Levels of aromatic DNA adducts in foundry workers and controls were followed at four annual samplings. During this time exposure to polycyclic aromatic hydrocarbons (PAH) decreased and the level of DNA adducts decreased accordingly. In the total group exposure was related to the level of adducts. Adduct levels correlated with urinary 1-hydroxypyrene (LOGU1OH), air benzo[a]pyrene, weekly working hours and daily cigarette consumption. In a multivariate model 1-hydroxypyrene had a consistent effect. Neither glutathione transferase M1 (GSTM1) nor cytochrome P450 1A1 (CYP1A1) genotypes had clear effects. Yet the individuals lacking GSTM1 had a stronger effect of LOGU1OH and some effect by other sources of PAH, such as charcoal broiled food, although all these variables were not significant in the multivariate model. The rare individuals with a CYP1A1 polymorphism MspI containing an amino acid change at isoleucine had an increased level of adducts. The results showed that the postlabelling method used was able to detect an increase in aromatic DNA adducts in leukocytes when exposure to benzo[a]pyrene in air was approximately 5 ng/m3. At such low levels smoking and charcoal broiled food may be important contributors to adducts.

Carcinogen-DNA adducts and somatic gene mutation at the hypoxanthine guanine phosphoribosyl transferase (HPRT) locus were evaluated in peripheral leukocytes of workers in an iron foundry with exposure to benzo[a]pyrene (B[a]P) and other polycyclic aromatic hydrocarbons (PAHs). During the two year study period, B[a]P exposure declined by approximately 40%, from a maximum of 60 ng/m3 in the first year to < 36 ng/m3 1 year later. A total of 64 persons were sampled in November/December of the two successive study years; 24 of them gave two samples one year apart. The biomarkers included carcinogen-DNA adducts in leukocytes (PAH-DNA measured by an immunoassay, aromatic-DNA by the 32P-postlabeling method) and HPRT mutation in lymphocytes. After adjusting for smoking, levels of PAH-DNA, aromatic-DNA and HPRT mutation frequency (Mf) increased with exposure among the 64 workers sampled during the 2 year period (P < or = 0.05). However, the markers showed a differential response to the change in exposure, consistent with their individual biology. For example, among the 24 workers sampled in both years, carcinogen-DNA adducts (which have a half-life on the order of several months) were markedly reduced from the first to the second year (PAH-DNA, 6.2 versus 2.3/10(8); aromatic-DNA, 2.5 versus 1.4/(8); P < 0.01). HPRT Mf (a longer-lived marker) was somewhat less affected by the decline in exposure (1.3 versus 0.8, P < or = 0.05). Moreover, in the second year several long-term workers had low levels of adducts, but elevated HPRT Mf. Thus, PAH-DNA and HPRT Mf were highly correlated in the first year (n = 17; r = 0.67; P < 0.01), but not in the second year or in the two years combined. However, when analysis was restricted to workers with detectable levels of adducts (who included the more highly exposed workers) the correlation was significant between PAH-DNA and HPRT (n = 17; r = 0.65; P = 0.005). In contrast, aromatic-DNA adducts and HPRT were not correlated in either year. These results suggest a molecular link between somatic gene mutation and PAHs; and they highlight the need in such molecular epidemiologic studies to consider the varying lifetimes of the individual markers.

Mutations were evaluated in workers in an iron foundry with exposure to polycyclic aromatic hydrocarbons (PAHs), measured by personal and area monitoring, ranging from < 5 to 60 ng/m3 of benzo[a]pyrene (B[a]P). Mutation at the hypoxanthine guanine phosphoribosyl transferase (HPRT) and glycophorin A (GPA) loci (measures of molecular effect in lymphocytes and erythrocytes respectively) were assessed to demonstrate their relationship to external exposure at lower levels than previously analyzed in foundry workers at this plant (< 50-200 ng/m3). The relationship between mutations and PAH-DNA adducts measured by immunoassay (as a measure of the biologically effective dose) was also investigated. The markers were analyzed for dose-response and interindividual variability. Workers were classified into three exposure categories (low, medium and high). PAH-DNA adduct values for the low, medium and high exposure groups were 5.19, 6.10 and 9.57 x 10(-8) nucleotides respectively (r = 0.28; P = 0.08). HPRT mutant frequencies (adjusted for age and cloning efficiency) for the low, medium and high exposure groups were 1.04, 1.13 and 1.82 x 10(-6) cells respectively and demonstrated an upward trend with increasing exposure that was of borderline significance (r = 0.46, P = 0.06). In contrast, HPRT mutations were highly correlated with PAH-DNA adducts (r = 0.67; P = 0.004). Interindividual variability in mutant frequencies ranged from 1.5- to 4.5-fold within the three exposure categories. With respect to GPA variants, NN frequency (Vf) in erythrocytes (which reflects chromosomal loss and duplication, recombination or gene conversion) was not positively correlated with PAH exposure. The level of N0 Vf (arising from small-scale structural mutations in the GPA gene or from larger-scale chromosomal rearrangements or deletions) increased slightly, but not significantly, over the three exposure groups from 8.2 to 10.7 to 11.8/10(6) cells (P = 0.32). Interindividual variation in GPA NN Vf ranged from 2- to 18-fold and in GPA N0 from 4- to 5-fold. NN and N0 Vf were highly correlated (P = 0.001) but no correlation was seen between GPA and HPRT or between GPA and PAH-DNA adducts. Thus, the most interesting and novel finding is that, even at relatively low exposures to PAH, HPRT mutations were increased in parallel with PAH-DNA adducts. The observed association between PAH-DNA adducts and HPRT gene mutation in humans is consistent with experimental data for PAHs. These results support the use of both biomonitoring and personal ambient monitoring in further molecular epidemiology studies.