Background: We recently reported two ESR1 fusions (ESR1-YAP1 and ESR1-PCDH11X) that drive endocrine therapy (ET) resistance and metastasis in estrogen receptor positive (ER+) metastatic breast cancer (MBC) (PMC6171747). Here, we report the functional properties of additional ESR1 fusions in ET-resistant MBC with an emphasis on the identification of a transcriptional signature designed to diagnose the presence of an active ESR1 fusion for targeted therapies directed against ESR1 fusion-driven biology. Methods:ESR1 fusions were detected by RNA-seq in ER+ MBC samples. ESR1 fusions were reproduced as cDNA constructs and expressed in ER+ breast cancer cell lines. Hormone-independent cell growth was detected by an Alamar blue assay and activated cell motility by a scratch wound assay. The transcriptional properties of ESR1 fusions was studied by RNA-seq followed by qPCR-based validation. Signature performance was evaluated using a ROC analysis on ER+ patient derived xenografts (PDX) harboring a variety of ESR1 somatic events. Results: All ESR1 fusions studied encoded the first six exons of ESR1 fused in-frame to diverse partner genes, thus replacing the ESR1 drug/ligand binding domain (LBD). Fusions involving a known transcription factor (TF) or coactivator (CoA) gene, including ESR1-YAP1, ESR1-SOX9 and ESR1-ARNT2 drove fulvestrant-resistant cell growth and hormone-independent cell motility. Other ESR1-e6 fusions, including ESR1-DAB2, ESR1-GYG1, ESR1-PCMT1 and ESR1-ARID1B did not induce these properties. From these examples, a functional rule is emerging whereby inter-chromosomal ESR1 translocations fused in-frame to 3’ partner genes with a positive role in transcription are active. Intra-chromosomal fusions with genes with no transcriptional roles are likely inactive. The ESR1-PCDH11X fusion is an exception, suggesting the need for continued functional study of non-TF/CoA partner ESR1-e6 fusions. RNA-seq of T47D cells expressing the full panel of gene fusions demonstrated an overlapping pattern of transcriptional activation focused on estrogen response and epithelial-to-mesenchymal transition (EMT) genes driven by active fusions. This gene signature was well-preserved in a PDX naturally expressing the ESR1-YAP1 fusion. Interestingly, further study showed that a series of ET-resistant PDXs bearing a variety of ESR1 LBD point mutations induced a similar pattern to the active ESR1 fusion signature suggesting overlapping transcriptional regulatory events between ESR1 fusions and ESR1 LBD mutations. The ESR1-D538G mutation conferred the most comparable gene dysregulation to ESR1 fusions. The Y537S/N and E380Q mutations also reproduced the signature driving hormone-independent growth but with exceptions. Two PDX lines bearing either a fully heterozygous Y537S or L536P mutations were surprisingly completely estrogen-dependent. Neither of these examples exhibited the ESR1 fusion gene signature, suggesting an unknown secondary event needed to fully express the phenotype of some ESR1 mutants. The gene signature distinguished ESR1 mutations (constitutively active fusions and point mutations) from wild-type ESR1, with a 92.0% Area Under Curve. Conclusion: Here, we show that ESR1 fusions are recurrent somatic mutations that lead to drug resistance and metastasis by transcriptional reprogramming. We describe a fusion gene signature that may be useful to determine whether an ESR1 fusion or mutation is transcriptionally active and is capable of driving hormone-independent growth and endocrine therapy resistance.

Citation Format: Xuxu Gou, Meenakshi Anurag, Jonathan T Lei, Purba Singh, Sinem Seker, Adrian V Lee, Dan R Robinson, Matthew J Ellis. Recurrent active ESR1 fusions render a diagnostic transcriptional signature in metastatic breast cancer [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS17-03.