Abstract
As the options for systemic treatment of malignant melanoma (MM) increase, the need to develop biomarkers to identify patients who might benefit from cytotoxic chemotherapy becomes more apparent. In preclinical models, oxaliplatin has activity in cisplatin-resistant cells. In this study, we have shown that oxaliplatin forms interstrand crosslinks (ICLs) in cellular DNA and that loss of the heterodimeric structure-specific endonuclease XPF-ERCC1 causes hypersensitivity to oxaliplatin in mammalian cells. XPF deficiency resulted in late S-phase arrest and persistence of double-strand breaks following oxaliplatin treatment. In a panel of 12 MM cell lines, oxaliplatin sensitivity correlated with XPF and ERCC1 protein levels. The knockdown of ERCC1 and XPF protein levels by RNA interference increased sensitivity of cancer cells to oxaliplatin; overexpression of exogenous ERCC1 significantly decreased drug sensitivity. Following immunohistochemical optimization, XPF protein levels were quantified in MM tissue samples from 183 patients, showing variation in expression and no correlation with prognosis. In 57 patients with MM treated with cisplatin or carboplatin, XPF protein levels did not predict the likelihood of clinical response. We propose that oxaliplatin should not be discarded as a potential treatment for MM on the basis of the limited activity of cisplatin in unselected patients. Moreover, we show that XPF-ERCC1 protein levels are a key determinant of the sensitivity of melanoma cells to oxaliplatin in vitro. Immunohistochemical detection of XPF appears suitable for development as a tissue biomarker for potentially selecting patients for oxaliplatin treatment in a prospective clinical trial.