The bioreductive anticancer
prodrug CI-1010 ((2R)-1-[(2-bromoethyl)amino]-3-(2-nitro-1H-imidazol-1-yl)-
2-propanol hydrobromide) is an alkylating
nitroimidazole which shows selective toxicity against hypoxic cells in murine
tumors, but causes extensive apoptosis in the outer retina in rodents and monkeys. This irreversible
retinal toxicity has terminated preclinical development of CI-1010. We have investigated whether such toxicity is due to physiological
hypoxia in the retina, and whether it is a general feature of
hypoxia-selective bioreductive drugs.
Retinal damage was quantified by morphometric analysis of histological sections following treatment of female C57Bl6 mice. Both CI-1010 and
tirapazamine (TPZ, 1,2,4-benzotriazin-3-amine 1,4-dioxide), a bioreductive
drug in Phase III clinical trial, caused a time and dose-dependent loss of photoreceptor cells of the outer retina following administration of single intraperitoneal doses. The lesion caused by TPZ was qualitatively similar to that with CI-1010, but was less severe at equivalent fractions of the maximum tolerated dose (as defined by lethality). With both bioreductive drugs, lesion severity was increased if animals breathed 10% O(2) for 3 h after
drug administration, while breathing 95% O(2)/5% CO(2) was protective. Other
hypoxia-selective bioreductive drugs tested (the
quinone porfiromycin, the
anthraquinone N-
oxide AQ4N and the
nitrogen mustard prodrugs SN 23816 and SN 25341) did not cause
retinal damage at their maximum tolerated doses. This study suggests that the
retinal toxicity of bioreductive drugs might be avoided by manipulation of tissue
hypoxia using 95% O(2)/5% CO(2), although this intervention could suppress antitumor activity. The finding that not all bioreductive drugs cause
retinal toxicity suggests this toxicity can be avoided through appropriate
drug design.