A Myc Activity Signature Predicts Poor Clinical Outcomes in Myc-Associated Cancers.

Abstract

Myc transcriptional activity is frequently deregulated in human cancers, but a Myc-driven gene signature with prognostic ability across multiple tumor types remains lacking. Here, we selected 18 Myc-regulated genes from published studies of Myc family targets in epithelial ovarian cancer (EOC) and neuroblastoma. A Myc family activity score derived from the 18 genes was correlated to <i>MYC</i>/<i>MYCN</i>/<i>MYCL1</i> expression in a panel of 35 cancer cell lines. The prognostic ability of this signature was evaluated in neuroblastoma, medulloblastoma, diffuse large B-cell lymphoma (DLBCL), and EOC microarray gene expression datasets using Kaplan-Meier and multivariate Cox regression analyses and was further validated in 42 primary neuroblastomas using qPCR. Cell lines with high <i>MYC, MYCN</i>, and/or <i>MYCL1</i> gene expression exhibited elevated expression of the signature genes. Survival analysis showed that the signature was associated with poor outcome independently of well-defined prognostic factors in neuroblastoma, breast cancer, DLBCL, and medulloblastoma. In EOC, the 18-gene Myc activity signature was capable of identifying a group of patients with poor prognosis in a "high-<i>MYCN</i>" molecular subtype but not in the overall cohort. The predictive ability of this signature was reproduced using qPCR analysis of an independent cohort of neuroblastomas, including a subset of tumors without <i>MYCN</i> amplification. These data reveal an 18-gene Myc activity signature that is highly predictive of poor prognosis in diverse Myc-associated malignancies and suggest its potential clinical application in the identification of Myc-driven tumors that might respond to Myc-targeted therapies. <i>Cancer Res; 77(4); 971-81. ©2016 AACR</i>.