The pathophysiology of hepatic veno-occlusive disease is poorly understood. These studies were undertaken to determine the initial cellular target and the role of glutathione detoxification of dacarbazine, a toxin implicated in hepatic veno-occlusive disease. Sinusoidal endothelial cells (SECs) and hepatocytes were isolated and plated in culture dishes. Dacarbazine (5-(3,3-dimethyl-triazeno) imidazole-4-carboxamide), 3 and 6 mM, was toxic to SECs but not to hepatocytes. Onset of toxicity occurred between 11 and 12 hr as determined by serial MTT assays and ethidium homodimer dye exclusion. Glutathione detoxification of dacarbazine in SECs was suggested by: (1) depletion of glutathione before onset of toxicity; (2) exacerbation of toxicity by buthionine sulfoximine (BSO) depletion of glutathione; and (3) protection by exogenous glutathione. Protection by exogenous glutathione may be by uptake of intact tripeptide rather than by extracellular hydrolysis: neither acivicin (inhibitor of gamma-glutamyltranspeptidase) nor BSO (inhibitor of gamma-glutamylcysteine synthetase) blocked the protective effect, and glutathione disulfide did not protect. The relative resistance to dacarbazine toxicity seen in hepatocytes is not due to more efficient GSH detoxification, because toxicity was not unmasked in hepatocytes cultures in medium lacking sulfur amino acid precursors of GSH. In conclusion, glutathione status may play an important role in the susceptibility to toxicity. Furthermore, the findings suggest that the SEC is the initial in vivo target of dacarbazine due to a relatively higher level of metabolic activation that more readily overcomes the available detoxification.