Targeting Drug-Induced Transcriptional Mechanisms to Guide Effective Cancer Therapy

Resistance to cancer therapy seems to be inevitable in many clinical settings. Increasing evidence has suggested that treatment-elicited molecular responses may play a role in some of the inducible mechanisms of resistance associated with cancer pathways, but a large-scale effort to explore this possibility is still limited. Here we present a systems-level pathway analysis of small-molecule perturbational profiles to discover potential transcriptional mechanisms of resistance and thus inform more effective treatment strategies for cancer. Instead of predicting synergistic drug combinations with non-overlapping patterns of gene reversal from our previous study, this framework aimed at identifying pathway-relevant cancer genes that can be transcriptionally inducible and thus targetable in response to specific small-molecule treatments. The perturbed pathways were benchmarked against those derived using a separate panel of small-molecule transcription profiles, and integrated with drug sensitivity and prior knowledge of pathways to predict transcriptionally inducible mechanisms of resistance. Comparison with recent large-scale screening data revealed that this treatment-induced transcriptional mechanism may not be a pervasive route in drug resistance, but has remained largely unexplored. We experimentally validated the predictions for two oncology drugs approved by US FDA in vitro, as further confirmed in a xenograft model to implicate afatinib-induced serine biosynthesis as a potential mechanism of resistance in lung cancer, with evidence supporting the rationale for our proposed framework. These data provide valuable insights into some of the small-molecule-induced transcriptional mechanisms of resistance to improve anticancer therapy.