Journal of Molecular Biomarkers and Clinical Trials

A Five-Gene Signature Inferred from Transcriptome Profiling of Homologous Recombination-Mediated DNA Repair Predicts Clinical Outcome of Patients with Cancer

Published on: 2017-12-17

Abstract

To prevent genomic instability due to DNA double strand breaks (DSBs), cells have evolved to repair DSBs in an error-free manner through the homologous recombination (HR) mechanism. Defective HR repair plays a causative role in the development of multiple human cancers, but defective HR repair can also sensitize cancer cells to the therapeutic effects of radiation therapy as well as to chemotherapy drugs that cause DSBs and importantly to the emerging class of PARP inhibitors (PARPi). As the HR pathway consists of a complex protein network, it has been challenging to predict patient outcomes or response to therapy by assessing HR repair status in tumors by interrogating genetic alterations of individual components in this network. We determined whether transcriptome profiling of HR-defective cells could provide global information on HR status. Based on the common gene expression changes identified in the HR-defective cells, we used statistical methods to identify core genes in the HR repair network that are associated with cancer patient survival. We identified and validated a five-gene signature HR repair network consisting of ADM, EXO1, LRP8, PRC1, and TRIP13 that was associated with overall cancer patient survival. These findings highlight that systems biology approaches can be used to analyze and define the core gene signature of a dysfunctional molecular pathway, which opens a new avenue for developing clinical applicable biomarkers to predict clinical outcomes

Keywords