Occupational exposures to Cd, Ni, and Cr modulate titers of antioxidized DNA base autoantibodies.

This study was undertaken to establish whether occupational exposures to derivatives of carcinogenic metals evoke inflammatory immune responses, as determined by the presence of elevated titers of antibodies (Ab) that recognize oxidized DNA bases. Sera obtained from the blood of steel welders (Delaware) and from workers of the Centra Ni-Cd Battery Factory (Poznań, Poland) were analyzed by the enzyme-linked immunosorbent assay. To determine specific and nonspecific binding, an oxidized thymidine [5-hydroxymethyl-2'-deoxyuridine (HMdU)] coupled to bovine serum albumin (HMdU-BSA) as well as mock-coupled BSA (M-BSA) were used as antigens for coating the wells of microtiter plates. Titers of anti-HMdU Ab were significantly elevated in the high Cd and Ni exposure groups (18.3 +/- 3.2 vs 10.8 +/- 2.1 A492/microliters; p < 0.05). The sera of the groups with low exposures to Cd and Ni also had enhanced titers of those Ab but those increases were not statistically significant. Interestingly, the Ab titers present in the sera of controls for Cd and Ni exposures appear to be constant regardless of the protein content. In contrast, both lightly and heavily exposed subjects exhibited Ab titers that increased with increasing protein content. When 12 randomly selected workers (4 from each of the control, lightly, and heavily exposed groups) were outfitted with personal monitors, anti-HMdU Ab titers of those workers showed a significant difference between the groups with light (< 100 micrograms/m3) and heavy (> 200 micrograms/m3) exposures to Cd (9.8 +/- 3.7 vs 22.1 +/- 3.7 A492/microliters; p < 0.01) and Ni (11.7 +/- 1.4 vs 31.0 +/- 1.8; p < 0.001).(ABSTRACT TRUNCATED AT 250 WORDS)


Introduction
Many carcinogenic metal compounds were shown to induce oxidative stress (reviewed in [1][2][3][4]. These include derivatives of Cr, Ni and Cd. Even in the absence of serum, the insoluble carcinogenic sulfides of Ni and Cd, as well as CaCrO4, can stimulate human neutrophils, which generate enhanced levels of H202 that are comparable to those mediated by the potent tumor promoter 12-O-tetradecanoylphorbol-13acetate (TPA) (5). Ni and Cd sulfides are known to accumulate in the lungs and are chemotactic for neutrophils (6)(7)(8)(9). Having the ability to activate phagocytic cells, those carcinogenic metal derivatives are able to induce oxidative stress (5), a hallmark of inflammation (4). Recently, we showed that Ni3S2 causes an oxidative response in Chinese hamster ovary (CHO) cells as well, which produce substantial amounts of H202 in response to incubation with Ni3S2 (10).
Recently, we have shown that people suffering from various types of inflammatory conditions elaborate antibodies (Ab) that recognize and bind to an antigen containing oxidized DNA bases (11)(12)(13). Titers of these Ab are significantly higher than those present in the sera of healthy controls. Chronic inflammation induces oxidative stress, which is known to cause oxidation of various cellular macromolecules, including DNA (4). The inflammatory conditions we showed to cause elaboration of high titers of antioxidized DNA base Ab include various types of lupus, psoriasis, and immune complex and neoplastic diseases (11)(12)(13). Interestingly, these Ab titers decline in people who are also treated with systemic cytotoxic or antiinflammatory drugs (12,13). In contrast, treatment of psoriasis with UVB causes an increase in Ab titers (13), which is consistent with UVB having proinflammatory properties.
In those studies, we used the oxidized thymidine 5-hydroxymethyl-2'-deoxyuridine (HMdU) coupled to bovine serum albumin (HMdU-BSA) as an antigen. The Ab that bind to HMdU-BSA were exclusively Environmental Health Perspectives  of the IgM isotype that are known to arise during acute inflammatory conditions. The goal of this study was to establish whether people who are occupationally exposed to various potentially carcinogenic metal derivatives also elaborate enhanced levels of Ab that bind to antigens containing oxidized DNA bases.

Experimental Procedures
Study Populations Ni and Cd Exposures. Male employees of the Centra Ni-Cd Battery Factory located in Poznari, Poland, from five different job categories, were divided into three groups according to their levels of occupational exposure to metals. Individuals who work in the chemistry and panel production departments without any type of body protection are considered heavily exposed to Cd oxide and Ni oxide dusts present at high ambient concentrations (group 1, heavy exposure; 10 subjects). Workers in the assembly and maintenance departments are considered less heavily exposed (group 2, light exposure; 9 subjects). Centra administrative personnel served as a control group (12 subjects).
A brief questionnaire was administered to all subjects to collect data on demographics, height, weight, current smoking habit, number of years worked in their respective departments, and number of hours worked in the 3 days preceding the interview.
Upon completion of the interview, blood was drawn from a peripheral vein into heparin-containing glass vacutainer tubes. All specimens were labeled using random codes and transferred for initial processing within 3 hr of sampling to the Laboratory at the Institute of Occupational Medicine in L6d£, Poland. Frozen plasma samples were shipped to New York University Medical Center and stored at -80°C until assay.
Exposure to Welding Fumes Containing Mn, Ni, Cr Members of a railroad workers union, who had been full-time professional metal arc welders for at least six months, participated in this study (27 volunteers). These workers either welded rails or structural steel on open railroad tracks or in enclosed shops and utilized a variety of welding electrodes. Frequent use of a high manganese-nickelchrome electrode in a number of applications was reported by most welders. No recent exposure to stainless steel was reported. No quantitative assessment of the actual exposure to welding fumes was attempted at this time. The 11 volunteer controls were office workers, field railroad supervisors, union representatives, janitors, and laboratory technicians. Data on demographic (age and race), body weight, general health, and smoking status were collected through a questionnaire administered by an interviewer at the time of blood withdrawal. Plasma samples were frozen and stored at -80°C until use.

Determination ofthe Actual Exposures to Cd and Nl
Personal exposure to airborne Cd and Ni was measured for 12 individuals, 4 in each of the following occupationally designated groups: controls, lightly or heavily exposed workers. Airborne particles were collected with lapel-mounted filter cassette monitors, 37 mm in diameter. Air was drawn through the cellulose-ester membrane filters at 2 1/min by a small, belt-mounted pump for one full work shift. Flow rates were checked at the beginning and end of each of the sampling periods. The sampling cassettes from eight workers were replaced once or twice during the shift when visual inspection revealed that the filters were heavily loaded. The cassettes were stoppered and returned to the Department of Environmental Medicine at New York University Medical Center for analysis. Experimental filters, as well as filter and control samples (spikes, blanks, and standards of the National Institute of Standards and Technology) were wet ashed, the solutions filtered, and the metal content measured by atomic absorption spectrometry. Residues contained negligible Ni and Cd, as determined by X-ray fluorescence spectrometry.

Preparation ofAntigens
Antigens, consisting of the riboside of 5hydroxymethyl uracil (HMU) coupled to bovine serum albumin (HMdU-BSA) and mock-coupled BSA (M-BSA), were prepared as previously described (11)(12)(13). Briefly, HMU riboside was coupled to BSA by periodate oxidation followed by borohydride reduction of the product, according to a method developed for coupling normal unoxidized nucleosides (14). Mock coupling was carried out under the same conditions but in the absence of HMU riboside. The crude products were purified  Where m is a slope, b is an intercept, X and Y are mean values of the x and y series (actual exposures to Cd or Ni in pg/m3 versus Ab titers in A492/pl in sera of Ni-Cd battery workers), while ox and (ay are standard errors of the x and y determinants. Results on a DG-P6 column (Bio-Rad, Melville, NY) and lyophilized.

Enzyme-linked Immunosorbent Assay
Assays were carried out in microtiter 96well plates, which were coated with either HMdU-BSA or M-BSA (10 pg/ml). This allowed determination of specific as well as nonspecific binding on the same plate. In addition to the sample sera and buffer (negative) control, a serum with a known high titer of anti-HMdU Ab was used as a positive control on all of the plates (at least three in HMdU-BSA-and three in M-BSAcoated wells). The presence of a positive control eliminates batch-to-batch variability of antigens and of the goat antihuman secondary Ab. In our previous work we found that anti-HMdU Ab are of the IgM isotype, therefore, we used goat antihuman IgM as the secondary Ab.
Sera were diluted 2.5x103 to 1x105 times and incubated in the antigen-coated wells at 37°C for 2 hr. Wells were washed three times with PBS containing 0.05% Tween-20, followed by incubation with goat antihuman IgM Ab (not affinity purified; Sigma Each serum was analyzed four to eight times at different concentrations and the results are presented as mean values of A492/pl undiluted serum ± SE. The reproducibility of this assay is 5.4 ± 0.5% (12).
Protein levels were measured in appropriately diluted sera, using bicinchoninic acid (BCA) and Cu2+ as reagents, according to the manufacturer's (Pierce Chemicals, Rockford, IL) conditions. The results of two to four determinations are expressed as mean values ± SE in tig/fil undiluted serum.

Statistical Evaluation
The difference between mean values was tested by the Student's t-test. Correlation coefficients (r) were calculated according to the following formulas: Cd and Nl Exposures Figure 1 shows the distribution of anti-HMdU Ab titers present in three groups assigned according to occupational exposure to Cd and Ni. Antibody titers of heavily exposed workers are higher than those of the lightly exposed group (18.3 ± 3.2 vs 12.1 ± 3.5 A492/pl serum), and are significantly (p<0.05) higher than controls (10.8 ± 2.1 A492/pl serum).
The plot of Ab titers (A492/Pl serum) versus specific activity (A492/pg protein) shows a high correlation between these two parameters with r = 0.98 ( Figure 2). However, the heavily exposed group exhibits a greater increase in Ab titers per unit of proteins present in those sera than do controls and lightly exposed people. below the control (2.5 vs. 10.8 A492/pl serum), while at the highest level (75 pg/pl) it is about 23 pg/pl, approximately a 9-fold increase. The heavily exposed group shows a near parallel shift in comparison to the lightly exposed group. At the low protein level (45 pg/pI), the Ab titer is close to that of the control, being about 9 A492/Pl serum. At the high protein level (75 pg/pl), the Ab titer is about 32 A492/pl serum, a 3.6-fold enhancement. When the workers are divided according to their actual exposures to Cd and Ni (as measured by personal monitors), the differences in Ab titers become more pronounced, even though only 12 workers were monitored. The light exposure is defined as 100 pg Cd or Ni/m3 or less, 3 while high exposure is above 200 pg/mi For comparison, an occupational control group is also included in Figure 4. The monitored low exposures to either Cd or Ni do not differ from those of the occupational controls. However, those exposed to very high doses of Cd or Ni have significantly elevated anti-HMdU Ab titers ( Table 1).
The correlation coefficients between Ab titers and actual exposure to Cd or Ni (determined by personal monitors) are rcd = 0.4699 and rNi = 0.7225, respectively.

Exposure to Welding Fumes
The sera of welders who were lightly exposed to welding fumes (Mn, Ni, Cr) also had enhanced mean anti-HMdU Ab titers as compared to the controls ( Figure  5); however, the result is not statistically significant. Some of those workers were taking antiinflammatory medications, which could have decreased the Ab titers. We found this to be the case in our previous work (12,13). Since medication information was not available for all of the study subjects, we could not take this parameter into account. Table 2 compares anti-HMdU Ab levels in the sera of workers occupationally exposed to Cd and Ni (Centra battery factory in Poznah, Poland) vs. those exposed to welding fumes (Delaware).

Discussion
The results presented demonstrate that occupational exposure to salts of carcinogenic metals induces an inflammatory response, as judged by the elevated titers of autoantibodies that recognize the oxidized DNA base derivative HMdU. Those Ab, which are of the IgM isotype, bind to more than one HMdU residue since free HMdU could not neutralize those Ab, while HMdU-BSA did (11). When workers are divided into groups according to their occupational exposure levels, a significant difference between the controls and those heavily exposed to Cd and Ni is apparent at p<0.05. Actual monitoring of the exposures using personal monitors demonstrates even greater differences, with p<0.01 for Cd and p<0.001 for Ni exposures. These differences are even more remarkable considering that only 12 workers selected at random (four from each occupationally exposed group: controls, light, and heavy exposures) were monitored in this pilot study. These results point to the importance of monitoring actual exposures, particularly since it reduces the possibility of making an error in classification of occupational exposures.
A careful examination of Figure 2 leads one to conclude that although the correlation between A492/pl serum and A492/pg protein is excellent (r = 0.98), the values of the highly exposed subjects deviate in such a way as to show that the rate of increase in anti-HMdU Ab titers is greater than in protein levels (i.e., specific activity is higher than in both controls and lightly exposed subjects).
When the correlation coefficient was determined separately for each of the three groups, the Ab titer did not change in controls regardless of the protein content ( Figure 3). In contrast, both low-and highexposure subjects exhibit Ab titers that increase with the increasing protein amounts. The difference between high-and low-exposure groups is not that much in the slope but rather in the y (Ab titer) intercept. However, the overall increase in specific activity of those Ab appears to be much greater in the case of light exposure to Cd and Ni (9-fold) than of the high exposure (3.6-fold). These results point to the possibility that analysis of changes in specific activities of anti-HMdU Ab provides a more sensitive indicator of occupational exposures. It could be that anti-HMdU Ab titers are constitutive and present at low levels at all times. However, exposure to proinflammatory agents modulates those Ab commensurate with severity of exposure by depleting existing Ab at very low exposures first, and then stimulating their production when increased formation of oxidized bases in DNA occurs.
Comparison of the control groups from the Poznaf and Delaware studies ( Table 2) points to the importance of determining the distribution of a new biomarker in each of the control populations. It is striking that controls from Delaware have about 2.5 times higher Ab titers (27.4 ± 5.4) than those from the Poznaii factory. The Ab titers of healthy controls analyzed in our previous study (12) were 14.9 ± 2.2, as compared to Poznani (10.8 ± 2.1) and L6dz (19.0 ± 2.6) (unpublished data). One possible explanation could be that there are differences in lengths of exposure to the sun and to air pollutants among various population groups, i.e., Lod£. The subjects used as controls for the welders' exposure are likely to spend more time outdoors than the other three populations. We previously have found that UVB exposure significantly enhances the anti-HMdU Ab titers of the patients with psoriasis who were treated with UVB modality (13). Also, a possibility exists that Mn present in the welding fumes counteracts the prooxidant effects of Ni and/or Cr. It was shown previously that Mn dust counteracts Niinduced toxicity and carcinogenicity (15). Whereas soluble Mn salts inhibit hydroxyl radical-mediated damage in the presence of H202 and superoxide anion radicals (16), conditions present during inflammation and responsible for the oxidation of bases in cellular DNA (4). These considerations could explain why the difference in anti-HMdU titers between control and exposed welder populations is statistically insignificant.
Occupational exposure to Cd and/or Ni was not associated with any of the following: history of smoking, age, weight, and years on the job. However, the small sample does not allow any definite conclusions. Exposure was correlated with antioxidized DNA base Ab titers as shown here Environmental Health Perspectives as well as production of autoantibodies that recognize brain glial fibrillary acidic protein (GFAP) and are against neuron-specific, neurofilament proteins (H El-Fawal and H Evans, personal communication). The plans are to compare the results obtained using an assay measuring anti-HMdU Ab with those measuring anti-GFAP and neurofilament protein Ab, to detect subgroups of subjects with a particular antibody pattern, for example high titers of both types of autoantibodies.
The preliminary results presented here combined with the sensitivity of the assay are indicative of the potential usefulness of anti-HMdU Ab titers as biomarkers of exposure to proinflammatory carcinogenic agents.