Transcriptional profiling of matched biopsies reveals molecular determinants of enzalutamide resistance.

Abstract

5058 Background: Castration-resistant prostate cancer (CRPC) is the lethal form of the disease. One of the principal therapies in CRPC is the potent androgen receptor (AR) signaling inhibitor enzalutamide (enza). Most patients benefit from enza, but disease progression is nearly universal. A variety of resistance mechanisms have been described by comparing enza-naïve and enza-resistant tumors. However, these results are largely from different groups of patients and do not provide information on the changes induced by enza within a given patient. Lineage plasticity—most commonly-exemplified by loss of AR signaling and switch from a luminal to an alternate differentiation program—is a particularly aggressive resistance mechanism. Importantly, lineage plasticity appears to be increasing in incidence since more widespread use of potent AR signaling inhibitors such as enza. To improve our understanding of resistance mechanisms induced by enza treatment, we analyzed the transcriptomes of matched metastatic CRPC patient biopsies obtained prior to treatment and at the time of disease progression. Methods: All biopsies were obtained as part of the Stand Up 2 Cancer/Prostate Cancer Foundation-funded West Coast Dream Team, a prospective, IRB-approved protocol focused on understanding the biology of metastatic CRPC. We identified 21 patients for whom matched tumor biopsies with RNA-seq were available prior to starting treatment with enza and at the time of progression while still taking enza. Results: Our RNA-seq analysis demonstrates that the majority of progression tumors cluster with their baseline pair, suggesting that enza does not markedly change the tumor transcriptome in most cases. Three of 21 patients showed evidence of lineage plasticity at progression by gene expression analysis. By analyzing the RNA-seq data, we identified pathways linked to stemness that were more activated in baseline tumors from patients whose progression tumors underwent lineage plasticity. Furthermore, we identified a gene signature enriched in these baseline tumors that was associated with risk of lineage plasticity after enza treatment. We determined that high expression of this signature was strongly associated with poor survival from the time of AR signaling inhibitor treatment in independent patient samples, suggesting this signature is linked to poor patient outcome. Conclusions: Enza-resistant tumors are heterogeneous. Most tumors do not undergo significant transcriptional changes at progression vs. baseline. Matching recent reports, approximately 15% of tumors underwent lineage plasticity upon progression. Our work implicates a gene program that may predispose tumors to enza-induced lineage plasticity. Finally, the gene signature we identified may be a marker of lineage plasticity risk and tumor aggressiveness in CRPC prior to the initiation of AR signaling inhibitor therapy.