Chromosome aberrations in chloralkali workers previously exposed to mercury vapor.

Chromosomeaberrations in chloral kali workers previously exposed to mercury vapor. Scand J Work Environ Health 1993;19:375-81. Chromosomeaberrations and micronucleiin peripheral lymphocyteswere studied in 29 male chloral kali workers previously exposed to mercury vapor and in two matched reference groups comprising 29 nitratefertilizer workersand 29 customsand policeofficers.The study was performedusing whole blood cultures with and without hydroxyurea and caffeine to inhibit deoxyribonucleic acid synthesis and repair, respectively. No significant differences in the frequenciesof chromosome aberrations and micronuclei were observed. However,a nonsignificantincrease in chromosome breaks and dicentrics was found in the subgroups with high urinary mercury peak levels or high cumulative mercuryexpo sure. An increased prevalence of "high" scores of chromatid breaks in the inhibited cultures, exceed ing the 75th percentile of all of the subjects studied, was observed for the chloralkali workers when compared with both reference groups. No evident cytogeneticeffects were observed among the chlo ralkali workers with the methods used in the present study.

Inhaled mercury vapor is oxidized to bivalent mercury in the blood. However, prior to oxidation, elemental mercury can penetrate into different tissues through biological membranes (I) and persist for many years after the cessation of exposure (2).
It has been reported that inorganic mercury inactivates the mitotic spindle in dividing cells, aneuploidy or polyploidy being the most evident genotoxic effect (3). In human studies, micronucleus assay indicated an accumulation of cytogenetic damage in lymphocytes after exposure to mercury vapor (4) or mercury fulminate (5). Previous studies on humans during ongoing exposure to inorganic mercury compounds showed conflicting results when other cytogenetic methods were used (6)(7)(8)(9).
Exposure to mercury vapor may constitute a health hazard to chloralkali workers. The aim of the present investigation was to study potential genotoxic effects of mercury vapor exposure among chloralkali workers after the cessation of exposure. The chromosome aberration method, with and without compounds inhibiting deoxyribonucleic acid (DNA) synthesis and repair, and the micronucleus assay were applied.

Subjects
The present study was part of a comprehensive study on possible adverse health effects in 77 male chlor- alkali workers previously exposed to mercury vapor. A group of 53 male workers from a nitrate fertilizer plant served as the referents. Details of the main study design have been presented elsewhere (10). Among these 77 workers, 30 highly exposed subjects were selected, on the basis of a long duration or high level of exposure. They were individually matched with 30 referents among the 53 nitrate fertilizer workers (reference group I). The matching criteria were current smoking status (smoker/nonsmoker) and age (±5 years). Attempts were also made to match the current consumption of tobacco as closely as possible in the matched pairs. The referents in group I were employed in the same industrial complex where the chloralkali plant was situated. An external reference group (reference group II), comprising male customs and police officers, was also established according to the same matching criteria. One subject in reference group I withdrew from the study. Therefore each of the three groups comprised 29 subjects.
Some characteristics of the subjects under study are presented in tables I and 2. The prevalence of infections, vaccinations, and X-ray examinations and the use of drugs during the previous two years were comparable in the groups and were thus not considered further. Two subjects in reference group II who were initially identified as nonsmokers were found to be smokers when the questionnaires were reevaluated after the statistical comparisons were performed (See the Discussion.) The current concentrations of mercury in whole blood and urine, and selenium in whole blood and urine were determined for the exposed subjects and reference group I with atomic absorption spectrometry (I I). The accuracy and precision of the measurements in whole blood were monitored by the use  Table 2. M erc u ry and sel enium con cen t rati ons in t he blo od and u rin e of th e su bjects at t h e t ime of t he st udy . (NO = not de t ermined) Con ce n t rat io n i n b lo od Con cen t rat io n i n u rine of Seronorrn" trace-element whole-blood qu ality assur ance materials , batch 904, 905 and 906 . The detecti on limit of the meth od was 3 nmol . I I for mercury and 0.02 um ol . I I for selenium. Human-quality assurance urine, Seronorm'" tra ce element, batc h 108, wa s used for the qu ality control of the urine ana lyses. The detection limit was 2 nmo l . I I for mercury a nd (LO005 urnol . I-I for sele nium. Selen ium was dete rmi ned because it is know n to interact with inorg anic merc ury in the body (1-2). The current con cen tration s we re low for both the ex posed subjects and refe rence group I.

Exposure assessment
Th e chloralka li plant under study started prod uc tion in 1947 and was closed at the end of 1987. The study was carried out in 1989. Biological monit oring of the employees through the dete rmination of urinary mercury began in 1948. Resu lts fro m 1434 urinary mercury measurements performed during the time of exposu re we re identified for the 29 exp osed subj ects. On the basis of the se results, an indi vidual cumulative "do se" estimate of urinary mercury was calculated as the sum of all individual mean annual urinary merc ury concen trat ions. Further detail s regarding the biologica l monitori ng of th e workers and the calcu lation of cumulative urina ry mercury have been given else where ( 10). The chloralkali workers had, on the ave rage, bee n exposed to merc ur y vapor for 11 .8 (median 10.3, 376 range 1.7-36.2) yea rs. The ir mean annual urinary mercury con centratio n wa s 6 11 (median 514 , range 23 1-292 1) nmol . I. J, and the mean c umulative urinary mercury co ncentratio n was 5827 (median 4843, range 1979-1 7 229) nmol · ) J. Expo sure had starte d on the average 22.9 (median 2 1.0, range 8.0-4 1.0) yea rs prior to the study and cea sed on an ave rage of 9.9 (median 2.0, range 1.0-33.0) years prior to the study. Se venteen subj ec ts had urinary mercury co ncen trations that exceeded 1500 nmol . I I, and for nine subj ects the urinary mercury conc entration exceeded 3000 nmo l . I I at lea st once during their period of exposure. Table 3 presents the dist ribu tion of the cumulative urinary mercu ry co ncentrations and the ann ua l mean urinary mercury concentratio n in re latio n to the number of yea rs ex posed.

Cytogenetic examinations
All of the blood samples we re collected and pro cessed o n the same day for the matched triplets. For the scoring of chromoso me aber rations, whole blood was cultured for 48-53 h in Ham F 10' " with Lglutamin, sup plemented with 20 % fe tal calf serum, 3% penicillin (5000 IV . ml I )/s trep tomycin (500 0 llg ' ml I) (all from Flow Lab oratories, Biggio, Switzerla nd ), and 0.8% phyto hema gg lutinin P-form (Difco , Mic higan , United States). Fo r the con venti onal cultures, 0.3 u g . ml I Demekolcin'" (Rikshospita let, Oslo, Norway) was added 3 h before the harvesting. Inhibition of the DNA synthes is and repa ir was ae-co rnplished in parallel cultures with hydroxyurea (Sigma, St Louis, United States) and caffeine (Sigma) both at a 7.5 . 10-3 M conce ntration, added together with Demekolc in" 3 h prior to the harvesting, as described previously ( 12). Th e preparations were stained with Giemsa (Merck, Darmstadt, Germany), and 100 mitoses for the conventional analysis and 50 mitoses for cultures for DNA synthesis and repair inhibition were scored for aberrations such as chro matid and chromosome break s, exchanges, dicentrics, rings, acentric fragment s, marker chromosomes, gaps, and aneuploid cells. The aneuploid cells included both hypo-and hyperd iploid cells. Aneuploidy and gaps were not included in the cells with aberratio ns. All of the counts have been given per 100 cells, except for dicentrics, for which the total number for each group has been given.
For the micronucleus assay, whole blood was cultured for 72 h in the same medium as described previously, the last 24 h with Cytochalasin B®(Sigma) in a concentration of 3 III . ml' ( 13). The cells were harvested and stained with Acridine Orange" (Merck) ( 14). For each subj ect 1000 binucl eated cells were scored.
The scori ng was equally di vided between three experienced scorers without knowledge of the exposure status. One matched triplet was always scored by the same person.

Statistical methods
The cytoge netic results of the ex posed group were compared with the results of the two reference groups separately with the use of the nonparametric Wilcoxo n' s matched-pair signed rank test.
Spearman' s rank correlation was calculated to study the univariate relationships between the cytogenetic variables and the exposure and life-style parameters . The level of signifi cance was set at 5% (twotailed).
Cytogenetic scores exceed ing the 75th percentile among all of the subjects under study were regarded as "high." This cut point was chosen a priori to secure an appropriate size for the study groups. If the 75th percentile included more than one value, the next unique cut point lower than the 75th percentile was chose n. Mantel-Haenszel ' s test for odds ratio (OR) with the 95% confidence interval was applied in order to study the distribution of "high" cytogenetic scores among the participants.
The statistical analyses were carrie d out on a personal computer using BMDP statistica l software, PC 90 versio n (BMDP Stati stical Software, Inc, Los Angeles, California).

Results
The results of the cytogenetic exa minations made with conventional method s are presented in table 4. No sig nificant differences were observe d for any of Scand J Work Environ Health 1993. vol 19. no 6 Table 3. Distr ib ut ion of the c umu lati ve level of urinary mercury and th e annual indiv idu al mean urinary mercury co nce ntration of 29 ch loralkal i wo rkers in relation to t he time exposed. the cytogenetic parameters when all of the exposed subject s or all of the subjects for whom the exposure had cease d within the last two years were compared with the results of the two separate reference groups. A statistica lly nonsignificant difference (P = 0.06) was observed for chromosome breaks and cells with micronuclei among the workers with a cumulative urinary mercury conce ntratio n of~5000 nmol . I I when they were compared with reference group II. The subjects with a urinary mercury peak level of~1500 nmol . I-I also had a nonsignificant-Iy higher number of chromosome breaks, and more cell s with micronucl ei than subjects in refere nce group II (P =0.05) . However, reference group I had an even higher number of cells with micronuclei than the exposed subjects with a urinary mercury peak of 1500 nmol . I -l or a cumulative urinary mercury value of~5000 nmol . I-I. The number of exposed subjects with dicentric chromo somes was small, but higher than in either reference group. A maximum of one dicentri c chromosome was observed per person. No significant intergroup differen ces were observed for chromatid and chromosome gaps, chromatid exchanges, rings, markers, and minutes (results not presented).
The mean values for all of the cytogenetic parameters were increased in cultures used for the DNA synthesis and repair inhib ition when they were compared with the result s of the conventional cultures, but no significa nt intergroup differences were observed between the exposed subjec ts or selected subgroups of the exposed subjec ts and the matched referents (table 5). However, the prevalence of "high" scores of chromatid breaks in the inhibi ted cultures was significantly higher among the exposed subjec ts than among either of the referenc e groups (table 6).
In the inhibited cultures the percentage of chromatid breaks among the exposed subj ects and reference group I was 12.5 versus 8.7 for the smokers and 6.6 versus 4.6 for the nonsmokers, respectively. The Table 4. Results of the cy toge neti c examinat ion with conve nti onal cell cu lt ures fo r all o f the exposed subjects and selected subgroups co mp ared wit h the two refe rence group s (aberrati ons per 100 ce ll s, mic ronuc lei per 1000 cell s). , P< 0.10, " P s O.05 in comparison wit h th e exposed. Table 5. Resul ts of the cy to ge net ic examinat ion for deoxy ribonuc leic acid synth esis and repair inh ibiti on in cultures f rom all of th e exposed subjects and selec ted s ubg roup s co mpared wi th th e two reference groups (aberratio ns per 100 ce ll s). mean current blood mercury and blood se lenium level s wer e com parable in these fo ur subgroup s, while the urinary mercu ry levels among the exposed smokers and nonsmokers was higher than among the ref- No statistica lly sig nifica nt associa tio n was observed for any of the cytogenetic end points and the Table 6. Number of subjects wi th "high" scores (cut off 75th percen ti le) among the exposed subjects and reference gro ups I and II, wit h the od ds rat ios and 95% co nfidence in tervals for the exposed gro up in co mparison wi th the two reference gro ups separate ly. exposure-related parameters (cumulati ve urinary mercury co ncentrati on, cur rent mercury concentration in the blood, urinary mercury concentration, and time since cessation of exposure) or life-style parameters (current tobacco and alcoho l consumption, age). The exposed group and the two reference groups were studied separately. Furthermore, no significant corr elations were found between any of the cytogenetic parameters scored in cult ures with and without DNA synthesis and repair inhibition.

Discussion
With regard to the comparability of the selected groups in the present study, reference group I was pote ntia lly expo sed to low levels of nitrate-containing dust but, to our knowledge, not to know n clastogenic agent s. The subjects in reference group II were not exposed to any chem ical agents at work. The con sistency of the resu lts in the two refer ence groups varied , and a higher score was observed for subjects in one of the reference group s than in the exposed group for some of the studied parameters. However, the figures were small and subject to high random variability. The two excess smokers in reference group II could have led to an underestimation of a difference with the expo sed group, but they had low scores for chromosome aberrations when compared with the nonsmokers in the same reference group. No significant differences between the chloralka-Ii workers previously exposed to mercury vapor and the referents were observed for any of the studied cytogenetic end points determined with con ventional method s. For chromosome breaks and dicentrics, howe ver, a nonsignificant incre ase was observed for the subgroup s with the high est exposure in the present study (tab le 4). Other cytogenetic studies among workers no longer expo sed to mercury vapor are not known to us.
Previous studi es on the effe cts of ongo ing exposure to inorganic mercury have shown conflicting result s (4-9). Versc haeve et al (6) observe d an increased prevalence of chromosome aberrat ions in 28 subjects exposed to various mercury compounds, but the resu lts were not confirmed in a later study (7) . Popescu et al (8) found an increa se in chromosome aberrations, but no increase in chromatid-type aberration s in four subjects with repeatedly high urinary mercury concentrations reaching 890 ug . I-I (4450 nmol . 1-1). At the time of the cytoge netic analysis the concentration varied between 100 and 420 ug . I-I (500 and 2100 nmo l . I-I). No smoking data were reported in that study . Mabille et al (9) did not observe any increased cytogenetic damage in 22 chloralkali worker s. Barregard et al (4) observed an association between the number of micronuclei and cumulative mercur y dose indicators. An increased prevalen ce of micronuclei was observed by Anwar & Gabal (5) in worker s exposed to mercury fulminate. If inor ganic mercury acts as a spindle poison in man, as shown for eukaryotes (3), this might be one of the exp lanations for an incre ased prevalence of micronuclei. Although the prevalence of micronuclei in the present study was significantly higher among the exposed subj ects with peak levels of urinary mercury exceeding 1500 nmol . I-I than among the matched referents in reference group II, the highest prevalence was observed for the subjects in reference group I. Furthermore, all of the exposed subjects had a higher prevalence of cells with micronuclei than the selected highly exposed subgroups. Therefore, the present results do not indicate any significant difference between the subgroups of exposed workers and the two independent reference groups for any of the studied cytogenetic end points either.
Increased cytogenetic damage , mainly of the chromatid type, was shown in the lymphocyte cultures for DNA synthesis and repair inhibition in studies with different type s of in vitro mutagen treatment compared with conventional cultures (15)(16)(17). In the inhibited cultures in the present study the prevalence of "high" scores of chromatid breaks (cut point : 75th percentil e) was significantly higher among the exposed subjects than among either of the reference groups. When other cut points were used, the same, but nonsignificant trends were observed.
Chromatid breaks may reflect ongoing expo sure to mutagens in addition to pr evious expo sure . It was shown for patient s with testicular ca ncer that chromatid breaks di sappeared fro m the peripheral lymphocyte population faster than any of the other aberration types after the ce ssation of chemotherapy. Furthermore no incre ase above th e refere nce le vel wa s detected after 19 months (18 ). The med ian for ce ssation of exposure in the present study was 24 months. As judged by the current blo od mercury concentration, no occupational exposure to mercury vapor was occurring at the time of the blood sampling, while the difference in the current urinary mercury concentration betw een the expose d subjects and the referents may reflect pre viou s exposure. Kosta et al (2) showed that sub stantial concentrations of mercury were present in form er mercury miners ma ny years after the ce ssation of mercury exposure. Tobacco smoke was probably the mo st importa nt current exposure to mutagens for the subjects under study . An incre ased pre valence of chromatid breaks in inhibited cultures has been previously shown for smokers when they were compared with nonsmokers (12) . In the pre sent study the highest prevalence of chromatid breaks in the inhibited cultures wa s ob served among pre viou sly expo sed smokers , foll owed by smoki ng referents and exposed non smokers, whil e the nonsmoking referents had the lowest prevalence.
The exposed workers in the present study had lower current concentrations of urinary se lenium than refe rence group 1. Selenium has been shown to coaccumulate with inorgani c me rcu ry in several tissues in for mer mercury miners (2). Studies among chl oralkali workers, both during ongoing (19,20) and pre viou s mercury vapor exp osure ( I I), sho wed reduced concentrations of urinary selenium wh en they were compared with referents. The po ssible mechanism of interactio n between seleniu m and ino rganic mercury is not fully understood (I ). However, selenium at nutritional levels is antimutagenic in different sys tems (2 1), and Morimoto et al (2 2) ob served that selenite pre vented the induction of sister chromatid exchanges by methyl mercur y in human whol e blood cultures . A high correlation bet ween sister chromatid exchanges and chromatid breaks in inhibited cultures has been previously shown for smoker s ( 12). Whether the interaction of inorganic mercury and se lenium may lead to increased chro matid breaks from current ex posure to mutagenic agents, such as smoki ng, is an interesting hypothesis which should be evaluated in further studies .
The mechanism of bivalent mercury toxicity at the cellular leve l is currently not well und erstood . Anoth er pos sibl e me chanism of action could, however , be a lower level of DN A repair acti vity in tissue s con taining bi valent mercury . Robison et al (23) showed that the persistence of bivalent mercury in Chinese ham ster ov ary cell s in culture can inhibit 380 DNA repair activity in cell s at concentrations and exposure time s for whi ch DNA single-strand breaks, cau sed by biv alent me rcury , could not be detected.