The Molecular Context of Vulnerability for CDK9 Suppression in Triple Wild-Type Melanoma

Abstract

Approximately half of melanoma tumors lack a druggable target and are unresponsive to current targeted therapeutics. One proposed approach for treating these therapeutically orphaned tumors is by targeting transcriptional dependencies (oncogene starvation), whereby survival factors are depleted through inhibition of transcriptional regulators. A drug screen identified a CDK9 inhibitor (SNS-032) to have therapeutic selectivity against wild-type (wt) BRAF^wt/NRAS^wt melanomas compared with BRAF^mut/NRAS^mut mutated melanomas. We then used two strategies to inhibit CDK9 in vitro-a CDK9 degrader (TS-032) and a selective CDK9 kinase inhibitor (NVP-2). At 500 nM, both TS-032 and NVP-2 demonstrated greater suppression of BRAF^wt/NRAS^wt/NF1^wt cutaneous and uveal melanomas than mutant melanomas. RNA sequencing analysis of eight melanoma lines with NVP-2 treatment demonstrated that the context of this vulnerability appears to converge on a cell cycle network that includes many transcriptional regulators, such as the E2F family members. The Cancer Genome Atlas human melanoma tumor data further supported a potential oncogenic role for E2F1 and E2F2 in BRAF^wt/NRAS^wt/NF1^wt tumors and a direct link to CDK9. Our results suggest that transcriptional blockade through selective targeting of CDK9 is an effective method of suppressing therapeutically orphaned BRAF/NRAS/NF1 wt melanomas.