Background: Glutathione S-transferases (GST) are known to play a role in the detoxification of carcinogens. Individual isoenzymes of the alpha-, mu-, and pi-class vary in substrate specificities, tissue distribution, and activities among individuals. GST-pi expression has been shown to be increased in preneoplastic and neoplastic lesions. GST-mu is known to play a role in detoxification of epoxides released from cigarette smoke, and individuals with low GST-mu activity have a relatively high risk to develop smoking-related lung and laryngeal cancer. The occurrence of a second primary tumor (SPT) in the whole respiratory and upper aerodigestive tract is an important factor for mortality in head and neck squamous cell carcinoma (HNSCC), and, at present, there are no markers that are available to predict which patient has increased chances of developing an SPT. Risk-assessment by use of biomarkers, particularly the ones that can be obtained with noninvasive techniques, are of great value in predicting prognosis and hence possibly more aggressive treatment and follow-up in selected patient groups.
Experimental design: In a nested case control study, 20 patients who had previous history of oral cancer were used; 10 of the 20 had developed an SPT, and the other 10 patients were minimally 7 years free of disease. The expression of GST-pi, GST-mu, and GST-alpha was immunohistochemically analyzed using apparently normal oral mucosa, free of tumor or dysplasia, obtained from the resection edges around the primary tumor. In another experiment, the three GST isoenzymes were immunohistochemically analyzed using exfoliated cells, obtained noninvasively from several sites of the upper aerodigestive tract of the apparently normal-looking mucosa of HNSCC patients (n = 25) and of control individuals (n = 10).
Results: The expression of all GST was significantly higher (p < 0.001) in the suprabasal and superficial layers of the mucosa at risk. Also, in cell scrapes of clinically healthy mucosa of HNSCC patients, we observed a significantly higher expression (p < 0.001) of GST-pi and GST-mu compared with their matched controls. For GST-alpha, we observed a more heterogenous expression pattern in these exfoliated cells.
Conclusions: Expression of GST-pi, -mu, and -alpha in normal tissue in the direct vicinity of the first tumor seems to have predictive value for the development of an SPT.