Identification of predictors of drug sensitivity using patient-derived models of esophageal squamous cell carcinoma.

Previous studies from the Cancer Cell Line Encyclopedia (CCLE) project have adopted commercial pan-cancer cell line models to identify drug sensitivity biomarkers. However, drug sensitivity biomarkers in esophageal squamous cell carcinoma (ESCC) have not been widely explored. Here, eight patient-derived cell lines (PDCs) are successfully established from 123 patients with ESCC. The mutation profiling of PDCs can partially recapture the tumor tissue actionable mutations from 161 patients with ESCC. Based on these mutations and relative pathways in eight PDCs, 46 targeted drugs are selected for screening. Interestingly, some drug and biomarker relationships are established that were not discovered in the CCLE project. For example, CDKN2A or CDKN2B loss is significantly associated with the sensitivity of CDK4/6 inhibitors. Furthermore, both PDC xenografts and patient-derived xenografts confirm CDKN2A/2B loss as a biomarker predictive of CDK4/6 inhibitor sensitivity. Collectively, patient-derived models could predict targeted drug sensitivity associated with actionable mutations in ESCC.