We have characterized the acquisition of resistance to cisplatin (DDP) in an in vivo solid tumor model. Human ovarian carcinoma cells (2008 cell line) were grown s.c. as xenografts in athymic mice. Tumors were selected with 3.0 mg/kg of DDP i.p. given once per week for four weeks. One week after the last dose of DDP, cell lines were generated from the tumors. Cells from these lines were then re-inoculated into athymic mice, and the DDP selection process was repeated. This procedure was continued for a total of four passages (16 doses). Although this chemotherapy did not affect the tumors' growth, cell lines derived from these tumors after the first passage displayed low-level resistance to DDP as determined by the concentration causing 50% inhibition of colony formation in a clonogenic assay. The mean resistance (+/- SD) of cell lines derived from tumors treated with four doses of DDP was 1.6 +/- 0.06 (n = 5). A minimum of only two doses of DDP was required to generate significant resistance (1.5 fold). The DDP resistance slowly increased with further selections so that after four passages with chemotherapy, the cells were 3.0-fold resistant. The DDP resistance was not stable; after three passages, resistance slowly declined over 104 days from 2.2- to 1.4-fold. Resistant cells obtained after both three and four passages did not have elevated glutathione as determined by flow cytometry of monochlorobimane-stained cells. After three passages, DDP-resistant cells were not resistant to CdCl2, suggesting that metallothioneins were also not elevated. A key biochemical change found in these cells was a decrease in DDP accumulation.(ABSTRACT TRUNCATED AT 250 WORDS)