PhD Student National University of Singapore Singapore, Singapore
Abstract: Radiotherapy (RT) is a primary treatment modality for most head and neck cancers (HNC). However, locoregional recurrence post-RT remains a pertinent clinical problem, given that salvage therapeutic options are often limited and ineffective, partly owing to the divergent tumor and microenvironmental factors underpinning radioresistance. Here, we investigated the utility of a high-throughput drug screening method – Quadratic Phenotypic Optimization Drug Screening Platform (QPOP) for the discovery of novel drug combinations in radioresistant (RR)-HNC models. In vitro isogenic models of RR-HNC were generated by exposing two wild-type (WT) HNC cell lines, FaDu and HK-1, to 90 Gy of 2 Gy irradiation (IR) (45 fr x 2 Gy/fr), and validated using clonogenic assays. QPOP is a complex system analytics platform that rapidly and efficiently ranks and compares all possible system-specific drug combinations from a large drug search set. Drug candidates include standard and targeted agents specific to HNC. Comparative QPOP analysis revealed that co-treatment of Panobinostat (HDAC inhibitor, Pano) and AZD7762 (CHK1 inhibitor) exhibited the most distinct and contrasting response between WT and RR cells. This combination resulted in synergistic growth inhibition, accompanied by enhanced DNA damage and reduced DNA repair specific to the RR cells. We interrogated if Pano can re-sensitize RR cells to ionizing radiation (IR). Interestingly, exposure to Pano before 4 Gy IR led to radiosensitization of RR cells, but not WT cells. We further showed that Pano potentiated the effects of IR and delayed tumor growth in an in vivo RR-HNC model, otherwise resistant to RT. Given that Pano is a pan-HDAC inhibitor, we performed genetic and chemical inhibition of major Class I and II HDACs and observed that HDAC6 inhibition not only phenocopies, but enhances the re-sensitization effects of Pano treatment. Mechanistically, we identified SP1 transcription factor as an intermediate element between HDAC6 activity and RAD51 promoter inhibition. We further proposed a specific epigenetic mechanism underlying RR pathogenesis and reversal. Finally, we identified an RR-associated, HDAC6i-sensitive biomarker panel and verified it in an independent cohort. Overall, our study underscores the potential of HDACi, in combination with CHK1 blockade or re-IR, for targeting RR-HNC, and demonstrates the utility of QPOP as a valuable translational tool for future biomarker-led clinical development.
