Mutagenicity in drug development: Interpretation and significance of test results

Abstract

The use of mutagenicity data has been proposed and widely accepted as a relatively fast and inexpensive means of predicting long-term risk to man (i.e., cancer in somatic cells, heritable mutations in germ cells). This view is based on the universal nature of the genetic material, the somatic mutation model of carcinogenesis, and a number of studies showing correlations between mutagenicity and carcinogenicity. An uncritical acceptance of this approach by some regulatory and industrial concerns is over-conservative, naive, and scientifically unjustifiable on a number of grounds: (1) Human cancers are largely life-style related (e.g., cigarettes, diet, tanning). (2) Mutagens (both natural and man-made) are far more prevalent in the environment than was originally assumed (e.g., the natural bases and nucleosides, protein pyrolysates, fluorescent lights, typewriter ribbon, red wine, diesel fuel exhausts, viruses, our own leukocytes). (3) “False-positive” (relative to carcinogenicity) and “false-negative” mutagenicity results occur, often with rational explanations (e.g., high threshold, inappropriate metabolism, inadequate genetic endpoint), and thereby confound any straightforward interpretation of mutagenicity test rsults. (4) Test battery composition affects both the proper identification of mutagens and, in many instances, the ability to make preliminary risk assessments. (5) In vitro mutagenicity assays ignore whole animal protective mechanisms, may provide unphysiological metabolism, and may be either too sensitive (e.g., testing at orders-of-magnitude higher doses than can be ingested) or not sensitive enough (e.g., short-term treatments inadequately model chronic exposure in bioassay). (6) Bacterial systems, particularly the Ames assay, cannot in principle detect chromosomal events which are involved in both carcinogenesis and germ line mutations in man. (7) Some compounds induce only chromosomal events and little or no detectable single-gene events (e.g., acyclovir, caffeine, methapyrilene). (8) In vivo mutagenicity assays are more physiological but appear to be relatively insensitive due to the inability to achieve sufficiently high acute plasma levels to mimic cumulative long-term effects. (9) Examination of the mutagenicity of naturally occurring analogs may indicate the irrelevance of a test compound's mutagenicity (e.g., deoxyguanosine, and the structurally related antiviral drug, acyclovir, have identical mutagenicity patterns). (10) Life-threatening or severe debilitating diseases (e.g., cancer, severe psychoses, severe crippling arthritis, sight-threatening diseases) may justify treatment with mutagenic or even carcinogenic therapeutic agents (benefit/risk considerations). (11) Some disease organisms are themselves mutagenic and/or carcinogenic, and treatment with a mutagen may both pose a lesser risk and confer therapeutic value (relative risk considerations: e.g., acyclovir treatment of genital herpes). The discussion of these 11 points, which represent our approach to assessing mutagenic risk of therapeutic compounds, is illustrated with specific examples.