Carzelesin (U-80244), one of the synthetic DNA minor groove binding cyclopropylpyrroloindole analogues, was selected for clinical development because of its high potency, promising antitumor activity in murine solid tumors and leukemia, and significant therapeutic efficacy against colon and rhabdomyosarcoma xenografts. In this Phase I study, carzelesin was given daily for 5 consecutive days to (a) determine the maximum tolerable dose (MTD) and the pattern of toxicity of this schedule; (b) define the pharmacokinetic profile of the parent, as was done for the intermediate compound U-76073 and the DNA-reactive agent U-76074; and (c) document any antitumor activity observed. Carzelesin was given as a 10-min infusion with a constant-rate infusion pump. Treatment was repeated every 4 weeks or when blood counts had recovered to normal values. The starting dose of 12 microgram/m2/day was escalated by 20-30% increments until the MTD (defined as the dose leading to grade 4 hematological or grade 3 nonhematological toxicity in at least two of six patients) was reached. Pharmacokinetic studies were planned on days 1 and 5 of the first cycle in at least two patients per dose level. Plasma levels of carzelesin, U-76073, and U-76074 were determined by high-performance liquid chromatography with UV detection and a detection limit of 0.5 ng/ml. Twenty-five patients were entered in the study, and 56 cycles were evaluable for hematological toxicity. Subsequent dose levels evaluated were 24, 30, 35, and 40 microgram/m2. Both neutropenia and thrombocytopenia were dose limiting and cumulative, with a high interpatient variability. Neutropenia occurred earlier (median time to neutrophil nadir and recovery, 15 and 29 days, respectively) than thrombocytopenia (median time to platelet nadir and recovery, 25 and >/=26 days, respectively); there were delays of treatment because of persisting thrombocytopenia in all patients treated at the MTD. At the MTD, the peak plasma concentrations of carzelesin were achieved at the end of the infusion and were higher than those found cytotoxic in vitro against tumor cell lines. Carzelesin was detectable up to a maximum of 1 h after the infusion. Smaller amounts of U-76073 were detectable for a maximum of 30 min only at the MTD, whereas U-76074 was never found. An 8-month partial remission was reported in one previously untreated patient with hepatocellular carcinoma at 40 microgram/m2. The MTD was fixed at 40 microgram/m2 daily; 35 and 30 microgram/m2 are the daily doses recommended for Phase II studies in good- and poor-risk patients. The daily regimen for 5 days seems to offer no advantage over the single intermittent schedule that has been selected for the Phase II program in Europe.