Methylation-associated silencing of S100A4 expression in human epidermal cancers.

S100A4 appears important for cancer metastasis and its overexpression is common in a variety of human malignancies, but its status in epidermal cancers remains lesser known. Likewise, E-cadherin downregulation and Wingless (Wnt) activation are frequent cancer-associated alterations, whereas their potential correlations with S100A4 expression in skin lesions have not been characterized. These issues were addressed in the present study using tissue microarray-based immunohistochemical staining, reverse transcriptase polymerase chain reaction and western blotting. Meanwhile, the underlying epigenetic mechanism leading to the altered S100A4 expression in epidermal tumors was elucidated. Immunohistochemistry revealed that S100A4 expression frequencies were 100% (8/8) in normal epidermis, 80.6% (25/31) in tumor-surrounding non-cancerous epidermis, 66.7% (10/15) in premalignant diseases, 8.3% (1/11) in Bowen's disease and 7.7-26.3% in different cancer tissues. The incidence of S100A4 detection in the normal and non-cancerous epidermis was significantly different from that of epidermal cancers (P = 0.000). Accordingly, human immortalized keratinocyte line HaCat but not skin squamous cell carcinoma (SCC) line colo16 was positive in S100A4 expression. S100A4 downregulation, E-cadherin reduction and Wnt activation coexisted in most of epidermal cancers but unnecessarily overlapped. Methylation DNA sequencing revealed methylation of four critical (cytosine and guanine separated by a phosphate or -C-phosphate-G-) CpG sites within S100A4 intron first in S100A4-negative colo16 cells and skin SCCs, and demethylator/5-aza-2'-deoxycytidine treatment efficiently recovered S100A4 expression in colo16 cells. Our findings demonstrate that S100A4 downregulation, as the consequence of DNA methylation, is closely correlated with skin tumor formation. Wnt activation and E-cadherin reduction and S100A4 down-regulation are paralleled molecular events in skin tumors, which may serve as the biomarkers for predicting epidermal cancer risk.