), overall response rate and clinical benefit rate (both in full, PI3K-activated, PI3K non-activated/unknown populations), safety (CTCAE 4.03), pharmacokinetics of BKM120 and fulvestrant, and patient-reported quality of life (EORTC QLQ-C30 and QLQ-BR23). Statistical methods: PFS and OS will be estimated using the Kaplan–Meier method for both full and PI3K pathway activated populations. Both primary and key secondary endpoints will be tested using a stratified log-rank test at levels α governed by a graphical gate-keeping procedure. A stratified Cox regression will be used to estimate the hazard ratio, along with two-sided 95% confidence interval. Target accrual: Estimated enrollment for BELLE-2 is 842 patients. Recruitment in this Global study is ongoing with planned enrollment of patients in North America, South America, Europe, Asia and Africa. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr OT2-3-09.

ER+ breast cancers typically respond to aromatase inhibitors (AIs), but a significant fraction exhibit de novo or acquired resistance. To discover mechanisms of resistance to AIs, we maintained four ER+, hormone-dependent human breast cancer cell lines in hormone-depleted conditions for several months until estrogen-independent populations emerged (termed long-term estrogen-deprived, LTED). A siRNA kinome screen identified 42 kinases that suppressed hormone-independent MCF-7/LTED cell growth ≥33% (p Activating mutations in PIK3CA, the gene encoding the p110α catalytic subunit of PI3K, are present in about 35% of ER+ breast cancers. We investigated whether PIK3CA mutations modify the tumor response to estrogen deprivation with letrozole in a pre-surgical clinical trial (NCT00651976). Twenty-one patients with stage I-II operable ER+/HER2− breast cancer were treated with letrozole for 10–21 days up to 24 h prior to surgery. Core tumor biopsies were obtained before treatment and at the time of surgical resection. Ki67 immunohistochemistry (IHC) was scored in pre- and post-letrozole specimens. The change in natural log of the Ki67 score between pre-and post-treatment samples was stratified by PIK3CA mutant vs. wildtype (WT) status. PIK3CA-mutant tumors exhibited a statistically lower reduction in Ki67 score compared to tumors with WT PIK3CA (p=0.03). These findings suggest that mutational activation of PI3K attenuates the response to estrogen deprivation. In some LTED cell lines, downregulation of ER with fulvestrant or siRNA inhibited estrogen-independent growth, suggesting that the unliganded ER can maintain proliferation of cells adapted to estrogen deprivation. Prior work suggested that PI3K-ER crosstalk promotes resistance to estrogen deprivation. However, inhibition of PI3K/AKT/TORC1 signaling did not alter ERS167 phosphorylation in LTED cells, and ER downregulation with fulvestrant did not inhibit PI3K/AKT, suggesting that PI3K and ER modulate non-overlapping pathways in cells adapted to estrogen deprivation. Treatment with BKM120 synergized with fulvestrant to inhibit the growth of 6/8 ER+ breast cancer cell lines in estrogen-depleted conditions. Treatment with single-agent fulvestrant or BKM120 in athymic mice devoid of estrogen supplementation partially inhibited the growth of ER+/PI3K-mutant MCF-7 xenografts, while the combination induced near-complete tumor regression. IHC analysis showed increased cleaved caspase-3+ cells in BKM120-treated tumors and decreased Ki67+ cells in fulvestrant-treated tumors, implying that the combination induced tumor regression by both inhibition of tumor cell proliferation and induction of apoptosis. These data suggest that upon progression on an aromatase inhibitor, patients with ER+/PI3K-mutant breast cancer would benefit from a combination of an ER downregulator and a PI3K pathway inhibitor. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr S3-4.

Summary Polyomavirus middle T (PyVmT) transforms cells through phosphatidylinositol 3-kinase (PI3K). Overexpression of ErbB2/ErbB3 heterodimers, potent activators of PI3K, occurs in PyVmT-driven mouse mammary tumors, but without known cause and consequence. Inhibition of ErbB2 with lapatinib, or genetically engineered loss of ErbB3, decreased PI3K signaling and tumor growth in mice, suggesting ErbB2/ErbB3 is required by PyVmT. ErbB2/ErbB3 formed signaling complexes containing PyVmT, p85 and Src. EZN-3920, a novel high-affinity oligonucleotide targeting ErbB3 transcripts, impaired ErbB3 expression in vivo, decreased PI3K signaling, cell survival, and growth of tumors, and improved response to lapatinib. Therefore, PyVmT utilizes ErbB2/ErbB3 overexpression to drive PI3K signaling. Furthermore, the challenge of ErbB3 as a kinase-inactive therapeutic target can be overcome by targeting ErbB3 mRNA transcripts in vivo. Significance Given the widespread use of, and knowledge gained from the MMTV-PyVmT transgenic mouse breast cancer model, understanding how PyVmT utilizes existing cell signaling programs is necessary to interpret how results impact our understanding/treatment of human breast cancers. We found that ErbB2/ErbB3 heterodimers were a necessary part of PyVmT signaling complexes involving Src and PI3K, and that PyVmT depends on ErbB3 in the breast epithelium for PI3K activity and tumor growth. Because ErbB3 harbors a weak/inactive tyrosine kinase, therapeutic anti-ErbB3 interventions present with challenges. Further, anti-ErbB3 antibodies that block ligand interaction with ErbB3 do not alter ligand-independent ErbB3 signaling. However. these challenges were overcome using high-affinity oligonucleotides targeting ErbB3 mRNA. Use of the oligonucleotide EZN-3920 effectively reduced ErbB3 expression, inhibited PyVmT tumor growth, and improved the response of tumors to lapatinib. These results support oligonucleotide knock-down of targets as potential therapeutic strategy. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P5-06-10.

Cancer therapies exert selective pressure on genomically heterogeneous breast cancers and may select for alterations in genes associated with drug resistance or metastasis. We evaluated this hypothesis by comparing sequential primary, post-neoadjuvant chemotherapy, and recurrent metastatic tumor samples from 5 patients with aggressive hormone receptor positive (HR+) breast cancer. Triplet samples (primary untreated, primary post-neoadjuvant chemotherapy, metastatic recurrence) were assessed in 5 patients with HR+ breast cancer. All patients were treated with cyclophosphamide and doxorubicin followed by paclitaxel or docetaxel, and adjuvant radiation and hormonal therapy. They all recurred during adjuvant antiestrogen treatment (12-43 months post-surgery). Metastatic sites included liver (2), ovary, brain, and pleural fluid. In these 5 tumor sets, we queried genomic alterations across 3,769 exons from 236 cancer genes using the FoundationOne™ NGS assay. Gene amplifications were frequently discordant among the triplet biopsies; all patients showed copy number alterations and 3 of 5 patients had two or more additional gene amplifications confirmed only in the biopsy of the metastasis. ESR1 (ERa) mutations were identified in the metastatic lesions of 2/5 patients but not in the diagnostic biopsy or the post-neoadjuvant therapy mastectomy specimen; both also harbored de novo and concurrent GATA3 truncations. One of the ESR1 mutations, Y537S, is known to result in an estrogen-independent, transcriptionally active receptor. The same patient showed amplification of 11q13 genes CCND1, EMSY, and FGF19, an amplicon known to be associated with endocrine resistance. Concordant FGFR1 and ZNF703 amplifications, both at 8p12, were observed de novo in 2 other patients; these alterations have beenassociated with the luminal B subtype and resistance to antiestrogens. FGFR1/ZNF703 alterations were mutually exclusive with ESR1 mutations. The 5th tumor contained a PIK3CA ‘hot spot’ mutation in all 3 sequential biopsies. This tumor also exhibited IGF1R, IRS2, MYC and MCL1 gene amplification only in the metastasis. Overall, IGF1R amplifications were observed in 3/5 patients. In addition, amplification of ZNF217 at 20q13.2, an amplicon with known association with poor prognosis, was observed de novo in 2 patients that exhibited rapid recurrence on adjuvant therapy (12 and 17 months). Acquired somatic alterations that deregulate ERα signaling may allow for escape from estrogen deprivation leading to treatment resistance. We speculate ESR1 mutations, rarely observed in primary breast tumors, are selected for during adjuvant hormonal treatment in aggressive HR+ breast cancer. Testing for de novo and acquired alterations throughout the clinical course of patients with ER+ cancers may identify targetable alterations, such as FGFR1 and IGF1R gene amplifications and others, against which targeted drugs are available. Studies are underway to validate these findings in additional patients with sequential samples and to investigate the function and therapeutic implications of ESR1 mutations. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr PD3-1.

Tumor-initiating cells (TICs) or cancer stem cells (CSCs) are resistant to chemotherapy and have been associated with metastatic recurrences and poor patient outcome. We have shown that the TGFb pathway plays a crucial role in the expansion of TICs and that inhibition of the type I TGFβ receptor (ALK5) decreases chemotherapy-resistant TICs. Further, a TGFβ-responsive gene signature is enriched in TNBCs upon treatment with chemotherapy and is predictive of poor prognosis (Bhola et al. JCI 2013). Herein we investigated the molecular and antitumor effects of TR1, a TGFβRII neutralizing antibody currently in phase I clinical trials, against TNBC. TR1 recognizes the TGFβ ligand binding epitope in the extracellular domain of TGFβRII; it blocks ligand binding in the nM range. Small molecule TGFβRI kinase inhibitors are another approach to inhibit TGFβ/SMAD signaling. However, they cross-react with other type I receptors (ALK1, 4, 7), some of which are important for normal physiology. Furthermore, TGFβRII can induce EMT in breast cancer cells independent of dimerization with TGFβRI. Therefore, we hypothesized that, by neutralizing TGFβRII, TR1 can abrogate the breast TIC population via inhibition of SMAD and non-SMAD signaling. Using the pCAGA12-luciferase reporter to assess SMAD-dependent transcriptional activity, TR1 decreased luciferase activity in TNBC cell lines (SUM159 and BT549) and steady-state levels of P-SMAD2, suggesting it inhibited autocrine TGFβ signaling. TR1 decreased the TIC fraction as identified by FACS analysis of cell line-specific TIC markers (ALDH for SUM159; CD44hi/PROCR+ for BT549; PROCR+/ESA+ for MDA231 cells). Although TR1 displayed no effects on TNBC growth in vitro, it decreased tumor growth in vivo. We used a phospho-kinase array (45 kinases) to elucidate the non-SMAD pathways altered by TR1. TR1 decreased phosphorylated EGFR (Tyr1068) and STAT3 (Tyr705) in TNBC cell lines. By flow cytometry, the cell-surface expression of EGFR was also lowered upon treatment with TR1. TR1 also reduced P-ERK1/2 and P-Akt (Ser473) levels, and lowered NFκB transcriptional activity as measured by luciferase reporter assay. Using a cytokine array (120 cytokines) and qRT-PCR we observed that TR1 decreased the expression and release of cytokines involved in TIC enrichment such as IL-6 and IL-8. Inhibition of EGFR (with gefitinib) and MEK1/2 (with selumetinib) decreased the TIC population while augmenting pCAGA luciferase activity. The combination of SMAD2 siRNA with each EGFR or MEK inhibitors significantly decreased the TIC population. We speculate based on these data that inhibition of both SMAD and non-SMAD signaling is necessary for complete abrogation of TICs. This may explain the efficacy of TR1 as a single agent. These findings suggest that targeting TGFβRII with TR1 exerts its anti-tumor effects via inhibition of both SMAD and non-SMAD (EGFR, ERK) pathways that are important to TIC survival. The results also encourage combination therapies of chemotherapy with TR1 to reduce tumor recurrences in patients with TNBC. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr PD5-1.

Janus-kinase 2 (JAK2) is a receptor-coupled tyrosine kinase which transmits cytokine-mediated signals to the STAT pathway to drive signals of proliferation and differentiation. JAK2/STAT signaling has been shown to play a role in promoting breast cancer ‘stem-ness’ and driving the proliferation of CD44+/CD24- basal-like breast cancer cells. We performed targeted next-generation sequencing (tNGS) using HiSeq2000 at a CLIA-certified laboratory on 68 residual TNBCs after neoadjuvant chemotherapy (NAC), which are putatively enriched for drug-resistant cells. JAK2 amplifications were identified in 7/68 (10.2%) post-NAC tumors. Gene copy estimation in amplified cases ranged from 7-10 copies. All cases were confirmed by a novel JAK2-FISH assay that we developed using BAC-derived probes targeting the JAK2 locus on chromosome 9, using CEN9 as a normalization factor. Using JAK2-FISH, we identified only one amplified tumor in an independent cohort of 30 untreated TNBCs (3%). Rates of JAK2 amplification in post-NAC residual cancers in this study were also higher than those reported by TCGA ( JAK2 amplifications in the post-NAC residual cancer were associated with poor RFS (median: 7 mo. vs 17 mo; HR: 3.36; p = 0.006) and OS (median: 11.3 mo. vs 27.6 mo; HR: 4.16; p = 0.002), and were significantly associated with higher gene expression of IL6, CXCR1 and SNAI1, among others. The coexistence of these alterations suggests an association of amplified JAK2 with an epithelial-to-mesenchymal transition and cancer-stem cell programs. To explore the clonal evolution of JAK2 lesions in breast cancer, we sequenced sequential samples (diagnostic biopsy, post-NAC residual disease and metastatic recurrence) of two TNBCs where JAK2 amplifications were detected in the residual disease. In each of these, JAK2 gain was noted at levels below the amplification threshold at diagnosis, but was focally amplified in both the residual tumor at the time of definitive surgery as well as in the subsequent metastatic recurrence. To our knowledge, this is the first report of JAK2 gene amplification in breast cancer. These data suggest a role for JAK2 amplifications in driving chemotherapeutic resistance and disease progression. We hypothesize these lesions are therapeutically targetable with currently approved JAK2 or pan-JAK inhibitors. Additional data to be presented will include analysis of tumor-infiltrating lymphocytes in JAK2-amplified tumors and molecular studies of the role of JAK2 overexpression in breast cancer cell lines as well as their sensitivity to JAK2 inhibitors in clinical development. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr S6-01.

Background: Modulation of estrogen synthesis or its receptor activity represents the treatment of choice for postmenopausal women with estrogen receptor-positive (ER+) breast cancer (BC). However, despite initial response to hormonal manipulation, most tumors ultimately develop endocrine resistance. The emerging evidence that ERα can signal in both ligand-dependent and independent manner supports the development of agents that are not only competitive ERα antagonists but also reduce steady state levels of the receptor and thus limit both modes of signaling. ARN-810 is a novel, non-steroidal ERα antagonist that induces proteasomal-mediated degradation of ERα in breast cancer cell lines at picomolar concentrations. This unique biological profile coupled with good oral pharmacokinetics (PK) produced tumor regression in both tamoxifen-sensitive and resistant BC models. Importantly, in the tamoxifen-resistant model, response to ER modulators like fulvestrant and tamoxifen was limited to tumor growth arrest. Trial Design: This is a safety and PK study of ARN-810 administered orally on a continuous daily dosing regimen with a PK lead-in. The incidence of dose limiting toxicity (DLT) will be assessed in the first cycle of treatment (28 days). Patients will be assigned to escalating doses of ARN-810 in cohorts of 3-6 patients per dose until determination of the maximum tolerated dose (MTD) and/or recommended Phase 2 Dose (RP2D) using standard 3+3 criteria. The MTD and/or RP2D will be defined as the dose with ≤1 out of 6 patients with DLT. Key eligibility criteria include postmenopausal women with locally advanced or metastatic ER+ (HER2-) BC, ECOG PS 0/1, and adequate organ function. Untreated or symptomatic brain metastases and endometrial disorders are exclusionary. At the MTD and/or RP2D, expansion cohorts of patients with disease progression after ≤2 prior hormonal therapies (≤1 in the metastatic setting) or prior therapy with fulvestrant will be enrolled. Plasma PK after first dose and at steady-state will be analyzed using non-compartmental methods. Pre- and post-treatment functional imaging with [18F]-fluoroestradiol (FES)-PET to assess tumor ER binding and paired tumor biopsies to assess for ER target modulation by IHC and gene expression will be performed. Objectives: The primary objective is to determine the MTD and/or RP2D of ARN-810. In addition to safety and PK, evaluation of biomarkers of pharmacodynamic (PD) response with FES-PET and ER target gene expression will be explored along with preliminary antitumor activity. Statistical Methods: Descriptive statistics will be used to summarize patient characteristics, treatment administration/compliance, safety and PK parameters, antitumor activity endpoints, and exploratory biomarkers. Status: To date, 3 patients have been enrolled in the first cohort at 100 mg/day. The target accrual for the dose escalation will depend upon the observed safety and PK/PD profile, which will determine the number of escalations. During dose expansion, 30 patients will be enrolled. More information about this study can be found at http://clinicaltrials.gov/ct2/show/NCT01823835?term = ARN-810. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr OT3-2-07.

Background: Neoadjuvant chemotherapy (NAC) induces a pathological complete response in approximately 30% of triple-negative or basal-like breast cancers (BLBC). However, patients with residual disease often recur after surgery, presumably due to persistent drug-resistant subpopulations with cancer stem cell (CSC)-like properties. Thus, elimination of this CSC compartment in BLBC has the potential to improve survival by reducing post-surgical metastatic recurrences. Methods: We sampled 49 post-NAC breast cancer samples including 22 BLBCs and quantified RNA for 355 cancer-related transcripts using Nanostring technology. Transcripts associated with a high Ki67 in the residual disease (a biomarker of early recurrence) were identified and bioinformatically examined for an association with drug resistance and a CSC phenotype. Loss of DUSP4, a negative feedback regulator of ERK1/2 and JNK1/2, was highly associated with a high post-NAC Ki67. We examined the role of loss of DUSP4 in promoting a drug-resistant, CSC phenotype. Results: Low DUSP4 expression in post-NAC tumors was associated with high ERK1/2 activation and BLBC gene expression. siRNA knockdown of DUSP4 enhanced resistance to anti-cancer chemotherapy (docetaxel, camptothecin). Alternatively, forced DUSP4 expression in breast cancer cell lines abrogated the activation of transcription factors downstream of ERK and JNK and sensitized cells to docetaxel-induced apoptosis. In highly metastastic BLBC cell lines, MEK inhibition with AZD6244 and JNK inhibition with SP600125 significantly reduced mammosphere formation and self-renewal. Inhibition of JNK and MEK reduced IL6 and IL8 expression, two cytokines known to expand the CSC compartment. Reconstitution of exogenous IL6 and IL8 after MEK inhibition restored mammosphere formation potential. Chromatin immunoprecipitation demonstrated that the oncogenic transcription factor ETS-1, an ERK1/2 substrate, binds the IL8 promoter in a MEK1/2-dependent manner. We are currently confirming the association of DUSP4 loss with Ki67, ERK activation, and IL6/IL8 expression in a cohort of 113 post-NAC triple negative breast cancers. Conclusions: Our data demonstrate that loss of DUSP4 in BLBC promotes ERK and JNK activation by impairing negative feedback of these pathways. Activation of ERK and JNK drives expression of IL6 and IL8 expression, possibly through ETS-1 activation. Thus, targeting these signaling pathways may eliminate residual cancer stem cells after neoadjuvant chemotherapy and improve cure rates in BLBC. Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr PD08-03.

Background: In basal-like breast cancers, drugs that either negatively modulate p63 and/or activate p73, such as cisplatin and paclitaxel, promote p73-dependent apoptosis. Further, the serine threonine kinase mTOR negatively regulates p73. Inhibition of mTOR with RAD001 upregulates p73, and results in p73-mediated apoptosis. Therefore, we conducted a phase II trial in patients with HER2−negative MBC to explore safety, tolerability and anti-tumor activity of the combination of paclitaxel, cisplatin and RAD001. Materials and Methods: We initiated an open-label, phase II multi-institutional study of weekly cisplatin (25 mg/m2), paclitaxel (80 mg/m2) and daily RAD001 (5 mg), given on a 28-day cycle. Treatment was continued until unacceptable toxicity or progression of disease. All toxicities were documented using the NCICTCv.4. Disease was assessed every 2 months. Results: A total of 55 patients were enrolled. Median age was 55 years; 62% of patients had prior chemotherapy with a median of 3 previous regimens in the metastatic setting. Sixty-three percent of patients had triple-negative disease, 81% patients had visceral disease, and 35% had bone metastases. Twenty-one patients are still on study. The most common toxicities are summarized in the table below. Neutropenia was the main cause for dose reductions, mainly after cycle 3. Only 1 patient developed febrile neutropenia. Of the 44 patients assessed for best response (RECIST) thus far, 11 had partial response, 21 stable disease, 9 disease progression, and 3 were not evaluable. Thirty eight patients have discontinued treatment so far; 30 due to disease progression, 5 due to toxicity, and 3 withdrew consent. Current median time to progression on the evaluable patients is 6 months. Discussion: The combination of RAD001, cisplatin and paclitaxel was overall very well tolerated despite cumulative pancytopenia. Significant antitumor activity in this heavily pre-treated patient population was seen. All patients’ primary diagnostic and/or metastatic tumor biopsies are currently being analyzed for the presence of PIK3CA and AKT1 mutations, and immunohistochemical expression of p53, p63, p73, and PTEN. Microarrays are being generated to determine whether time to progression is superior in patients with basal-like breast cancers. Microarrays will be mined to identify a pretreatment profile that mirrors a low p63/high p73 gene expression signature. Correlation of mutational analysis and gene signatures with clinical benefit will be presented at the meeting Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr PD09-06.

Characterization of breast cancer subtypes at the protein-level is largely unexplored and is a powerful approach to identify novel biomarkers that may have diagnostic and prognostic utility. Shotgun proteomics analyzes mixtures at the peptide-level to generate MS/MS spectra that are then used to identify the peptides and proteins from which they were derived. We developed a method in our laboratory utilizing SDS-PAGE followed by in-gel digestion coupled with reversed-phase (RP) liquid chromatography-tandem mass spectrometry (RP-LC-MS/MS) for proteomic analysis of LCM-acquired cells. This method was then applied to a clinical sample set consisting of basal, Her2-overexpressing and luminal A frozen tumor tissues as determined by prior microarray studies. Three tumors per group were dissected in triplicate, for a total of 27 samples. Each dissection consisted of approximately 10,000 cells, corresponding to 3-5 μg protein. Following LCM, the thermoplastic membranes containing the captured cells were peeled from the cap, suspended in SDS loading buffer and gently heated for 10 minutes. The solubilized proteins were electrophoresed 2 cm into a 10-20% Tricine gel followed by in-gel trypsin digestion and peptide extraction. Replicate LC-MS/MS analysis of the tryptic peptides from each of the 27 samples was performed on a Thermo LTQ XL ion trap mass spectrometer. The resultant tandem mass spectra were searched against the human IPI database using the Myrimatch search algorithm and the results filtered using IDPicker. A spectral counting approach was applied to the data-dependent data to compare quantitative differences among the identified proteins. A total of 91,646 spectra were confidently identified, corresponding to 1671 protein groups across all biological and technical replicates. An average of 545, 543, and 610 protein groups were identified in the basal, ERBB2, and luminal A subtypes, respectively, from an equivalent of approximately 800 cells. Spectral count differences between datasets were analyzed using both a Quasi-likelihood Poisson regression model, developed in-house, and the Limma package in Bioconductor. Using a False Discovery Rate < 0.05, a total of 90 proteins showed statistically significant differences in expression among the breast tumor subtypes. Hierarchical cluster analysis generated a heat map of the protein expression patterns. Novel proteins as well as proteins corresponding to genes previously shown by gene array studies to be specific to each tumor type, including NES, ERBB2 and LGALS3BP, were represented among the differentially expressed proteins. These results demonstrate the feasability of this method to identify differential expression of protein biomarkers among breast cancer subtypes. These biomarkers should yield new insights into the pathogenesis of breast cancer and could be rapidly adapted into clinical diagnostics to guide therapeutic decisions for individual patients. The possible role in breast cancer for the novel proteins identified is currently under investigation. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P6-04-11.

In order to discover mechanisms of escape from hormone dependence, we established four ERα-positive human breast cancer cell lines after long-term estrogen-deprivation (LTED). Growth of MCF-7/LTED and HCC-1428/LTED cells was inhibited by the ER downregulator fulvestrant and by ER siRNA, suggesting continued dependence upon estrogen-independent ER signaling. Fulvestrant also prevented the emergence of hormone-independent MCF-7, HCC-1428, and BT-474 cells, but not ZR75- 1, MDA-361, and T47D cells. Gene expression analysis revealed that ER modulates a transcriptional program directed by the E2F proteins, a family of transcription factors derepressed by CDK4/6-mediated phosphorylation of Rb. A set of 1,003 genes with expression altered ≥1.5-fold (P Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr PD05-05.

ER+ breast cancers typically respond to treatment with aromatase inhibitors (AIs), but a significant fraction exhibit de novo or acquired resistance. To model AI resistance, we cultured four ER+, hormone-dependent human breast cancer cell lines under hormone-depleted conditions for several months until hormone-independent populations emerged (termed long-term estrogen-deprived, LTED). LTED cells outgrew parental counterparts under hormone-depleted conditions. While two LTED lines showed increased ER levels and response to estradiol, two LTED lines showed the opposite. Therefore, ER+ breast cancer cells did not consistently overcome hormone deprivation by increasing sensitivity to low estrogen levels.We analyzed protein lysate arrays with 625 antibodies against signaling proteins to discover common mechanisms of hormone-independent growth. All LTED lines showed increased phosphorylation of p70S6K, an mTOR substrate, compared to parental controls. An siRNA library screen targeting 779 kinases revealed that downregulation of kinases linked to the phosphatidylinositol-3 kinase (PI3K) signaling pathway (serum/glucocorticoid-regulated kinase 1; insulin receptor; p110α/PI3K) inhibited the growth of MCF-7/LTED but not parental MCF-7 cells under hormone-depleted conditions. Immunoblot analysis for P-AKT and P-p70S6K showed that PI3K/AKT/mTOR signaling was increased in all LTED lines. Pathway analysis of gene expression microarray data showed significant alteration of genes involved in insulin-like growth factor-I (IGF-I) signaling in 3/4 LTED lines. Receptor tyrosine kinase array analysis revealed increased activation of receptors for IGF-I (IGF-IR) and/or insulin (InsR) in 3/4 LTED lines. These receptors transduce signals to potently activate PI3K. Treatment with the IGF-IR/InsR kinase inhibitor AEW541 decreased P-AKT in 3/4 LTED lines. Treatment with inhibitors of IGF-IR/InsR (AEW541), PI3K/mTOR (BEZ235), or mTOR (RAD001) suppressed the monolayer and anchorage-independent growth of 3/4, 4/4, and 4/4 LTED lines, respectively, and prevented the emergence of hormone-independent cells. Treatment with PI3K and mTOR inhibitors induced apoptosis in 3/4 LTED lines. Finally, we analyzed levels of PI3K pathway markers (P-AKT, P-S6, P-GSK3α/β) in primary tumor lysates from 65 ER+ breast cancer patients using reverse-phase protein arrays. Hierarchical clustering yielded two groups of tumors with high or low PI3K activation. Following adjuvant anastrozole therapy, PI3K-high patients had significantly shorter disease-free survival compared to PI3K-low patients. With a median follow-up of 9.4 months, recurrence rates were 16% and 0%, respectively (p Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 403.

Background: The EGFR and the PI3K pathways have been implicated in breast cancer's prognosis and mechanisms of resistance to conventional therapies. Preclinical data have shown synergy between EGFR and PI3K/Akt pathway blockade. Therefore, we conducted a phase Ib trial to determine the maximum tolerated dose (MTD) and antitumor activity of Erlotinib, an oral EGFR tyrosine kinase inhibitor, and Everolimus (RAD001), an oral mTOR inhibitor, given daily to patients with metastatic breast cancer.Materials and Methods: This was an open-label, phase I dose-escalation study. Everolimus was started at 2.5 mg/d and Erlotinib was started at 100 mg/d on a 28-day treatment cycle. Doses were escalated over 6 levels (Everolimus 2.5, 5 or 10 mg daily with Erlotinib 100 or 150 mg daily). Treatment was continued until consistent dose-limiting toxicity (DLT) was observed or until progression of disease. All toxicities were documented using the NCI CTC v.3. Disease assessment was done every 2 months after initiation of therapy.Results: A total of 14 patients were enrolled on the first 2 dose levels (Erlotinib 100 mg with Everolimus 2.5 mg [level I] and 5 mg [level II] daily). Eight patients were enrolled on level I and 6 patients were enrolled on level II. The median age of all patients enrolled was 55 years of age, and 92% of them had visceral disease. Half of the patients had ER+ disease, 35% had HER2 + disease, and 15% had triple-negative disease. The median number of previous chemotherapy regimens in the metastatic setting was 4. Two patients on the level II cohort discontinued trial due to grade 3 stomatitis prior to their first disease assessment. Of the remaining 13 patients, 11 progressed at the time of their first disease assessment, and 1 patient on the dose level I cohort had a partial response that lasted for 7 months. The most common toxicities were rash (16%), transaminase elevation (15%), stomatitis (13%), fatigue (7%) and diarrhea (5%). Grade 3 and 4 toxicities were overall uncommon (3.5%), and mostly consisted of stomatitis, transaminase elevation and rash. Stomatitis was the only DLT, and was seen at dose level II. The MTD was determined to be Erlotinib 100 mg and Everolimus 2.5 mg daily.Discussion: In patients with metastatic HER2-negative breast cancer the combination of Erlotinib and Everolimus was overall well tolerated, but clinically ineffective from a tumor activity point of view in this heavily pre-treated patient population. Stomatitis was one of the most common toxicities and also the DLT in the cohort treated with dose level II. In view of the plethora of novel active agents currently in clinical trials in breast cancer, we do not believe the above combination warrants further testing in this disease. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3094.