Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Asbestos fibers are known to be toxic for mesothelial cells, the progenitor cell of the asbestos-induced tumor mesothelioma, although the mechanism of toxicity is not understood. Two major possible mechanisms have been considered: that fibers are harmful because of reactive oxygen species generated either by the fibers or by the ingesting cells, or that fibers are harmful because of mechanical damage to the cell (1). For either of these possible mechanisms of asbestos-induced injury, phagocytosis of the fibers may be a necessary first step. Phagocytosis may be necessary for the intracellular generation of reactive species either from the fibers themselves or in the process of phagocytosis. Phagocytosis may be necessary for fibers to interact with and damage chromosomes or the mitotic apparatus. On the other hand, phagocytosis may not be necessary if the fibers exert toxic alterations at the plasma membrane (2). The study of the role of phagocytosis has been quite limited by the difficulty of measuring uptake of fibers and by the inability to enhance phagocytosis of the fibers selectively without increasing the overall dose of fibers.

Phagocytosis by mesothelial cells is particularly difficult to quantify because, in the extremely thin cell (< 1 µm), it is difficult to identify intracellular fibers and distinguish them from extracellular fibers on top of or underneath the cell. In an earlier study, we found that even confocal microscopy was not able to identify fibers as intracellular; we required a lipid-soluble fluoroprobe to visualize the plasma membrane so that membrane-surrounded fibers could be identified as internal (3). Additional techniques that have been applied to this question include scanning electron microscopy combined with electron backscatter and transmission electron microscopy (4, 5), or transmission electron microscopy alone (6). We have previously developed a relatively simple and rapid assay for phagocytosis, by counting intracellular fibers in cells first exposed to trypsin to remove adherent fibers (3). We confirmed that the trypsin did remove all adherent fibers and validated this assay by comparing it with the fluorescence confocal microscopic assay. Here we have used it to address the role of phagocytosis in asbestos-induced toxicity.

We have previously shown that a selective increase in phagocytosis of asbestos fibers can be achieved by adsorption of vitronectin (VN) onto the fibers (3). When coated with VN, the major adhesive protein of serum, asbestos fibers are recognized and internalized via mesothelial cell integrins that act as VN receptors. The integrin-mediated phagocytosis leads to a selective increase in internalization of VN-coated fibers without an increase in adherence of fibers, similar to what has been shown for internalization of another integrin ligand, adenovirus (7). By coating fibers with VN, we realized we could introduce more fibers into the cell without increasing the overall dose of asbestos.

Previous studies have yielded conflicting data on the cytotoxic effects of asbestos on mesothelial cells. For example, increases in intracellular oxidation have not been found in studies of asbestos-exposed mesothelial cells (8, 9), although upregulation of antioxidant enzymes and oxidative changes in DNA suggest a role for oxidative damage to the cell and to cellular DNA (10, 11). In addition, DNA strand breakage, an expected toxic effect of oxidative damage to DNA, has not been found in previous studies of asbestos-exposed mesothelial cells (8, 12), although other DNA lesions and indirect measures of DNA damage have been reported (13).

Thus, we chose to expose mesothelial cells to asbestos to learn whether we could measure increases in intracellular oxidation and DNA strand breakage as well as increases in cytotoxicity such as apoptosis and cell-cycle arrest. We then used VN-coated asbestos fibers to learn whether a selective increase in fiber phagocytosis would increase these measures of asbestos toxicity. We then addressed whether these effects were due to the increased fiber uptake by either using cytochalasin B to inhibit phagocytosis of uncoated as well as VN-coated fibers or using RGD-containing peptides to inhibit the integrin-mediated phagocytosis of only the VN-coated fibers.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Fibers and Reagents

Crocidolite asbestos was obtained from the National Institute of Environmental Health and Safety (Research Triangle Park, NC). The fibers had been previously characterized at a mean length of 19 µm, range of 1 to 560 µm, and mean width of 0.25 µm, and used in our previous studies (3, 16). Control particles included: wollastonite, a relatively nonpathogenic calcium silicate fiber (mean length 12 µm, range 2 to 80 µm, mean width 2 µm) (Nyglos I; NYCO Minerals, Willsboro, NY), and glass beads (mean diameter 1.6 ± 0.3 µm; Duke Scientific Corp., Palo Alto, CA).

Cytochalasin B, camptothecin, acridine orange, propidium iodide (PI), and dimethyl sulfoxide (DMSO) were obtained from Sigma Chemical Company (St. Louis, MO). Hydrogen peroxide (H₂O₂) was obtained from Fisher Scientific (Pittsburgh, PA). Actinomycin D was obtained from Merck and Co. (West Point, PA). GRGDSP peptides, the control GRGESP peptides, and trypsin- ethylenediaminetetraacetic acid (EDTA) were obtained from GIBCO BRL (Gaithersburg, MD).

Proteins

VN and fibronectin (FN) were purified from outdated human plasma by heparin affinity chromatography (17) and used at a concentration found to support maximal mesothelial cell adhesion to plastic dishes. Mouse laminin was purchased from GIBCO BRL. Bovine serum albumin (BSA) (Fraction V; Sigma) was confirmed by immunoblot to have no detectable VN (3).

Asbestos Coating

Crocidolite fibers, wollastonite, or glass beads were incubated with proteins before addition to cells, as we have described (3). Spun in a high-speed bench-top centrifuge and resuspended in 200 µl of phosphate-buffered saline (PBS), fibers or beads (usually 100 µg) were incubated with either VN (30 µg/ml), BSA (30 µg/ml), or FN (90 µg/ml) for 1 h during vortexing at room temperature (RT), washed and spun, blocked with BSA (10 mg/ml, fraction V; Sigma) for 10 min at RT, washed and spun two times, sonicated for 5 s (power 60 W; Branson 450 sonifier; Branson, Danbury, CT), and added to mesothelial cells to provide the desired concentration of fibers on the cell surface (5 to 10 µg/cm²). Wollastonite was used at twice the concentration of asbestos to obtain similar fiber counts; glass beads were used at the same concentration as asbestos.

Cell Culture

Rabbit pleural mesothelial cells were harvested as previously described (18). Mesothelial cells were grown in standard media: RPMI 1640/Dulbecco's modified Eagle's medium, 10 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid (Hepes), 10% heat-inactivated fetal calf serum (FCS) (Hyclone Laboratories, Logan, UT), 2 mM L-glutamine, 100 U/ml penicillin, and 100 µg/ml streptomycin, and studied in experimental media: standard media without FCS. Cells were confirmed to be mesothelial by positive staining for cytokeratin and Wilms' tumor antigen (anti-WT ascites HC17 provided by Dr. Steven Albelda, and Dr. Ulrich Rodeck, The Wistar Institute, Philadelphia, PA) and negative staining for factor VIII. For all experiments, cells were used between passages 3 and 7.

Fiber Counting

An assay of fiber uptake has been described by us previously in which incubation of mesothelial cells with trypsin/EDTA releases adherent fibers and gives internalized fiber counts matching those by confocal microscopy (3). The percentage of cells with four or more fibers per cell seen on darkfield were counted as a measure of fiber uptake. Free-floating and adherent mesothelial cells exposed to asbestos fibers for 4 h were harvested and exposed to trypsin (0.25%, wt/vol) and EDTA (0.5 mM) for 10 min with gentle rotation on a shake table at RT to remove the adherent fibers. Once washed with PBS, the cells were examined for internalized fibers in a blinded fashion under darkfield microscopy.

General Experimental Procedure

Pleural mesothelial cells were plated at 40,000 cells per well onto laminin-coated 24-well plates (10 µg/ml laminin for 2 h) (Falcon; Becton Dickinson, Franklin Lakes, NJ) and allowed to adhere overnight in standard media. The next day, at approximately 60 to 80% confluence, the cells were washed to remove serum proteins before fibers were added to the desired concentration (usually 5 µg/cm²) in experimental, serum-free media. Potential blockers such as cytochalasin (5 µg/ml) or RGD-peptides (0.2 mg/ml) were added 1 h before asbestos. After 4 to 18 h, as stated, floating cells were collected and combined with adherent cells detached with trypsin (0.25%) and EDTA (0.5 mM) before processing.

Dichlorofluorescein Assay of Intracellular Oxidation

For measurement of intracellular oxidation induced by fibers, cells were incubated with an oxidation-sensitive fluorescent probe, 2',7',-dichlorofluorescin (DCFH)-diacetate (DA) (Molecular Probes, Inc., Eugene, OR). The probe diffuses freely into cells where it becomes trapped after deacylation as the nonfluorescent DCFH. After oxidative stress, the DCFH is oxidized to the fluorescent dichlorofluorescein (DCF) by a variety of reactive oxygen species and is sensitive to a general level of oxidative stress (19). DCFH-DA, aliquotted in DMSO at a stock solution of 5 M and stored in a dessicator in the dark at 20°C, was diluted in PBS immediately before the experiment. Adherent mesothelial cells were exposed to asbestos for 4 h, detached with trypsin/EDTA, combined with the floating cells, and incubated with DCFH-DA (5 µM) at 37°C for 1 h before and continuously during flow cytometric analysis. PI was added before flow collection to allow exclusion of cells that were permeable and thus would not retain the DCF probe. Data are reported as the percentage of PI-negative cells with a shift of fluorescence to the right of 95% of the control cells. After measurements were obtained, an excess of H₂O₂ (500 µM) was added to each tube to determine maximum fluorescence and thereby assure equivalent loading of DCF.

DNA Strand Breakage as Measured by Fluorescent Assay of DNA Unwinding

DNA strand breaks increase the speed of unwinding of DNA in alkaline solution. Because ethidium bromide binds preferentially to double-stranded DNA, its binding in alkaline conditions can define the amount of unwinding and thus the degree of DNA strand breakage (single- or double-stranded). This strategy has been used to develop a highly sensitive assay for DNA strand breaks, the fluorescence assay of DNA unwinding (20, 21). For this assay, pleural mesothelial cells (2.0 × 10⁶) were plated onto 15-cm dishes (Corning, Corning, NY) to adhere overnight to 60 to 80% confluence and exposed to protein-coated or uncoated asbestos (10 µg/cm²) or wollastonite (20 µg/cm²) or to H₂O₂ (100 µM) for 4 h. In some experiments, cytochalasin (5 µg/ml), RGD- or RGE-containing peptides (0.2 mg/ml), or DMSO (0.5%) was added 1 h before the fibers or H₂O₂. Cells were detached by trypsin-EDTA and added to the floating cells, and the suspension was divided equally among three sets of tubes labeled T, P, and B. T represents total double-stranded DNA not unwound; P, the DNA exposed to unwinding conditions to determine rate of unwinding; and B, fully unwound DNA in single-stranded form. For cell lysis and chromatin disruption, all tubes received 200 µl of solution B (0.25 M mesoinositol, 10 mM sodium phosphate, and 1 mM MgCl₂; pH 7.2) followed without mixing by 200 µl of solution C (9 M urea, 10 mM NH₄OH, 2.5 mM trans-1,2-diaminocyclohexane-N,N,N',N'-tetraacetic acid, and 0.1% sodium dodecyl sulfate) with incubation for 10 min on ice. For initiation of alkaline conditions (pH  12.8) to promote unwinding, tubes P and B received 100 µl of solution D (0.55 vol solution C in 0.2 N NaOH) and 100 µl of solution E (0.45 vol solution C in 0.2 N NaOH) and were maintained on ice for 30 min (20). To completely denature all DNA to single-stranded DNA, tube B was also sonicated for 3 s in the alkaline solution. To neutralize the pH ( 11.0) and stop unwinding after the 30-min incubation, tubes P and B received 400 µl of solution F (1 M glucose and 14 mM -mercaptoethanol). During this 30-min incubation, tube T was exposed to the combined solutions D, E, and F (600 µl) so that there was no period of alkaline unwinding. Finally, each tube was sonicated for 2 s to homogenize the samples and each tube received 1.5 ml solution G (6.7 µg/ml ethidium bromide in 13.3 mM NaOH). Tubes were maintained on ice until fluorescence was determined using a spectrofluorometer (excitation, 520 nm; emission, 590 to 600 nm). The percent of double-stranded DNA remaining in the samples after unwinding was given by %D = [(P  B)*100]/(T  B).

Green Fluorescent Protein Annexin V Assay for Apoptosis

Entry into apoptosis is associated with exposure of phosphatidylserine on the outer leaflet of the plasma membrane, a process that can be detected by the binding of annexin V, a member of a family of proteins that bind to acidic phospholipids. For detection of exposed phosphatidylserine in apoptotic cells, we constructed, purified, and characterized a green fluorescent protein (GFP)- annexin V fusion protein as described (22).

After exposure to fibers or other conditions, mesothelial cells were collected and centrifuged (1,000 rpm, 10 min). The cell pellet was resuspended in Hepes buffer (Hanks', 15 mM Hepes, and 2 mM CaCl₂), and stained with GFP-annexin V fusion protein (3 µg/ml in Hepes buffer) for 10 min on ice. PI (15 µg/ml, Sigma) was added just before analysis using a FACSort flow cytometer (Becton Dickinson, San Jose, CA), with acquisition and data analysis performed using CELLQuest software (Becton Dickinson) as described (16). A total of 10,000 events per sample was acquired to ensure adequate mean data. Early apoptotic cells are not permeable (GFP-annexin V +, PI ) whereas late apoptotic and necrotic cells are both permeable and are thus indistinguishable (GFP-annexin V +, PI +).

Morphologic Analysis of Apoptosis

For quantification of apoptosis by morphologic criteria, cells were stained with acridine to identify the condensed nuclei of apoptotic cells, and with PI to distinguish early from late apoptosis, as described (16).

Cell-Cycle Analysis

Cell-cycle progression was analyzed by flow cytometry for both DNA synthesis by bromodeoxyuridine (BrdU) incorporation and DNA content by PI fluorescence (23, 24). Mesothelial cells were plated at 750,000 cells per 10-cm plate (Falcon) and allowed to adhere overnight in standard media. In the morning, at 60% confluence, cells were washed and the media changed to 1% serum, which we found could support cell growth while minimizing adsorption of the serum proteins on asbestos. Cells were then exposed to asbestos fibers at 5 µg/cm² for different times (6, 12, or 24 h) and to BrdU (10 µM; Sigma) for the final 2 h. In some experiments, to learn whether the accumulation of cells in a phase represented a complete arrest or a delay, cells were exposed to fibers for 24 h and BrdU for the final 12 h. After incubation, cells were detached, spun at 2,000 rpm for 5 min, washed with PBS, and, to minimize clumping during fixation, were exposed dropwise to ice-cold 70% ethanol during vortexing before incubating in the ethanol overnight. After a wash in PBS and centrifugation, cells were resuspended in 1 ml of 0.1 M HCl containing 0.25% Triton X-100, incubated on ice for 10 min, and washed with distilled water. Cells were spun, resuspended in 1 ml of water, boiled for 10 min, cooled on ice for 10 min, spun, washed in PBS with Triton 0.25% (vol/vol; PBS/Triton), and spun again. Cells were then incubated with anti-BrdU (1:50 in PBS/Triton; Dako Corporation, Carpinteria, CA) for 30 min on ice. Cells were then washed and incubated with fluorescein-labeled sheep antimouse secondary antibody (1:100 in PBS/ Triton; Sigma) for 30 min on ice. Finally, cells were stained with PI (25 µg/ml) prepared in PBS containing RNase A (100 µg/ml; Sigma) for 30 min at 37°C. Cells were analyzed for BrdU and PI fluorescence using FACSort flow cytometry (Becton Dickinson) with linear fluorescence for PI (DNA content) and log fluorescence for BrdU (BrdU incorporation). At least 15,000 cells were analyzed for each sample by CellQuest software. The quadrants for cell cycle phases, G1/S, S, and G2/M, were selected by analysis of control cells for each experiment, and the percentage of cells in each phase was quantified.

Statistics

Differences were determined among data by analysis of variance with Tukey's test for post hoc analysis (16). Significant difference was considered at P < 0.05. Data are averages of mean data and are expressed as mean ± one standard error of the mean (SEM), unless otherwise indicated.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

VN-coating increased phagocytosis of asbestos fibers compared with no coating (Table 1) or BSA- and FN-coating (data not shown). Blockade of actin polymerization with cytochalasin (5 µg/ml) decreased uptake of all fibers. Integrin blockade with RGD-containing peptides (0.2 mg/ml) decreased uptake of VN-coated fibers but not of uncoated fibers, and thereby was the concentration selected for these experiments. At higher concentrations (> 0.5 mg/ml), RGD-containing peptides induced cell rounding, leading to a decrease in uptake of all fibers (data not shown). DMSO, up to a concentration of 1% (data not shown), had no effect on fiber uptake.

Intracellular oxidation of mesothelial cells was induced after incubation for 4 h with BSA-coated asbestos (Figure 1) and uncoated asbestos (data not shown), as measured by a shift of fluorescence of the oxidative probe. VN-coated asbestos fibers induced a greater oxidative shift than did BSA-coated fibers (Figure 1). Glass beads did not induce any fluorescent shift, despite the fact that they were effectively phagocytosed (VN-coated, 83 ± 8% of cells with more than four intracellular beads; BSA-coated, 73 ± 13% cells; P > 0.05). Cytochalasin inhibited intracellular oxidation for both VN-coated and uncoated fibers (Figure 2) in a manner that paralleled the inhibition of fiber uptake (Table 1). After each study, maximal oxidative shift was determined in the mesothelial cells by exposing them to excess H₂O₂ (500 µM); the maximal fluorescence shift ranged between 75 and 95% of cells, showing similar loading of the oxidative probe.

View larger version (26K): [in this window] [in a new window]   Figure 1.   Intracellular oxidation of mesothelial cells is increased by VN-coated asbestos more than by BSA-coated asbestos. Adherent cells were exposed to no fiber or to BSA- or VN-coated crocidolite or glass beads (5 µg/cm2) for 4 h before harvesting and incubating for 1 h with DCFH-DA (5 µM) before and during analysis by flow cytometry. PI-positive cells were excluded to avoid including cells with leakage of the fluoroprobe. The percentage of PI-negative cells with fluorescence greater than 95% of the control cells is shown. *Greater than no fiber or glass beads; **greater than shift due to BSA-coated crocidolite asbestos; P < 0.05, n = 4 experiments.

View larger version (24K): [in this window] [in a new window]   Figure 2.   Cytochalasin inhibits the intracellular oxidation of mesothelial cells exposed either to uncoated or VN-coated asbestos. Cells were exposed to uncoated or VN-coated crocidolite (5 µg/ cm2) for 4 h with or without cytochalasin B (5 µg/ml). Cells were harvested, incubated with DCFH-DA (5 µM) for 1 h before analysis by flow cytometry in the presence of DCFH-DA and PI (15 µg/ml). PI-positive cells were excluded due to leakage of the fluoroprobe. The percentage of PI-negative cells with fluorescence greater than 95% of the control cells is shown. *Different from cells exposed to uncoated asbestos; **different from cells exposed to asbestos without cytochalasin; P < 0.05, mean ± SD, n = 4 experiments.

DNA strand breaks were detected in mesothelial cells incubated with asbestos fibers for 4 h, as measured by the ethidium bromide fluorescence, alkaline unwinding assay (Figure 3A). VN-coated asbestos induced more DNA strand breakage than did uncoated or BSA-coated asbestos. Wollastonite, the control fiber, failed to induce significant DNA strand breakage, whether VN-coated or uncoated. The role of intracellular oxidation in asbestos-induced DNA damage was confirmed by the ability of DMSO, a hydroxyl radical scavenger, to block the DNA damage due to both crocidolite and VN-coated crocidolite (Figure 3B). The role of phagocytosis was investigated using cytochalasin and RGD peptides. Cytochalasin completely blocked DNA damage induced by uncoated or BSA- or VN-coated asbestos (Figure 4). RGD peptides blocked DNA damage due to VN-coated fibers only (Figure 5). Neither cytochalasin nor RGD peptides affected the DNA damage due to H₂O₂ (100 µM for 4 h) (Figure 5).

View larger version (40K): [in this window] [in a new window]   Figure 3.   DNA strand breaks are increased by VN-coated asbestos more than by BSA-coated asbestos and are due to reactive oxygen species. (A) Cells were exposed to crocidolite (10 µg/cm2) or wollastonite (20 µg/cm2) coated with different proteins for 4 h, harvested, and exposed to alkaline conditions to promote unwinding of DNA (see MATERIALS AND METHODS). The amount of double-stranded DNA remaining was measured by ethidium bromide fluorescence and is expressed as percent of total double-stranded DNA. (B) In additional experiments, DMSO (0.5%) was added 1 h before the fibers. *Different from cells exposed to no crocidolite; **different from cells exposed to uncoated crocidolite; means ± SEM, P < 0.05, n = 3 experiments, done in triplicate.

View larger version (56K): [in this window] [in a new window]   Figure 4.   Cytochalasin blocks DNA strand breaks induced by asbestos. Mesothelial cells were exposed to cytochalasin (5 µg/ml) for 1 h before the addition of uncoated or BSA- or VN-coated crocidolite asbestos (10 µg/cm2) for an additional 4 h, then harvested and exposed to alkaline conditions to promote unwinding. The amount of double-stranded DNA remaining was measured by ethidium bromide fluorescence. *Different from cells exposed to no crocidolite; **different from cells exposed to uncoated crocidolite; means ± SEM, P < 0.05, n = 3 experiments, performed in triplicate.

View larger version (45K): [in this window] [in a new window]   Figure 5.   RGD peptides block DNA strand breaks due to VN-coated asbestos. Mesothelial cells were exposed to cytochalasin (5 µg/ml) or RGD- or RGE-containing peptides (0.2 mg/ml) for 1 h before the addition of no asbestos or BSA- or VN-coated crocidolite asbestos (10 µg/cm2) for an additional 4 h. The double-stranded DNA remaining after alkaline unwinding is shown. *Different from the cells exposed to the same protein-coated asbestos without the blocker; means ± SEM, P < 0.05, n = 6 experiments.

Early apoptosis of mesothelial cells was greater after exposure to VN-coated crocidolite asbestos than after uncoated or BSA- or FN-coated asbestos (Figure 6). Late apoptosis or necrosis, as defined by cells that stained both with GFP-annexin V and with PI, was also greater for cells exposed to VN-coated crocidolite asbestos (9.4 ± 0.6% SEM) compared with either uncoated or other protein- coated asbestos (5.4 ± 0.6%) or to no fiber (2.9 ± 0.3%) (P < 0.05; n = 10 experiments). Apoptosis determined by morphologic assessment of nuclear condensation was also higher for cells exposed to VN-coated asbestos (14.0 ± 1.0% SEM) than for those exposed to uncoated asbestos (11.0 ± 2.0%) or to no fiber (4.0 ± 1.0%) (P < 0.05; n = 3 experiments).

View larger version (34K): [in this window] [in a new window]   Figure 6.   Apoptosis of mesothelial cells is increased by VN-coated asbestos more than by other asbestos. Cells were exposed to no fibers or to crocidolite asbestos (5 µg/cm2) or wollastonite (10 µg/cm2) coated with different proteins for 18 h and studied by annexin V staining for early apoptosis (annexin V +, PI ). *Different from no asbestos; **different from uncoated asbestos; P  0.05, n = 10 experiments.

RGD-containing peptides, but not the control RGE peptides, inhibited the apoptosis induced by VN-coated crocidolite (Figure 7). RGD peptides had no effect on apoptosis induced by uncoated or BSA-coated fibers (Figure 7) or by nonfiber stimuli such as H₂O₂, actinomycin D (100 nM), or camptothecin (10 µg/ml) (data not shown). Cytochalasin could not be used because it itself induced apoptosis in these 18-h experiments.

View larger version (28K): [in this window] [in a new window]   Figure 7.   RGD peptides block apoptosis of cells exposed to VN-coated asbestos. Cells plated on laminin-coated wells were exposed to no peptide or to RGD or control RGE peptides (0.2 mg/ ml) for 1 h before the addition of asbestos (5 µg/cm2) for an additional 18 h. Floating and adherent cells were harvested, stained with GFP-annexin V and PI, and analyzed by flow cytometry. *Different from apoptosis due to VN-coated asbestos with control peptide or no peptide; P < 0.01, n = 6 experiments.

Cell-cycle studies showed that at 12 h, cells exposed to VN-coated asbestos fibers had a greater percentage of cells in G2/M than did cells exposed to the same concentration of uncoated or BSA- or FN-coated fibers (Figure 8). There was no difference in the percentage of cells in S phase (control, 18.2 ± 14.1%; BSA-and FN-coated asbestos, 17.5 ± 4.5%; VN-coated asbestos, 16.2 ± 8.4%; mean ± standard deviation [SD]) but there was a significant difference in G1 (control, 52.6 ± 15.5%, BSA- and FN-coated asbestos, 53.0 ± 4.5%; VN-coated asbestos, 44.2 ± 7.3%; mean ± SD). By 24 h, the accumulation in G2/M was no longer significantly different for cells exposed to VN-coated asbestos. Cells exposed for 24 h to VN-coated asbestos using a longer BrdU incubation showed the reappearance of a population of cells in G1 (a second G1) confirming that the accumulation in G2/M did not represent a permanent arrest (data not shown).

View larger version (35K): [in this window] [in a new window]   Figure 8.   Accumulation of cells in the G2/M phase of the cell cycle for mesothelial cells exposed to asbestos. Cells were exposed for 6, 12, or 24 h to uncoated or BSA-, FN-, or VN-coated crocidolite asbestos (5 µg/cm2), with addition of BrdU (10 µM) for the final 2 h. Cells were then harvested, stained with PI for DNA content and stained for BrdU incorporation as described in MATERIALS AND METHODS, and analyzed by flow cytometry. *Different from cells exposed to uncoated asbestos; P < 0.05, n = 4 experiments, means ± SD.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We have shown a crucial role for phagocytosis of asbestos by mesothelial cells for a wide range of cytotoxic effects, including intracellular oxidation, DNA strand breakage, apoptosis, and cell-cycle delay in G2/M. Increasing fiber phagocytosis by VN-coating led to increases in asbestos-induced toxicity. Decreasing fiber phagocytosis was often able to decrease or prevent the asbestos-induced toxicity. The importance of phagocytosis was best established for DNA strand breakage because the toxicity could be blocked both by cytochalasin for all fibers and by integrin blockade with RGD peptides for the VN-coated fibers. This study is unique in its correlation of a wide variety of cytotoxic effects with a selective increase in fiber phagocytosis.

In previous studies, it has not been clear whether phagocytosis of asbestos fibers, especially by "nonprofessional" phagocytic cells such as mesothelial cells, is necessary for cytotoxicity. Early studies first demonstrated that mesothelial cells were capable of phagocytosis, both in vitro (25) and in vivo (26). However, the study of the functional importance of phagocytosis has been limited by the difficulty of quantifying internal fibers and by the inability to increase fiber uptake selectively. In studies using electron microscopy to identify intracellular fibers in Syrian hamster embryo cells or in a tracheal epithelial cell line, Hesterberg and colleagues correlated the observed increases in phagocytosis (of long glass fibers versus short fibers [4], and of chrysotile versus crocidolite [5]) with a greater toxicity. In a recent study using flow cytometry to sort cells with more associated fibers from those with fewer, Takeuchi and associates correlated more fiber uptake with a greater degree of aneuploidy (39). In our study, we have explored the role of phagocytosis by manipulating phagocytosis independently. We have both increased and decreased the intracellular dose of one asbestos fiber type in primary mesothelial cells without altering the total dose and have shown the effect of phagocytosis on multiple measures of cellular toxicity.

For "professional" phagocytic cells exposed to asbestos, reactive oxygen species have been identified as a major product, presumably via the respiratory burst that accompanies phagocytosis (27, 28). For "nonprofessional" phagocytic cells such as mesothelial cells, however, the respiratory burst is not a feature of phagocytosis (28) and intracellular oxidation has been more difficult to demonstrate. Indeed, several studies have failed to show generation of reactive oxygen species by asbestos-exposed mesothelial cells, by using either electron paramagnetic resonance (8), chemiluminescence (29), or a DCF microtiter plate assay (9) although, more recently, generation of reactive nitrogen species has been identified (30). We were able to show that, by 4 h of exposure to asbestos, mesothelial cells showed a definite fluorescent shift of the DCF, a shift that may represent formation of either reactive oxygen or nitrogen species because DCF is known to respond to both (19, 31). The fluorescent shift was not due to the process of phagocytosis itself, because glass beads, which were avidly phagocytosed, did not induce a shift. The fluorescent shift was not due to integrin-VN interactions because the VN-coated glass beads also had no effect on fluorescence. Instead, the fluorescent shift appeared to require asbestos fiber internalization because it could be inhibited by cytochalasin (Figure 2) to the same extent that fiber uptake was inhibited (see Table 1). Our findings of the importance of fiber uptake in the DCF fluorescence shift support the observations of Choe and colleagues that cytochalasin B could inhibit asbestos-induced production of the nitric oxide radical by primary rat mesothelial cells (30). The possible contribution of adherent fibers to a fluorescent shift was not possible to assess in our study because adherent cells were detached before incubation with the DCF fluoroprobe, thus removing adherent fibers. Nonetheless, in the assays for DNA damage and apoptosis, the adherent fibers were present during the entire assay and yet both DNA damage and apoptosis could be blocked by blocking fiber phagocytosis alone. It is likely from these experiments, then, that internalized fibers, not external fibers and not any oxidative burst of mesothelial cells themselves, generate the reactive species detected by this fluoroprobe in mesothelial cells.

DNA strand breaks in mesothelial cells due to asbestos were detected for the first time using an alkaline unwinding assay, a direct assay of DNA damage both for single- and double-strand breaks. Earlier studies attempting to measure DNA damage in asbestos-exposed mesothelial cells by similar assays, such as alkaline elution and another alkaline unwinding assay, failed to show strand breakage (8, 12). Previous evidence for DNA breakage in asbestos-exposed mesothelial cells has relied on more indirect measures, such as the demonstration of repair by unscheduled DNA synthesis (13), of oxidized bases (11), and of induction of the PARP enzyme (14). More recently, an assay of DNA breakage, the comet gel electrophoresis assay, has shown asbestos-induced DNA breaks in Met5A cells, a simian virus 40-transformed mesothelial line (9), and has been confirmed by us in normal mesothelial cells (32). In this study, we applied a sensitive alkaline unwinding assay previously shown to detect asbestos-induced damage in epithelial cells (21). Perhaps, compared with the earlier alkaline elutriation and unwinding assays, this ethidium bromide fluorescence assay is more sensitive; it is reported to be able to detect DNA damage comparable to only one strand break per chromosome (20). Using this assay, we were able to confirm that asbestos induces DNA strand breaks in mesothelial cells via oxygen radicals. We then used this assay to show that this damage required asbestos fiber phagocytosis. Only phagocytosed fibers were capable of injuring the DNA because adherent fibers were present for the entire 4-h assay but when phagocytosis was inhibited, DNA damage was almost completely blocked. When phagocytosis was increased by VN-coating the crocidolite fibers, DNA damage was increased by an amount that could be blocked by RGD peptides, which specifically block uptake of VN-coated fibers (Table 1) (3). Finally, the effect of the blockers, cytochalasin and RGD peptides, was specific for asbestos-induced DNA damage because neither blocker affected the DNA strand breaks induced by H₂O₂. We conclude that asbestos fibers induce DNA strand breaks and that the damage required phagocytosis of the fibers.

We and others have previously shown that asbestos fibers induce apoptosis of mesothelial cells (33, 16), presumably due to oxidant injury and to DNA damage (16). In this study, we show that increased uptake of fibers via integrins leads to an increase in cell toxicity as measured by apoptosis. We could not use cytochalasin in this assay to show the effect of fiber uptake because in this study, in contrast to our previous study, cytochalasin itself induced apoptosis. RGD peptides, however, were able to block the effect of the VN-induced increase in apoptosis without affecting the apoptosis due to uncoated or BSA-coated asbestos.

The final cytotoxic effect we examined was that of alteration of the cell cycle, a characteristic response to DNA damage. Both a G1/S delay and G2/M delay have been described after exposure of mesothelial cells to crocidolite asbestos (34). We did not identify a G1/S delay, although we did identify a G2/M delay, at least for the VN-coated asbestos fibers. The inability to show a G2/M delay for uncoated asbestos as well was likely due to the low serum conditions, which we used to avoid VN-coating of all the fibers. An arrest at G2/M has been most closely associated with the types of damage that would interfere most with mitosis, particularly double-stranded DNA breaks or abnormalities in the spindle apparatus (35). We find that increased asbestos phagocytosis led to an accumulation of cells in G2/M; this cell-cycle disturbance at G2/M may result from the greater degree of DNA strand breakage we showed or from an interaction of fibers with the mitotic apparatus. Asbestos fibers, unlike other DNA damaging agents such as irradiation, are a constant injurious presence to the cell that ingests them. In certain cells, DNA damage may be completely or sufficiently repaired to allow progress through mitosis; alternatively, DNA damage may be persistent and induce cells to enter an apoptotic route (36).

The role of VN in altering fiber uptake and thus its toxicity may be of importance in vivo. In our previous study we showed that the VN in a variety of biologic fluids, such as serum, bronchoalveolar lavage fluid, and pleural liquid, coats crocidolite asbestos fibers and induces increased fiber phagocytosis by mesothelial cells (3). In this study, instead of biologic fluids we used purified VN to increase fiber uptake while avoiding the confounding effect of the myriad proteins in biologic fluids. In the in vivo setting, however, the exposure of fibers to biologic fluids may lead to increased fiber uptake and cytotoxicity. If so, the consequences of that increased toxicity would depend on the cells' ability to survive, either by repair or adaptation to damage, or by responding to survival signals from the environment. If cells with damaged DNA manage to survive and reenter the cell cycle, they may begin to accumulate genetic damage and thereby initiate a multistep progression toward malignancy (37).

One consequence of this study was to emphasize the need to consider the role of phagocytosis in interpreting experimental studies of asbestos toxicity. The effect of interventions in experiments on fiber toxicity may be confounded by effects on intracellular delivery of the fibers. For example, various aspects of fiber shape and composition that are thought to be important for fiber toxicity may exert their effects in part by altering phagocytosis. As mentioned, Hesterberg and colleagues found that, in their experiments, long glass fibers were ingested more readily than short fibers, and chrysotile fibers were ingested more than crocidolite fibers; in both cases, the difference in phagocytosis may have explained the observed differences in toxicity (4, 5). Various blockers used to identify pathways of fiber-induced toxicity may, in fact, operate by blocking phagocytosis. We, for example, found that catalase at high doses interfered with crocidolite asbestos phagocytosis (3). Phagocytosis involves complex interactions with the cell-surface and signaling pathways, as well as the cytoskeleton and phagosomes (38). Therefore, interventions used to investigate fiber toxicity may unexpectedly alter phagocytosis and thereby alter intracellular fiber delivery. By taking account of phagocytosis, we can better understand the intrinsic characteristics of fibers that determine toxicity separate from those that alter fiber uptake.

In summary, we have investigated asbestos fiber phagocytosis as an independent variable for toxicity to normal mesothelial cells, using techniques to increase phagocytosis selectively (VN-coating) and to decrease phagocytosis (actin or integrin blockade). We conclude that, at least over the duration of these experiments, phagocytosis of asbestos fibers is necessary for asbestos-induced cytotoxicity to mesothelial cells.