Purpose
Deregulation of the Myc transcription factor family and its downstream targets plays a crucial role in human cancers. Expression profiling studies have identified prognostic Myc-regulated gene-sets for several cancers, however the generation of a succinct Myc activity signature for risk-stratification in multiple cancer types would facilitate development of novel targeted therapies for Myc-driven cancers.
Experimental Design
18 genes were selected from published Myc-regulated gene lists based on a priori biological knowledge and on analysis of neuroblastoma and epithelial ovarian cancer (EOC) microarray datasets. Signature scores were determined using real-time PCR (qPCR) and Taqman-low-density arrays for a panel of 35 cancer cell lines and 42 primary neuroblastomas. The prognostic value of the 18-gene Myc activity signature was analysed in neuroblastoma, medulloblastoma, diffuse large B-cell lymphoma (DLBCL) and EOC microarray datasets using Kaplan-Meier analysis.
Results
In cancer cell lines, the signature score positively correlated with overall MYC/MYCN/MYCL1 expression. Microarray dataset analysis showed that a high score was predictive of poor outcome independently of MYCN amplification (P<0.001) in neuroblastoma, but not in the overall cohort of EOC. However, a high score conferred a poor prognosis in specific EOC molecular subtypes and in other Myc-driven cancers including medulloblastoma (P=0.012), DLBCL (P=0.019), and breast cancer (P=0.019). The prognostic performance of the signature remained significant in both overall and non-MYCN amplified neuroblastoma, when evaluated with qPCR.
Conclusion
This study identified an 18-gene signature reflecting Myc transcriptional activity and carrying prognostic power across multiple Myc-driven cancers. Use of the signature in different platforms confirmed its robustness, and suggests its clinical applicability for identifying targeted therapies.