Generation of oxygen radicals by minerals and its correlation to cytotoxicity.

Occupational exposure to mineral dust causes pneumoconiosis and other diseases. A cytotoxicity assay to predict the potential of minerals to cause disease would be of great value as a prevention strategy. This study compares the ability of several minerals to generate the more potent oxidizing agent, hydroxyl radical (.OH), and their cytotoxicity and lipid peroxidation potentials. Crystalline silica, the most potent cytotoxic and pathogenic mineral studied, showed the least ability to generate .OH radicals while inducing the maximal lipid peroxidation. Coal mine dust, showing the maximal ability to generate .OH radicals, was the least cytotoxic in bioassays of toxicity and induction of lipid peroxidation. Based on these results, it would appear that the ability of minerals to induce lipid peroxidation provides a better correlation with known cytotoxicity and pathogenicity of minerals than does their ability to generate oxygen radicals.


Introduction
Inhalation of inorganic minerals and coal mine dust can produce pneumoconiosis, a lung disease characterized by an initial pulmonary inflammatory response in the lung associated with an increase in cell-mediated secretions of cytokines, growth factors, lysosomal enzymes, hydrogen peroxide, and superoxide anion (1)(2)(3)(4). Several studies suggest that many of these factors work together to cause cell injury and the progression of the disease process (3,4). The possible mechanisms leading to this fibrotic lung disease are under intense investigation. To this end, we have devoted considerable effort during the last several years toward understanding the surface properties and chemical structure of minerals and their relation to oxygen radical generation and related cytotoxicity (5)(6)(7).
Several recent studies indicate that fracturing of crystalline silica, which may occur in occupations such as sandblasting, tunneling, rock drilling, or silica flour mill operations, may result in the generation of silicon oxygen radicals on the cleavage planes of silica and that freshly fractured silica can react with aqueous medium to generate hydroxyl (-OH) radicals (5)(6)(7). Studies have shown that freshly fractured silica is cytotoxic, induces the release of cytosolic enzymes from alveolar macrophages, and This paper was presented at the Conference on Oxygen Radicals and Lung Injury held 30 August-2 September 1993 in Morgantown, West Virginia. enhances the generation of OH radicals during phagocytosis (5,(8)(9)(10). In addition to crystalline silica, our earlier studies also suggest that freshly fractured coal exhibits surface reactivity not found in aged coal (11,12). The concentration of surfacebased radicals in freshly fractured silica and coal decreases with aging in air.
The objective of the present study is to evaluate the potential role of several occupational mineral dusts in the generation of -OH radicals from hydrogen peroxide (H202) and to correlate in vitro cytotoxicity and lipid peroxidation potential with oxygen radical generation. In an attempt to generate the oxygen radical and measure the radical efficiently, we used H202 and dust mixtures in the presence of an -OH radical trap and electron spin resonance (ESR) as a direct spectroscopic technique for measuring the -OH radical. The -OH radical generated from the H202 dust mixtures forms a stable radical adduct DMPO-OH with the spin trap 5,5dimethyl-1-pyrroline-N-oxide (DMPO) with a characteristic 1:2:2:1 hyperfine quartet signal. The signal peak heights of this stable adduct provide a quantitative measure of the radical generated. To correlate the oxygen radical-generating potential with cytotoxicity, we measured red blood cell hemolysis, lactate dehydrogenase (LDH), f-N-acetylglucosaminidase (P-NAG) and ,-glucuronidase (P-GLUC) release from alveolar macrophages (AM) exposed to dusts. The potential to induce lipid peroxidation as a measure of free radical-induced toxicity was determined using a noncellular in vitro assay.

Materials and Methods
Crystalline silica (min-U-sil) was obtained from Pennsylvania Sand and Glass Corp. (Pittsburgh, PA). All other minerals were obtained from primary milling and mining facilities in the United States. Bituminous coal mine dust was obtained from Blacksville, WV, coal mines, which represented the Pittsburgh coal seam. All the minerals were classified to < 5 pm size with the aid of an Accucut Particle Classifier (Donaldson-Majal Division, St. Paul, MN). Scanning electron microscopy (SEM) analysis with the aid of an automated image analysis system (LeMont), and X-ray spectrometric analysis on samples prepared on Nucleopore filters were made at xIOOO magnification to determine the elemental composition and particle diameter. Surface area measurements were made by the nitrogen adsorption technique according to the method of Brunauer et al. (13).

Electron Spin Resonance Measurements
Electronic spin resonance (ESR) measurements were made with the aid of a spin trap DMPO to monitor the generation of -OH radical from all the minerals. The reaction mixtures in a final volume of 1 ml contained 10 mg dust, 100 p1 1 M DMPO, and 100 pl 0.1 M H202. The reaction was initiated in a 5-ml plastic syringe by the addition of H202, mixed well in a vortex for 10 sec and filtered through a 0.45-tim nylon Acrodisc filter attached to the syringe. For the ESR measurements, 250-pl flat quartz cells were used as a sample container and measurements made on a Varian E109 ESR operating at x-band (-9.4 GHz) Environmental Health Perspectives ill frequency. A microwave power of 50 mW and modulation amplitude of 2 G were found to be adequate for the optimal development of signal peak heights without compromising resolution of the spectrum.

Cytotoxicty Measurements
Cytotoxic potential of all the minerals was measured by determining cellular membrane damage, i.e., hemolysis of red blood cells, and release of cytosolic enzyme LDH from AM, as well as the potential of the dusts to induce the selective release of lysosomal enzymes ,B-GLUC and 1-NAG. In addition, the potential to induce lipid peroxidation by all the dusts was measured as an index of free radical-induced cellular damage.

Hemolysis
Hemolytic potential of the minerals was measured based on the method by Harrington et al. (14) using a 2% suspension of sheep erythrocytes for incubation with the dusts (1 mg/ml) for 1 hr at 370C. The amount of hemoglobin released was measured in the supernatant after centrifugation by monitoring the absorbance at 540 nm. Percentage of hemoglobin released as a result of dust interaction was calculated as the ratio of absorbance value obtained from cells lysed with Triton X-100.

Isolation ofAlveolar Ma phages
For the enzyme release studies, AM were harvested from pathogen-free Sprague-Dawley rats by pulmonary lavage using calcium and magnesium-free Hank's balanced salt solution. Alveolar macrophages were centrifuged at 5OOg for 5 min at 40C and resuspended in HEPES-buffered medium containing 5 mM glucose. Trypan blue dye exclusion tests and microscopic estimates were made to determine the cell viability and differential cell populations (15). Viable AM were found to be the predominant cell type in all the lavages.

Enzyme Measurements
Alveolar macrophages (2 x 1 06/ml) were incubated with (1 mg/ml) and without dusts for 2 hr in a shaking water bath at 37°C. Incubation was terminated by rapid cooling and centrifugation. The supernatant was separated and used in all the three enzyme assays. LDH was estimated in a total reaction mixture of 3 ml containing phosphate buffer, pH 7.4, 100 pl enzyme supernatant, 0.07 mg/ml NADPH, and 100 pl of 0.0007 M sodium pyruvate (16). The reaction was initiated by the addition of sodium pyruvate and LDH activity was  (18). These lysosomal enzymes were also expressed as the percent of enzyme released as a result of dust interaction by comparing with the total enzyme present in the cells without dust treatment.

Surface Iron Measurements
Surface iron was measured using a spectrophotometric method according to Roth et al. (19). Minerals (10 mg) were treated with 5.5 ml 0.3 M sodium citrate, 1 M sodium bicarbonate, and 100 mg sodium dithionate and heated for 30 min at 80°C. Mixtures were centrifuged and supernatant diluted to 100 ml. Aliquots of the supernatant were treated with 1 ml 10% hydroxylamine hydrochloride and 2 ml 0.5% aphenanthroline for 5 min, diluted to 100 ml, and absorbance read at 508 nm. Iron standards were treated similarly and graphed for the conversion of surface iron present in mineral samples.

Lipid Peroxidation
Lipid peroxidation potential by all minerals was monitored by measuring the malondialdehyde (MDA) generated during the incubation of dust with linoleic acid for 1 hr in a buffered medium without the addition of any promoters. The reaction was terminated by the addition of 0.3 ml 5 N HCI and 0.625 ml 40% trichloroacetic acid (20). After adding and mixing 0.625 ml 2.0% thiobarbuturic acid to the reaction mixture, the tubes were heated in a water bath at 950C for 20 min. The thiobarbuturic acid reactive substance developed a color which was measured at 535 nm after cooling and centrifugation for 10 min at 600g. Malondialdehyde produced was calculated from a standard graph.

Results
Physical and Chemical Characteristics of Dusts The major physical and chemical characteristics of the dusts used in this study are presented in Table 1. The mass median aerodynamic diameter (MMAD) of all the dusts was < 5 pm. The particle diameter never exceeded 7 pm in dust samples and less than 2% of the dust exceeded the limit of 5 pm. However, several dusts showed large proportions of particles in the lower size range of less than 1 pm. These particle size differences are well reflected in the surface area measurements showing a wide variability between dusts (Table 1).

OH Radical Generation
All the minerals reacted with H202 in the presence of DMPO and produced DMPO-OH radical adducts showing typical 1:2:2:1 quartet signals. All the ESR signals were identical and showed only major differences in peak intensities. Figure 1 shows the typical ESR spectra obtained from crystalline silica (min-U-sil), talc, bentonite, and coal. The spectra shown for crystalline silica (  Maoneti Field (Gaus)  Figure 2 illustrates the relative potential of all minerals in equal mass concentration (10 mg/ml) in the generation of OH radicals from 10 mM H2O2 in the presence of 100 mM DMPO. It is evident from the data that coal mine dust generated the strongest -OH radical signal, while crystalline silica generated the weakest 0OH radical signal. Surface iron concentrations of coal mine dust and crystalline silica showed a direct correlation to the -OH radical generation (Figure 3). However, talc, with the highest concentration of surface iron, exhibited a relatively weak ESR signal compared to coal (Figure 2).

Cytotoxicity Studies
The hemolytic potential of all the minerals in equal mass concentration (1 mg/ml) is shown in Figure 4. Bentonite, kaolin, and crystalline silica showed the greatest hemolytic activity, whereas feldspar and coal were the least hemolytic. Release of cytosolic enzyme, LDH, and selective release of lysosomal enzymes, ,-GLUC and ,B-NAG, were maximal for kaolin, crystalline silica, and talc, and least for bentonite and coal in equal mass concentrations ( Figure 5). On the other hand, the combined cytotoxicity index of all three enzymes and hemolysis showed kaolin, feldspar, and crystalline silica most cytotoxic and coal and talc least cytotoxic in equal mass concentrations ( Figure 6). Figure 7 illustrates data on the lipid peroxidation potential of the different minerals in equal mass concentrations. The data illustrate that feldspar, crystalline silica, and talc induce the greatest lipid peroxidation, and kaolin, coal, and bentonite induce the least lipid peroxidation. It is interesting to note that coal and talc with greater concentrations of surface iron, and coal and bentonite with the greatest OH radical generation potential, were minimally effective in inducing lipid peroxidation. Figure 8 presents the combined cytotoxicity index based on equal surface area (0.1 Inm) of the minerals. The combined cytotoxicity index was calculated by adding all the three enzyme data (% released) and hemolysis percent. The data was then normalized for coal with a cytotoxicity index of 1. Coal was found to be least cytotoxic The data indicate a good correlation between the cytotoxicity index and MDA produced.
in the combined cytotoxicity on an equal mass basis. Similarly, the lipid-peroxidation data on an equal surface area basis were also normalized to 1 for coal. The data indicate a good positive correlation between cytotoxicity index and lipid peroxidation.

Discussion
In this study we evaluated the ability of several common occupational minerals to generate oxygen radicals from H,0,. The OH radical generation potential of the dusts was then compared with the cellular toxicity and potential to induce lipid per-OXidation. This investigation is important because several studies in recent years have implicated iron on the minerals as the causative agent inducing the generation of oxygen radicals leading to lipid peroxidation, cellular injury, and the disease process (21). The data presented here demonstrates that the concentration of surface iron in minerals is directly related to the potential of the dust to generate OH radicals. However, OH radical generation by these higher iron-containing minerals does not appear to be directly related to the cytotoxicity of minerals or to their potential to induce lipid peroxidation. In contrast, lipid peroxidation potential shows a better correlation with the known pathogenicity of dusts. Among the six minerals studied, crystalline silica (the most potent toxic and pathogenic mineral) showed the least ability to produce OH radicals from H202. This inability of OH radical generation by crystalline silica correlated well with the surface iron concentration, while in cytotoxicity measurements, silica was not the most cytotoxic compared to feldspar, bentonite, and kaolin (Figures 3-5). However, when lipid peroxidation (a specific irreversible cell membrane injury induced by OH radicals) was measured, crystalline silica induced the maximal lipid peroxidation. The significance of this lipid peroxidation potential parallels the cytotoxicity and pathogenicity of silica.
On the other hand, coal mine dust, which showed the maximal ability to generate -OH radicals with a corresponding high concentration of surface iron, is generally considered least cytotoxic and pathogenic in the absence of crystalline silica and other minerals. This correlates with the data presented here on the ability of coal to induce toxicity and lipid peroxidation. Therefore, the potential to produce the -OH radical alone may not be directly related to the ability of minerals to induce damage. Other factors and events involved in cytotoxicity, coupled with OH radical generation, are likely to be important in inducing the cell injury by "lipid peroxidationI.

Conclusion
Based on the current evidence, it appears that neither in vitro toxicity studies nor the ability of the minerals to generate the more potent oxidizing agents, such as OH from H 20., can directly predict the potential for toxicity and resultant pulmonary fibrosis. There appears to be a better correlation between the ability of the minerals to induce lipid peroxidation and klnown toxicity and fibrogenicity in humans.