Generation of Oxygen Radicals by Minerals and Its Correlation to Cytotoxicity Val Vallyathan Pathology Section, Division of Respiratory Disease Studies, National Institute for Occupational Safety and Health, Morgantown, West Virginia 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. - Environ Health Perspect 102(Suppl 10):1 1 1-1 15 (1994) Key words: oxygen radicals, cytotoxicity, pneumoconiosis, hemolysis, enzymes, surface iron, lipid peroxidation Introduction Inhalation of inorganic minerals and coal mine dust can produce pneumoconiosis, a lung disease characterized by an initial pul- monary inflammatory response in the lung associated with an increase in cell-mediated secretions of cytokines, growth factors, lysosomal enzymes, hydrogen peroxide, and superoxide anion (1-4). Several studies suggest that many of these factors work together to cause cell injury and the pro- gression of the disease process (3,4). The possible mechanisms leading to this fibrotic lung disease are under intense investiga- tion. To this end, we have devoted consid- erable effort during the last several years toward understanding the surface proper- ties and chemical structure of minerals and their relation to oxygen radical generation and related cytotoxicity (5-7). Several recent studies indicate that frac- turing of crystalline silica, which may occur in occupations such as sandblasting, tunnel- ing, rock drilling, or silica flour mill opera- tions, 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 (5-7). 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. Address correspondence to Val Vallyathan, Pathology Section, NIOSH, 1095 Willowdale Road, MS 211, Morgantown, WV 26505. Telephone (304) 291-4582. Fax (304) 284-5467. enhances the generation of OH radicals during phagocytosis (5,8-10). 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 surface- based 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 occu- pational mineral dusts in the generation of -OH radicals from hydrogen peroxide (H202) and to correlate in vitro cytotoxic- ity 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 5,5- dimethyl-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 corre- late 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 radi- cal-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 repre- sented 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 spec- trometric 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 nitro- gen adsorption technique according to the method of Brunauer et al. (13). Electron Spin Resonance Measurements Electronic spin resonance (ESR) measure- ments 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 con- tained 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 con- tainer and measurements made on a Varian E109 ESR operating at x-band (-9.4 GHz) Environmental Health Perspectives ill V. VALLYATHAN frequency. A microwave power of 50 mW and modulation amplitude of 2 G were found to be adequate for the optimal devel- opment of signal peak heights without compromising resolution of the spectrum. Cytotoxicty Measurements Cytotoxic potential of all the minerals was measured by determining cellular mem- brane 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 lyso- somal enzymes ,B-GLUC and 1-NAG. In addition, the potential to induce lipid per- oxidation 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% suspen- sion 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 centrifu- gation 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 of Alveolar Ma phages For the enzyme release studies, AM were harvested from pathogen-free Sprague- Dawley rats by pulmonary lavage using cal- cium 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 predomi- nant cell type in all the lavages. Enzyme Measurements Alveolar macrophages (2 x 1 0 6/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 super- natant 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 Table 1. Physical and chemical characteristics of minerals. Surface/area, Surface iron, Mineral/coal MMAD,apm m2/g % < 1pm % Sib mg/ 00 mg Coal 1.78 7.4 18 2.3 0.119 Feldspar 2.21 2.2 7 15.2 0.097 Kaolin 0.75 12.5 29 0 0.029 Bentonite 0.85 9.8 18 0.8 0.062 Talc 1.67 9.3 19 3.7 0.314 Silica 3.5 4.7 11 99 0.028 'MMAD, mass median aerodynamic diameter. bPercent concentration of silica by number as determined by X-ray spectrometric analysis of 1000 or more particles. calculated as the percent of enzyme released by comparing with the total enzyme released from cells lysed with Triton X-100. [-NAG was assayed accord- ing to the method of Lockhard and Kennedy (17) and [-GLUC was assayed according to the method of Sellinger et al. (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 spec- trophotometric 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 super- natant 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 malondi- aldehyde (MDA) generated during the incubation of dust with linoleic acid for 1 hr in a buffered medium without the addi- tion 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 reac- tion mixture, the tubes were heated in a water bath at 950C for 20 min. The thio- barbuturic acid reactive substance devel- oped 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 charac- teristics 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 sur- face 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 typi- cal ESR spectra obtained from crystalline silica (min-U-sil), talc, bentonite, and coal. The spectra shown for crystalline silica (Figure la), talc (Figure 1 b), and bentonite (Figure 1 c) are presented 10 times bigger than the coal spectra (Figure Id). Crystalline silica generated the least amount, whereas coal generated the maxi- mal amount of OH radicals. Controls with mixtures of DMPO and dusts and DMPO with H202 showed barely detectable sig- nals. The splitting constants of hyperfine couplings were characteristic of -OH radical adducts of DMPO. To confirm the -OH radical generation, we added a competitive -OH radical scavenger, ethanol, in increas- ing concentrations to the reaction mixture. The addition of 33% ethanol in the mix- tures with dusts and 0.1 M DMPO and 10 mM H202 resulted in a six-line spectrum characteristic of DMPO ethanolyl radical adducts. OH radical scavengers such as catalase and metal chelators such as deferox- amine, inhibited approximately 90% of the Environmental Health Perspectives 112 CYTOTOXICITYAND OXYGEN RADICAL GENERATION BY MINERALS TAlC (10 mg) + DMPO + H202 xlO B 3345 3370 ( 35 Magonetic Field (Gauss) I 370 335 Maoneti Field (Gaus) Figure 1. Typical ESR spectra obtained from (A)10 mg crystalline silica; (B), 10 mg talc; (C), 10 mg bentonite; (D), and 10 mg coal; with 10 mM H202 in the presence of 0.1 M DMPO. Note the gain used for the 10-mg coal sample, which produced a signal 18-fold stronger than the bentonite generating the maximal *OH radical generation on a equal mass basis. 1700 U) so - 4 0 20d li Si FELD TALC KAOL BDNT COAL MINERALS(1Omg/ml) Figure 2. Bar graph illustrating the relative intensities of DMPO-OH adduct signals generated by reacting 10 mg/ml minerals with 10 mM H202 in the presence of 100 mM DMPO as a spin trap. Note that the coal gen- erated signals 18-fold greater than the bentonite. -OH radical generation by dusts from H 22 Figure 2 illustrates the relative potential of all minerals in equal mass concentration (10 mg/ml) in the generation of OH radi- cals 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 crys- talline silica generated the weakest 0OH radical signal. Surface iron concentrations 00 z 0 N Ui) c; a VI so 0 VI 20-3 0 FELD COAL TALC St KAOL BDNT MINERALS(lmg/ml) Figure 4. Bar graph showing the percent hemolysis potential of minerals on an equal mass basis (1 mg/ml). Note that bentonite and kaolin are the most cytotoxic and feldspar and coal are the least cytotoxic minerals. so 7 0 80 5 0 C) N20- 10-3 BDNT COAL FEW TALC Si KAOL MINERALS(lmg/ml) Figure 5. Bar graph illustrating the relative release of enzymes (LDH, ,-GLUC, 5-NAG) from alveolar macro- phages on an equal mass basis (1 mg/ml). In this com- bined cytotoxicity index of enzymes, bentonite is the least cytotoxic and kaolin and silica are the most cyto- toxic minerals. Si KAOL BDNT FEW COAL TALC Figure 3. Surface iron concentration in the minerals. 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 sur- face 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, crys- talline silica, and talc, and least for ben- tonite 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 min- erals in equal mass concentrations. The data illustrate that feldspar, crystalline sil- ica, and talc induce the greatest lipid per- oxidation, 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 Volume 102, Supplement 10, December 1994 min-U-sul (10 mg) + DMPO + H202 x10 A 33o" 3370 s" Magn.oti FIel (Gaus) Bentonite (10 mg) + OMPO + H202 x10 CI 33s" 2" 70 e Maneric Field (Oues) 113 V. VALLYATHAN COAL TALC BENT Si FELT KAOL MINERALS)( 1 mg/mi) Figure 6. Bar graph representation of the combined toxicity index of hemolysis and LDH, f-GLUC, and f- NAG release on an equal mass basis. C.. 7 0 0 4 4 -3 0 KAOL COAL BENT TALC Si FELD MINERALS( 1 Omg/ml) Figure 7. Lipid peroxidation potential of the minerals. coal and bentonite with the greatest OH radical generation potential, were mini- mally effective in inducing lipid peroxida- tion. Figure 8 presents the combined cyto- toxicity 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 nor- malized for coal with a cytotoxicity index of 1. Coal was found to be least cytotoxic 6 uEAR B~~~~~~~~~~EST FIT S5 D 95% CONFIDENCE INTERVAL SILICA S 3 2 TALC ,@BNONITE KAOLIN 1 2 3 4 COMBINED CYTOTOXICITY INDEX/O. lm Figure 8. Comparison of the relative cytotoxicity index of minerals (hemolysis and enzyme release) on an equal surface area basis (0.1 mi2) vs lipid peroxidation. The data indicate a good correlation between the cyto- toxicity 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 indi- cate 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 peroxida- tion, 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 cytotox- icity of minerals or to their potential to induce lipid peroxidation. In contrast, lipid peroxidation potential shows a better corre- lation with the known pathogenicity of dusts. Among the six minerals studied, crys- talline silica (the most potent toxic and pathogenic mineral) showed the least ability to produce OH radicals from H 202. This inability of OH radical generation by crys- talline silica correlated well with the surface iron concentration, while in cytotoxicity measurements, silica was not the most cyto- toxic 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 signi- ficance 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 gener- ate -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 pre- sented 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 toxi- city and fibrogenicity in humans. REFERENCES 1. Craighead JE, Kleinerman J, Abraham JL, Gibbs AR, Green FHY, Harley RA, Ruettner JR, Vallyathan NV, Juliano EB. Diseases associated with exposure to silica and nonfibrous sili- cate minerals. Arch Pathol Lab Med 112:673-720 (1988). 2. Davis GS. Pathogenesis of silisosis: current concepts and hypothesis. Lung 164:139-154 (1986). 3. Davis GS. The pathogenesis of silicosis: state of the art. Chest 89:166-169 (1986). 4. Heppleston AG. Silicotic fibrogenesis: a concept of pulmonary fibrosis. Ann Occup Hyg 26:449-462 (1986). 5. Vallyathan V, Shi X, Dalal NS, Irr W, Castra;nova V. Generation of free radicals from freshly fractured silica dust. Potential role in acute silica-induced lung injury. Am Rev Respir Dis 138:1213-1219 (1988). 6. Shi X, Dalal NS, Vallyathan V. ESR evidence for the hydroxyl radical formation in aqueous suspension of quartz particles and 114 Environmental Health Perspectives CYTOTOXICITYAND OXYGEN RADICAL GENERATION BY MINERALS its possible significance to lipid peroxidation in silicosis. J Toxicol Environ Health 25:237-245 (1988). 7. Dalal NS, Shi X, Vallyathan V. Role of free radicals in the mechanisms of hemolysis and lipid peroxidation by silica: com- parative ESR and cytotoxicity studies. J Toxicol Environ Health 29:307-316 (1988). 8. Vallyathan V, Mega JF, Shi X, Dalal NS. Enhanced generation of free radicals from phagocytes induced by mineral dusts. Am J Respir Cell Mol Biol 6:404-413 (1992). 9. Kuhn DC, Demers LM. Influence of mineral dust surface chemistry on eicosanoid production by the alveolar macrophages. J Toxicol Environ Health 35:39-50 (1992). 10. Castranova V, Vallyathan V, Van Dyke K, Dalal NS. Use of chemiluminescence assays to monitor the surface characteristics and biological reactivity of freshly fractured versus aged silica. In: Effects of Mineral Dusts on Cells, NATO ASI Series, Vol 30 (Mossman BT, Begin RO, eds). New York:Springer-Verlag, 1989;18 1-18 1. 11. Dalal NS, Suryan MM, Vallyathan V, Green FHY. Electron spin resonance detection of reactive free radicals in fresh coal dust and quartz and its implication to pneumoconiosis and sili- cosis. In: Respirable Dust in Mineral Industries: Health Effects Characterization and Control (Frantz RL, Ramani RV, eds). University Park, PA:Pennsylvania State University Press, 1988;20-24. 12. Dalal NS, Suryan MM, Vallyathan V, Green FHY, Jafari B, Wheeler R. Detection of reactive free radicals in fresh coal mine dust and their implication for pulmonary injury. Ann Occup Hyg 33:79-84 (1989). 13. Brunauer SP, Emmett PH, Teller ET. Absorption of gases in multi-molecular layers. J Am Chem Soc 60:309-319 (1938). 14. Harington JS, Miller K, Macnal G. Hemolysis by asbestos. Environ Res 4:95-117 (1971). 15. Phillips HJ. Dye exclusion tests for cell viability. In: Tissue Culture Methods and Applications (Kruse PR, Patterson K, eds). New York:Academic Press, 1973;406-408. 16. Reeves WJ, Fimignari GM. An improved procedure for the preparation of crystalline lactic dehydrogenase from hog heart. J Biol Chem 238:3853 (1963). 17. Lockard VG, Kennedy RE. Alterations in rabbit alveolar macrophages as a result of traumatic shock. Lab Invest 35: 501-506 (1976). 18. Sellinger OZ, Beufay H, Jacques P, Doyan A, Deduve C. Tissue fractionation studies, intracellular distribution and prop- erties of P-N-acetyl-glucosaminidase and ,B-galactosidase in rat liver. Biochem J 74:450 (1960). 19. Roth CB, Jackson ML, Syers JK. Clays and clay minerals. 17:253-267 (1969). 20. Hunter FE Jr, Gelicki JM, Hoffsten PE, Weinstein J, Scott A. Swelling and lysis of rat liver mitochondria induced by ferrous ions. J Biol Chem 238:828-832 (1963). 21. Ghio AJ, Kennedy TP, Whorton AR, Crumbliss AL, Hatch GE, Hoidal JR. Role of surface complexed iron in oxidant gen- eration and inflammation induced by silicates. Am J Physiol 263:511-518 (1992). Volume 102, Supplement 10, December 1994 115