O(6)-Alkylguanine-DNA alkyltransferase (AGT) repairs O(6)-alkylating DNA adducts generated by alkylating therapeutic agents. Therefore, AGT activity may be an important marker of tumor and normal tissue sensitivity to chemotherapeutic agents and a predictor for the success of chemotherapeutic regimens. It is rapidly inactivated by O(6)-benzylguanine (BG) that mimics its substrates, O(6)-methylguanine and O(6)-chloroethylguanine DNA adducts. In a Phase I clinical trial, BG was given in increasing doses (from 10 to 120 mg/m(2)) by 1-h infusion. We previously reported depletion of AGT activity, and in this report, we demonstrate the relationship between degradation of BG-inactivated AGT protein and the depletion of AGT activity in peripheral blood mononuclear cells (PBMCs) and tumor samples obtained by computed tomography-guided cutting needle biopsy from patients prior to BG and either 2 or 18 h after BG. In PBMCs, BG inactivated AGT activity by over 95-100% at the end of a 1-h infusion, and depletion was maintained for 18 h. In contrast, AGT protein remained almost unchanged for up to 18 h after BG, suggesting that inactivated AGT proteins remain immunoreactive and are not rapidly degraded in PBMCs. In patient tumor biopsies, AGT activity was depleted approximately 90% 2 h after BG. Tumor AGT protein levels were reduced to approximately 40% of pretreatment values when detected by either Western blot or immunohistochemistry staining. In tumor samples obtained 18 h after BG, >95% inactivation of tumor AGT activity was observed at BG doses of 36-80 mg/m(2), and complete depletion of tumor AGT activity occurred at 120 mg/m(2) BG. However, residual AGT protein (5-10% of baseline) was detectable in all tumor samples. Therefore, the degradation of BG-inactivated AGT protein appeared to be much more rapid in tumors than that in PBMCs, which may impact on AGT regeneration rates as well. Because degradation of BG-inactivated AGT takes place slowly, antibody-based measurements of AGT protein correlate poorly with depletion of AGT activity immediately after BG. Thus, biochemical activity measurements remain the appropriate monitor of AGT during therapeutic modulation. These data provide the first and conclusive evidence of differential degradation rates of inactivated AGT in PBMCs and tumors of patients after treatment with BG and suggest that immunoreactive AGT measurements in PBMCs are a poor surrogate for AGT activity in tumor tissue.