There is considerable concern about an enhanced risk of lung tumor development upon exposure of humans to polycyclic aromatic hydrocarbons (PAHs), like benzo[a] pyrene (B[a]P), in combination with induced lung cell proliferation by toxic agents like ozone. We studied this issue in wild-type (WT) C57BL/6 mice, the cancer prone nucleotide excision repair-deficient Xeroderma pigmentosum complementation group A mice (Xpa-/-) and the even more sensitive Xpa-/-/p53+/- mice. The mice were treated with B[a]P through the diet at a dose of 75 p.p.m., in combination with intermittent ozone exposures (0.8 p.p.m.). First, a dose-range finding study with WT and Xpa-/- mice was conducted to determine the optimal ozone concentration giving high cell proliferation and low toxic side effects. We show by BrdU incorporation that cell proliferation in the lung was induced by ozone, with an optimal concentration of 0.8 p.p.m., which was subsequently used in the (sub)chronic studies. In the subchronic study, in which lacZ mutant frequency and BPDE-DNA adduct formation were measured, the mice were treated for 13 weeks with B[a]P and/or ozone, whereas in the chronic study this treatment protocol was followed by a 6 month period on control feed and filtered air. As expected, oral B[a]P exposure appeared to be highly carcinogenic to Xpa-/- and Xpa-/-/p53+/- mice and to a lesser extent to WT mice. A high incidence of forestomach tumors and some tumors of the esophagus were found. In the lung, a clear genotoxic effect of B[a]P was found as shown by the presence of BPDE-DNA adducts. However, these DNA adducts in combination with induction of cell proliferation did not result in increased lacZ mutations, nor in lung tumor formation not even in the highly sensitive Xpa-/- and Xpa-/-/p53+/- mice. The implication of these findings for tumor risk assessment will be discussed.