The antitumor effect of cyclophosphamide (CPM) was investigated against SCCVII murine tumors and PC-3 human xenografts in combination with the hypoxia-selective cytotoxin 4-[3-(2-nitro-1-imidazolyl)-propylamino]-7-chloroquinoline hydrochloride (NLCQ-1). The in vivo-in vitro and the tumor regrowth assays were used, respectively, as end points. In certain cases the hypoxia-selective cytotoxin tirapazamine (TPZ) was included for comparison purposes. In the SCCVII/C3H model, bone marrow toxicity studies were performed in parallel by using a modified CFU-GM assay. In the SCCVII/C3H model, when NLCQ-1 (10 mg/kg i.p.) was given 1 h before cyclophosphamide (CPM; 75-200 mg/kg i.p.), dose-modification factors (DMFs) of 1.9 and 1.0 were achieved for the antitumor effect and bone marrow toxicity, respectively. The corresponding DMF values obtained with TPZ (23 mg/kg) given 2.5 h (optimal time) before CPM were 1.3 and 1.0, respectively. Thus, therapeutic indices (T.I.) of 1.9 and 1.3 were achieved with NLCQ-1 and TPZ, respectively. In the PC-3/athymic nude mouse model, NLCQ-1 (10 mg/kg) given 90 min before CPM (36 mg/kg), qd x 4, increased tumor regrowth delay by 8.7 days compared to CPM alone, at 16-fold the original tumor size. The corresponding log cell kill was 0.86 and -0.03 for NLCQ-1 + CPM and CPM alone, respectively. In general, NLCQ-1 in combination with nontoxic but inactive CPM doses (36 or 54 mg/kg, qd x 4) elicited good antitumor activity without subsequent additive systemic toxicity, whereas NLCQ-1 had minimal effect in combination with the active but toxic (> 10% mean net weight loss) CPM dose of 80 mg/kg. These results suggest a potential use of NLCQ-1 in the clinic as an adjuvant to chemotherapy with CPM.