Adozelesin,
bizelesin and
carzelesin are synthetic cyclopropylpyrroloindole (
CPI) analogs, a class of potent
antineoplastic agents modeled on the antitumor
antibiotic CC-1065, that specifically bind to the minor groove of
DNA and preferentially alkylate AT-rich regions. These compounds were evaluated against fresh human
tumors in a human
tumor colony-forming assay (HTCFA) to assess and to compare their relative antitumor spectra, concentration-response relationships and schedule-dependence. Human
tumor colony-forming units were treated with
adozelesin and
bizelesin at concentrations of 0.02, 0.1 and 0.5 ng/ml as a continuous exposure for 14 days, and to 0.2, 1.0 and 5.0 ng/ml as a 1 h exposure.
Carzelesin concentrations were 0.04, 0.2 and 1 ng/ml as a continuous exposure, and 0.6, 3.0 and 15.0 ng/ml as a 1 h exposure. A response was scored if there was 50% or less colony survival. The three analogs had similar antitumor activity against colon
carcinoma, kidney
carcinoma and
melanoma colony-forming units.
Adozelesin also displayed activity against both breast and
non-small cell lung carcinoma colony-forming units, and
carzelesin was active against ovarian
carcinoma colony-forming units. Significantly positive concentration-response relationships were apparent with all three agents. Responses increased from below 15% at the lowest concentration to above 45% at the highest concentration for the three drugs on all schedules (p < 0.01). At the highest concentration, the overall response rate was significantly higher (p < 0.01) with
carzelesin on the continuous schedule (71%) compared to the 1 h schedule (46%). However, overall response rates for
adozelesin and
bizelesin were similar on both schedules (1 h/continuous:
adozelesin, 67/58%;
bizelesin, 49/44%), indicating that
adozelesin and
bizelesin are less schedule dependent than
carzelesin in the HTCFA. These results demonstrate that the CPIs have broad-spectrum activity against human
tumor colony-forming units in the HTCFA at very low concentrations, as well as differences with regard to schedule dependence which may help guide the optimal clinical development of these agents.