- Email: [email protected]
Cytotoxicity of refractory ceramic fibres to Chinese hamster ovary cells in culture
Toxic. in Vitro Vol. 6, No. 4, pp. 317-326, 1992 Printed in Great Britain. All rights reserved
0887-2333/92 $5.00+ 0.00 Copyright © 1992 Pergamon Press Ltd
CYTOTOXICITY OF REFRACTORY CERAMIC FIBRES TO CHINESE HAMSTER OVARY CELLS IN CULTURE G. A.
HART,M. M. NEWMAN,W. B. BUNN a n d T. W. HESTERBERG Mountain Technical Center, Littleton, CO 80127, USA
(Received 20 October 199I; revisions received 17 January 1992)
Abstract--The toxicity/oncogenicity of refractory ceramic fibres have been tested in chronic inhalation studies in rodents. Because these studies are time consuming and expensive, there is a need to develop and validate short-term models to screen fibres for their toxicological potential. In the present study, the toxic effects of four different compositions of refractory ceramic fibres were determined using Chinese hamster ovary cells grown in culture. These refractory ceramic fibres were the same size-selected fibres that had been used in animal inhalation studies, thus facilitating a direct comparison of findings in the two systems. Chinese hamster ovary cells were treated with refractory ceramic fibres 24 hr after seeding into 60-mm culture dishes in Ham's FI2 medium with 10% serum. Inhibition of cell proliferation and colony formation were determined after 3-5 days of fibre exposure. Crocidolite and chrysotile asbestos were used as positive controls. Concentration-dependent inhibition of both cell proliferation and colony formation was observed after treatment with refractory ceramic fibres. The LCs0 for the different refractory ceramic fibres ranged from 10 to 30 #g/cm 2. The LCs0s for crocidolite and chrysotile were 5 #g/cm 2 and 1/~g/cm2, respectively. To assess the genotoxic potential of these fibres, fibre-exposed Chinese hamster ovary cell cultures were stained with acridine orange and scored for the incidence of micronuclei and other nuclear abnormalities. The incidence of nuclear abnormalities for refractory ceramic fibres at 20 pg/cm ~ ranged from 20 to 40%. Toxic endpoints of the in vitro studies were compared with those of the chronic animal inhalation studies. The latter included induction of lung fibrosis and pleural and airway tumours. A correlation was observed between the in vitro and in vivo toxicological potencies of the respective four refractory ceramic fibres: the fibres that were most toxic in vitro were also the most toxic in the chronic animal inhalation studies. A direct relationship was also observed, both in vitro and in vivo, between average fibre length and the severity of the toxic effect.
INTRODUCTION Awareness of the h u m a n health hazards associated with asbestos exposure has led to the development of m a n y new m a n - m a d e a n d naturally occurring fibres for use as asbestos substitutes. The importance of testing all fibrous products for their possible pathogenic potential is generally recognized. However, m a n y fibrous materials currently in use and at the p r o d u c t - d e v e l o p m e n t stage have not been evaluated. A n i m a l inhalation testing is currently the only accepted laboratory model for determining the potential h u m a n health h a z a r d of respirable particles (Hesterberg et al., 1991c). Because this form of testing is time c o n s u m i n g a n d expensive, short-term tests are needed to screen fibres for their potential toxicity. A n u m b e r of short-term screening systems have been suggested, such as short-term animal assays (Warheit et al., 1991), in vitro assays ( M o s s m a n a n d Sesko,
CD = crocidolite; CFE = colony-forming efficiency; CH =chrysotile; CHO=Chinese hamster ovary; CM = complete medium; FCS = foetal calf serum; I C P = inhibition of cell proliferation; MN = micronuclei; NA = nuclear abnormality; OM = optical microscopy; PN = polynuclei; RCF = refractory ceramic fibre; SEM = scanning electron microscopy; SHE = Syrian hamster embryo; TEM = transmission electron microscopy.
Abbreviations:
1990), a n d a battery of in vitro assays c o m b i n e d with short-term animal inhalation tests (Hesterberg et al., 1992). Any short-term test system proposed as a means of screening fibres for their toxicological potential must be validated by d e m o n s t r a t i n g that the effects observed using the screening system correlate with h u m a n epidemiological data a n d / o r animal inhalation data. A recent chronic rodent inhalation study of the toxicity and tumorigenicity of refractory ceramic fibres (RCFs; Hesterberg et al., 1991a) has provided the animal data necessary to m a k e such correlations. C o n d u c t i n g in vitro research in parallel with in vivo studies also provides a n additional perspective on toxic mechanisms at the cellular level. Thus, the present study was conducted: (1) to develop and validate in vitro assays that could be useful as part of a battery of short-term toxicity screening tests for fibres; a n d (2) to contribute to a more complete u n d e r s t a n d i n g of the mechanisms of fibre-cell interactions. In the present study, Chinese h a m s t e r ovary (CHO) cells were exposed in vitro to the same four sizeselected refractory ceramic fibres that were used in the in vivo study. Toxic e n d p o i n t s assessed in the in vitro test system [inhibition of cell proliferation (ICP), inhibition of colony-forming efficiency (CFE) and the induction of micronuclei ( M N ) a n d polynuclei (PN)] 317
318
G . A . HART et al. Table I. Chemical composition of RCFs tested, as supplied by the Thermal Insulation Manufacturers' Association Fibre Repository Percentage of mass of Oxide
RCFI
RCF2
RCF3
RCF4*
SiO 2 AIzO3 F%O 3 Cr20 ~ TiO2 CaO MgO Na20 K20 ZrO 2
47.7 48.0 0.97 0.03 2.05 0.07 0.08 0.54 0.16 0.11
50.0 35.0 < 0.05 <0.01 0.04 0.05 0.01 <0.3 <0.01 15.0
50.8 48.5 0.16 <0.01 0.02 0,04 <0.01 0.19 <0.01 0.23
47.7 48.0 0.97 0.03 2.05 0.07 0.08 0.54 0.16 0.11
*RCF4, which was RCF1 heated to 1300 C to simulate its use in the workplace, is composed primarily of crystalline materials: approximately 27% of its mass is cristobalite (crystalline silica) and the remainder is mullite. The other RCFs are composed of amorphous materials.
were compared with toxic endpoints in the animal inhalation study (induction of lung fibrosis and incidence of thoracic tumours). The present study is limited in that only three parameters of toxic response are evaluated and only one cell type is used. However, a correlation was observed between the in vivo pathogenesis and the in vitro toxicity of RCFs. This study is the first phase of investigations that will expand to include other cell types and evaluate other in vitro endpoints, in an attempt to develop a battery of in vitro tests that will be useful in screening fibrous materials for their toxic potential. MATERIALS AND METHODS
Test fibres. Four different compositions of refractory ceramic test fibres were obtained from the TIMA Fibre Repository*: RCF1 (kaolin), RCF2 (zirconia), RCF3 (high purity) and RCF4 (after service). These are size-selected, respirable test fibres that were used in rodent inhalation studies (Hesterberg et al., 1991,b,c,d). Details of the chemical compositions of the fibres were provided by the TIMA Repository xand are listed in Table 1. UICC (International Union Against Cancer) crocidolite (CD) and UICC chrysotile (CH) asbestos were obtained from
*Thermal Insulation Manufacturers" Association Fiber Repository.
Dr V. Timbrell (Pneumoconiosis Research Unit, Llandough Hospital, Penarth, Glamorgan, UK). Fibre dimensions (Table 2) were determined using scanning electron microscopy (SEM) for the RCFs and CD and transmission electron microscopy (TEM) for CH. Electron micrographs in Plate la and b illustrate the two extremes of average particle dimensions of the RCFs. Plate la (RCF3) is representative of the three longer fibres, RCF1, 2 and 3 (average lengths of 21.5, 16.7 and 24.3 #m, respectively); Plate lb illustrates the dimensions of the shortest fibre, RCF4 (average length 9.2,um). The densities of fibres were also determined by SEM, TEM and optical microscopy (OM). Electron microscopy data on fibre numbers were used to calculate fibre exposure as numbers of fibres/cm2. OM fibre counts of the RCFs were used to confirm the SEM data; the two asbestos fibres were not counted with OM because most of these fibres are too small to be seen using optical microscopy. For OM fibre counts, 1 ml aqueous fibre suspension (100 #g of fibre/ml) was allowed to settle in a 60-mm culture dish (5Ftg/cm2). The fibres (particles with length/ diameter ratio of >~3) were counted at a magnification of × 400 using an ocular grid. Cell cultures and in vitro assays. The CHO-K 1 Cell line was derived from CHO tissue (Puck, 1958) and has an epithelia-like morphology. Although the tissue of origin is not known to be a target tissue of fibre-induced disease, CHO cells were selected for these studies because they divide rapidly and were observed to internalize asbestos fibres readily. Frozen cells were obtained from American Type Tissue Collection (Rocky±lie, MD, USA) and were thawed and plated in tissue-culture flasks with complete medium (CM) consisting of Ham's F I 2 supplemented with 10% foetal calf serum (FCS), 2 mM-L-glutamine, and l ml Fungi-Bact solution/100ml medium (all from Irvine, Santa Ana, CA, USA). To create a large bank of uniform stock cells for experiments, cultures were incubated (37"C, 5% CO2 and 100% humidity), allowed to replicate for several days (three-four doublings), and then harvested with 0.05% trypsin and 0.02% ethylenediaminetetraacetic acid in Hanks' balanced salt solution (Irvine). The stock cells were then suspended in freezing medium, consisting of CM supplemented with an additional 10% FCS and
Table 2. Physical characterization, by electron microscopy*, of fibres tested Test fibre Abbreviation
Fibre dimensions (pm) Fibre type
RCFI RCF2 RCF3 RCF4
Kaolin Zirconia High purity After service
CD CH
UICC Crocidolite UICC Chrysotile
No. of fibres/ng × 1000 RCFs 4.7 4.6 3.1 6.3 Asbestos 2400 4400
Diameter (mean _+ SD)
Length (mean _+ SD)
1.03 + 0.73 1.11 _+ 0.82 1.22 + 0.98 1.43 -- 0.79
21.5 ± 16.12 16.7 ± 15.03 24.3 ± 18.82 9.2 ,+ 7.08
0.21 ± 0.12 0.12 ± 0.07
1.81 ± 1.94 1.65 ± 1.83
*Transmission electron microscopy was used to analyse CH. Scanning electron microscopy was used to analyse RCFs and CD.
Cytotoxicity of RCF to CHO cells
319
120
50 I A 1OO| ,~ T
40
o=~8o
-~ -~ 30
[,
..'.AChrysotile "
:
RCF1
Crocidolite
.:
~
.c.. -/ "/
/~
RCF4
.o
r~ o 20
-,;i
Crocidolite
" ~ . ~
RCF2
Chrysotile
RCF1 RCF3 I 10
0
I 20
Concentration
I
I 30
40
( # g / s q . cm)
12% dimethyl sulphoxide (Sigma Chemical Co., St Louis, MO, USA), frozen at - 8 0 ° C and later transferred to liquid nitrogen for storage. For experiments, stock cells were thawed and plated in culture flasks (Falcon, N J, USA) in CM and allowed to grow for 2-4 days. I C P assay. For ICP assays, 60-ram culture dishes were seeded with 100,000 cells in 5 ml CM/dish. After 24 hr of incubation, fibres were added to cultures in 1 ml CM/dish. Dishes were gently agitated to disperse fibres over the bottom. Negative control cultures received 1 ml CM/dish. Each exposure group was set up in triplicate. After 3 days of exposure, incubated as described above, cells were harvested using trypsin (as described above) and counted using a Coulter counter (Coulter Electronics Hialeah, FL, USA). Relative proliferation was determined by dividing the number of cells present in each exposed culture by the number of cells present in negative control cultures. C F E assay. For the CFE assay, 60-mm culture dishes were seeded with 200 cells in 5 ml complete medium and exposed as described for ICP above.
120
._. .~---.
i 0
5
I 10
I 15
Concentrotion
Fig. 1. Inhibition of proliferation of CHO cells cultured for 3 days with various RCFs and with crocidolite and chrysotile asbestos (concentrations expressed as/~g/cm2).
o
o,
T
I 20
[ 25
I 30
( F g / s q . cm)
Fig. 3. Induction of nuclear abnormalities (percentage of cells with micronuclei and/or polynuclei) in CHO cells cultured for 3 days with various RCFs and with crocidolite and chrysotile asbestos. Each exposure was set up in triplicate. After 5 days of exposure, colonies were stained with 0.4% (w/v) Giemsa in buffered methanol (Sigma). Colonies with > 10 cells were counted using a stereoscope at low power. CFE was determined by dividing the number of colonies in each exposed culture by the number of colonies in unexposed cultures. M N and P N induction. For these assays, cultures were prepared as described for ICP above but mitomycin C (MMC, 2/~M)was also included as a positive control. After 2 days of exposure, culture dishes were fixed with methanol-acetic acid (3 : 1, v/v) and stained with 0.01% acridine orange (from Sigma; as described in Clark, 1981). Using a microscope with epifluorescence, the percentage of cells containing micronuclei (MN) and/or polynuclei (PN) was determined for each culture dish. An MN was defined as a nucleus that appeared to be less than half the size of the normal-sized nuclei; PN was defined as a cell having bi- or multiple nuclei or a lobed nucleus (in many cases the different forms of PN were difficult to distinguish). A minimum of 100 cells/dish and 2-3 dishes/exposure group were scored for (1) MN but no other visible nuclear abnormalities, (2) both MN and PN, or (3) PN without MN. Calculations included the percentage of cells with MN (including both uninucleate and polynucleate cells), the percentage of PN (with or without MN), and the percentage of cells with either type of nuclear abnormality (NA), that is, having MN and/or PN.
RESULTS 20
Crocidolite
~ ' t ~
RCF2 RCF3 RCF1
O O
I 10
I 20
Concentration
~ 30
I 40
( # g / s q . cm)
Fig. 2. Colony-forming efficiencyof CHO cells cultured for 5 days with various RCFs and with crocidolite asbestos•
The results of the three in vitro assays (Figs 1-3) all show the same relative toxicities for the four RCFs when concentration is based on/~g/cm2: in each case RCF4 is the least toxic, RCF2 is intermediate, and RCF1 and RCF3 are the most toxic. In all three assays, CD was more toxic than the RCFs; in the ICP (Fig. 1) and MN/PN (Fig. 3) assays, CH was more
320
G.A. HART et al. 120
lO0
o--- o
80
u 60 ~-.~
~>~ 40 g 2O o
I
I
I
I
I
I
I
25
50
75
lOO
125
1.'50
175
Fibres/sq.
cm
I 200
x 1000
Fig. 4. I n h i b i t i o n of p r o l i f e r a t i o n of C H O cells cultured for 3 d a y s with various R C F s and with crocidolite a n d chrysotile asbestos ( c o n c e n t r a t i o n s expressed as fibres/cm2). E r r o r bars indicate the S E M from four or five trials.
toxic than CD (CH was not included in the CFE assay). Although some of the error bars (SEM, standard error of the means) overlap, in each individual test performed with these fibres (more than 14 individual tests in all), the same relative toxic±ties resulted. When the concentration of the RCFs is expressed as numbers of fibres/cm 2 (Fig. 4), similar relative toxic±ties of the different RCFs is observed, although the spread between RCFI and RCF3 is increased. Data from the asbestos types are not plotted in this manner since both the length and diameter are an order of magnitude less than those of RCF fibres, making a fibre-to-fibre comparison inappropriate.
Plates 2-4 were all photographed using fluorescent light with the exception of Plate 3b and d, which were photographed using phase-contrast light. Nuclear abnormalities (NA) observed in all the fibre-exposed cultures appeared qualitatively similar (Plates 3 and 4). Some examples of NA include: PN that appear to be clearly more than one nucleus (Plate 4a and b) as well as PN that appear to be bi- (Plate 3c and d) or multi-lobed (Plate 3a and b); micronuclei with one apparently normal nucleus or with PN (both in Plate 4a). The phenomenon of multiple nuclei aligned along the length of a fibre (Plate 4b) was not as prominent with asbestos as with RCF exposure and is likely to result from the higher frequency of long fibres in the RCFs than in the asbestos. The relationship between intracellular fibres and nuclear distortion is unmistakable in all of the photographs, and is especially apparent when phase-contrast and fluorescent views of the same field are compared as in Plate 3a,b and c,d. Both wholly and partially internalized fibres are visible (Plate 4c). The frequencies of MN and PN induced by each of the fibres and by mitomycin C are shown in Table 3. A difference was observed in the types of NA induced by MMC compared with fibres: 2/~M-MMC induced a 29% incidence of MN and a 10% incidence of PN, whereas all of the fibre types induced roughly half as many MN as PN. In the fibre-exposed cultures, the incidence of both types of nuclear abnormalities (NA; percentage of cells with MN and/or PN) was strongly concentration dependent, whereas the incidence of MN alone, although elevated above background levels, was generally weakly concentration dependent. CH at 5/~g/cm2 induced NA in 49% of
Table 3. Incidence of abnormal nuclei in Chinese hamster ovary cells cultured with RCFs or asbestos Percentage of Test fibre RCF1
RCF2
RCF3
RCF4
Crocidolite Chrysotile
Exposure concentration (l~g)
No. of tests
0 5 10 20 0 5 10 20 0 5 10 20 0 5 10 20 0 5 0 1
Mitomycin C
2 5 0 2/~M
*One or more micronuclei; poly- or uninucleate. "t'More than one nucleus or lobed nucleus.
Micronuclei* (Mean ± SEM)
Polynucleit (Mean ± SEM)
3±0 5±0 8±1 10±1 2±0 4±0 6±1 8±0 2±0 6±1 8±1 8±1 3±0 5±1 5±1 9±2 2±0 9±2 2 10 18 7 2 29
2±1 11+1 18+2 26±3 2±1 8±0 13±3 22±3 2±1 11±2 17±4 28±4 2±1 9±1 11±2 17±3 0±1 21±0 I 17 29 46 1 10
Plate 1. Electron photomicrographs of (a) RCF3 and (b) RCF4 fibres.
321
Plate 2. CHO cells growing in culture medium only. Note uniformity of nuclear size, shape and number/cell. Acridine orange stain, photographed with fluorescent light. Original magnification × 400. Plate 3(a,b). Caption opposite. 322
Plate 3(c,d) Plate 3. C H O cells exposed to crocidolite asbestos, photographed with fluorescent light (a,c) and with phase contrast optical microscopy (b,d). Note the broad range of nuclear size, shape and number/cell. A polynucleate cell with two very small micronuclei is visible in the centre of the field (Plate 3a); fibres are visible in this cell in Plate 3b, and appear to be associated with the nuclear lobes. The central cell in Plate 3c has a C-shaped polynucleus; In Plate 3d a packet of fibres can be seen in the center of the "C". Acridine orange stain. Original magnification x 1000.
323
Plate 4. CHO cells exposed to RCF2, original magnification x400 (a-b), and RCF3, original magnification x 1000 (c). Note the variety of nuclear configurations: uninucleate, binucleate and polynucleate cells, with and without micronuclei. Alignment of nuclei along the length of a long fibre is visible in Plate 4b; this was a common phenomenon in cultures exposed to RCFs. Fibres are visible in all three photographs, Acridine orange stain, photographed with fluorescent light. 324
Cytotoxicity of RCF to CHO cells
325
Table 4. Comparisonof the in vitro and in vivo toxic endpoints for RCFs Toxic endpoints In vitro
In vivo
LCso (,ug/cm 2) Exposure group
ICP
CFE
Wagner score* at 6 months
Percentage with lung tumourst
Percentage with mesotheliomast
Negative control RCF1 RCF2 RCF3 RCF4
-13 17 13 30
-10 18 12 22
1.0 4.0 3.0 4.0 2.7
1.6 13.0 6.9 14.6 3.2
0.0 1.5 2.3 1.5 0.8
*1 =normal lung; 2 = m a c r o p h a g e infiltration; 3=inflammation and bronchiolization; 4 = m i n i m a l but irreversible fibrosis; 5,6,7 and 8 = increasingly severe fibrosis. tNumber of animals with tumour, expressed as a percentage of the number of animals at risk.
the cells; CD at 5 Ftg/cm2 induced NA in 28% of the cells. The NA incidence for RCFs at 20/~g/cm z ranged from 22 to 33% (Fig. 3).
the test fibres. Secondly, specificity is evident in this model, in the consistent relative response of these cells to the six different fibres. Thirdly, cellular uptake of fibres and alterations of nuclear morphology are clearly major CHO cell responses. The observed DISCUSSION fibre-induced nuclear abnormalities fell into two basic The toxic endpoints of the in vitro studies were categories (MN and PN) and the incidence of each compared with those of the animal inhalation studies was consistent for the six fibres in three or more using the same RCF test fibres (Table 4). The LCs0 separate assays. The mechanism of MN induction is the in vitro concentration that resulted in 50% as may be different from that of PN induction. MN are many cells or colonies in fibre-treated cultures as in believed to form when chromosomes or fragments are unexposed control cultures. Wagner grades (defined separated from the migrating masses of replicate in Table 4) were used to quantitate the lung pathol- chromosomes during mitosis. Previous studies have ogy observed in the chronic inhalation study of RCFs shown a correlation between the incidences of MN where: 1 indicates a normal lung; 2 indicates macro- and aneuploidy in three different cell lines (CHO phage infiltration; 3 represents inflammation and cells, Sincock et al., 1982; Syrian hamster embryo bronchiolization; 4 indicates minimal but irreversible [SHE] cells, Oshimura et al., 1984; human lymphofibrosis; and 5, 6, 7 and 8 represent increasingly severe" cytes, Migliore and Nieri, 1991). PN may result from fibrosis (Hesterberg et al., 1991b). In the animal the failure of cytokinesis following mitosis. Finally, inhalation studies, three to six animals per group the incidence of total NAs was approximately the were scored every 3-6 months during the 24-month inverse of the proliferation curves, indicating that exposure to fibre aerosol and at the end of the lifetime disruption of mitosis may be the major or immediate study. Only the 6-month scores are given for com- cytotoxic effect, rather than the disruption of cytoparison here, because all fibre-exposed lung scores plasmic metabolic processes. However, validation of plateaued near grade 4 soon after the 6 month time this hypothesis would require cell viability testing. Differences in the length of the four RCFs probpoint (air-control lungs remained at level 1). Also indicated in Table 4 are the percentages of animals ably account for the differences in both the in vitro that developed lung tumours and mesotheliomas by cytotoxicity and the in vivo lung pathology observed the end of the RCF inhalation study. The relative in the animal inhalation study. Many previous in vitro toxicity of the different RCFs seen in vitro generally studies also report a relationship between fibre length correlates with that observed in vivo: in each case, and cytotoxicity. Brown et al. (1979) demonstrated RCF4 is least toxic, RCF2 is intermediate, and RCF! that glass fibres were more toxic to cultured macroand RCF3 are more toxic (except in the case of the phages and to V79-4 cells when average fibre length was greater than 10ttm; Lipkin (1980) observed a percentage mesotheliomas seen in vivo where RCF1 and RCF3 results were intermediate, with RCF2 relationship between fibre length and the reduction of showing most toxicity and RCF4, least). cell proliferation in cultures of P388DI cells exposed The present study provides results that contribute to several populations of glass fibres; Tilkes and Beck towards both the development of short-term screen- (1980 and 1983a,b) showed a direct relationship ing tests that can be used to assess the toxic potential between fibre length and both lactose dehydrogenase of fibrous dusts, and to our understanding of the release (macrophages and ascites) and reduction in mechanisms of fibre toxicity at the cellular level. proliferation (ascites) in cell cultures exposed to First, the CHO cell test system is highly responsive to fibrous glass. A mechanism to explain the relationship between fibre exposure in all three parameters investigated; concentration-dependent reductions in cell prolifer- fibre length and cytotoxicity could be related to cell ation and cell colony formation as well as increases diameter. For example, when the flattened, adherent in NAs were observed following exposure to each of cell becomes spherical as it enters metaphase prior to
326
G.A. HART et al.
cell division, internalized fibres that are longer than the diameter of the m e t a p h a s e cell might p u n c t u r e the cell, causing cell death. Fibre length has also been shown to be i m p o r t a n t in the induction ofcytogenetic effects by internalized fibres (Hesterberg et al., 1986; O s h i m u r a et al., 1984) and in the t r a n s f o r m a t i o n of cultured cells (Hesterberg a n d Barrett, 1984). In a study t h a t d e m o n s t r a t e d direct interaction between intracellular asbestos fibres a n d migrating c h r o m o somes during a n a p h a s e in cultured S H E cells, Hesterberg a n d Barrett (1985) suggested that long fibres are m o r e easily entangled in the migrating c h r o m o s o m e s or spindle a p p a r a t u s t h a n are short fibres. The correlation between in vitro a n d in vivo toxic e n d p o i n t s is encouraging. However, more research is needed to determine which in ~,itro test systems a n d procedures will best correlate with animal i n h a l a t i o n studies a n d h u m a n epidemiology. Again, fibres tested in long-term animal inhalation studies will be needed for validation of the sensitivity and specificity of these short-term assays. Such fibres should be well characterized chemically (chemical composition, crystallinity a n d surface chemistry) a n d physically (length, width, total surface area, n u m b e r of fibres/unit mass). F u r t h e r m o r e , the test fibres should include a range of d e m o n s t r a t e d in vivo toxicities, from non-toxic to strongly toxic. In vitro testing should include other cell types a n d other in vitro parameters (i.e. durability, cytotoxicity and mutagenesis). Several p a r a m e t e r s of in vivo pathogenesis c a n n o t at this time be m e a s u r e d in vitro, for example lung deposition, lung clearance and long-term biopersistence. Some work has been done in the third area, by incubating fibres in flow-through systems using synthetic physiological solutions (Law et al., 1990 and 1991). However, such systems do not contain all o f the factors that might be i m p o r t a n t in the biopersistence of fibres in the lung, such as enzymes, lysosomal secretions a n d mechanical removal by phagocytic cells. These limitations might be overcome by including short-term animal inhalation studies in a battery o f toxicity screening tests. Acknowledgements--Funding was provided by MTC Research and Development. We thank Bill Miiller and Barry Fitzpatrick for the preparation of test fibres, R. Hamilton, J. Strothers and F. D'Ovidio for electron microscopy, and Lisa M. Kathman for technical assistance. REFERENCES
Brown R. C., Chamberlain M. and Skidmore J. W. (1979) In vitro effects of man-made mineral fibers. Annals of Occupational Hygiene 22, 175 179. Clark G. (1981) Staining Procedures. 4th Ed. Williams and Wilkins, Baltimore, MD. Hesterberg T. W. and Barrett J. C. (19843 Dependence of asbestos- and mineral dust-induced transformation of mammalian cells in culture upon fibre dimension. Cancer Research 44, 2170-2180. Hesterberg T. W. and Barrett J. C. (1985) Induction by asbestos fibers of anaphase abnormalities: mechanism for aneuploidy induction and possibly carcinogenesis. Carcinogenesis 6, 473-475.
Hesterberg T. W., Butterick C. J., Oshimura M., Brody A. R. and Barrett J. C. (1986) Role of phagocytosis in Syrian hamster cell transformation and cytogenetic effects induced by asbestos and short and long glass fibers. Cancer Research 46, 5795-5802. Hesterberg T. W., Hart G. A. and Bunn W. B. (1992) In vitro toxicology of fibres: mechanistic studies and possible use for screening assays. In Contemporary Issues in Fiber Toxicology. Edited by D. Warheit. Academic Press, New York. In press. Hesterberg T. W., Mast R., McConnell E. E., Chevalier J., Bernstein D. M., Bunn W. B. and Anderson R. (1991a) Chronic inhalation toxicity of refractory ceramic fibers in Syrian hamsters. In Proceedings of the International Workshop on Mechanisms of Fibre Carcinogenesis. Edited by R. C. Brown et al. pp. 531-538. Plenum Press, New York. Hesterberg T. W., Mast R., McConnei1 E. E., Vogel O., Chevalier J., Bernstein D. M. and Anderson R. (1991b) Chronic inhalation toxicity and oncogenicity study of refractory ceramic fibers in Fisher 344 rats. Toxicologist II, 254. Hesterberg T. W., Vu V., McConnell E. E., Chase G. R., Bunn W. B. and Anderson R. (1991 c) Use of animal models to study man-made fiber carcinogenesis. In Current Communications in Molecular Biology: Molecular Mechanisms of Fiber Carcinogenesis. Edited by B. Brinkley, J. Lechner and C. Harris. Cold Spring Harbor Laboratory, New York. In press. Law B., Bunn W. B. and Hesterberg T. W. (1990) Solubility of polymeric organic fibers and man-made vitreous fibers in gambles solution. Inhalation Toxicology 2, 321-339. Law B., Bunn W. and Hesterberg T. W. (1991) Dissolution of natural mineral and man-made vitreous fibres in karnovsky's and formalin fixatives. Inhalation Toxicology 3, 309-321. Lipkin L. E. (1980) Cellular effects of asbestos and other fibers: correlations with in vivo induction of pleural sarcoma. Environmental Health Perspectives 34, 91 102. Migliore L. and Nieri M. (1991) Evaluation of twelve potential aneuploidogenic chemicals by the in vitro human lymphocyte micronucleus assay. Toxicology in Vitro 5, 32%336. Mossman B. T. and Sesko A. M. (1990) In vitro assays to predict the pathogenicity of mineral fibers. Toxicology 60, 53-61. Oshimura M., Hesterberg T. W. and Barrett J. C. (1984) Correlation of asbestos-induced cytogenetic effects with cell transformation of Syrian hamster embryo cells in culture. Cancer Research 44, 5017-5022. Puck T. T. (1958) Genetics of somatic mammalian cells. III. Long-term cultivation of euploid cells from human and animal subjects. Journal of Experimental Medicine 108, 945. Sincock A. M., Delhanty J. D. A. and Casey G. (19823 A comparison of the cytogenetic response to asbestos and glass fibre in Chinese hamster and human cell lines. Demonstration of growth inhibition in primary human fibroblasts. Mutation Research 101, 257 268. Tilkes F. and Beck E. G. (1980) Comparison of lengthdependent cytotoxicity of inhalable asbestos and manmade mineral fibres. In Biological Effects of Mineral Fibres. IARC Scient. Publ. no. 30. pp. 475~483. International Agency for Research on Cancer, Lyon. Tilkes F. and Beck E. G. (1983a) Macrophage functions after exposure to mineral fibres. Environmental Health Perspectives 51, 67 72. Tilkes F. and Beck E. G. (1983b) Influence of well-defined mineral fibres of proliferating cells. Environmental Health Perspectives 51, 275 279. Warheit D. B., Carakostas M. C., Hartsky M. A. and Hansen J. F. (19913 Development of a short-term inhalation bioassay to assess pulmonary toxicity of inhaled particles: comparisons of pulmonary responses to carbonyl iron and silica. Toxicology and Applied Pharmacology 107, 350 368.
0887-2333/92 $5.00+ 0.00 Copyright © 1992 Pergamon Press Ltd
CYTOTOXICITY OF REFRACTORY CERAMIC FIBRES TO CHINESE HAMSTER OVARY CELLS IN CULTURE G. A.
HART,M. M. NEWMAN,W. B. BUNN a n d T. W. HESTERBERG Mountain Technical Center, Littleton, CO 80127, USA
(Received 20 October 199I; revisions received 17 January 1992)
Abstract--The toxicity/oncogenicity of refractory ceramic fibres have been tested in chronic inhalation studies in rodents. Because these studies are time consuming and expensive, there is a need to develop and validate short-term models to screen fibres for their toxicological potential. In the present study, the toxic effects of four different compositions of refractory ceramic fibres were determined using Chinese hamster ovary cells grown in culture. These refractory ceramic fibres were the same size-selected fibres that had been used in animal inhalation studies, thus facilitating a direct comparison of findings in the two systems. Chinese hamster ovary cells were treated with refractory ceramic fibres 24 hr after seeding into 60-mm culture dishes in Ham's FI2 medium with 10% serum. Inhibition of cell proliferation and colony formation were determined after 3-5 days of fibre exposure. Crocidolite and chrysotile asbestos were used as positive controls. Concentration-dependent inhibition of both cell proliferation and colony formation was observed after treatment with refractory ceramic fibres. The LCs0 for the different refractory ceramic fibres ranged from 10 to 30 #g/cm 2. The LCs0s for crocidolite and chrysotile were 5 #g/cm 2 and 1/~g/cm2, respectively. To assess the genotoxic potential of these fibres, fibre-exposed Chinese hamster ovary cell cultures were stained with acridine orange and scored for the incidence of micronuclei and other nuclear abnormalities. The incidence of nuclear abnormalities for refractory ceramic fibres at 20 pg/cm ~ ranged from 20 to 40%. Toxic endpoints of the in vitro studies were compared with those of the chronic animal inhalation studies. The latter included induction of lung fibrosis and pleural and airway tumours. A correlation was observed between the in vitro and in vivo toxicological potencies of the respective four refractory ceramic fibres: the fibres that were most toxic in vitro were also the most toxic in the chronic animal inhalation studies. A direct relationship was also observed, both in vitro and in vivo, between average fibre length and the severity of the toxic effect.
INTRODUCTION Awareness of the h u m a n health hazards associated with asbestos exposure has led to the development of m a n y new m a n - m a d e a n d naturally occurring fibres for use as asbestos substitutes. The importance of testing all fibrous products for their possible pathogenic potential is generally recognized. However, m a n y fibrous materials currently in use and at the p r o d u c t - d e v e l o p m e n t stage have not been evaluated. A n i m a l inhalation testing is currently the only accepted laboratory model for determining the potential h u m a n health h a z a r d of respirable particles (Hesterberg et al., 1991c). Because this form of testing is time c o n s u m i n g a n d expensive, short-term tests are needed to screen fibres for their potential toxicity. A n u m b e r of short-term screening systems have been suggested, such as short-term animal assays (Warheit et al., 1991), in vitro assays ( M o s s m a n a n d Sesko,
CD = crocidolite; CFE = colony-forming efficiency; CH =chrysotile; CHO=Chinese hamster ovary; CM = complete medium; FCS = foetal calf serum; I C P = inhibition of cell proliferation; MN = micronuclei; NA = nuclear abnormality; OM = optical microscopy; PN = polynuclei; RCF = refractory ceramic fibre; SEM = scanning electron microscopy; SHE = Syrian hamster embryo; TEM = transmission electron microscopy.
Abbreviations:
1990), a n d a battery of in vitro assays c o m b i n e d with short-term animal inhalation tests (Hesterberg et al., 1992). Any short-term test system proposed as a means of screening fibres for their toxicological potential must be validated by d e m o n s t r a t i n g that the effects observed using the screening system correlate with h u m a n epidemiological data a n d / o r animal inhalation data. A recent chronic rodent inhalation study of the toxicity and tumorigenicity of refractory ceramic fibres (RCFs; Hesterberg et al., 1991a) has provided the animal data necessary to m a k e such correlations. C o n d u c t i n g in vitro research in parallel with in vivo studies also provides a n additional perspective on toxic mechanisms at the cellular level. Thus, the present study was conducted: (1) to develop and validate in vitro assays that could be useful as part of a battery of short-term toxicity screening tests for fibres; a n d (2) to contribute to a more complete u n d e r s t a n d i n g of the mechanisms of fibre-cell interactions. In the present study, Chinese h a m s t e r ovary (CHO) cells were exposed in vitro to the same four sizeselected refractory ceramic fibres that were used in the in vivo study. Toxic e n d p o i n t s assessed in the in vitro test system [inhibition of cell proliferation (ICP), inhibition of colony-forming efficiency (CFE) and the induction of micronuclei ( M N ) a n d polynuclei (PN)] 317
318
G . A . HART et al. Table I. Chemical composition of RCFs tested, as supplied by the Thermal Insulation Manufacturers' Association Fibre Repository Percentage of mass of Oxide
RCFI
RCF2
RCF3
RCF4*
SiO 2 AIzO3 F%O 3 Cr20 ~ TiO2 CaO MgO Na20 K20 ZrO 2
47.7 48.0 0.97 0.03 2.05 0.07 0.08 0.54 0.16 0.11
50.0 35.0 < 0.05 <0.01 0.04 0.05 0.01 <0.3 <0.01 15.0
50.8 48.5 0.16 <0.01 0.02 0,04 <0.01 0.19 <0.01 0.23
47.7 48.0 0.97 0.03 2.05 0.07 0.08 0.54 0.16 0.11
*RCF4, which was RCF1 heated to 1300 C to simulate its use in the workplace, is composed primarily of crystalline materials: approximately 27% of its mass is cristobalite (crystalline silica) and the remainder is mullite. The other RCFs are composed of amorphous materials.
were compared with toxic endpoints in the animal inhalation study (induction of lung fibrosis and incidence of thoracic tumours). The present study is limited in that only three parameters of toxic response are evaluated and only one cell type is used. However, a correlation was observed between the in vivo pathogenesis and the in vitro toxicity of RCFs. This study is the first phase of investigations that will expand to include other cell types and evaluate other in vitro endpoints, in an attempt to develop a battery of in vitro tests that will be useful in screening fibrous materials for their toxic potential. MATERIALS AND METHODS
Test fibres. Four different compositions of refractory ceramic test fibres were obtained from the TIMA Fibre Repository*: RCF1 (kaolin), RCF2 (zirconia), RCF3 (high purity) and RCF4 (after service). These are size-selected, respirable test fibres that were used in rodent inhalation studies (Hesterberg et al., 1991,b,c,d). Details of the chemical compositions of the fibres were provided by the TIMA Repository xand are listed in Table 1. UICC (International Union Against Cancer) crocidolite (CD) and UICC chrysotile (CH) asbestos were obtained from
*Thermal Insulation Manufacturers" Association Fiber Repository.
Dr V. Timbrell (Pneumoconiosis Research Unit, Llandough Hospital, Penarth, Glamorgan, UK). Fibre dimensions (Table 2) were determined using scanning electron microscopy (SEM) for the RCFs and CD and transmission electron microscopy (TEM) for CH. Electron micrographs in Plate la and b illustrate the two extremes of average particle dimensions of the RCFs. Plate la (RCF3) is representative of the three longer fibres, RCF1, 2 and 3 (average lengths of 21.5, 16.7 and 24.3 #m, respectively); Plate lb illustrates the dimensions of the shortest fibre, RCF4 (average length 9.2,um). The densities of fibres were also determined by SEM, TEM and optical microscopy (OM). Electron microscopy data on fibre numbers were used to calculate fibre exposure as numbers of fibres/cm2. OM fibre counts of the RCFs were used to confirm the SEM data; the two asbestos fibres were not counted with OM because most of these fibres are too small to be seen using optical microscopy. For OM fibre counts, 1 ml aqueous fibre suspension (100 #g of fibre/ml) was allowed to settle in a 60-mm culture dish (5Ftg/cm2). The fibres (particles with length/ diameter ratio of >~3) were counted at a magnification of × 400 using an ocular grid. Cell cultures and in vitro assays. The CHO-K 1 Cell line was derived from CHO tissue (Puck, 1958) and has an epithelia-like morphology. Although the tissue of origin is not known to be a target tissue of fibre-induced disease, CHO cells were selected for these studies because they divide rapidly and were observed to internalize asbestos fibres readily. Frozen cells were obtained from American Type Tissue Collection (Rocky±lie, MD, USA) and were thawed and plated in tissue-culture flasks with complete medium (CM) consisting of Ham's F I 2 supplemented with 10% foetal calf serum (FCS), 2 mM-L-glutamine, and l ml Fungi-Bact solution/100ml medium (all from Irvine, Santa Ana, CA, USA). To create a large bank of uniform stock cells for experiments, cultures were incubated (37"C, 5% CO2 and 100% humidity), allowed to replicate for several days (three-four doublings), and then harvested with 0.05% trypsin and 0.02% ethylenediaminetetraacetic acid in Hanks' balanced salt solution (Irvine). The stock cells were then suspended in freezing medium, consisting of CM supplemented with an additional 10% FCS and
Table 2. Physical characterization, by electron microscopy*, of fibres tested Test fibre Abbreviation
Fibre dimensions (pm) Fibre type
RCFI RCF2 RCF3 RCF4
Kaolin Zirconia High purity After service
CD CH
UICC Crocidolite UICC Chrysotile
No. of fibres/ng × 1000 RCFs 4.7 4.6 3.1 6.3 Asbestos 2400 4400
Diameter (mean _+ SD)
Length (mean _+ SD)
1.03 + 0.73 1.11 _+ 0.82 1.22 + 0.98 1.43 -- 0.79
21.5 ± 16.12 16.7 ± 15.03 24.3 ± 18.82 9.2 ,+ 7.08
0.21 ± 0.12 0.12 ± 0.07
1.81 ± 1.94 1.65 ± 1.83
*Transmission electron microscopy was used to analyse CH. Scanning electron microscopy was used to analyse RCFs and CD.
Cytotoxicity of RCF to CHO cells
319
120
50 I A 1OO| ,~ T
40
o=~8o
-~ -~ 30
[,
..'.AChrysotile "
:
RCF1
Crocidolite
.:
~
.c.. -/ "/
/~
RCF4
.o
r~ o 20
-,;i
Crocidolite
" ~ . ~
RCF2
Chrysotile
RCF1 RCF3 I 10
0
I 20
Concentration
I
I 30
40
( # g / s q . cm)
12% dimethyl sulphoxide (Sigma Chemical Co., St Louis, MO, USA), frozen at - 8 0 ° C and later transferred to liquid nitrogen for storage. For experiments, stock cells were thawed and plated in culture flasks (Falcon, N J, USA) in CM and allowed to grow for 2-4 days. I C P assay. For ICP assays, 60-ram culture dishes were seeded with 100,000 cells in 5 ml CM/dish. After 24 hr of incubation, fibres were added to cultures in 1 ml CM/dish. Dishes were gently agitated to disperse fibres over the bottom. Negative control cultures received 1 ml CM/dish. Each exposure group was set up in triplicate. After 3 days of exposure, incubated as described above, cells were harvested using trypsin (as described above) and counted using a Coulter counter (Coulter Electronics Hialeah, FL, USA). Relative proliferation was determined by dividing the number of cells present in each exposed culture by the number of cells present in negative control cultures. C F E assay. For the CFE assay, 60-mm culture dishes were seeded with 200 cells in 5 ml complete medium and exposed as described for ICP above.
120
._. .~---.
i 0
5
I 10
I 15
Concentrotion
Fig. 1. Inhibition of proliferation of CHO cells cultured for 3 days with various RCFs and with crocidolite and chrysotile asbestos (concentrations expressed as/~g/cm2).
o
o,
T
I 20
[ 25
I 30
( F g / s q . cm)
Fig. 3. Induction of nuclear abnormalities (percentage of cells with micronuclei and/or polynuclei) in CHO cells cultured for 3 days with various RCFs and with crocidolite and chrysotile asbestos. Each exposure was set up in triplicate. After 5 days of exposure, colonies were stained with 0.4% (w/v) Giemsa in buffered methanol (Sigma). Colonies with > 10 cells were counted using a stereoscope at low power. CFE was determined by dividing the number of colonies in each exposed culture by the number of colonies in unexposed cultures. M N and P N induction. For these assays, cultures were prepared as described for ICP above but mitomycin C (MMC, 2/~M)was also included as a positive control. After 2 days of exposure, culture dishes were fixed with methanol-acetic acid (3 : 1, v/v) and stained with 0.01% acridine orange (from Sigma; as described in Clark, 1981). Using a microscope with epifluorescence, the percentage of cells containing micronuclei (MN) and/or polynuclei (PN) was determined for each culture dish. An MN was defined as a nucleus that appeared to be less than half the size of the normal-sized nuclei; PN was defined as a cell having bi- or multiple nuclei or a lobed nucleus (in many cases the different forms of PN were difficult to distinguish). A minimum of 100 cells/dish and 2-3 dishes/exposure group were scored for (1) MN but no other visible nuclear abnormalities, (2) both MN and PN, or (3) PN without MN. Calculations included the percentage of cells with MN (including both uninucleate and polynucleate cells), the percentage of PN (with or without MN), and the percentage of cells with either type of nuclear abnormality (NA), that is, having MN and/or PN.
RESULTS 20
Crocidolite
~ ' t ~
RCF2 RCF3 RCF1
O O
I 10
I 20
Concentration
~ 30
I 40
( # g / s q . cm)
Fig. 2. Colony-forming efficiencyof CHO cells cultured for 5 days with various RCFs and with crocidolite asbestos•
The results of the three in vitro assays (Figs 1-3) all show the same relative toxicities for the four RCFs when concentration is based on/~g/cm2: in each case RCF4 is the least toxic, RCF2 is intermediate, and RCF1 and RCF3 are the most toxic. In all three assays, CD was more toxic than the RCFs; in the ICP (Fig. 1) and MN/PN (Fig. 3) assays, CH was more
320
G.A. HART et al. 120
lO0
o--- o
80
u 60 ~-.~
~>~ 40 g 2O o
I
I
I
I
I
I
I
25
50
75
lOO
125
1.'50
175
Fibres/sq.
cm
I 200
x 1000
Fig. 4. I n h i b i t i o n of p r o l i f e r a t i o n of C H O cells cultured for 3 d a y s with various R C F s and with crocidolite a n d chrysotile asbestos ( c o n c e n t r a t i o n s expressed as fibres/cm2). E r r o r bars indicate the S E M from four or five trials.
toxic than CD (CH was not included in the CFE assay). Although some of the error bars (SEM, standard error of the means) overlap, in each individual test performed with these fibres (more than 14 individual tests in all), the same relative toxic±ties resulted. When the concentration of the RCFs is expressed as numbers of fibres/cm 2 (Fig. 4), similar relative toxic±ties of the different RCFs is observed, although the spread between RCFI and RCF3 is increased. Data from the asbestos types are not plotted in this manner since both the length and diameter are an order of magnitude less than those of RCF fibres, making a fibre-to-fibre comparison inappropriate.
Plates 2-4 were all photographed using fluorescent light with the exception of Plate 3b and d, which were photographed using phase-contrast light. Nuclear abnormalities (NA) observed in all the fibre-exposed cultures appeared qualitatively similar (Plates 3 and 4). Some examples of NA include: PN that appear to be clearly more than one nucleus (Plate 4a and b) as well as PN that appear to be bi- (Plate 3c and d) or multi-lobed (Plate 3a and b); micronuclei with one apparently normal nucleus or with PN (both in Plate 4a). The phenomenon of multiple nuclei aligned along the length of a fibre (Plate 4b) was not as prominent with asbestos as with RCF exposure and is likely to result from the higher frequency of long fibres in the RCFs than in the asbestos. The relationship between intracellular fibres and nuclear distortion is unmistakable in all of the photographs, and is especially apparent when phase-contrast and fluorescent views of the same field are compared as in Plate 3a,b and c,d. Both wholly and partially internalized fibres are visible (Plate 4c). The frequencies of MN and PN induced by each of the fibres and by mitomycin C are shown in Table 3. A difference was observed in the types of NA induced by MMC compared with fibres: 2/~M-MMC induced a 29% incidence of MN and a 10% incidence of PN, whereas all of the fibre types induced roughly half as many MN as PN. In the fibre-exposed cultures, the incidence of both types of nuclear abnormalities (NA; percentage of cells with MN and/or PN) was strongly concentration dependent, whereas the incidence of MN alone, although elevated above background levels, was generally weakly concentration dependent. CH at 5/~g/cm2 induced NA in 49% of
Table 3. Incidence of abnormal nuclei in Chinese hamster ovary cells cultured with RCFs or asbestos Percentage of Test fibre RCF1
RCF2
RCF3
RCF4
Crocidolite Chrysotile
Exposure concentration (l~g)
No. of tests
0 5 10 20 0 5 10 20 0 5 10 20 0 5 10 20 0 5 0 1
Mitomycin C
2 5 0 2/~M
*One or more micronuclei; poly- or uninucleate. "t'More than one nucleus or lobed nucleus.
Micronuclei* (Mean ± SEM)
Polynucleit (Mean ± SEM)
3±0 5±0 8±1 10±1 2±0 4±0 6±1 8±0 2±0 6±1 8±1 8±1 3±0 5±1 5±1 9±2 2±0 9±2 2 10 18 7 2 29
2±1 11+1 18+2 26±3 2±1 8±0 13±3 22±3 2±1 11±2 17±4 28±4 2±1 9±1 11±2 17±3 0±1 21±0 I 17 29 46 1 10
Plate 1. Electron photomicrographs of (a) RCF3 and (b) RCF4 fibres.
321
Plate 2. CHO cells growing in culture medium only. Note uniformity of nuclear size, shape and number/cell. Acridine orange stain, photographed with fluorescent light. Original magnification × 400. Plate 3(a,b). Caption opposite. 322
Plate 3(c,d) Plate 3. C H O cells exposed to crocidolite asbestos, photographed with fluorescent light (a,c) and with phase contrast optical microscopy (b,d). Note the broad range of nuclear size, shape and number/cell. A polynucleate cell with two very small micronuclei is visible in the centre of the field (Plate 3a); fibres are visible in this cell in Plate 3b, and appear to be associated with the nuclear lobes. The central cell in Plate 3c has a C-shaped polynucleus; In Plate 3d a packet of fibres can be seen in the center of the "C". Acridine orange stain. Original magnification x 1000.
323
Plate 4. CHO cells exposed to RCF2, original magnification x400 (a-b), and RCF3, original magnification x 1000 (c). Note the variety of nuclear configurations: uninucleate, binucleate and polynucleate cells, with and without micronuclei. Alignment of nuclei along the length of a long fibre is visible in Plate 4b; this was a common phenomenon in cultures exposed to RCFs. Fibres are visible in all three photographs, Acridine orange stain, photographed with fluorescent light. 324
Cytotoxicity of RCF to CHO cells
325
Table 4. Comparisonof the in vitro and in vivo toxic endpoints for RCFs Toxic endpoints In vitro
In vivo
LCso (,ug/cm 2) Exposure group
ICP
CFE
Wagner score* at 6 months
Percentage with lung tumourst
Percentage with mesotheliomast
Negative control RCF1 RCF2 RCF3 RCF4
-13 17 13 30
-10 18 12 22
1.0 4.0 3.0 4.0 2.7
1.6 13.0 6.9 14.6 3.2
0.0 1.5 2.3 1.5 0.8
*1 =normal lung; 2 = m a c r o p h a g e infiltration; 3=inflammation and bronchiolization; 4 = m i n i m a l but irreversible fibrosis; 5,6,7 and 8 = increasingly severe fibrosis. tNumber of animals with tumour, expressed as a percentage of the number of animals at risk.
the cells; CD at 5 Ftg/cm2 induced NA in 28% of the cells. The NA incidence for RCFs at 20/~g/cm z ranged from 22 to 33% (Fig. 3).
the test fibres. Secondly, specificity is evident in this model, in the consistent relative response of these cells to the six different fibres. Thirdly, cellular uptake of fibres and alterations of nuclear morphology are clearly major CHO cell responses. The observed DISCUSSION fibre-induced nuclear abnormalities fell into two basic The toxic endpoints of the in vitro studies were categories (MN and PN) and the incidence of each compared with those of the animal inhalation studies was consistent for the six fibres in three or more using the same RCF test fibres (Table 4). The LCs0 separate assays. The mechanism of MN induction is the in vitro concentration that resulted in 50% as may be different from that of PN induction. MN are many cells or colonies in fibre-treated cultures as in believed to form when chromosomes or fragments are unexposed control cultures. Wagner grades (defined separated from the migrating masses of replicate in Table 4) were used to quantitate the lung pathol- chromosomes during mitosis. Previous studies have ogy observed in the chronic inhalation study of RCFs shown a correlation between the incidences of MN where: 1 indicates a normal lung; 2 indicates macro- and aneuploidy in three different cell lines (CHO phage infiltration; 3 represents inflammation and cells, Sincock et al., 1982; Syrian hamster embryo bronchiolization; 4 indicates minimal but irreversible [SHE] cells, Oshimura et al., 1984; human lymphofibrosis; and 5, 6, 7 and 8 represent increasingly severe" cytes, Migliore and Nieri, 1991). PN may result from fibrosis (Hesterberg et al., 1991b). In the animal the failure of cytokinesis following mitosis. Finally, inhalation studies, three to six animals per group the incidence of total NAs was approximately the were scored every 3-6 months during the 24-month inverse of the proliferation curves, indicating that exposure to fibre aerosol and at the end of the lifetime disruption of mitosis may be the major or immediate study. Only the 6-month scores are given for com- cytotoxic effect, rather than the disruption of cytoparison here, because all fibre-exposed lung scores plasmic metabolic processes. However, validation of plateaued near grade 4 soon after the 6 month time this hypothesis would require cell viability testing. Differences in the length of the four RCFs probpoint (air-control lungs remained at level 1). Also indicated in Table 4 are the percentages of animals ably account for the differences in both the in vitro that developed lung tumours and mesotheliomas by cytotoxicity and the in vivo lung pathology observed the end of the RCF inhalation study. The relative in the animal inhalation study. Many previous in vitro toxicity of the different RCFs seen in vitro generally studies also report a relationship between fibre length correlates with that observed in vivo: in each case, and cytotoxicity. Brown et al. (1979) demonstrated RCF4 is least toxic, RCF2 is intermediate, and RCF! that glass fibres were more toxic to cultured macroand RCF3 are more toxic (except in the case of the phages and to V79-4 cells when average fibre length was greater than 10ttm; Lipkin (1980) observed a percentage mesotheliomas seen in vivo where RCF1 and RCF3 results were intermediate, with RCF2 relationship between fibre length and the reduction of showing most toxicity and RCF4, least). cell proliferation in cultures of P388DI cells exposed The present study provides results that contribute to several populations of glass fibres; Tilkes and Beck towards both the development of short-term screen- (1980 and 1983a,b) showed a direct relationship ing tests that can be used to assess the toxic potential between fibre length and both lactose dehydrogenase of fibrous dusts, and to our understanding of the release (macrophages and ascites) and reduction in mechanisms of fibre toxicity at the cellular level. proliferation (ascites) in cell cultures exposed to First, the CHO cell test system is highly responsive to fibrous glass. A mechanism to explain the relationship between fibre exposure in all three parameters investigated; concentration-dependent reductions in cell prolifer- fibre length and cytotoxicity could be related to cell ation and cell colony formation as well as increases diameter. For example, when the flattened, adherent in NAs were observed following exposure to each of cell becomes spherical as it enters metaphase prior to
326
G.A. HART et al.
cell division, internalized fibres that are longer than the diameter of the m e t a p h a s e cell might p u n c t u r e the cell, causing cell death. Fibre length has also been shown to be i m p o r t a n t in the induction ofcytogenetic effects by internalized fibres (Hesterberg et al., 1986; O s h i m u r a et al., 1984) and in the t r a n s f o r m a t i o n of cultured cells (Hesterberg a n d Barrett, 1984). In a study t h a t d e m o n s t r a t e d direct interaction between intracellular asbestos fibres a n d migrating c h r o m o somes during a n a p h a s e in cultured S H E cells, Hesterberg a n d Barrett (1985) suggested that long fibres are m o r e easily entangled in the migrating c h r o m o s o m e s or spindle a p p a r a t u s t h a n are short fibres. The correlation between in vitro a n d in vivo toxic e n d p o i n t s is encouraging. However, more research is needed to determine which in ~,itro test systems a n d procedures will best correlate with animal i n h a l a t i o n studies a n d h u m a n epidemiology. Again, fibres tested in long-term animal inhalation studies will be needed for validation of the sensitivity and specificity of these short-term assays. Such fibres should be well characterized chemically (chemical composition, crystallinity a n d surface chemistry) a n d physically (length, width, total surface area, n u m b e r of fibres/unit mass). F u r t h e r m o r e , the test fibres should include a range of d e m o n s t r a t e d in vivo toxicities, from non-toxic to strongly toxic. In vitro testing should include other cell types a n d other in vitro parameters (i.e. durability, cytotoxicity and mutagenesis). Several p a r a m e t e r s of in vivo pathogenesis c a n n o t at this time be m e a s u r e d in vitro, for example lung deposition, lung clearance and long-term biopersistence. Some work has been done in the third area, by incubating fibres in flow-through systems using synthetic physiological solutions (Law et al., 1990 and 1991). However, such systems do not contain all o f the factors that might be i m p o r t a n t in the biopersistence of fibres in the lung, such as enzymes, lysosomal secretions a n d mechanical removal by phagocytic cells. These limitations might be overcome by including short-term animal inhalation studies in a battery o f toxicity screening tests. Acknowledgements--Funding was provided by MTC Research and Development. We thank Bill Miiller and Barry Fitzpatrick for the preparation of test fibres, R. Hamilton, J. Strothers and F. D'Ovidio for electron microscopy, and Lisa M. Kathman for technical assistance. REFERENCES
Brown R. C., Chamberlain M. and Skidmore J. W. (1979) In vitro effects of man-made mineral fibers. Annals of Occupational Hygiene 22, 175 179. Clark G. (1981) Staining Procedures. 4th Ed. Williams and Wilkins, Baltimore, MD. Hesterberg T. W. and Barrett J. C. (19843 Dependence of asbestos- and mineral dust-induced transformation of mammalian cells in culture upon fibre dimension. Cancer Research 44, 2170-2180. Hesterberg T. W. and Barrett J. C. (1985) Induction by asbestos fibers of anaphase abnormalities: mechanism for aneuploidy induction and possibly carcinogenesis. Carcinogenesis 6, 473-475.
Hesterberg T. W., Butterick C. J., Oshimura M., Brody A. R. and Barrett J. C. (1986) Role of phagocytosis in Syrian hamster cell transformation and cytogenetic effects induced by asbestos and short and long glass fibers. Cancer Research 46, 5795-5802. Hesterberg T. W., Hart G. A. and Bunn W. B. (1992) In vitro toxicology of fibres: mechanistic studies and possible use for screening assays. In Contemporary Issues in Fiber Toxicology. Edited by D. Warheit. Academic Press, New York. In press. Hesterberg T. W., Mast R., McConnell E. E., Chevalier J., Bernstein D. M., Bunn W. B. and Anderson R. (1991a) Chronic inhalation toxicity of refractory ceramic fibers in Syrian hamsters. In Proceedings of the International Workshop on Mechanisms of Fibre Carcinogenesis. Edited by R. C. Brown et al. pp. 531-538. Plenum Press, New York. Hesterberg T. W., Mast R., McConnei1 E. E., Vogel O., Chevalier J., Bernstein D. M. and Anderson R. (1991b) Chronic inhalation toxicity and oncogenicity study of refractory ceramic fibers in Fisher 344 rats. Toxicologist II, 254. Hesterberg T. W., Vu V., McConnell E. E., Chase G. R., Bunn W. B. and Anderson R. (1991 c) Use of animal models to study man-made fiber carcinogenesis. In Current Communications in Molecular Biology: Molecular Mechanisms of Fiber Carcinogenesis. Edited by B. Brinkley, J. Lechner and C. Harris. Cold Spring Harbor Laboratory, New York. In press. Law B., Bunn W. B. and Hesterberg T. W. (1990) Solubility of polymeric organic fibers and man-made vitreous fibers in gambles solution. Inhalation Toxicology 2, 321-339. Law B., Bunn W. and Hesterberg T. W. (1991) Dissolution of natural mineral and man-made vitreous fibres in karnovsky's and formalin fixatives. Inhalation Toxicology 3, 309-321. Lipkin L. E. (1980) Cellular effects of asbestos and other fibers: correlations with in vivo induction of pleural sarcoma. Environmental Health Perspectives 34, 91 102. Migliore L. and Nieri M. (1991) Evaluation of twelve potential aneuploidogenic chemicals by the in vitro human lymphocyte micronucleus assay. Toxicology in Vitro 5, 32%336. Mossman B. T. and Sesko A. M. (1990) In vitro assays to predict the pathogenicity of mineral fibers. Toxicology 60, 53-61. Oshimura M., Hesterberg T. W. and Barrett J. C. (1984) Correlation of asbestos-induced cytogenetic effects with cell transformation of Syrian hamster embryo cells in culture. Cancer Research 44, 5017-5022. Puck T. T. (1958) Genetics of somatic mammalian cells. III. Long-term cultivation of euploid cells from human and animal subjects. Journal of Experimental Medicine 108, 945. Sincock A. M., Delhanty J. D. A. and Casey G. (19823 A comparison of the cytogenetic response to asbestos and glass fibre in Chinese hamster and human cell lines. Demonstration of growth inhibition in primary human fibroblasts. Mutation Research 101, 257 268. Tilkes F. and Beck E. G. (1980) Comparison of lengthdependent cytotoxicity of inhalable asbestos and manmade mineral fibres. In Biological Effects of Mineral Fibres. IARC Scient. Publ. no. 30. pp. 475~483. International Agency for Research on Cancer, Lyon. Tilkes F. and Beck E. G. (1983a) Macrophage functions after exposure to mineral fibres. Environmental Health Perspectives 51, 67 72. Tilkes F. and Beck E. G. (1983b) Influence of well-defined mineral fibres of proliferating cells. Environmental Health Perspectives 51, 275 279. Warheit D. B., Carakostas M. C., Hartsky M. A. and Hansen J. F. (19913 Development of a short-term inhalation bioassay to assess pulmonary toxicity of inhaled particles: comparisons of pulmonary responses to carbonyl iron and silica. Toxicology and Applied Pharmacology 107, 350 368.