Combining factors on protein and gene level to predict radioresponse in head and neck cancer cell lines

Background: Radiotherapy is the main therapy for head and neck squamous cell carcinoma (HNSCC); however, treatment resistance and local recurrence are significant problems, highlighting the need for predictive markers. In this study, we evaluated selected proteins, mutations, and single nucleotide polymorphisms (SNPs) involved in apoptosis, cell proliferation, and DNA repair alone or combined as predictive markers for radioresponse in 42 HNSCC cell lines.

Methods: The expression of epidermal growth factor receptor, survivin, Bax, Bcl‐2, Bcl‐X_L, cyclooxygenase‐2 (COX‐2), and heat shock protein 70 was analyzed by ELISA. Furthermore, mutations and SNPs in the p53 gene as well as SNPs in the MDM2, XRCC1, and XRCC3 genes were analyzed for their relation to radioresponse. To enable the evaluation of the predictive value of several factors combined, each cell line was allocated points based on the number of negative points (NNP) system, and the NNP sum was correlated with radioresponse.

Results: Survivin was the only factor that alone was significantly correlated with the intrinsic radiosensitivity (IR; r = 0.36, P = 0.02). The combination of survivin, Bax, Bcl‐2, Bcl‐X_L, COX‐2, and the p53 Arg72Pro polymorphism was found to most strongly correlate with radioresponse (r = 0.553, P < 0.001).

Conclusion: These data indicate that the IR of 42 HNSCC cell lines can be predicted by a panel of factors on both the protein and gene levels. Moreover, among the investigated factors, survivin was the most promising biomarker of radioresponse.