S100-RAGE signaling in the pathogenesis of Head and Neck Squamous Cell Carcinoma

An aberrant expression of RAGE (receptor for advanced glycation end products) and its ligands, especially the S100-calgranulins has been demonstrated in squamous cell carcinoma of the upper aerodigestive tract. However, while S100-RAGE signaling is commonly linked to the induction and maintenance of a cancer promoting inflammation, the question whether RAGE-signaling is causally linked with neoplastic transformation of keratinocytes in mucosal epithelia has not been addressed so far. The presence of S100-calgranulin positive infiltrating immune cells as well as S100-calgranulin expression in tumor cells was assessed in oropharyngeal squamous cell carcinoma (OPSCC) on tissue microarrays (TMAs) containing tumor biopsies from 188 human patients and compared with the amount of CD66b positive myeloid inflammatory cells on the same TMAs. To address the causal role of S100-RAGE signaling in the onset of oral carcinogenesis, the well established mouse model of 4-nitroquinoline-1-oxide (4-NQO) induced carcinogenesis was used to investigate tumor development in control and RAGE deficient (Rage-/-) mice as well as mice deficient in S100a9 (S100a9-/-). While patient tumors varied strongly with regard to the amount of S100-calgranulin and CD66b positive immune cells as well as the expression pattern of S100-calgranulins in tumor cells, these features did not correspond with clinico-pathological parameters or prognosis. In the onset of oral and esophageal cancer, driven by 4-NQO induced genotoxic stress and in the absence of an additional inflammatory stimulus, both Rage and S100a9 expression was dispensable for tumor development. In both cohorts, mice developed tumors in the esophagus and tongue with similar incidence rates and comparable multiplicity. Also a detailed analysis of tumor sections by histological and immunohistochemical staining revealed no difference in size or histological architecture of 4-NQO induced lesions, tumor cell proliferation and the number of inflammatory immune cells in the tumor microenvironment. S100a8 and S100a9 were induced upon 4-NQO treatment independent of the presence of RAGE, which may in part be explained by induced transcript and protein levels of the Toll-like receptor 4 (Tlr4) in carcinogen treated tissue, suggesting that signaling via the S100-TLR4 axis may compensate for the lack of RAGE in early stages of tumor development. In summary, these data point out that the impact of S100-RAGE signaling is critically depending on the context. While important in inflammation associated cancer, S100-RAGE signaling is dispensable in cancer caused by genotoxic stress without a promoting inflammation. With regard to therapy and prevention, this illustrates the need of a clear stratification for the presence of a driving inflammation