ENVIRONMENTAL RESEARCH 56, 57--67 (1991)
The Toxicity, in Vitro, of Silicon Carbide Whiskers GERALD L . V A U G H A N , JACQUELINE JORDAN, AND SUSAN K A R R
Zoology Department, University of Tennessee, Knoxville, Tennessee 37916 Received December 4, 1990 To mouse cells in culture, SiC whiskers (SiCW) and asbestos are similarly cytotoxic, disrupting cell membranes and killing cells. Both shorten cell generation time, increase the rate of DNA synthesis, increase total ceil DNA content, and cause a loss in growth control often associated with malignant cellular transformation. Within the narrow size range of materials examined, the amount of damage appeared to be more a function of the number of whiskers present than of their size. Silicon carbide whiskers, if mishandled, may pose a serious health hazard to humans. © 1991 AcademicPress, Inc.
INTRODUCTION Interest in the biological effects of manmade mineral whiskers and fibers relates to the fact that there are a variety of serious disease states associated with exposure to similar naturally occurring materials such as zeolite and asbestos. These conditions range from simple eye, skin, and pulmonary irritations to lifethreatening pulmonary fibrosis, pneumoconeosis, and cancer. Although conditions associated with pulmonary fibrosis are, perhaps, of greater cost and consequence to society, the major concern of research appears to have centered on carcinogenesis. A variety of fibrous mineral materials, including asbestos, silicon carbide whiskers (SiCW), alumina, and glass, have been examined. Stanton et al. (1981), correlated particle dimension and geometry with carcinogenicity for 72 fibrous and nonfiberform materials and concluded that durability, size, and shape rather than physiochemical properties were the major determinants of malignant cellular transformation. He found the best correlation with fibers longer than 8.0 and 0.25 ~m or less in diameter. Glass, for instance, is thought to have little or no cytotoxic, genotoxic, or carcinogenic characteristics in powdered form or as coarse fibers. Small glass fibers, on the other hand, are as effective in transforming cells in vitro as are some forms of asbestos (Lechner et al., 1983; Oshimura et al., 1984; Sincock, 1977; Stanton et al., 1977). Reduction in diameter from 1.0 to 0.1 I~m elicits an order of magnitude increase in frequency of cytogenetic aberrations produced by fiberglass (Oshimura et al., 1984). Milling the same fibers to dust abolishes the effect. Asbestos exists in a variety of forms which differ in the severity of effects they have on cells with chrysotile being the most and amosite the least effective (Lechner et al., 1983; Lemaire et al., 1982; Oshimura et al., 1984; Sincock, 1977). Again, milling or, in some cases, chemical leaching, reduces or abolishes the response. At least a portion of the differences between the toxicities of the various asbestos forms, including those that have been leached, may stem from differences in durability or surface structure of the fiber. These factors 57 0013-9351/91 $3.00 Copyright© 1991by AcademicPress, Inc. All fightsof reproductionin arty formreserved.
58
VAUGHAN, JORDAN, AND KARR
should be considered in risk assessment for whisker/fiber materials (Mossman and Gee, 1989; Mossman et al., 1990). In the last few years a variety of fiberform materials such as single-crystal mineral whiskers, carbon filaments, and spun ceramic insulating materials have entered the marketplace. As these are nonfibrous materials, most are classed as nuisance dusts. The emergent biological effects of these materials when produced as fibers or whiskers deserve serious attention. In this report we discuss results from experiments conducted as part of a broad investigation of one of these materials, SiCW. The goal of this portion of the study has been, using in vitro methods, to make an initial determination of cytotoxicity for SiCW. We conclude that, although it is relatively innocuous in nonfiber particulate form, SiC is likely to pose a serious health hazard as a fine whisker. Indeed, in this form, it appears to have cellular effects similar to those of crocidolite asbestos (Vaughan, 1989). MATERIALS AND METHODS Test Materials
Two silicon carbide whisker samples were used in this study, one of 0.8 ~m diameter (SiCW-1) manufactured by Tateho, Japan, and one of 1.5 ~xm diameter (SiCW-2) manufactured by American Matrix, Inc., Knoxville, TN. A sample of crocidolite asbestos was kindly provided by Dr. L. W. Ortiz of Los Alamos National Laboratory, Los Alamos, NM. Dimensional characteristics of each of these materials were determined from direct measurement of at least 100 images in scanning electron micrographs (1000--2000X magnification factor). Size standards (latex beads, 2.68 ~m diameter) were included for scale in each micrograph. These measurements, made to the nearest 0.05 p~m are summarized in Table 1. SiOz dust, purchased from Sigma Chemical Company, St. Louis, MO, at >99% purity was boiled in 1.0 N HCI, washed four times in distilled H20, and dried before use. Samples of each test material were suspended by sonication in phosphate buffered saline (PBS) (grams per liter: NaC1, 8:00; KC1, 0.20; KHzPO 4, 0.20; NazHPO 4 • 7H20, 2.16; adjusted to pH = 7.4) at a concentration of 1.0 mg/ml and were sterilized by autoclaving before use. Since these materials are not soluble and settle to the surface of test plates, and since depth of fluid in various culture vessels varies, concentration is always expressed as mass of material per surface area, or as number of particles per cell or surface area.
TABLE1 DIMENSIONAL CHARACTE~STICS OF TEST MATE~ALS Material
Diameter ~
Length a
Aspect ratio(L × D -1)
SiCW-1 SiCW-2 Asbestos
0.8 (0.3) 1.5 (0.6) 0.4 (0.2)
18.1 (14.3) 19.0 (11.0) 4.3 (0.2)
23.3 (18.7) 15.3 (11.2) 14.3 (3.8)
Note. Numbers in parentheses represent the standard deviation.
Dimensions given in micrometers.
T O X I C I T Y OF SILICON CARBIDE W H I S K E R S
59
Cell Culture BALB/3T3 embryonic mouse cells (clone A31) used in this study were obtained from American Type Culture Collection, RockviUe, MD and maintained in minimum essential medium (MEM) with Earles salts supplemented with 10% calf serum and antibiotics (penicillin-streptomycin) at 37°C in an atmosphere of 5% CO2 in air. Cultures were divided 1:3 by trypsinization upon reaching 80% confluence. To retard spontaneous loss of contact inhibition, cells were allowed to form a complete monolayer only when required by experimental design. Trypsin and components of the culture medium were procured from GIBCO, Grand Island, NY. Determinations of cell titer for various experiments were made with a Coulter (Model ZM) electronic cell counter or by hemocytometer. Cytotoxicity as Determined by Dye Exclusion Trypan blue is excluded from healthy living cells but enters moribund cells or those with leaky membranes, staining them blue. Dye exclusion is commonly used as an indicator of viability in tissue culture (Kaltenbach et al., 1958). Crocidolite, SiCW, or an appropriate quantity of saline vehicle was added to suspensions of cells just prior to their distribution to fresh petri dishes to give 5.0, 10.0, 15.0, or 20.0 p~g/cm2 of plate surface. Dye exclusion tests were performed 24 hr later. For this purpose the medium (containing dead cells) was decanted and combined with the remainder of cells removed from the plate by trypsinization. Cells were collected by centrifugation, resuspended in saline (0.5 ml), and mixed with an equal volume of isotonic saline containing trypan blue (0.4%). Results were recorded as a ratio of the number of stained cells to the total counted (at least 200 individuals per slide). Ten replicate cultures were analyzed for each condition of treatment. Cytotoxicity as Indicated by 51Cr Release When cells are allowed to accumulate radioactive chromium (51Cr) and then placed in growth medium without the radioactive label, the amount of 51Cr that leaks relative to the control is a reflection of membrane damage or cell death (Zawydiwski and Duncan, 1978). Cells were seeded at 3.0 × 10 3 cells per well in a 92-well multiwell plate and incubated 24 hr to achieve approximately 50% confluency before exposure to 51Cr (25 p~Ci/ml) in fresh complete medium at 37° C for 3 hr. The labeled medium was removed, and the cells were washed five times then exposed to complete medium containing SiCW or asbestos at a final concentration of 5.0 p~g/cm2. Control wells contained complete medium only. After 18 hr, a sample of supernatant was taken for determination of leaked 51Cr by liquid scintillation (Beckman Model LS 8100). The cells, still attached to the plates, were washed twice with complete medium after which they were covered with distilled water which was frozen and thawed three times to ensure total cell lysis. Radioactivity in the lysate was measured to determine the amount of label remaining in the cells. Control values were subtracted from experimental values and the results were expressed as percentage radiochromium leakage. Colony-Forming Efficiency To measure the effects of the test materials on proliferative ability, cells were
60
VAUGHAN, JORDAN, AND KARR
seeded at 300 cells per 60-mm culture dish and incubated for 24 hr to allow attachment before treatment with the appropriate test material at varying concentrations. After a 7-day incubation to allow for development of colonies, plates were fixed in methanol and stained with Giemsa stain. Colonies were counted manually by inverted phase microscopy. Only colonies of 50 or more cells were scored. Control cultures with untreated cells of similar passage were assayed concurrently.
Tritiated Thymidine Incorporation Assay The rate of DNA syntheses in cells exposed to whiskers or fibers was determined by measuring the incorporation of [3H]thymidine (New England Nuclear, 100 ixCi/mM) into DNA. Cells were exposed to test materials suspended in complete medium at concentrations ranging from 0.0 to 2.0 ~g/cm 2 and maintained in culture. At various times, cultures derived from those which had been exposed to test materials were exposed to [3H]thymidine (2.0 ~xCi/ml) in MEM for 2 hr after an initial determination that uptake and incorporation were linear over a 5-hr period. Labeled medium was removed by aspiration and the cells were washed twice with ice-cold PBS, then harvested by scraping or trypsinization. The amount of [3H]thymidine which had been incorporated into DNA was determined in trichloracetic acid-extracted cells by liquid scintillation (Beckman Model LS230).
Incidence of Binuclear Cells Cells were seeded at 1 × 10 6 cells per 25-cm z flask in complete medium and allowed to attach for 24 hr as described above, then treated with PBS or the appropriate test material at 5 ~g/cm 2. After incubation for 48 hr, 200 cells per flask were scored for multinuclearity by phase contrast microscopy.
Cellular Transformation Transformation frequency, reflected in loss of contact inhibition, was determined for treated and untreated cultures maintained concurrently. Flasks were seeded at 2.0 × 105 cells per 25-cm 2 flask in complete medium, allowed to attach overnight, then exposed to the appropriate whisker or fiber at 5 ixg/cm 2 for 24 hr. Cultures were split 1:3 upon reaching 80% confluency. One flask at each passage was used to propagate the culture which was maintained for 10 passages. The other two flasks at each passage were incubated 10-14 days, fixed in methanol, stained with Giemsa and scanned for focal colonies of transformed cells with a stereomicroscope at 60x total magnification. To be scored positive, foci had to exhibit layered cell growth and the swirling pattern characteristic of transformed colonies (Freshney, 1987).
Cellular DNA Content In whisker-exposed cells, total DNA content was estimated using the method of Keck (1956) based upon colorimetric evaluation of the colored reaction product of indole and DNA.
TOXICITY OF SILICON CARBIDE WHISKERS
61
Data Analysis Analysis of variance (ANOVA) and Student's two-tailed and single-group t tests were used for statistical comparisons and analyses of experimental results. RESULTS Within 24 hr of being added to cell cultures, many, perhaps a majority, of the whiskers/fibers that can be detected by phase contrast microscopy are found associated with the cells, attached to cell surfaces, or internalized. As demonstrated by the micrographs in Fig. 1, SiCW whiskers too large to be engulfed are found penetrating cell surfaces, often entering the cell on one side, and exiting on the other as if the cell were "skewered." It seems likely that transmembrane particles, compromising membrane integrity, are responsible for much of the cytotoxicity we observe to be associated with asbestos and SiCW. This possibility is circumstantially supported by the results of cytotoxicity assays which are based upon evaluation of membrane selectivity. The dye, trypan blue, is excluded by healthy living cells but not by moribund or dead cells which stain blue. Based on dye exclusion, crocidolite asbestos and SiCW-1 exhibited similar levels of dosedependent cytotoxicity within the first 24 hr (Fig. 2). Similar results were obtained with the radiochromium release assay wherein cells were allowed to accumulate 51Cr and then placed in growth medium without the label where they are exposed to the test materials. As cells die, or are damaged, plasma membranes deteriorate and lose natural selective permeability. As a consequence, 5~Cr leaks to the medium surrounding the cell. As shown in Fig. 3, cells in cultures exposed to SiCW-1 or crocidolite at 5.0 ~g/cm 2 release 20-30% of label in excess of controls, while SiO2 induces an excess loss of approximately 8.0%. Dye exclusion and radiochromium release assays do not report cells which survive the first 24 hr to be compromised and die later. Analysis of colony-forming efficiency, however, measures cellular proliferative capacity. Cells must not only survive the first 24 hr but must also be sufficiently viable to produce individual colonies of at least 50 cells each. Colony formation efficiency curves for crocidolite asbestos, SiCW-1, and SiCW-2 are shown in Fig. 4 where colony-forming efficiency is plotted as a function of whisker/fiber concentration. On a mass per surface area basis, SiCW-1 is slightly more toxic than equal quantities of crocidolite. The larger SiCW is less cytotoxic than SiCW-1 (P < 0.01) but not significantly different from asbestos (P > 0.05). This finding as to the relative toxicities of SiCW-1 and SiCW-2 seems to follow from the common understanding that fibers of smaller diameter are more toxic than larger ones. On the other hand, consider the obvious fact that a given mass of small diameter fibers will contain more particles than the same mass of larger fibers of similar length. From the known density of SiC (3.2) and the dimensions given in Table 1, we have estimated the number of whiskers per gram for both SiCW samples. In Fig. 5 we present cytotoxicity as a function of the number of fibers to which the cells are exposed rather than mass. On a numerical basis, we found no significant difference (P < 0.05) between the cytotoxicities of 0.8 and 1.5 p~m SiCW of similar lengths. In addition to being immediately cytotoxic, SiCW-1, SiCW-2, and crocidolite
62
VAUGHAN, JORDAN, AND KARR
TOXICITY OF SILICON CARBIDE WHISKERS
63
induce, within eight generations of exposure, changes in cellular growth habits and structure generally held to be characteristic of cellular transformation. As determined by the appearance, within culture, of cells that have lost contact inhibition, transformation frequencies resulting from exposure to SiCW relative to saline controls are shown in Fig. 6. Not shown in Fig. 6, crocidolite asbestos produced a transformation frequency of 0.000069 (SD = 0.000046). As with the cytotoxicity values in Fig. 4, these values are a function of dosage expressed as mass per area. When the data for SiCW-2, the larger whisker, are recalculated so that a comparison of transformation frequencies can be made on the basis of the number of particles rather than mass (Fig. 7), there is, as with colony-forming efficiency, little difference between the effects of SiCW-1 and SiCW-2. Perhaps the most significant alteration in cells exposed to SiCW and asbestos occurred at the level of the genome. In a crude fashion this was initially observed as an increase in the number of binucleate cells (six to eight fold). Lemaire et al., (1982) have suggested that exposure to asbestos and other fiberform materials alters the rate of synthesis of DNA, and that increases in DNA synthesis rate are proportional to the carcinogenicity of the material. We found that, although DNA synthesis rates for fiber-/whisker-exposed cells were generally elevated relative to controls, often by a factor of as much as 2.5, the results were inconsistent and not reported here. Difficulties in interpretation resulted from the fact that cultures transformed by fiber/whisker treatment were often more mitotically active, with generation times ranging from the control value of approximately 20 hr to as low as 10 hr. It appears that this observation was not considered by LeMaire et al. Significant increases in total cellular DNA content, however, were consistently observed 10-20 generations after treatment. Cells treated with SiCW and crocidolite contained an excess of DNA (Table 2) ranging from approximately 40% to near 70%. This observation is consistent with those in other systems where transformed cells and those cells associated with malignancies have elevated levels of DNA (Vanderlaan et al., 1983; Schieck et al., 1987). DISCUSSION There is ample evidence that, within a narrow size range, durable whisker and fiberform dusts generally manifest significant cytotoxic effects. Based upon the experimental results presented here, the silicon carbide and crocidolite samples we have tested are no exceptions to the rule. The SiCW materials we examined were similar in physical dimensions with the exception that SiCW-2 was approximately twice the diameter of SiCW-1 and the crocidolite sample (Table 1). Judging from dye exclusion measurements (Fig. 2) and radiochromium leak rates (Fig. 3) for SiCW-1 and crocidolite, the two materials are similarly cytotoxic. Of the FIG. 1. Scanning electron micrographs of two cell types 24 hr after exposure to SiCW in vitro. Photo la depicts cultured dog tracheal epithelial cells transfixed by SiCW-2, and Photo lb shows particles of SiCW-2 which have become attached to or engulfed by a mouse 3T3 cell in culture. Cell samples were fixed with glutaraldehyde followed by osmium tetroxide in sodium cacodylate buffer, dehydrated, and critical point dried. Dried samples were coated with gold/palladium and photographed using an ETEC Antoscan scanning electron microscope. The horizontal bar represents 5.0 p.m for scale comparison.
64
V A U G H A N , JORDAN, A N D KARR (dye
L
exclusion)
2 o o
/
x
//x
lO
-A-
d_ x
/,,! I F z © © 0
3O 20 10 0
-
0
5
10
15
20
CONCENTRATION (ug/cm 2) FIG. 4. Cell proliferative ability, measured by colony-forming efficiency, is shown as a function of mass-based particulate concentration. Colony-forming efficiency is expressed as a percentage of control values. Vertical bars, in this figure, represent standard error of the mean.
SiCW-2. It seems likely that, within the narrow size range we examined, cellular response probably depends more on the number of particles a cell encounters than on relative size. Although the processes involved in cellular transformation and those in the development of malignancy are not necessarily identical, they are similar. When, therefore, a material is shown to be an agent capable of transforming cells, there
>(9 Z uJ O LL
LL UJ © Z
(% control) 00' 8O 70 6O 5O
LL
>Z 0 J
0 O
SiCW-
-o--
SiCW-2
1
.z
4o
0
-o-
"
30 20
~ _ ~
lO
[
0
,
O
+
,
1
,
2
,
i ~
3
,
4
i
5
l ,
i
6
,
J
7
CONCENTRATION ( 105 whiskers/crn 2) F I 6 . 5. Colony-forming efficiency data for SiCW-1 and SiCW-2 from Fig. 4 are replotted as a
function of the number of particles per area rather than the mass of particles per area. Vertical bars represent standard deviation about the mean.
66
V A U G H A N , JORDAN, AND KARR ~..
t .oo
o x z 0 ~,4
SlOW conc. 5.0 ug/cm 2
0.80
0,60
IX
0
LL d) Z 4~ [[ F-
o~
0.40
0.20
O.00 SALINE CONTROL SiCW O.8uM DIA
WHISKER
SiCW
1.SUM DIA
TYPE
FIG. 6. The frequency of cellular transformation is given for cultures of mouse 3T3 cells exposed to SiCW-I, SiCW-2, or saline (control). The particulate dose was 5.0 mg/cm 2. Vertical bars are standard deviation about the mean.
is cause for concern. In light of the general hypothesis that cell transforming and carcinogenic potentials in durable whisker/fiberform materials are largely functions of particle size and aspect ratio (Stanton, et al., 1981), it was not unexpected that SiCW-I and crocidolite would be more effective than SiCW-2 in cell transformation capability. When, however, transformation frequency for SiCW is shown as a function of the relative number of whiskers rather than mass (Fig. 7), there is no difference between SiCW-1 and SiCW-2. This supports the probability that, as with cytotoxicity, the effects of these materials within this size frame depend more upon the number of fibers than on the size. Finally, reinforcing our conclusion that SiCW may pose a serious health hazard, is the fact that exposure to SiCW and crocidolite caused a large increase in cellular DNA content relative to untreated controls. Taken as a whole, such changes as we describe are common in malignant tumors and as a concomitant of carcinogeninduced cellular transformation (Vanderlaan et al., 1983). ~.
1.00
b Z
080
Z 0 b-
O.60
0 LL
0.40
O3
Z II
0.20
F0.00 CONTROL
SiCW- 1
SJOW-2
FIG. 7. Frequency of cellular transformation is presented as in Fig. 6 except that the data for SiCW-2 are recalculated to reflect a comparison based upon the number of particles rather than their mass.
TOXICITY OF SILICON CARBIDE WHISKERS
67
TABLE 2 TOTAL DNA CONTENT, MOUSE CELLS (BALB/3T3) Treatment SiCW-1 SiCW-2 Asbestos Saline
DNA/cell (× 10-11 g) 3.48 3.54 4.37 2.49
(0.81)* (0.98)* (1.21)* (0.58)
Excess
n
39% 42% 75% --
32 29 18 45
Note. Numbers in parentheses represent the standard deviation. * Significantly different from control (P < 0.05, ANOVA).
ACKNOWLEDGMENTS The necessary funding and support for this research were provided by the University of Tennessee, Advanced Composite Materials Corporation, and American Matrix, Inc. The authors are also grateful for the personal support and encouragement of Dr. Sam Weaver, Third Millenium Technologies, Inc.
REFERENCES Fre shney, R. I. (1987). "Culture of Animal Cells: A Manual of Basic Technique." R. Lis s, New York. Hoch-Ligeti, Sass, B., Sobel, H. J., and Stewart, H. L. (1983). J. Nat. Cancer Inst. Endocardial tumors in rats exposed to durable fibrous materials. 71(5). Kaltenbach, J. P., Kaltenbach, M. H., and Lyons, W. B. (1958). Nigrosin as a dye for differentiating live and dead ascites cells. Exp. Cell Res. 15, 112-117. Keck, K. (1956). An ultramicro technique for the determination of deoxypentose nucleic acid. Arch. Biochem Biophys. 53,446--451. Lechner, J. F., Haugen, A., Trump, B. F., Tokiwa, T., and Harris, C. C. (1983). Effects of asbestos and carcinogenic metals on cultured human brochial epithelium. Hum. Carcinog. 51, 561-585. Lemaire, I,, Gingras, D., Lemaire, S. (1982). Tbymidine incorporation by lung fibroblasts as a sensitive assay for biological activity of asbestos. Environ. Res. 28, 399--409. Mossman, B. T., Bignon, J., Corn, M., Seaton, A., and Gee, J. B. L. (1990). Asbestos: Scientific developments and implications for public policy. Science 247, 294-301. Mossman, B. T. and Gee, J. B. L., (1989). Asbestos-related diseases. N. Engl. J. Med. 320, 17211730. Osbimura, M., Hesterberg, T. W., Tsutsui, T., and Barret, J. C. (1984). Correlation of asbestosinduced cytogenetic effects with cell transformation of Syrian hamster embryo cells in culture. Cancer Res. 44, 5017-5022. Schieck, R., Taubert, G., and Krug, H. (1987). Dry mass, DNA and nonhistone protein determinations in lung cancer cells. Histochem. J. 19, 504-508. Sincock, A. M. (1977). Preliminary studies of the in vitro cellular effects of asbestos and fine glass dusts. Origins Hum. Cancer Cold Spring Harbor Lab. Stanton, M. Layard, M., Tegeus, A., Miller, E., May, M. and Kent, E. (t977). Carcinogenicity of fibrous glass: Pleural response in the rat in relation to fiber dimension. J. Nat. Cancer Inst. 58(3), 587-603. Stanton, M. F., Layard, M., Tegeris, A., Miller, E., May M., Morgan, E., and Smith, A. (1981). Relation of particle dimension to carcinogenicity in amphibole asbestoses and other fibrous minerals. J. Nat. Cancer Ins. 67(5), 965-975. Vanderlaan, M. Steel, V., and Nettesheim, P. (1983). Increased DNA content as an early marker of transformation in carcinogen-exposed rat tracheal cell cultures. Carcinogenesis 4(6), 721-727. Vaughan, G. L. (1989). Biological consequences of exposure to silicon carbide whiskers. In "91st Annual Meeting of American Ceramic Society, Indianapolis, April 1989." Zawydiwski, R., Duncan, G. R., (1978). Spontaneous 51Cr release by isolated rat hepatocytes: An indicator of membrane damage. In Vitro 14, 707-714.
The Toxicity, in Vitro, of Silicon Carbide Whiskers GERALD L . V A U G H A N , JACQUELINE JORDAN, AND SUSAN K A R R
Zoology Department, University of Tennessee, Knoxville, Tennessee 37916 Received December 4, 1990 To mouse cells in culture, SiC whiskers (SiCW) and asbestos are similarly cytotoxic, disrupting cell membranes and killing cells. Both shorten cell generation time, increase the rate of DNA synthesis, increase total ceil DNA content, and cause a loss in growth control often associated with malignant cellular transformation. Within the narrow size range of materials examined, the amount of damage appeared to be more a function of the number of whiskers present than of their size. Silicon carbide whiskers, if mishandled, may pose a serious health hazard to humans. © 1991 AcademicPress, Inc.
INTRODUCTION Interest in the biological effects of manmade mineral whiskers and fibers relates to the fact that there are a variety of serious disease states associated with exposure to similar naturally occurring materials such as zeolite and asbestos. These conditions range from simple eye, skin, and pulmonary irritations to lifethreatening pulmonary fibrosis, pneumoconeosis, and cancer. Although conditions associated with pulmonary fibrosis are, perhaps, of greater cost and consequence to society, the major concern of research appears to have centered on carcinogenesis. A variety of fibrous mineral materials, including asbestos, silicon carbide whiskers (SiCW), alumina, and glass, have been examined. Stanton et al. (1981), correlated particle dimension and geometry with carcinogenicity for 72 fibrous and nonfiberform materials and concluded that durability, size, and shape rather than physiochemical properties were the major determinants of malignant cellular transformation. He found the best correlation with fibers longer than 8.0 and 0.25 ~m or less in diameter. Glass, for instance, is thought to have little or no cytotoxic, genotoxic, or carcinogenic characteristics in powdered form or as coarse fibers. Small glass fibers, on the other hand, are as effective in transforming cells in vitro as are some forms of asbestos (Lechner et al., 1983; Oshimura et al., 1984; Sincock, 1977; Stanton et al., 1977). Reduction in diameter from 1.0 to 0.1 I~m elicits an order of magnitude increase in frequency of cytogenetic aberrations produced by fiberglass (Oshimura et al., 1984). Milling the same fibers to dust abolishes the effect. Asbestos exists in a variety of forms which differ in the severity of effects they have on cells with chrysotile being the most and amosite the least effective (Lechner et al., 1983; Lemaire et al., 1982; Oshimura et al., 1984; Sincock, 1977). Again, milling or, in some cases, chemical leaching, reduces or abolishes the response. At least a portion of the differences between the toxicities of the various asbestos forms, including those that have been leached, may stem from differences in durability or surface structure of the fiber. These factors 57 0013-9351/91 $3.00 Copyright© 1991by AcademicPress, Inc. All fightsof reproductionin arty formreserved.
58
VAUGHAN, JORDAN, AND KARR
should be considered in risk assessment for whisker/fiber materials (Mossman and Gee, 1989; Mossman et al., 1990). In the last few years a variety of fiberform materials such as single-crystal mineral whiskers, carbon filaments, and spun ceramic insulating materials have entered the marketplace. As these are nonfibrous materials, most are classed as nuisance dusts. The emergent biological effects of these materials when produced as fibers or whiskers deserve serious attention. In this report we discuss results from experiments conducted as part of a broad investigation of one of these materials, SiCW. The goal of this portion of the study has been, using in vitro methods, to make an initial determination of cytotoxicity for SiCW. We conclude that, although it is relatively innocuous in nonfiber particulate form, SiC is likely to pose a serious health hazard as a fine whisker. Indeed, in this form, it appears to have cellular effects similar to those of crocidolite asbestos (Vaughan, 1989). MATERIALS AND METHODS Test Materials
Two silicon carbide whisker samples were used in this study, one of 0.8 ~m diameter (SiCW-1) manufactured by Tateho, Japan, and one of 1.5 ~xm diameter (SiCW-2) manufactured by American Matrix, Inc., Knoxville, TN. A sample of crocidolite asbestos was kindly provided by Dr. L. W. Ortiz of Los Alamos National Laboratory, Los Alamos, NM. Dimensional characteristics of each of these materials were determined from direct measurement of at least 100 images in scanning electron micrographs (1000--2000X magnification factor). Size standards (latex beads, 2.68 ~m diameter) were included for scale in each micrograph. These measurements, made to the nearest 0.05 p~m are summarized in Table 1. SiOz dust, purchased from Sigma Chemical Company, St. Louis, MO, at >99% purity was boiled in 1.0 N HCI, washed four times in distilled H20, and dried before use. Samples of each test material were suspended by sonication in phosphate buffered saline (PBS) (grams per liter: NaC1, 8:00; KC1, 0.20; KHzPO 4, 0.20; NazHPO 4 • 7H20, 2.16; adjusted to pH = 7.4) at a concentration of 1.0 mg/ml and were sterilized by autoclaving before use. Since these materials are not soluble and settle to the surface of test plates, and since depth of fluid in various culture vessels varies, concentration is always expressed as mass of material per surface area, or as number of particles per cell or surface area.
TABLE1 DIMENSIONAL CHARACTE~STICS OF TEST MATE~ALS Material
Diameter ~
Length a
Aspect ratio(L × D -1)
SiCW-1 SiCW-2 Asbestos
0.8 (0.3) 1.5 (0.6) 0.4 (0.2)
18.1 (14.3) 19.0 (11.0) 4.3 (0.2)
23.3 (18.7) 15.3 (11.2) 14.3 (3.8)
Note. Numbers in parentheses represent the standard deviation.
Dimensions given in micrometers.
T O X I C I T Y OF SILICON CARBIDE W H I S K E R S
59
Cell Culture BALB/3T3 embryonic mouse cells (clone A31) used in this study were obtained from American Type Culture Collection, RockviUe, MD and maintained in minimum essential medium (MEM) with Earles salts supplemented with 10% calf serum and antibiotics (penicillin-streptomycin) at 37°C in an atmosphere of 5% CO2 in air. Cultures were divided 1:3 by trypsinization upon reaching 80% confluence. To retard spontaneous loss of contact inhibition, cells were allowed to form a complete monolayer only when required by experimental design. Trypsin and components of the culture medium were procured from GIBCO, Grand Island, NY. Determinations of cell titer for various experiments were made with a Coulter (Model ZM) electronic cell counter or by hemocytometer. Cytotoxicity as Determined by Dye Exclusion Trypan blue is excluded from healthy living cells but enters moribund cells or those with leaky membranes, staining them blue. Dye exclusion is commonly used as an indicator of viability in tissue culture (Kaltenbach et al., 1958). Crocidolite, SiCW, or an appropriate quantity of saline vehicle was added to suspensions of cells just prior to their distribution to fresh petri dishes to give 5.0, 10.0, 15.0, or 20.0 p~g/cm2 of plate surface. Dye exclusion tests were performed 24 hr later. For this purpose the medium (containing dead cells) was decanted and combined with the remainder of cells removed from the plate by trypsinization. Cells were collected by centrifugation, resuspended in saline (0.5 ml), and mixed with an equal volume of isotonic saline containing trypan blue (0.4%). Results were recorded as a ratio of the number of stained cells to the total counted (at least 200 individuals per slide). Ten replicate cultures were analyzed for each condition of treatment. Cytotoxicity as Indicated by 51Cr Release When cells are allowed to accumulate radioactive chromium (51Cr) and then placed in growth medium without the radioactive label, the amount of 51Cr that leaks relative to the control is a reflection of membrane damage or cell death (Zawydiwski and Duncan, 1978). Cells were seeded at 3.0 × 10 3 cells per well in a 92-well multiwell plate and incubated 24 hr to achieve approximately 50% confluency before exposure to 51Cr (25 p~Ci/ml) in fresh complete medium at 37° C for 3 hr. The labeled medium was removed, and the cells were washed five times then exposed to complete medium containing SiCW or asbestos at a final concentration of 5.0 p~g/cm2. Control wells contained complete medium only. After 18 hr, a sample of supernatant was taken for determination of leaked 51Cr by liquid scintillation (Beckman Model LS 8100). The cells, still attached to the plates, were washed twice with complete medium after which they were covered with distilled water which was frozen and thawed three times to ensure total cell lysis. Radioactivity in the lysate was measured to determine the amount of label remaining in the cells. Control values were subtracted from experimental values and the results were expressed as percentage radiochromium leakage. Colony-Forming Efficiency To measure the effects of the test materials on proliferative ability, cells were
60
VAUGHAN, JORDAN, AND KARR
seeded at 300 cells per 60-mm culture dish and incubated for 24 hr to allow attachment before treatment with the appropriate test material at varying concentrations. After a 7-day incubation to allow for development of colonies, plates were fixed in methanol and stained with Giemsa stain. Colonies were counted manually by inverted phase microscopy. Only colonies of 50 or more cells were scored. Control cultures with untreated cells of similar passage were assayed concurrently.
Tritiated Thymidine Incorporation Assay The rate of DNA syntheses in cells exposed to whiskers or fibers was determined by measuring the incorporation of [3H]thymidine (New England Nuclear, 100 ixCi/mM) into DNA. Cells were exposed to test materials suspended in complete medium at concentrations ranging from 0.0 to 2.0 ~g/cm 2 and maintained in culture. At various times, cultures derived from those which had been exposed to test materials were exposed to [3H]thymidine (2.0 ~xCi/ml) in MEM for 2 hr after an initial determination that uptake and incorporation were linear over a 5-hr period. Labeled medium was removed by aspiration and the cells were washed twice with ice-cold PBS, then harvested by scraping or trypsinization. The amount of [3H]thymidine which had been incorporated into DNA was determined in trichloracetic acid-extracted cells by liquid scintillation (Beckman Model LS230).
Incidence of Binuclear Cells Cells were seeded at 1 × 10 6 cells per 25-cm z flask in complete medium and allowed to attach for 24 hr as described above, then treated with PBS or the appropriate test material at 5 ~g/cm 2. After incubation for 48 hr, 200 cells per flask were scored for multinuclearity by phase contrast microscopy.
Cellular Transformation Transformation frequency, reflected in loss of contact inhibition, was determined for treated and untreated cultures maintained concurrently. Flasks were seeded at 2.0 × 105 cells per 25-cm 2 flask in complete medium, allowed to attach overnight, then exposed to the appropriate whisker or fiber at 5 ixg/cm 2 for 24 hr. Cultures were split 1:3 upon reaching 80% confluency. One flask at each passage was used to propagate the culture which was maintained for 10 passages. The other two flasks at each passage were incubated 10-14 days, fixed in methanol, stained with Giemsa and scanned for focal colonies of transformed cells with a stereomicroscope at 60x total magnification. To be scored positive, foci had to exhibit layered cell growth and the swirling pattern characteristic of transformed colonies (Freshney, 1987).
Cellular DNA Content In whisker-exposed cells, total DNA content was estimated using the method of Keck (1956) based upon colorimetric evaluation of the colored reaction product of indole and DNA.
TOXICITY OF SILICON CARBIDE WHISKERS
61
Data Analysis Analysis of variance (ANOVA) and Student's two-tailed and single-group t tests were used for statistical comparisons and analyses of experimental results. RESULTS Within 24 hr of being added to cell cultures, many, perhaps a majority, of the whiskers/fibers that can be detected by phase contrast microscopy are found associated with the cells, attached to cell surfaces, or internalized. As demonstrated by the micrographs in Fig. 1, SiCW whiskers too large to be engulfed are found penetrating cell surfaces, often entering the cell on one side, and exiting on the other as if the cell were "skewered." It seems likely that transmembrane particles, compromising membrane integrity, are responsible for much of the cytotoxicity we observe to be associated with asbestos and SiCW. This possibility is circumstantially supported by the results of cytotoxicity assays which are based upon evaluation of membrane selectivity. The dye, trypan blue, is excluded by healthy living cells but not by moribund or dead cells which stain blue. Based on dye exclusion, crocidolite asbestos and SiCW-1 exhibited similar levels of dosedependent cytotoxicity within the first 24 hr (Fig. 2). Similar results were obtained with the radiochromium release assay wherein cells were allowed to accumulate 51Cr and then placed in growth medium without the label where they are exposed to the test materials. As cells die, or are damaged, plasma membranes deteriorate and lose natural selective permeability. As a consequence, 5~Cr leaks to the medium surrounding the cell. As shown in Fig. 3, cells in cultures exposed to SiCW-1 or crocidolite at 5.0 ~g/cm 2 release 20-30% of label in excess of controls, while SiO2 induces an excess loss of approximately 8.0%. Dye exclusion and radiochromium release assays do not report cells which survive the first 24 hr to be compromised and die later. Analysis of colony-forming efficiency, however, measures cellular proliferative capacity. Cells must not only survive the first 24 hr but must also be sufficiently viable to produce individual colonies of at least 50 cells each. Colony formation efficiency curves for crocidolite asbestos, SiCW-1, and SiCW-2 are shown in Fig. 4 where colony-forming efficiency is plotted as a function of whisker/fiber concentration. On a mass per surface area basis, SiCW-1 is slightly more toxic than equal quantities of crocidolite. The larger SiCW is less cytotoxic than SiCW-1 (P < 0.01) but not significantly different from asbestos (P > 0.05). This finding as to the relative toxicities of SiCW-1 and SiCW-2 seems to follow from the common understanding that fibers of smaller diameter are more toxic than larger ones. On the other hand, consider the obvious fact that a given mass of small diameter fibers will contain more particles than the same mass of larger fibers of similar length. From the known density of SiC (3.2) and the dimensions given in Table 1, we have estimated the number of whiskers per gram for both SiCW samples. In Fig. 5 we present cytotoxicity as a function of the number of fibers to which the cells are exposed rather than mass. On a numerical basis, we found no significant difference (P < 0.05) between the cytotoxicities of 0.8 and 1.5 p~m SiCW of similar lengths. In addition to being immediately cytotoxic, SiCW-1, SiCW-2, and crocidolite
62
VAUGHAN, JORDAN, AND KARR
TOXICITY OF SILICON CARBIDE WHISKERS
63
induce, within eight generations of exposure, changes in cellular growth habits and structure generally held to be characteristic of cellular transformation. As determined by the appearance, within culture, of cells that have lost contact inhibition, transformation frequencies resulting from exposure to SiCW relative to saline controls are shown in Fig. 6. Not shown in Fig. 6, crocidolite asbestos produced a transformation frequency of 0.000069 (SD = 0.000046). As with the cytotoxicity values in Fig. 4, these values are a function of dosage expressed as mass per area. When the data for SiCW-2, the larger whisker, are recalculated so that a comparison of transformation frequencies can be made on the basis of the number of particles rather than mass (Fig. 7), there is, as with colony-forming efficiency, little difference between the effects of SiCW-1 and SiCW-2. Perhaps the most significant alteration in cells exposed to SiCW and asbestos occurred at the level of the genome. In a crude fashion this was initially observed as an increase in the number of binucleate cells (six to eight fold). Lemaire et al., (1982) have suggested that exposure to asbestos and other fiberform materials alters the rate of synthesis of DNA, and that increases in DNA synthesis rate are proportional to the carcinogenicity of the material. We found that, although DNA synthesis rates for fiber-/whisker-exposed cells were generally elevated relative to controls, often by a factor of as much as 2.5, the results were inconsistent and not reported here. Difficulties in interpretation resulted from the fact that cultures transformed by fiber/whisker treatment were often more mitotically active, with generation times ranging from the control value of approximately 20 hr to as low as 10 hr. It appears that this observation was not considered by LeMaire et al. Significant increases in total cellular DNA content, however, were consistently observed 10-20 generations after treatment. Cells treated with SiCW and crocidolite contained an excess of DNA (Table 2) ranging from approximately 40% to near 70%. This observation is consistent with those in other systems where transformed cells and those cells associated with malignancies have elevated levels of DNA (Vanderlaan et al., 1983; Schieck et al., 1987). DISCUSSION There is ample evidence that, within a narrow size range, durable whisker and fiberform dusts generally manifest significant cytotoxic effects. Based upon the experimental results presented here, the silicon carbide and crocidolite samples we have tested are no exceptions to the rule. The SiCW materials we examined were similar in physical dimensions with the exception that SiCW-2 was approximately twice the diameter of SiCW-1 and the crocidolite sample (Table 1). Judging from dye exclusion measurements (Fig. 2) and radiochromium leak rates (Fig. 3) for SiCW-1 and crocidolite, the two materials are similarly cytotoxic. Of the FIG. 1. Scanning electron micrographs of two cell types 24 hr after exposure to SiCW in vitro. Photo la depicts cultured dog tracheal epithelial cells transfixed by SiCW-2, and Photo lb shows particles of SiCW-2 which have become attached to or engulfed by a mouse 3T3 cell in culture. Cell samples were fixed with glutaraldehyde followed by osmium tetroxide in sodium cacodylate buffer, dehydrated, and critical point dried. Dried samples were coated with gold/palladium and photographed using an ETEC Antoscan scanning electron microscope. The horizontal bar represents 5.0 p.m for scale comparison.
64
V A U G H A N , JORDAN, A N D KARR (dye
L
exclusion)
2 o o
/
x
//x
lO
-A-
d_ x
/,,! I F z © © 0
3O 20 10 0
-
0
5
10
15
20
CONCENTRATION (ug/cm 2) FIG. 4. Cell proliferative ability, measured by colony-forming efficiency, is shown as a function of mass-based particulate concentration. Colony-forming efficiency is expressed as a percentage of control values. Vertical bars, in this figure, represent standard error of the mean.
SiCW-2. It seems likely that, within the narrow size range we examined, cellular response probably depends more on the number of particles a cell encounters than on relative size. Although the processes involved in cellular transformation and those in the development of malignancy are not necessarily identical, they are similar. When, therefore, a material is shown to be an agent capable of transforming cells, there
>(9 Z uJ O LL
LL UJ © Z
(% control) 00' 8O 70 6O 5O
LL
>Z 0 J
0 O
SiCW-
-o--
SiCW-2
1
.z
4o
0
-o-
"
30 20
~ _ ~
lO
[
0
,
O
+
,
1
,
2
,
i ~
3
,
4
i
5
l ,
i
6
,
J
7
CONCENTRATION ( 105 whiskers/crn 2) F I 6 . 5. Colony-forming efficiency data for SiCW-1 and SiCW-2 from Fig. 4 are replotted as a
function of the number of particles per area rather than the mass of particles per area. Vertical bars represent standard deviation about the mean.
66
V A U G H A N , JORDAN, AND KARR ~..
t .oo
o x z 0 ~,4
SlOW conc. 5.0 ug/cm 2
0.80
0,60
IX
0
LL d) Z 4~ [[ F-
o~
0.40
0.20
O.00 SALINE CONTROL SiCW O.8uM DIA
WHISKER
SiCW
1.SUM DIA
TYPE
FIG. 6. The frequency of cellular transformation is given for cultures of mouse 3T3 cells exposed to SiCW-I, SiCW-2, or saline (control). The particulate dose was 5.0 mg/cm 2. Vertical bars are standard deviation about the mean.
is cause for concern. In light of the general hypothesis that cell transforming and carcinogenic potentials in durable whisker/fiberform materials are largely functions of particle size and aspect ratio (Stanton, et al., 1981), it was not unexpected that SiCW-I and crocidolite would be more effective than SiCW-2 in cell transformation capability. When, however, transformation frequency for SiCW is shown as a function of the relative number of whiskers rather than mass (Fig. 7), there is no difference between SiCW-1 and SiCW-2. This supports the probability that, as with cytotoxicity, the effects of these materials within this size frame depend more upon the number of fibers than on the size. Finally, reinforcing our conclusion that SiCW may pose a serious health hazard, is the fact that exposure to SiCW and crocidolite caused a large increase in cellular DNA content relative to untreated controls. Taken as a whole, such changes as we describe are common in malignant tumors and as a concomitant of carcinogeninduced cellular transformation (Vanderlaan et al., 1983). ~.
1.00
b Z
080
Z 0 b-
O.60
0 LL
0.40
O3
Z II
0.20
F0.00 CONTROL
SiCW- 1
SJOW-2
FIG. 7. Frequency of cellular transformation is presented as in Fig. 6 except that the data for SiCW-2 are recalculated to reflect a comparison based upon the number of particles rather than their mass.
TOXICITY OF SILICON CARBIDE WHISKERS
67
TABLE 2 TOTAL DNA CONTENT, MOUSE CELLS (BALB/3T3) Treatment SiCW-1 SiCW-2 Asbestos Saline
DNA/cell (× 10-11 g) 3.48 3.54 4.37 2.49
(0.81)* (0.98)* (1.21)* (0.58)
Excess
n
39% 42% 75% --
32 29 18 45
Note. Numbers in parentheses represent the standard deviation. * Significantly different from control (P < 0.05, ANOVA).
ACKNOWLEDGMENTS The necessary funding and support for this research were provided by the University of Tennessee, Advanced Composite Materials Corporation, and American Matrix, Inc. The authors are also grateful for the personal support and encouragement of Dr. Sam Weaver, Third Millenium Technologies, Inc.
REFERENCES Fre shney, R. I. (1987). "Culture of Animal Cells: A Manual of Basic Technique." R. Lis s, New York. Hoch-Ligeti, Sass, B., Sobel, H. J., and Stewart, H. L. (1983). J. Nat. Cancer Inst. Endocardial tumors in rats exposed to durable fibrous materials. 71(5). Kaltenbach, J. P., Kaltenbach, M. H., and Lyons, W. B. (1958). Nigrosin as a dye for differentiating live and dead ascites cells. Exp. Cell Res. 15, 112-117. Keck, K. (1956). An ultramicro technique for the determination of deoxypentose nucleic acid. Arch. Biochem Biophys. 53,446--451. Lechner, J. F., Haugen, A., Trump, B. F., Tokiwa, T., and Harris, C. C. (1983). Effects of asbestos and carcinogenic metals on cultured human brochial epithelium. Hum. Carcinog. 51, 561-585. Lemaire, I,, Gingras, D., Lemaire, S. (1982). Tbymidine incorporation by lung fibroblasts as a sensitive assay for biological activity of asbestos. Environ. Res. 28, 399--409. Mossman, B. T., Bignon, J., Corn, M., Seaton, A., and Gee, J. B. L. (1990). Asbestos: Scientific developments and implications for public policy. Science 247, 294-301. Mossman, B. T. and Gee, J. B. L., (1989). Asbestos-related diseases. N. Engl. J. Med. 320, 17211730. Osbimura, M., Hesterberg, T. W., Tsutsui, T., and Barret, J. C. (1984). Correlation of asbestosinduced cytogenetic effects with cell transformation of Syrian hamster embryo cells in culture. Cancer Res. 44, 5017-5022. Schieck, R., Taubert, G., and Krug, H. (1987). Dry mass, DNA and nonhistone protein determinations in lung cancer cells. Histochem. J. 19, 504-508. Sincock, A. M. (1977). Preliminary studies of the in vitro cellular effects of asbestos and fine glass dusts. Origins Hum. Cancer Cold Spring Harbor Lab. Stanton, M. Layard, M., Tegeus, A., Miller, E., May, M. and Kent, E. (t977). Carcinogenicity of fibrous glass: Pleural response in the rat in relation to fiber dimension. J. Nat. Cancer Inst. 58(3), 587-603. Stanton, M. F., Layard, M., Tegeris, A., Miller, E., May M., Morgan, E., and Smith, A. (1981). Relation of particle dimension to carcinogenicity in amphibole asbestoses and other fibrous minerals. J. Nat. Cancer Ins. 67(5), 965-975. Vanderlaan, M. Steel, V., and Nettesheim, P. (1983). Increased DNA content as an early marker of transformation in carcinogen-exposed rat tracheal cell cultures. Carcinogenesis 4(6), 721-727. Vaughan, G. L. (1989). Biological consequences of exposure to silicon carbide whiskers. In "91st Annual Meeting of American Ceramic Society, Indianapolis, April 1989." Zawydiwski, R., Duncan, G. R., (1978). Spontaneous 51Cr release by isolated rat hepatocytes: An indicator of membrane damage. In Vitro 14, 707-714.
