Effects of mineral fibers on the expression of genes whose product may play a role in fiber pathogenesis.

To determine which factors are useful for the risk assessment of man-made fibers, we examined the gene expression of proinflammatory cytokines, growth factors, manganese superoxide dismutase (MnSOD), and inducible nitric oxide synthase (iNOS) in mineral fiber-exposed rats by means of reverse transcription-polymerase chain reaction (RT-PCR). Male Wistar rats received a single intratracheal instillation of either saline (control) or two types of fibers (2 mg of Union Internationale Centre le Cancer (UICC) chrysotile or alumina silicate refractory ceramic fiber [RCF]). Expression of interleukin-1 alpha (IL-1 alpha), interleukin-6 (IL-6), tumor necrosis factor alpha (TNF-alpha), platelet-deriving growth factor-A, (PDGF-A), platelet-deriving growth factor-B (PDGF-B), transforming growth factor beta 1 (TGF-beta 1), basic fibroblast growth factor (bFGF), MnSOD, and iNOS mRNA from lung and lipopolysaccharide (LPS)-stimulated alveolar macrophages (AM) were assessed by RT-PCR. Among these factors, IL-1 alpha, TNF-alpha, IL-6, bFGF, and iNOS would be the possible parameters for the risk assessment of fibers. In a follow-up study, we investigated the time course (3 days, 1 week, 1 month, and 3 months) of expression of IL-1 alpha and TNF-alpha by LPS-stimulated AM exposed to mineral fibers in vivo. Male Wistar rats were instilled intratracheally with saline or fibers (2 mg of Union Internationale Contre le Cancer UICC crocidolite or potassium octatitanate whisker [TW]). The expression of IL-1 alpha mRNA by fibers was greatest in TW, crocidolite, chrysotile, and RCF-instilled rat AM, in that order. The increase of IL-1 alpha and TNF-alpha mRNA in AM peaked at 1 month and 3 days after exposure to crocidolite or TW, respectively. The expression of IL-1 alpha by fibers (crocidolite, chrysotile, TW, and RCF) may be a good indicator of the pathologic potential of fibers.


Introduction
Release of oxidants by these cells may lead to lung injury (3). Manganese superoxide dismutase (MnSOD) and inducible nitric oxide synthase (iNOS) are biomarkers for superoxide (4). Growth factors like platelet-derived growth factor A (PDGF-A), platelet-derived growth factor B (PDGF-B), and transforming growth factor P1 (TGF-1) signal interstitial fibroblasts to replicate and modulate their production of connective tissue proteins (5). Fibroblast growth factors modulate potent growth of cells and neovascularization (6).
These internal tissue cells and cytokine cascades could account for the chronic nature of the inflammation. The accumulation of inflammatory cells, fibroblasts, and connective tissue matrices leads to lung remodeling such as thickening of alveolar and bronchiole walls.
Even though the mechanism is not completely understood, evidence suggests that various factors are related to each fibrotic process in the lung (Figure 1). It is important to investigate among these factors parameters useful for the risk assessment of MMF and the kinetics of their expression in the process of the lung remodeling.

Fiber Preparation
The fibers used in this study were Union Internationale Contre le Cancer (UICC) crocidolite asbestos (crocidolite), UICC chrysotile asbestos ([chrysotile]; potassium octatitanate whisker (TW), and alumina silicate refractory ceramic fibers (RCF) (1). The crocidolite preparation, measured using scanning electron microscopy, had a geometric mean diameter of ( 0.20 pm (SD 1.5) and a geometric mean length of 1.3 pm (SD 2.3). For chrysotile, geometric mean diameter and geometric mean length were 0.085 pm (SD 1.4) and 0.7 pm (SD 1.9), respectively. For TW, they were 0.41 pm (SD 1.5) and 2.8 pm (SD 2.0), respectively. For RCF, they were 1.2 pm (SD 1.7) and 9.6 pm (SD 1.9), respectively. Intratracheal Instillation Study 1 Ten-week-old male Wistar rats, in groups of five per treatment, were intratracheally instilled with saline or fiber (2 mg chrysotile or RCF). One month after tracheal instillation, bronchoalveolar lavage (BAL) was performed using the left lung. The cells recovered from BAL were plated in tissue culture plates and allowed to attach for 1 hr at Environmental Health Perspectives * Vol 105, Supplement 5 * September 1997  Figure 2. Semiquantification of IL-la mRNA and 3-actin in serial thermocycle. IL-la and P-actin was co-amplified in LPS-stimulated rat AM exposed to saline or chrysoytile. The number of thermocycles used allowed quantitation without saturation. IL-i a and the 1-actin PCR products quantitated by densitometry analysis using NIH image 1.56 were plotted against the number of amplification cycles.
37°C with RPMI-1640 medium containing 10% fetal bovine serum (7). Adhered AM were adjusted to a concentration of 1 x 105/ml, and were stimulated with 10 pg/ml lipopolysaccharide [LPS]; Sigma Chemical Co., St. Louis, MO) (7). Cells were cultured for 2 and 6 hr on cell culture plates at 37°C in a CO2 incubator. After incubation, mRNA was extracted using a Quick Prep kit (Pharmacia Biotech, Uppsala, Sweden). RNA was also extracted from the right lung using the guanidinium thiocyanate-phenol-chloroform method (8). cDNA Synthesis, Polymerase, and Chain Reaction RNA was used for the synthesis of singlestrand cDNA using Moloney murine leukemia virus-derived reverse transcriptase (Perkin Elmer, Norwalk, CT). Equal amounts of cDNA from each sample were then used for amplification by specific primers for rat ILIa, IL-6, TNF-a, PDGF-A, PDGF-B, TGF-f13, bFGF, MnSOD, and iNOS. The amplification was performed with a Thermocycler (Astech, Fukuoka, Japan) under the following conditions: 94°C for denaturation for 45 sec, 60°C for 45 sec for annealing, and 72°C for 2 min for extension. ,-Actin was co-amplified as an internal standard to quantitate polymerase chain reaction (PCR) amplification of mRNA. The number of thermocycles used allowed quantitation without saturation (9) (Figure 2). Four to five cDNAs per treatment group that expressed 1-actin uniformly were entered into the study.
Detection of the fragments amplified by the PCR was made by electrophoresis on a 2% agarose gel and visualized by ethidium bromide staining. The gel was photographed with Polaroid Type 665 positive/negative film (Polaroid, Cambridge, MA) over ultraviolet (UV) light at the same exposure and developing time. The bands of the positive film were scanned and the density of each PCR product was measured using National Institutes of Health (NIH) image 1.56 software (written by W. Rasband, NIH, Bethesda, MD). The ratio of specific gene product to J-actin product was used for further analysis. Statistcl Analysis Data are expressed as mean ± SEM. Comparisons were performed using the Student's t test with a paired test used for paired data. Correlation coefficients in ILla expression by fibers were obtained by Spearman's rank-order method; p values less than 0.05 were considered significant.

Inratracheal Instillation Study 2
Wistar rats in groups of five animals per treatment were intratracheally instilled with saline or fibers (2 mg of crocidolite or TW. After exposure for 3 days, 1 week, 1 month, and 3 months, AM were recovered from the left lung and stimulated with 10 pg/ml LPS for 2 hr. Expression of IL-la and TNF-a by LPS-stimulated AM was assessed by reverse transcription (RT)-PCR as in intratracheal instillation study 1.

Gene Expression One Month after Intratracheal Instillaton
We examined rat ILl-a, IL-6, TNF-a, PDGF-A, PDGF MnSOD, and iNOS from LPS-stimulated AM and right lung. When compared with the control (saline-instilled) group, rats exposed to chrysotile and RCF had significantly increased levels of IL-la and TNFa mRNA in AM (Figure 3). TNF-a mRNA in the lung increased only in chrysotileexposed rats. TNF-a, iNOS, and bFGF mRNA in the lung increased significantly in chrysotile-exposed rats. IL-6 mRNA in AM was significantly increased in chrysotile-exposed rats.
rlme-Cou Expression ofIL-la and TNF-a Total cell counts in the saline-instilled group were approximately 5 x 105 cells ( Figure 2). In the crocidoliteand TWinstilled groups, total cell counts increased approximately 3-fold in 3 days compared with those in the control group. Subsequently, the total cell counts from these groups decreased at 1 month, then increased again 3 months after instillation. Most of the bronchoalveolar lavage fluid (BALF) consisted of AM. Contents of neutrophils in BALF were around 25% at 3 days and at 1 week after crocidolite and TW instillation (Figure 4). Figure 5A shows the time course of gene expression of IL-la by RT-PCR in LPSstimulated AM after 3 days, 1 week, 1 month, and 3 months of exposure to saline, crocidolite, or TW. From the densitmetric analysis, the level of IL-la mRNA in AM peaked at 1 month after instillation of crocidolite or TW ( Figure SB). Levels ofTNFa mRNA in AM peaked at 3 days after the instillation of crocidolite or TW ( Figure 6).

IL-la mRNA Expression by Fibers
As the level of IL-la mRNA peaked at 1 month after instillation, we compared the expression of IL-la by fibers (crocidolite, chrysotile, TW, and RCF) 1 month after instillation. The expression of IL-la mRNA by fibers was greatest in TW-, crocidolite-, chrysotile-, and RCF-instilled rat AM in that order (Figure 7), which was significant (p < 0.05) by Spearman's rank correlation coefficient.

Discussion
It has been suggested that MMF with widths and lengths similar to asbestos are most likely associated with the induction of pulmonary fibrosis and lung cancer (10). To assess molecular factors that may be useful for risk assessment of MMF, we examined the gene expression of proinflammatory cytokines, growth factors,  MnSOD, and nitric oxide synthase (NOS) in mineral fiber-exposed rats using RT-PCR. RT-PCR has multiple advantages (9): a) its exquisite sensitivity allows the detection of extremely rare mRNA in small numbers of cells in a semiquantitative manner; b) it differs from the assay system of proteins, and the same methodology can be applied in analyzing the expression of many genes; and c) cDNA can be used for future studies. However, RT-PCR has several limitations. It can give useful information about transcription but does not provide information about translation, posttranscriptional processes, or cytokine exportation. In addition, because of exponential amplification over repeated cycles and differences in reaction efficiencies, RT-PCR is useful for detecting relative differences in the amount of a given mRNA. A number of studies describe semiquantitative RT-PCR techniques whereby relative mRNA production was estimated to be between samples; comparison of amplified target molecules co-amplified constitutively expressed mRNA (11). Housekeeping genes such as P-actin or D-glyceraldehyde-3-phosphate dehydrogenase (G3PDH) often are selected as endogenous internal standard in such studies (11). This housekeeping gene may be increased under some conditions.
In our experiment, P-actin expression was not altered significantly in the samples from different groups, based on AM cell numbers ( Figure 2). Since the endogenous internal standard is usually a housekeeping gene, it frequently is found in higher concentration than the target message. This limitation was avoided by using different PCR cycles for both products. Because Figure 5. Time-course of expression of IL-la mRNA in LPS-stimulated rat AM exposed to mineral fibers. (A) Ethidium bromide staining of PCR products separated in 2% agarose gel. The 623-bp products for IL-la and the 357-bp products for f-actin are indicated. LPS-stimulated AM from four to five rats per treatment per time point were examined. (B) Levels of IL-1 a mRNA in AM exposed to mineral fibers. Results are expressed as the ratio of IL-1 a to P-actin (mean ± SEM).
To find more sensitive and specific   (15). Short-term inhalation of RCF C, control, Cro, crocidolite; Chr, chrysotile; TW: potas-resulted in markedly increased IL-10 prosium octatitanate whisker; RCF; alumina silicate tein expression after stimulation with LPS ceramic fibers. Results are expressed in ratio of IL-la (15). In vivo exposure of AM to LPS to fP-actin (mean ± SEM).
increased proinflammatory cytokine mRNA, although the kinetics of upregulasynthetic RNA sequence or a synthetic tion varied (16). For these reasons, we DNA sequence that is not present in the tar-examined mRNA expression in 2 and 6 hr get sample (12). Nevertheless, the house-LPS-stimulated AM. As a cautionary note, keeping gene is very useful in controlling for results from ex vivo LPS-stimulated AM differences in RNA loading and for assess-may not necessarily indicate a role for a ing differences in the quality of RNA (12). stimulated cytokine in the pathogenesis of inflammation associated with exposure to fibers in vivo.
In intratracheal study 1, AM exposed to chrysotile or RCF were found to have upregulated IL-la, TNF-a and IL-6 mRNA transcripts in response to LPS. These are proinflammatory cytokines with both inflammatory and fibrogenic activities such as attraction of inflammatory cells, production of superoxide and collagenases, and proliferation of fibroblasts (17). TNFa mRNA and protein have been detected in the lung from patients with idiopathic pulmonary fibrosis (18) and in lungs from mice with pulmonary fibrosis elicited by exposure to bleomycin or silica (19). Increased release of IL-I from AM has been reported after asbestos exposure by inhalation or intratracheal instillation (20). In inflammatory reactions, IL-6 could act as not only a proinflammatory cytokine because of its ability to induce the expression of cellular adhesion molecules on monocytes and the facilitation of their infiltration into the lung, but also as an antiinflammatory cytokine that inhibits the production of TNF and IL-1 (21). An increase has been reported in IL-6 released by bronchoalveolar cells from rats treated with asbestos or coal mineral dust (22).
Oxidants produced by inflammatory cells are thought to lead to lung injury in pulmonary fibrosis. Nitric oxide synthase (NOS) produces reactive species such as nitric oxide (NO0) and peroxynitrite anion. NO and peroxynitrate are also cytotoxic to host parenchymal cells (4). In our model, levels of iNOS mRNA increased in-lungs exposed to chrysotile. Among the FGFs, bFGF stimulated the replication of endothelial cells in vitro and new microvessel growth in vivo (23). Asbestos exposure induced lavaged cells to secrete a fibroblast growth factor from 1 to 24 weeks after exposure in rats (24). In lungs exposed to chrysotile, we found increased levels of bFGF mRNA. Based on these results, IL-la in the AM, TNF-a in the AM and the lung, IL-6 in the AM, and iNOS and bFGF in the lung would be the possible parameters of risk assessment of man-made fibers in this model. Accordingly, we set out to further investigate the time course of expression of ILla and TNF-a mRNA from exposed to crocidolite and TW.
As previously reported for rats (24), total cell count of BALF at 1 month after chrysotile or TW instillation was almost the same as the control group and increased at 3 months in the present study. Crocidolite or TW instillation resulted in pulmonary inflammation as evidenced by increased numbers of BALF neutrophils and macrophages at 3 days and 1 week after the exposure. Consistent with the acute inflammation, levels of TNF-a mRNA were greatest at 3 days after the exposure and decreased thereafter. In contrast to TNF-a, levels of IL-la mRNA peaked at 1 month after crocidolite or TW expossure.
There have been many attempts to predict the toxicity of mineral fibers based on cytotoxic potentials of fibers using a variety of cell types in vitro. In an in vitro study using AM, TW caused the highest level of TNF-a production among fibers (1). This is consistent with the present study on TNF-a mRNA expression 3 days after instillation. TNF-a mRNA expression in TW-instilled animals was higher than that in crocidolite-instilled animals. Lee et al. (25) reported that crocidolite was the most potent fibrogenic agent and was 10 times more fibrogenic than potassium octatitanate (Fybex) in terms of exposure concentration In our study, the expression of IL-la by fiber challenge (TW > crocidolite > chrysotile > RCF) may correlate with the reported pathologic potential of fibers (25,26). Accordingly, our approach may be useful for evaluating the potential toxicity of newly developed man-made fiber. Further investigations using other fibers are necessary to confirm the general applicability of the method.
Both TNF-a and IL-la are proinflammatory cytokines. It is necessary to demonstrate that acute inflammation completely predicts the chronic change induced by fibers. As TNF-a plays a key role in lung remodeling (27), further investigations on the correlation between proinflammatory cytokines and the order of the toxicity of fibrous materials are also required. Along this line, correlations between gene expression and pathologic changes induced by fibrous materials may prove to be a powerful approach for assessing health risks due to fiber exposure.