Targeting kinome reprogramming in ESR1 fusion-driven metastatic breast cancer.

Abstract

1085 Background: Genomic analysis has recently identified multiple ESR1 gene translocations in estrogen receptor-alpha positive (ERα+) metastatic breast cancer (MBC) that encode chimeric proteins whereby the ESR1 ligand binding domain is replaced by C-terminal sequences from many different gene partners. Transcriptionally active ESR1 fusions promoted hormone-independent cell growth, motility and resistance to endocrine therapy. The diversity of partner genes creates a considerable diagnostic challenge and no targeted treatments exist for ESR1 translocated tumors. Thus, we have established a transcriptional signature to diagnose the presence of an active ESR1 fusion (PMID: 34711608) and developed novel targeted therapies against ESR1 fusion-driven biology. Methods: Fifteen ESR1 fusion cDNA constructs were expressed in ER+ breast cancer cell lines by lentiviral transduction. Cell growth was assayed by Alamar blue assay. A mass spectrometry (MS)-based Kinase Inhibitor Pulldown Assay (KIPA) and tandem mass tag-based proteomics were performed to identify ESR1 fusion-driven druggable kinases for subsequent pharmacological inhibition. Results: KIPA profiling demonstrated an increase of multiple receptor tyrosine kinases including RET in T47D cells expressing active ESR1 fusions. Inhibition of RET by repurposing an FDA-approved drug significantly suppressed ESR1 fusion-driven cell growth in vitro, suggesting that despite marked diversity in the 3’ partners, common kinase activities were elevated and targetable. Proteogenomic profiling, including whole exome sequencing, RNA sequencing, and MS-based proteomics and phosphoproteomics were further performed on 22 ER+ patient-derived xenograft (PDX) tumors, which demonstrated different degrees of estradiol dependence. These integrated “omic” profiles defined targetable genes/pathways and predict tumor subsets that could be responsive to kinase inhibition therapy from this biologically heterogeneous panel of PDX tumors. WHIM18, a PDX naturally harboring the ESR1-YAP1 fusion showed elevated level of RET and CDK4/6 pathways. The tumor volumes were significantly reduced by the RET inhibitor. CDK4/6 inhibitor treatment showed similar tumor reductions to RET inhibition. Interestingly, WHIM9 PDX that expressed wild-type ESR1 conferred a comparable kinome profile to WHIM18. The tumor growth was significantly suppressed by RET or CDK4/6 inhibition. Therefore, pharmacological experiments validated proteogenomics-predicted drug response in two tested ER+ PDX models. Conclusions: Proteogenomics characterization of PDX tumors can drive clinical trial hypotheses. Here, we reveal therapeutic kinase vulnerabilities in ESR1 fusion-driven tumors as exemplified by RET inhibition, which will lay the framework for future clinical trials.