The biochemical basis for bleomycin-induced pulmonary toxicity was studied in vitro and in vivo with an improved HPLC system. The in vitro metabolism of bleomycin A2 to desamido-bleomycin A2 was measured in tissue homogenates from a species sensitive (mice) and relatively resistant (rabbits) to the pulmonary fibrogenic properties of bleomycin. Lung tissue from mice lacked detectable bleomycin hydrolase activity, whereas rabbit lung tissue homogenates had high levels of the enzyme activity, equaling that seen in rabbit kidneys and spleen. Injection of radiolabeled bleomycin A2 into mice demonstrated that only a small percentage of the total dose was taken up by any organ and that extensive metabolism of this drug occurred within 1 hr in liver, kidneys, and spleen but not in lungs in vivo. In addition, metabolites other than desamido-bleomycin A2 were prominent, and their relative amounts increased with time. Mice injected subcutaneously with bleomycin A2 developed pulmonary fibrosis, while animals treated with equivalent doses of desamido-bleomycin A2 did not, indicating that this metabolite is not as toxic to the lungs as is the parent compound. These results provide direct evidence that metabolism plays a major role in determining the toxic potential of bleomycin to the lungs.Bleomycin (BLM) is a glycoprotein that is extensively used in combination with other anticancer agents because of its relative lack of hematological and gastrointestinal toxicity. However, pulmonary toxicity is common with BLM and limits its therapeutic utility (1). Life-threatening, irreversible pulmonary fibrosis is associated with high cumulative doses of BLM; thus, it is important to understand the biochemical basis of its lung toxicity. Furthermore, BLM may be a useful prototype for studying the general mechanism by which other substances, including anticancer drugs, cause pulmonary toxicity.Cellular toxicity from BLM exposure is believed to result from interactions between the drug and DNA (2 Previous experiments (4-7) indicated that repeated injections of mice with 5-10 mg of BLM per kg of body weight for 4 weeks or more produced pulmonary fibrosis as detected by an increase in lung hydroxyproline content and by histological analyses, whereas a similar injection protocol failed to yield fibrosis in rabbits (8). The experiments reported here examined the relative BLM hydrolase activities in various organs of these two species to evaluate the hypothesis of Umezawa et al. (3). A modification of our recently described sensitive HPLC system for detection of BLM B2 and desamido-BLM B2 (9) was used. In addition, the metabolism of radiolabeled BLM A2 in organs of mice in vivo and the pulmonary toxicity of the only previously described metabolite of BLM A2, desamido-BLM A2 (BLM dA2), were determined.MATERIALS AND METHODS Drugs. BLM (Blenoxane) and Cu-free BLM A2 that were used as HPLC standards were obtained from W. T. Bradner (Bristol Laboratories, Syracuse, NY). The Cu-free metabolite of BLM A2, BLM dA2 (lot TN-A-029), that was us...