Several recent studies have examined the possibility of producing tumor-specific cytotoxicity with various enzyme/ prodrug combinations. The enzymes are targeted to tumor cells either with antibodies (ADEPT, antibody directed enzyme prodrug therapy) or with viruses (VDEPT). The goal of the present study was to identify an appropriate enzyme for use in activating the prodrug 7-ethyl-10-[4-(1-piper-idino)-1-piperidino]carbonyloxycamptothe cin (CPT-11). In this study, we compared the efficiency of CPT-11 metabolism by rabbit and human carboxylesterases in in vitro and in situ assays. Although the rabbit and human enzymes are very similar (81% identical; 86% homologous) and the active site amino acids are 100% identical, the rabbit enzyme was 100-1000-fold more efficient at converting CPT-11 to SN-38 in vitro and was 12-55-fold more efficient in sensitizing transfected cells to CPT-11. In vivo, Rh30 rhabdomyosarcoma cells expressing the rabbit carboxylesterase and grown as xenografts in immune-deprived mice were also more sensitive to CPT-11 than were control xenografts or xenografts expressing the human enzyme. Each of the three types of xenografts regressed when the mice were treated with CPT-11 given i.v. at 2.5 mg of CPT-11/kg/daily for 5 days/week for 2 weeks [(dx5)2] (one cycle of therapy), repeated every 21 days for a total of three cycles. However, following cessation of treatment, recurrent tumors were detected in seven of seven mice bearing control Rh30 xenografts and in two of seven mice bearing Rh30 xenografts that expressed the human enzyme. No tumors recurred in mice bearing xenografts that expressed the rabbit carboxylesterase. We conclude that rabbit carboxylesterase/CPT-11 may be a useful enzyme/prodrug combination.