A core NRF2 gene set defined through comprehensive transcriptomic analysis predicts selective drug resistance and poor multi-cancer prognosis.

Abstract

The NRF2-KEAP1 pathway plays an important role in the cellular response to oxidative stress but may also contribute to metabolic changes and drug resistance in cancer. However, despite its pervasiveness and important role, most of NRF2 target genes are defined in context specific experiments and analysis, making it difficult to translate from one situation to another. Our study investigates whether a core NRF2 gene signature can be derived and used to represent NRF2 in various contexts, allowing better reproducibility and understanding of NRF2 activation. We define a core set of 14 upregulated NRF2 target genes from seven RNA-sequencing datasets that we generated and analyzed. This NRF2 gene signature was validated using analyses of published datasets and gene sets. An NRF2 activity score based on expression of these core target genes correlates with resistance to drugs such as PX-12 and necrosulfonamide but not to paclitaxel or bardoxolone methyl. We validated these findings in our KEAP1 knockout cancer cell lines. Finally, our NRF2 score is prognostic for cancer survival and validated in additional independent cohorts for lung adenocarcinoma and also novel cancers types not associated with NRF2-KEAP1 mutations such as clear cell renal carcinoma, hepatocellular carcinoma, and acute myeloid leukemia. These analyses define a core NRF2 gene signature that is robust, versatile, and useful for evaluating NRF2 activity and for predicting drug resistance and cancer prognosis. Using this gene signature, we uncover novel selective drug resistance and cancer prognosis associated with NRF2 activation.