Cyclooxygenase-2 expression in murine and human nonmelanoma skin cancers: implications for therapeutic approaches

Inflammatory stimuli result in the production of cutaneous eicosanoids, which are known to contribute to the process of tumor promotion. Cyclooxygenase (COX), the rate-limiting enzyme for the production of prostaglandins (PG) from arachidonic acid, exists in at least two isoforms, COX-1 and COX-2. COX-1 is constitutively expressed in most tissues and plays various physiological roles, whereas increased COX-2 expression is known to occur in several types of epithelial neoplasms. Enhanced PG synthesis is a potential contributing factor in UVB-induced nonmelanoma skin cancers (NMSC). Increased COX-2 staining occurs in murine skin neoplasms after chronic exposure to carcinogenic doses of UVB. In this study, immunohistochemical and Western blot analyses were employed to assess longitudinally COX-2 expression in a standard mouse UVB complete carcinogenesis protocol and in human basal cell carcinomas (BCC) and squamous cell carcinomas (SCC). During UVB irradiation of mice, COX-2 expression consistently increased in the hyperplastic skin, the benign papillomas and the SCC. COX-2 expression was also increased in human actinic keratoses, SCC and BCC as well as in murine SCC and BCC. The pattern of COX-2 expression was quite variable, occurring in a patchy distribution in some lesions with staining confined mainly to suprabasal cell layers. In general, COX-2 expression progressively became more extensive in benign papillomas and well-differentiated murine SCC. The staining was predominantly cytoplasmic and perinuclear in some focal areas in tissue stroma around both murine and human tumors. Western blot analysis confirmed negative COX-2 expression in normal skin, whereas acute UVB exposure resulted in increased enzyme expression, which continued to increase in developing papillomas and SCC. Because of the evidence indicating a pathogenic role for eicosanoids in murine and human skin neoplasms, we performed studies to assess the anti-inflammatory and anticarcinogenic effects of green tea extracts, which are potent antioxidants. Acute exposure of the human skin to UVB (minimum erythema dose x 4) caused a transient enhancement of the COX-2 expression, which reverted to baseline within hours; however, in murine skin the expression persisted for several days. Pretreatment with the topically applied green tea extract (1 mg/cm2) largely abrogated the acute COX-2 response to UVB in mice or humans. In summary, enhanced COX-2 expression serves as a marker of epidermal UVB exposure for murine and human NMSC. These results suggest that COX-2 inhibitors could have potent anticarcinogenic effects in UVB-induced skin cancer.