Cytogenetic and hematological effects in plastics workers exposed to styrene.

F. Cytogenetic and hemato logical effects in plastics workers exposed to styrene. Scand J Work Environ Health 1989;15:136-41. For 20 glass-reinforced plastics workers exposed to styrene and 22 unexposed referents, the frequency and sizedistribution of micronuclei weredetermined for lymphocytes stimulated with phytohemagglutinin or pokeweed mitogen, and white blood cell counts were made. Furthermore, chromosome aberrations werescored for II of theexposedsubjects and 15of the referents. The meanlevelof styrenein the breath ing zone of the workers was 56 mg/m '. Workers exposed to styrene did not show an increase in any of the cytogenetic end points studied whentheeffect of ageand smoking wasallowed for in a multipleregres sion analysis. A significant 30 % increase in the number of peripheral monocytes was observed for the exposed workers. No correlations between the cytogenetic and hematological tests on one hand and the length of exposure time on the other could be detected.

During the last few years we ha ve been abl e to develop the micronucleus assay for lymphocytes by preserving the cytoplasm of the cells (13), which increases the precision in scoring micronuclei, and by measuring m icronucleus sizes , thus separating different inducing agent s (14,15). Furthermore, we ha ve been using both ph ytohemagglutinin (PHA) and pokeweed (PWM) as mitogens which probably stimulate different sub groups of lymphocyte s (15,16).
The carcinogenic or ganic solvent benzene induces cyto genetic effects in humans and ma y also , due to a to xic depression o f the bone marrow, cau se lymphopenia in peripheral blood (17). In contrast, exposure to other organic solvents may cause lymphocytosis (18,19). Whether styrene exposure may affect the per ipheral lymphocyt e level is not known . Thus differential white blood cell counts were included in the examination .
The objectives of the present study were to inve stigate whether low-le vel expo sure to styrene causes cyto geneti c effects or disturbances in th e per ipheral white blood cell counts. Furthermore, we attempted to investigate whether styrene preferably indu ces larger micronuclei and whether the type of mitogen used, PHA or PWM , is of importance for styrene-induced micronuclei frequencies and size ratios.

Subjects and methods
Production and exposure classification Th e exposed workers of the study came from a plant producing tanks of glass -reinfor ced pol yester pla stics. Th e framework of the tanks consists of cylindric tubes (diameter 600-2500 mm) produced in a winding machine. Gables, tube flares, and other additional parts of the tanks are made by the spraying of glass and polyester compounds onto molds. Finally, the tanks are put together in an assembly hall . Thi s part of the work includes hand-operated application of pol yester. The polyester is dissolved in styrene, which has the double fun ction of being a sol vent and a reactive monomer. A peroxide in itiates the pol ymerization reaction.
All the workers in the plant are expo sed to styrene. In addition there is low-level exposure to acetone and methylene chloride, both o f which are used for the cleaning of tools and skin.
Th e work room air concentrations of sty rene ha ve been determined by personal sampling since 1974 . Result s from a total sa mpling time o f 703 h are given in tabl e I. The mean sam pling time per sample was 60 h. The mean concentration during the period 1985-1986 was 56 rng/rn J. As the differences in the exposure levels between the different kinds of work have been small and there is a large rotation of work tasks among the workers, we have not tried to classify different work operation s with regard to exposure levels.
Respira to ry protection devices were used during spray ing operations and during work with the polyester inside the tank, otherwise not.
For 43 urine samples collected immediately afte r work in Octob er 1985, the mean excreti on of the styrene metabolites mandelic acid and phen ylglyoxylic acid was 128 (ran ge <6-317) mmo l/mol creatinine.

Subjects and sampling
All 18 pro cess operators and two foremen in the glassreinforced plastics plant, emplo yed for at least one month , pa rticipated in the stud y. All but one were males. Their mean age was 40.5 (ran ge  years, and their mean time of employment was 8.1 (range 0.1-25.4) years.
Twen ty-two male workers from a factory producing gelatin, without any significant chemical exposure, served as a reference group. The ir mean age was 42.3 (ran ge  years. The referents were not exposed to any industri al air pollutants. They were mainl y work ing with internal transport , as electr ician s, technicians, and process supervisors. All the subjects were interviewed by a physician (LH) with regard to occupational and med ical history, especially concerning smoking habit s (cigarettes per day; I g of pipe tobacco being considered equivalent to one cigarette) , viral infections , dru g int ake , radio graph ic examinations, and exposure to ionizing radiation during the last year.

Methods
Venous blood sampling. Venous blood samples were obtained from all the examined subje cts between 0800 and 0900 on Monday mornings. Sample s from all the subjects could not be cultured simultaneously. Thus the samples were taken on three consecut ive Monday mornings. At each sampling time about half of the subjects were referents . All blood samples were coded and immediately tra nsferred to the laborato ry for analy-SIS .
Micronucleus assay. Buffy coat leukoc ytes were cultured for 80 h in Roswell Park Memorial Institute (RPM I) 1640 medium with 15 070 fetal calf serum (Flow), as described by Hogstedt (13). Parall el cultures were set up with either PHA (Gibco; I ml/IOO ml) or PWM (Gibco ; I ml/IOO ml). It has earlier been shown that the appropriat e incubation time is the same for both mito gens (15). The cells were prepared according to the method described by Hogstedt (13).
In the first step I 000 lymphocytes from each individual were analyzed for the presence of intracellular micron uclei. In the second step, a number of cells corr espond ing to 10 micronuclei were scored , and the size of the cell nuclei and the micronuclei was measured according to the method described by Ho gstedt & Karlsson (14). The size of the micronucleus was expressed as the ratio between the sur face of the micronucl eus and the cor responding main nucleus of the cell. Th is procedure ha s been found necessar y as the different phases of the cell cycle cause different sizes of both cell nuclei and micronuclei. In no case was more than 3 000 lymphocytes scored in the second step . We calcul ated the estimated frequency of micronucl ei by adding the denominators and nominators, respectively, from both steps of the analysis. All scoring was performed by one observer. The interobservational precision of scoring micronuclei was very high; the correl at ion coefficient for the two steps of  Chromosome aberrations in lymphocytes. Ten drops of whole blood were incubated at 37°C with PHA in 10 ml of medium containing RPMI 1640 and 15 0/ 0 fetal ca lf seru m . The incubatio n time was 48 h. Colcemid (0.1 ug /ml) was added I h before the harvest. Hypotonic treatment was carried out with potassium chloride (0.075 molll) for 15 min at room temperature, and the cells were fixed in methanollacetic acid (3: I). The chromosome preparations were stained with Giemsa. For each individual, 100 metaphases were scored according to the classification system recommended by the International System for Human Cytogenetic Nomenclature (20). The results have been pre sented as "gaps" and "breaks" (comprising chromatid and isochromatid breaks, pericentric inversions, ring s, and dicentrics).
Due to poor preparations and mitotic inhibition , slides from nine exposed subjects and seven referents were not scored. Five of the II exposed subjects examined and nine of the 14 examined referents were smo kers.
Differential white blood cell counts. In each blood sample 10 000 cells were analyzed by a Technicon H-6000C''' automated flow cytochemistry cell counter (21). The precision for this technique is very good, as shown in a previous study (22).

Statistical methods. For intraindividual comparisons
of the frequencies and size rat ios of micronuclei , between PHA and PWM cultures, Wilcoxon matchedpairs signed-rank test and the Pearson correlation coefficient were used. For interobservational variation of the micronuclei scoring, Spearman's rank order correlation coefficient was calculated.
The styrene-exposed group and the referents were compared with regard to the biological effect variables, age and smoking habits being allowed for in the multiple regression analysis. Furthermore, in a multiple regression analysis of the biological effe ct variables performed only for the styrene-exposed group, the importance of employment time in the plant was tested with age and smoking bein g allowed for.
As the frequencies of chromosome aberrations (ran ge 0-5 /100 cells) were very low, the average square root transformation [(Y-x + -Jx + I)12] was applied to these counts in order to stabilize the variance (23). Graphic methods showed that the log-transformation was suitable for both th e frequencies of micronuclei and the white blood cell counts. The individual geometric mean va lues of the size ratios for micronuclei were analyzed untransformed .
Significance levels of 5 % or less have been considered significant. All the stati stical tests were two-tailed.

Results
With PWM stimulatio n the mean frequency of micronuclei, for all the examined subjec ts, was 6.5 o/(X l, compared to 4.4 D/OO with PHA. Thi s difference was stati sticall y significant (P < 0.000 1; Wilco xon mat ched-pairs signed-rank test). The correlation coefficient between the two log-transformed scores was 0.37 (Pe arson' s r, P =0.02).
For lymphocytes stimulated with eithe r PHA or PWM , no statistically significant difference was found in the frequencies of micronuclei between the exposed and reference groups when allowance was made for age and smo king (table 2). For the cultures stimulated with PHA, but not for those stimulated with PWM, the micronucleus frequencies showed a significant positive association with age (P < 0.0001) . No association with smoking habits was found.
There was no significant difference in the individual mean micronucleus size ratios between the lymphocytes stimulated with PHA or PWM (P = 0.29 Wilcoxon mat ched -pairs signed-rank test) . Furthermore, ther e was no correlation between the two scores (Pearson's r=0.15, P =OA).
Th e individual mean micronucleus size ratios were not associated with expo sure to styrene, irre spective of the mito gen used (table 2). For the cultures stimulated with PHA , but not for tho se stimulated with PWM, the size ratios showed a significa nt association with age (P =0.004). No association with smo king habits was found.
Furthermore, no association between either the micronucleus frequencies or the size ratios and the length of employment was found for the exposed gro up.
There was no statistically signifi cant difference for chr omosome breaks between the exposed and reference groups when allowance was made for age and smoking (table 3). On the other hand, significantly (P = 0.02) more gaps were observed for the referents. No association between either the chromosome breaks or gap s and the length of employment was found for the exposed group.
The number of peripheral monocytes, but not that of any other leukocytes, was increased in the exposed gro up when it was compared with the reference group (table 4). The mean difference was 30 % and statistically significa nt (P = 0.002) . No association between an y leuko cyte count and length of employment was found for the exposed group.

Discussion
The main result of th is study was that workers exposed to styrene did not show an increa se in micronucleus frequen cies or size ratios, irrespecti ve of wheth er PHA or PWM was used as the mitogen , or chromosome a ber ra tions . Table 2. Mean num ber of micron uclei and the in div idual mean values of the size ratios fo r mi cro nuclei in lym phoc ytes sti mulated with phyto hemaggl ut inin (PHA) or pokeweed (PWM) in 20 sty rene-exposed workers and 22 refere nt s. The unad ju ste d 95 % con fide nce int erval (95 % CI) is also gi ven. The si gn if ic ance level for the numbers of mi cron ucle i refer s to the diff erence in the me ans after log-tr ansf orm at ion and for the untransformed in dividu al mean size ratios, in bo th case s adjusted for age and sm oking . D ue to tec hn ical problems , the sco ring of chromoso me aberrat io ns was per fo rm ed for o nly 26 of th e 42 subjects . The loss was not systematic wit h resp ect to the exp osed subjects and referents a nd thus did not intr od uce a ny bias.
In the pre sen t study the referen ts had significa ntly mo re gaps th an the expo sed wo rkers . Ho wever, a ll the sco res were low and withi n th e "normal" limit s of th e lab o rat ory. As the referents were not chemica lly exposed , th e discrepan cy in the num ber of ga ps was co nside red a spurio us findin g.
The mean exposure levels of styre ne in the air were co nsidera bly higher (477 a nd 197 mg / m ' in so me of the pr eviou s " pos itive" st udies (4,6) in co mpariso n with th e expos ure level of 56 mg /m ' of th e pr ese nt study . However , an increas ed number of chro mos ome a berratio ns (8) a nd micronuclei (9, 10) has been o bserve d a lso fo r workers exposed to styrene levels in the sam e ran ge (55,55, a nd 101 mg/m') as th at of the present st udy. On the other hand, Maki-Paakkanen (12) co uld no t show a n increase in ch romosom e aberratio ns, siste r chromatid exc ha nges , or micr onuclei in wo rkers exposed to a mean level of 98 mg/m ' styrene . T hus no co nsistent dose-r esponse relationship betwee n styrene exposure level and cytoge netic end points ha ve been dem o nstrated so far. Furthermo re, no association betw een cytogenetic effects and length of exposure time has been fo und (12). Thi s inability to demon strat e dose-re sp on se relati on sh ips may depen d on th e lo w pr ecisio n of cytoge netic meth ods or th e problem of identifying th e relevant exposure parameter. Th e mean level o f expos ure is a ver y insensitive par am eter. High occasio nal exp osure peak s, whic h pr o ba bly vary in bo th amplitude a nd frequen cy betwee n different plants, may be of greater im po rt ance fo r cytog enetic effects th an th e tim e-weighted ave rage level. In additio n the age d ist ribution of lymphocytes in peripheral blood may di ffer betwe en su bjects . Thus a blood sample with many yo ung lymphocyte s pro ba bly re flects a more recent expos ure tha n a sam ple with man y old er lymphocytes. T he relevant expos ure pa ram eter ma y th erefore var y between different exposed subjects, and this possibility contri bu tes to the di fficulti es in displaying dose -response relatio nship s.  It has previously been shown in vitro th at th e size dis tributio n o f micr onuclei dep en ds o n the type o f indu cing age nt (14,24). In an in vivo study o f piperazineexposed wo rkers we have a lso ob served a d ifferent size distribution between the exposed and refer ence groups ( 15). Ho wever , no such di fferenc e co uld be found in th e present st udy of styre ne-exposed workers.
It is well estab lished th at the re is a n association between age and th e frequen cy of micronuclei in PHAactiva ted lym phocytes (13 , 15, 22) . As in the previo us study of p iperazine-expo sed wor kers, the re was also an age effect on th e size distribution in the present st udy . Thus it is apparent that not onl y the numbers o f micronuclei , but also th e sizes inc rease with age .
Thi s age effect on size ma y depend on the increased lagging of sex-linked chro mosomes by age (25).

The frequencies of micronuclei induced in vitro by
Xvrays and mitomycin C in lympho cytes activated with PWM were found to be significantly higher than in PHA cultures (16). This was also the case in the in vivo study of piperazine-expo sed workers (15). We have suggested that the two mito gens activate lympho cyte subgroups of diffe rent sensitivity to mutagens. Such an effect was, however, not observed for the styreneexposed workers.
Regardl ess of exposure th e micronucleus frequencies in the PWM-cultured lymphocytes were significantly higher than in the PHA-activated cells. This differ ence cannot be at tributed to a difference in variance (table 2). Ther efore, it is likely that also the "normal" level of micronuclei is higher with PWM than with PHA.
It is well known that exposure to benzene may cause lymphopenia (17). On the other hand, workers exposed to other solvents have shown per ipheral lymphocytosis (18,19). In addition workers exposed to various chemical s, including organi c solvents, in a chemical plant showed lymphocytosis, and also increased levels of eosinophils and basophils but not monocytes (22). The present study does not indicate that low-level exposure to styr ene affects the lymphocyte cell count. It is not ob vious whether the observed, marginal increase in monocytes is a spur ious finding or reflects the styrene exposure. Howe ver , it should be emphasized that using the automated cytochemistry counting technique makes it possible to detect much smaller differences in the cell counts than older techn iques allowed. Th e intraassay variation for mon ocyte counts has been shown to be less than 4 010 (22).
It could be question ed whether or not an extended analysis of the leukocytes ought to be performed in all cytogenetic studi es, includin g measuring the subgroups of lymphocytes before starting the cultures. The rat ionale for this pro cedure is that not onl y the type of mito gen ma y alter the composit ion of lympho cytes studied , but also differ ent types of chemical exposu res may have thi s effect in vivo . It is interesting to not e that we not on ly found an excess o f micronuclei in a piperazine-expo sed group of workers , but also an increased number of lymphocytes (15). Furthermore, in a recent study of gasoline pum p repair workers, we obtained the same results (Hog stedt, unpublished results). It could not be ruled out that positive cytogenetic findings could be att ribu ted to differences in the compo sition of the lymphocyte subgroups studied. Moreover, the exposure-r elated differences in the peripheral leuko cyte cell counts do not necessarily reflect increas es in the production of leukoc ytes. They may rather be the result of an altered dist ribution of white blood cells bet ween dif ferent compartments of the bod y.