Exposure to genotoxins may compromise DNA integrity in male reproductive cells, putting future progeny at risk for developmental defects and diseases. To study the usefulness of sperm DNA damage as a biomarker for genotoxic exposure, we have investigated cellular and molecular changes induced by benzo[a]pyrene (B[a]P) in human sperm in vitro, and results have been compared for smokers and non-smokers. Sperm DNA obtained from five smokers was indeed more fragmented than sperm of six non-smokers (mean % Tail DNA 26.5 and 48.8, respectively), as assessed by the alkaline comet assay (P < 0.05). B[a]P-related DNA adducts were detected at increased levels in smokers as determined by immunostaining. Direct exposure of mature sperm cells to B[a]P (10 or 25 muM) caused moderate increases in DNA fragmentation which was independent of addition of human liver S9 mix for enzymatic activation of B[a]P, suggesting some unknown metabolism of B[a]P in ejaculates. In vitro exposure of samples to various doses of B[a]P (with or without S9) did not reveal any significant differences in sensitivity to DNA fragmentation between smokers and non-smokers. Incubations with the proximate metabolite benzo[a]pyrene-r-7,t-8-dihydrodiol-t9,10-epoxide (BPDE) produced DNA fragmentation in a dose-dependent manner (20 or 50 muM), but only when formamidopyrimidine DNA glycosylase treatment was included in the comet assay. These levels of DNA fragmentation were, however, low in relation to very high amounts of BPDE-DNA adducts as measured with (32)P postlabelling. We conclude that sperm DNA damage may be useful as a biomarker of direct exposure of sperm using the comet assay adapted to sperm, and as such the method may be applicable to cohort studies. Although the sensitivity is relatively low, DNA damage induced in earlier stages of spermatogenesis may be detected with higher efficiencies.