Use of a navigable interface for integrated whole genome and transcriptome sequencing as a platform for pursuit of therapeutic targets in advanced prostate cancers.

Abstract

225 Background: Metastatic, castration-resistant prostate cancer (mCRPC) is commonly the deadly form of PC. Among these, a subset of tumors are androgen-indifferent with the most aggressive often manifesting variant histology, including neuroendocrine or small cell changes. Neuroendocrine PC can be de novo (NEPC) or develop in response to therapy as treatment emergent (CRPC-NE). Currently effective durable treatments for NEPC are lacking. Hence, we sought to identify additional targets in CRPC/NEPC using an integrative platform of whole genome (WGS) and transcriptome sequencing (RNAseq). Methods: WGS was performed on55 tumor/normal pairs (CRPC-Ad, n= 32; CRPC-NE, n=13; de novo NEPC, n=7; metastatic hormone naïve, PC n=3) from 48 patients. RNAseq data was available in a subset of 21 samples. We employed the Isabl GxT analytic platform and manually curated single base substitution (SBS, COSMIC v3) molecular signatures and structural variants (SV) that involved tumor suppressor genes and oncogenes. Results: We observed 184 events in cancer-associated genes and targets in 38 cases. Non-canonical ETS fusions were identified in 2 CRPC-Ad patients ( MSMB-ERG and YWHAE-ETV4). Other rare events included SVs affecting ALK ( SLC45A3-ALK) and FGFR1 amplification in 1 patient each. Pathogenic germline alterations in 15% of patients with equal frequency in each clinicopathological state. These variants included genes such as BRCA1, BRCA2, and ATM, and other genes of uncertain relevance for prostate cancer ( e.g., PPM1D and MUTYH). SBS genomic signatures associated with homologous recombination deficiency (HRD) were observed in 15% of the patients (7 cases): 3 harbored germline BRCA1/2mutations, 2 with somatic BRCA2 mutations, and 2 without alteration in BRCA1/2 (1 of these CRPC-Ad had a complex SV disrupting RAD51B) without apparent enrichment for any histology, and a majority of both histologies were enriched in Mismatch repair (MMR)-associated SBS. One subject CRPC-NE and amphicrine character, which displayed a complete response to immune checkpoint blockade, harbored driver mutations in AR and CTNNB1, and homozygous loss of MSH2/6. Further, molecular signatures of potential clinical relevance were detected at varying contributions and included CDK12-type genomic instability (CRPC-Ad, n=2) (4%) and MMR deficiency with POLD1 proofreading (CRPC-Ad) who also experienced a durable response to pembrolizumab. Conclusions: WGS/RNAseq in CRPC and NEPC elucidates genomic signatures associated with HRD and MMR, complex SVs in oncogenes, and non-canonical ETS fusions. Expansion of our analysis is underway with enhanced integration of clinical metadata and RNAseq for rational trial design for aggressive variant CRPC and NEPC.