Glutathione S-
transferases (
GSTs) are
enzymes involved in the detoxification of
xenobiotics and are divided into four subclasses, alpha, mu, pi, and theta, with different although overlapping substrate specificities. Most human gastrointestinal
tumors contain increased amounts of GST-pi and GST
enzyme activity. The relationship between GST parameters and
tumor and patient characteristics, including overall survival, were studied retrospectively in 100 primary colorectal
adenocarcinomas. Levels of
GST-alpha, GST-mu, GST-pi, and GST
enzyme activity were not related to the Dukes stage, differentiation grade, localization, histological type and diameter of the
tumor, or gender and age of the patient. Fifty-seven patients died (median survival, 21 months; range, 1-65 months) during follow-up, and 43 patients were still alive at the closing date of the study (median follow-up, 68 months; range, 60-87 months). Optimal dichotomization and uni- and multivariate analyses were done with the Cox proportional hazard model. Multivariate analysis with all clinicopathological parameters revealed higher Dukes stage (hazard ratio, 2.7; P < 0.001) and older age (hazard ratio, 2.8; P = 0.001) to be the only independent prognostic variables for overall survival. In contrast to
GST-alpha and GST-mu, high levels of GST-pi (hazard ratio, 3.1; P = 0.002) and GST
enzyme activity (hazard ratio, 2.0; P = 0.020) in the
tumors were found to have a significant prognostic value independent from the clinicopathological parameters when added separately to this Cox model. Thus, this study indicates that GST subclass levels in colorectal
adenocarcinomas are not related to clinicopathological parameters and that the GST-pi level and GST
enzyme activity have a prognostic value for the overall survival of the patients.