Nitracrine N-oxides: effects of variations in the nature of the side chain N-oxide on hypoxia-selective cytotoxicity.

Abstract

The tertiary amine N-oxide (nitracrine-N-oxide, 1b) of the 1-nitroacridine nitracrine is a bis-bioreductive agent showing very high hypoxic selectivity (approximately 1000-fold) against tumour cells in culture, but only modest activity against the hypoxic subfraction of tumours in vivo. Because the hypoxic selectivity of 1b was considered to depend significantly on the rate of enzyme-mediated reduction of the N-oxide group, this paper reports the preparation and evaluation of a series of analogues in which the environment of this group was modified. Three analogues contained more weakly basic N-oxides, while two others had varying degrees of steric bulk around the N-oxide. In all but one case (an aromatic N-oxide), the N-oxides were much less cytotoxic (10- to 300-fold) than the corresponding tertiary amines towards AA8 Chinese hamster cells under aerobic conditions. Both the N-oxides and the corresponding amines were more cytotoxic to an ERCC-1 mutant defective in nucleotide excision repair, indicating that DNA alkylation was the cytotoxic event. However, there was no apparent correlation of these parameters with structure. All of the aliphatic N-oxides, with the exception of the aromatic N-oxide example, showed substantial (70- to 800-fold) hypoxic selectivity against AA8 cells in a clonogenic assay. While the weakly basic derivatives were the least selective, there was no apparent relationship between hypoxic selectivity and the steric environment of the N-oxide. Selectivity for hypoxic cells in culture is shown to depend on the hypoxic selectivity of the corresponding tertiary amine (reflecting O2-inhibitable reduction of the 1-nitro group) and the differential in aerobic toxicity between amine and N-oxide (a measure of the potential toxicity increase achievable by reducing the N-oxide). Four analogues whose structures fairly represented the range of steric and electronic modifications of the N-oxide site were evaluated against the hypoxic subfraction of cells in KHT tumours in vivo, but were inactive. These results suggest that either such modifications do not exert significant effects on N-oxide reduction, or that the rate of such reduction is not a factor limiting the in vivo activity of the parent analogue 1b.