GRHL2 is a key lineage determining factor which collaborates with FOXA1 to establish a targetable collateral pathway in the setting of endocrine therapy-resistant breast cancer

The estrogen receptor (ER) is expressed in the majority of luminal breast cancers and inhibition of its transcriptional activity with selective estrogen receptor modulators, selective estrogen receptor degraders and/or aromatase inhibitors is a standard approach used in the management of this disease. Despite the positive clinical impact of these interventions, de novo and acquired resistance limits the therapeutic lifespan of these classes of drugs. Considering what is known about the complex mechanisms that contribute to the development of resistance it is likely that further development of ER-modulators will yield only incremental improvements. Thus, with the view that resistance is inevitable, we undertook the development of a new approach to treat ER-positive breast cancer by identifying and exploiting targetable vulnerabilities that emerge in endocrine therapy resistant disease. Genomic discovery platforms, including DNASeq, ChIPSeq and RNASeq were used to assess the epigenome, targeting global transcription factor binding profile, and transcriptome in cellular models of endocrine therapy sensitive and resistant disease. DNASeq was first used to identify the chromatin state, with a focus on differences, between these two models. Motif enrichment analysis indicated FOXA1 was a candidate transcription factor influencing the chromatin architecture, which was consistent with previously published studies. This led to the examination of the FOXA1 chromatin binding profile in these models. FOXA1 has previously been described to bind at enhancers. Furthermore, the relative transcription activity of specific enhancers has been shown to be indicated by the epigenomic marks on histones flanking transcription factor binding sites. For this reason, we assessed the specific pattern of histone 3 lysine 4 methylation to confirm enhancer status and histone 3 lysine 27 acetylation as an indicator of transcriptional activity. The specific patterns and distribution of FOXA1 binding was then integrated with this epigenomic information to reveal a subset of enhancers that became activated and another subset that gained enhanced activation in the tamoxifen resistant setting relative to the tamoxifen sensitive model from which it was derived. These results were integrated with the differential transcriptome, or genes shown to be differential expressed based on RNASeq, in the TAMR model as compared to its parental cell line, MCF7-WS8, and confirmed that the active enhancers were in fact associated with genes that were expressed more highly, on average, in the TAMR model. Motif enrichment at these two subgroups of enhancers indicated that another transcription factor, GRHL2, likely interacts with FOXA1 at these active sites. These results were again integrated with the “differential transcriptome” based on RNASeq and confirmed that the active enhancers, and indicated an even stronger enrichment of genes that were expressed more highly, on average, in the TAMR model relative to the MCF7-WS8 model. The GRHL2 transcriptome was then further defined by both GRHL2 ChIPSeq as well as RNASeq by comparing TAMR samples in which downregulation of GRHL2 expression had been achieved via siRNA as compared to control siRNA sequences. The collection of these data defined a subset of genes, the GRHL2 dependent transcriptome, that demonstrated increased expression in TAMR. The results of these cell lines studies were corroborated by assessing the transcriptome of xenograft mouse models of endocrine therapy sensitive and resistant disease. Integrative analysis of these data identified a collateral ER-independent signaling pathway in endocrine therapy resistant tumors that converges upon and modulates the FOXA1 and GRHL2 cistrome/transcriptome. This SuperSeries is composed of the SubSeries listed below. MCF7-WS8 and TAMR cells were treated analyzed in duplicate (for DNASeq) or triplicate (for ChIPSeq, RNASeq) following various treatments. For DNASeq, MCF7-WS8 and TAMR cells were grown in hormone free media for 2 days, followed by 1 day of 4-hydroxytamoxifen treatment. For RNASeq of MCF7-WS8 and TAMR cells at baseline, cells were grown in hormone free media for 2 days, followed by 1 day of 4-hydroxytamoxifen treatment. For ChIPSeq, MCF7-WS8 and TAMR cells were grown in hormone free media for 3 days, following 45 min of vehicle or hormone treatment (estradiol or 4-hydroxytamoxifen). For RNASeq following siRNA treatment, TAMR cells were transfected with various siRNA to GRHL2 or control siRNA for 72 hours in hormone free media. For RNASeq of xenograft tumors, MCF7-WS8 and TAMR tumors were analyzed in quadruplicate following various hormone treatments. For all tumors, female nu/nu mice (~6 weeks in age) were ovariectomized under isoflurane anesthesia; a slow release estradiol pellet (0.72mg/60 days from Innovative Research of America) or tamoxifen treatment pellet (5mg/60 days from Innovative Research of America) was implanted in the scapular region during the same procedure. The next day an approximate 8 mm3 of tumor issues (derived from a sectioning a freshly harvested tumor of 0.8 – 1cm3 volume) was engrafted orthotopically (right axial mammary fatpad) under anesthesia. Tumors were measured 3X weekly, concurrent with weight and behavior monitoring. For MCF7-WS8 tumors, tumors were grown for 30 days when they reached a size of ~0.15-0.2 cm3 volume ( l x w2 x 0.5) animals were then randomized to treatment with vehicle (0.1cc corn oil sc 3x weekly) or tamoxifen (40mg/kg 3x weekly). Animals were euthanized and tissues preserved following a total of 6 weeks of tumor growth. For TAMR tumors, tamoxifen treated animals were randomized to vehicle (0.2cc corn oil sc weekly), or fulvestrant (200mg/kg sc weekly). Animals were euthanized and tissues preserved following a total of 8 weeks of tumor growth. Refer to individual Series