This study was designed to evaluate the occupational exposure to PAH, as assessed by biological monitoring techniques, in workers employed at a graphite electrode manufacturing plant and in reference subjects with presumed lower exposure to PAH. Current exposure to PAH was assessed by determining the urinary concentration of 1OHP (biomarker of internal dose), whereas biological effect monitoring focused on the evaluation of primary DNA damage extent as evaluated by the comet assay in PBL (biomarker of biologically effective dose). Moreover, in this study we analysed genetic polymorphisms in the CYP1A1 (Msp I and Ile/Val sites), EPHX (exons 3 and 4), and GSTM1 genes to evaluate the impact of these metabolic genotypes (biomarker of individual susceptibility) on the levels of urinary 1OHP and the extent of primary DNA damage.
Urinary 1OHP levels were significantly higher in exposed workers than in matched controls. This finding is consistent with the results of other studies aimed to determine the effects of occupational exposure to PAH on urinary 1OHP concentrations [3, 43–45]. The statistically significant increase in urinary 1OHP levels observed in control smokers as compared to control non smokers was not confirmed in the group of exposed workers. This aspect could be explained in terms of saturating dose, probably due to the large work-related effects, at which no further effect can be seen at higher doses.
In the comet assay, DNA strand breakage is quantified from geometric (e.g. migration distance) and fluorescence (e.g. per cent of fluorescence migrated in the tail) measurements both by eye (i.e. comet score) or computerized image analysis. In the presence of DNA damage, the distribution of tail parameters results strongly skewed to the right and the two sides of the distribution have different spreads. In this situation the mean is strongly influenced by extreme observations and, as the resulting large standard deviation, appears to be rather uninformative. In this situation, skewed not symmetrical distributions could be defined more accurately by the median value (i.e. 50th percentile) or the values corresponding to the 75th/95th percentile . This statistical approach in the comet assay was adopted in several researches [47–49].
In epidemiological studies, enrolled subjects should be characterized for a given aspect in a simple manner, preferably with a single number. In this molecular epidemiology approach we have chosen to describe the extent of DNA damage only in terms of tail intensity, a parameter describing the percentage of DNA migrated in the tail which is relatively unaffected by threshold settings in the computerized imaging .
As regard the extent of primary DNA damage, the comet assay has been employed only in few studies aimed to evaluate the PAH genotoxic effects in exposed workers [45, 50–55], with both positive and negative findings. DNA single-strand breakage did not differ between 99 potroom workers at an aluminum reduction plant and 55 unexposed referents . Occupational exposure to PAH did not result in increased DNA strand breaks coke oven workers as compared to unexposed subjects . The analysis of DNA damage did not show significant differences between 42 primary aluminium industry workers and 16 local residents with no occupational exposure to PAH . No effect of occupational exposure was observed in 50 coke oven workers as compared to 50 control workers not exposed to PAH in the extent of DNA damage . Exposure to PAH caused a significantly higher single strand DNA breakage in lymphocytes and granulocytes of 24 workers from automobile emission inspection companies and 28 workers from a waste incinerating company as compared to 43 matched unexposed subjects . The extent of primary DNA damage evaluated with the comet assay was found to be 3.13 times higher for graphite-electrode-producing plant workers (n = 29) when compared with controls (n = 32) . Thus, the findings of the present work are in line with the results of the unique study considering graphite-electrode-producing plant workers  and support the evidence that occupational exposure to PAH during graphite electrode manufacturing can result in primary DNA damage (strand breakage as evaluated with the comet assay).
Smoking habit did not increase any of the DNA damage parameters, either in the exposed than in the reference group. The effect of smoking as a potential confounder in occupational studies has been recently evaluated in a meta-analysis study of the available, conflicting results obtained with the comet assay . The authors concluded that an effect of smoking could not be formally demonstrated when the evaluation of DNA damage was based on image analysis.
Homozygous variant carriers of the CYP1A1 polymorphisms (Msp I and Ile/Val) are extremely rare in Caucasians . The frequencies (all subjects) that we found (2.65 and 0.53%, for Msp I and Ile/Val, respectively) agree with these observations. The EPHX allele frequencies found in this study are similar to those reported in non-Hispanic whites . The frequency of the GSTM1 null genotype (43.4%) among the studied population (all subjects) agrees with the frequencies reported in the literature for the Caucasian population indicating that 40÷50% of the considered subjects lack of GSTM1 activity [59, 60].
Our results indicate that polymorphisms in EPHX exons 3 and 4 is associated to higher urinary concentrations of 1OHP, whereas the presence of variants in the CYP1A1 gene (i.e. Msp I and Ile/Val mutations) as well as the presence of the GSTM1 null genotype showed to have no effect on urinary 1OHP excretion. The increased urinary excretion of 1OHP in subjects having a high EPHX activity agrees with previously published results . Whereas, the absence of a relationship between urinary levels of 1OHP and the presence of the GSTM1 null genotype is not unexpected, as 1OHP is mainly excreted as glucuronide conjugate .
None of the genotypes analyzed (CYP1A1, EPHX, and GSTM1) had any significant influence on primary DNA damage as evaluated by the comet assay. However, it was reported that DNA damage by benzo(a)pyrene (i.e. benzo [a]pyrene diolepoxide-DNA adducts) in PAH-exposed coke oven workers is influenced by smoking habits and GSTM1 polymorphisms [63, 64]. Thus, it could be of interest in the future, in workers exposed to high concentration of PAH, to compare damage caused by benzo(a)pyrene (such as BPDE-DNA adducts) and DNA strand breakage (as evaluated with the comet assay) also in relation to genetic polymorphisms.
No significant correlations were observed between urinary levels of 1OHP and the extent of DNA strand breakage (Spearman's correlation coefficients: r = 0.338, r = 0.157, and r = 0.205, for tail length, tail intensity, and tail moment, respectively) in this study. The absence of significant correlations could be explained in terms of different persistence for the considered biomarkers. In fact, urinary metabolites (i.e. 1OHP) mirror the exposure during the last workshift and some days before , whereas the alkaline comet assay measures temporary strand breaks that happen before the DNA repair systems or cell turnover occur .