TNF-α inhibits asbestos-induced cytotoxicity via a NF-κB-dependent pathway, a possible mechanism for asbestos-induced oncogenesis

Asbestos Is Cytotoxic to Human Mesothelial Cells in Culture.

Light microscopy observation of HM in tissue culture, suggested that asbestos was cytotoxic (data not shown). This observation was verified by using both the lactate dehydrogenase (LDH) and the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assays. These assays indicated that crocidolite was cytotoxic to HM in a dose-dependent manner (Fig. 1 A and B). This interpretation was further verified by flow cytometry (Fig. 1C). For example, ≈66% and 40% of the HM were viable after exposure to crocidolite at 0.5 or 5.0 μg/cm², respectively, for 24 h, and the percentage of viable HM was ≈40–50% and 0–1%, respectively, 1 week after treatment (data not shown). This finding raised the question of how asbestos could cause malignant transformation of HM when, in tissue culture, asbestos killed HM in a dose-dependent manner.

Asbestos Induces Expression of TNF-α and of the TNF-α Receptor TNF-R1 in HM.

In experiments performed during RNA microarray analyses to determine gene expression profiles in MM, the most notable effect was that HM exposed to crocidolite (5 μg/cm² for 2 h) induced TNF-α and TNF-α receptor mRNA (4.07- and 3.5-fold induction, respectively, over untreated controls; data not shown). We verified the induction of TNF-α mRNA in HM exposed to crocidolite by using quantitative real-time PCR (Fig. 2A). Moreover, we found that HM exposed to crocidolite released TNF-α in the tissue culture medium (Fig. 2B). Immunohistochemistry showed that, after crocidolite exposure, most HM in each of three separate primary cultures expressed TNF-R1 (data not shown). Western blot experiments confirmed that TNF-R1 was induced in HM after asbestos exposure in a dose-dependent manner (Fig. 2C).

TNF-α Significantly Reduced Asbestos Cytotoxicity.

We tested the hypothesis that TNF-α modulated asbestos-induced cytotoxicity. We found that when HM were pretreated with TNF-α, asbestos cytotoxicity was significantly reduced (Fig. 3A). Addition of 0.5 μg/ml anti-human TNF-α antiserum to the tissue culture medium of HM 1 h before TNF-α treatment abolished the anticytotoxicity effect caused by TNF-α. A matched IgG control had no protective effects confirming specificity (Fig. 3A). These results were verified by flow cytometry assays, measuring HM viability by annexin V and propidium iodide staining. The percentage of double-negative-staining viable cells was significantly higher in HM exposed to both TNF-α and asbestos compared with HM exposed to asbestos (P < 0.05; data not shown). MTT assays confirmed that TNF-α protected HM from crocidolite cytotoxicity (Fig. 3B).

TNF-α Induces Activation of NF-κB in HM.

Although TNF-α has the potential of inducing programmed cell death in certain cell types (31), usually cells survive exposure to TNF-α due to potent activation of the NF-κB pathway. The protective effect observed in Fig. 3 suggested that TNF-α could have induced NF-κB in HM; therefore, we tested whether TNF-α induced the translocation of the NF-κB subunit p65 (RelA) from the cytoplasm into the nucleus, an effect linked to NF-κB activation and to cell survival (32). Western blotting revealed that the amount of NF-κB p65 was negligible in the nucleus of HM before TNF-α exposure (Fig. 4A, lane 1) and markedly increased after treatment with TNF-α (Fig. 4A, lanes 2–6). EMSA revealed that TNF-α induced DNA-binding activity of NF-κB (Fig. 4B), supporting the results shown in Fig. 4A.

TNF-α Inhibits Asbestos-Induced Cytotoxicity Through NF-κB Signaling.

We used the NF-κB inhibitor Bay11-7082 to test whether NF-κB was directly related to the anticytotoxicity effect of TNF-α. This molecule is an irreversible inhibitor of IκB phosphorylation, therefore increasing stabilization of IκB and thus inhibiting NF-κB nuclear translocation and activation (33–35). We pretreated HM with or without 5 μM Bay11-7082 for 1 h followed by 30 min of exposure to TNF-α (10 ng/ml). Western blot analysis revealed that Bay 11-7082 inhibited the nuclear translocation of the NF-κB p65 subunit induced by TNF-α, thus blocking the NF-κB pathway in HM (data not shown). LDH assays showed that inhibition of the NF-κB pathway by Bay11-7082 suppressed the anticytotoxicity effect induced by TNF-α in HM (Fig. 5). The results were similar to those obtained by using anti-TNF-α antiserum (Figs. 3A and 5). These findings suggested that TNF-α prevents asbestos-induced programmed cell death of HM through the activation of NF-κB signaling. Addition of Bay11-7082 and TNF-α did not induce cell death, confirming that this pathway is not induced by TNF-α in HM, even when the NF-κB pathway is inactivated.

To verify these results, we down-regulated NF-κB p65 by using small hairpin-mediated RNA interference. The micro RNA precursor targeting p65 was expressed by a lentiviral vector, which also expressed GFP (see Materials and Methods). We found that 48 h after infection GFP became detectable at high levels in >95% of cells (data not shown), indicating that we had obtained high levels of transduction. The amount of p65 (RelA) in both nuclear and cytoplasmic extracts decreased significantly in short hairpin RNA (shRNA)–RelA infected cells compared with cells infected with the control virus (Fig. 6A). These infected HM were treated with TNF-α and asbestos, and cytotoxicity was measured by using the LDH assay. In the shRNA–RelA infected group, the anticytotoxicity effect of TNF-α was abolished (Fig. 6B). These data confirmed that TNF-α inhibits asbestos-induced cytotoxicity via a NF-κB-dependent mechanism.

Cytogenetic Analysis.

The two primary HM cultures tested produced very similar results. Metaphases from untreated control HM did not show any structural abnormalities (data not shown). Rare metaphases were found in HM treated with crocidolite and nearly all exhibited very poor chromosome morphology and premature separation of chromatids, such that they were no longer connected at the centromere (Fig. 7A). In contrast, HM pretreated with TNF-α before asbestos exposure had numerous metaphases, and most of them were without mitotic irregularities and with good chromosome morphology (Fig. 7B). In summary, cytogenetics was consistent with the experiments reported in Figs. 1 and 3, indicating that asbestos was cytotoxic and that pretreatment of HM with TNF-α enhanced survival of HM exposed to asbestos.

Cell Cultures.

HM were from patients who accumulated pleural fluid because of nonmalignant diseases. HM were identified morphologically and characterized by immunostaining for calretinin and pancytokeratin as described (7). HM were cultured in DMEM with 20% FBS and were used at passages 4–6. HM from three different donors were used in all experiments, and each experiment was done in triplicate, except for cytogenetic analyses in which primary HM from two donors were used.

Materials.

Human recombinant TNF-α, anti-human TNF-α antiserum (IgG fraction of antiserum developed in rabbit), normal rabbit purified Ig control, Bay11-7082, and (E)-3-(4-methylphenylsulfonyl)-2-propenenenitrile were from Sigma.

Asbestos Preparation.

Crocidolite asbestos fibers were obtained from the Union Internationale Contre le Cancer (47) and processed as described in ref. 9 (average length, 3.2 ± 1.0 μm; average diameter, 0.22 ± 0.01 μm). Fibers were baked at 150°C for 18 h, suspended in Hanks’ balanced salt solution at 2 mg/ml, triturated 10 times through a 22-gauge needle, and autoclaved. Crocidolite fibers from Union Internationale Contre le Cancer also were characterized previously and shown to be carcinogenic, to cause the release of reactive oxygen species, and to cause DNA damage (9).

LDH Cytotoxicity Assays.

Cytotoxicity was assessed with an LDH-cytotoxicity detection kit (Roche Diagnostics), which measures LDH activity released from the cytosol of damaged cells. HM were seeded in a 96-well tissue culture plate at a density of 1 × 10⁴ cells per well and incubated in DMEM with 20% FBS. The next day, we changed the medium to 1% FBS and added 100 μl of crocidolite dilutions at different concentrations per well. After 24 h, 100 μl of supernatant per well was harvested and transferred into a new 96-well, flat-bottom plate. LDH substrate (100 μl) was added to each well and incubated for 30 min at room temperature (RT) protected from light. The absorbance of the samples was measure at 490 nm with an ELISA reader. Cytotoxicity was calculated with the formula: % cytotoxicity = (experimental value − low control) × 100/(high control − low control), where low control is assay medium plus cells and high control is assay medium (plus 2% Triton X-100) plus cells.

MTT Assays.

The cell proliferation kit I (MTT; Roche Diagnostics) was used to check viability. The assay is based on the cleavage of the yellow tetrazolium salt MTT to purple formazan crystals by metabolically active cells.

Flow Cytometry Assays.

After exposure to different amounts of crocidolite, HM were harvested, washed twice with cold PBS, and suspended in binding buffer (10 mM Hepes, pH 7.4/140 mM NaCl/2.5 mM CaCl₂) at a concentration of 1 × 10⁶ cells per ml. The binding buffer [100 μl (1 × 10⁵ cells)] was transferred to a new tube, and 5 μl of annexin V-Biotin (BD Biosciences) was added for 15 min at RT in the dark. HM were washed, suspended in 100 μl of binding buffer with 5 μg/ml streptavidin-FITC (BD Biosciences), and stained with propidium iodide (Sigma) at a final concentration of 5 μg/ml for 15 min at RT. Binding buffer (400 μl) was added to the HM, and the cells were analyzed by flow cytometry.

Quantitative Real-Time PCR.

HM were exposed to asbestos (5 μg/cm²) for 6, 12, and 24 h. Total cellular RNA of each sample was isolated with the RNeasy kit (Qiagen, Valencia, CA) and treated with RNase-free DNase. For each sample, 3 μg of total RNA was reverse transcribed by using the First-Strand cDNA synthesis kit (Fermentas, Burlington, ON, Canada). 18S rRNA was used as the internal control. PCRs for TNF-α and 18S rRNA were performed in triplicate in 25-μl total reaction volumes using the SYBR green master PCR mix (Applied Biosystems). The GeneAmp 5700 sequence detection system (PerkinElmer–Applied Biosystems) was used by following standard procedures. The following primers were used: TNF-α, 5′-GCTTGTTCCTCAGCCTCTTCTC-3′ and 5′-CTCAGCTTGAGGGTTTGCTACA-3′, and 18S, 5′-TGATTAAGTCCCTGCCCTTTGT-3′ and 5′-TCAAGTTCGACCGTCTTCTCAG-3′. No RT-PCR control was run alongside each sample. Relative TNF-α mRNA expression levels for each sample were normalized to the internal control 18S rRNA.

Nuclear Extract Preparation and EMSA.

Cell pellets were resuspended in buffer A (10 mM Hepes, pH 7.9/10 mM KCl/0.1 mM EDTA/0.1 mM EGTA/1 mM DTT/0.5 mM PMSF) and placed on ice for 15 min, and 25 μl of 10% Nonidet P-40 was added to each tube followed by vortexing for 10 s. Separation of nuclei was assessed by using a phase microscope. The samples were centrifuged for 5 min at 9,000 × g at 4°C. The nuclear pellet was lysed with 30 μl of lysis buffer (Sigma), and the protein concentration was determined with the Bradford assay. Five micrograms of the nuclear extracts was used for EMSA. The gel shift assay was as described in ref. 48. NF-κB consensus oligonucleotides were from Santa Cruz Biotechnology and labeled with [γ-³²P]ATP (Amersham Pharmacia) using T4 polynucleotide kinase. Unincorporated nucleotides were removed by Chroma Spin column (BD Biosciences Clontech). The ³²P-labeled NF-κB probe was incubated with 5 μg of nuclear extract in gel shift binding buffer (Promega). Unlabeled oligonucleotide was incubated with nuclear extract for 10 min before the addition of the probe as a specific competitor. The DNA–protein complex was analyzed in 5% nondenaturing polyacrylamide gel. Lanes with samples were devoid of loading dyes, which can interfere with the binding. Dried gels were exposed to x-ray film.

Western Blots.

Total cell protein extraction was prepared by using lysis buffer (Sigma). For each sample, 40 μg of protein was used. The proteins were separated on a 10% polyacrylamide gel and transferred to poly(vinylidene difluoride) membranes. The membranes were blocked in Tris-buffered saline containing 0.05% Tween 20 and 5% skim milk at 4°C overnight. The membranes were then probed with the first antibody at RT for 2 h. Anti-NF-κB p65, anti-TNF-α, anti-TNF-R1, and anti-Histone H1 antibodies were from Santa Cruz Biotechnology; anti-GAPDH was obtained from Chemicon International (Temecula, CA). The membrane was washed four times with Tris-buffered saline containing 0.05% Tween 20 and incubated with a horseradish peroxidase-conjugated secondary antibody (Pierce), and the signal was detected by chemiluminescence (Pierce). To check for the secretion of TNF-α by HM, the cell culture medium was concentrated with an Amicon centrifugal filter (Millipore). A rabbit anti-human TNF-α polyclonal IgG (Santa Cruz Biotechnology) was used as the first antibody, and biotinylated anti-rabbit IgG (Vector Laboratories) was used as the secondary antibody. The membrane was then incubated with the horseradish peroxidase-conjugated stretavidin. Bands were visualized by enhanced chemiluminescence (Pierce).

RNA Interference Assay.

The lentiviral vector, pLentiLox37 (pLL), expressing eGFP is described in ref. 49. pLL-shRelA and pLL-shNS, expressing RelA-specific and nonspecific shRNA oligonucleotides, were constructed by ligating to the HpaI and XhoI restriction sites of pLL the following DNA oligonucleotides (sense): shRelA, tGGATTGAGGAGAAACGTAAttcaagagaTTACGTTTCTCCTCAATCCtttttttc; shNS, tCAGTCGCGTTTGCGACTGGttcaagagaCCAGTCGCAAACGCGACTGtttttttc (the 19-nt target sequences are in uppercase and inserts were verified by sequencing). High-titer lentiviral preparations were produced by calcium phosphate transfection of pLL vectors (10 μg) and the pMDL (6.5 μg), pRev (2.5 μg), and pENV(VSVG) (3.6 μg) plasmids, encoding packaging proteins, into 293T cells (49). Supernatants were harvested at 36–60 h, virions were concentrated by ultracentrifugation at 23,000 rpm for 1.5 h with a Beckman SW28 rotor and stored at −80°C. HM were seeded in six-well plate at ≈70% confluency. The next day, the cells were infected with 10 μl of shRNA–RelA lentivirus or control virus shRNA-nonspecific. The virus was added directly into the medium of the cell culture. The plate was spun 1 h at 1,300 × g, and the cells were incubated at 37°C for 48 h. The expression of the reporter gene GFP was verified by fluorescence microscopy. Then the cells were treated with 10 ng/ml TNF-α for 30 min. HM were harvested, and cytoplasmic and nuclear proteins were extracted. NF-κB was assessed by Western blot with an antibody (Santa Cruz Biotechnology) specific for the p65 (RelA) NF-κB subunit.

Cytogenetic Analysis.

Chromosome analyses were conducted on two primary HM. Metaphases were arrested by exposing cultured cells overnight to colcemid (0.01 μg/ml). HM were detached with trypsin and treated with 0.075 M KCl hypotonic solution for 20 min, followed by fixation in a 3:1 mixture of methanol/acetic acid. Metaphase slide preparations were made by a steam-drying technique. Chromosomes were G-banded and analyzed.

Statistical Analysis.

Multiple group comparisons were performed with ANOVA followed by Fisher’s least significant difference testing. Statistical significance between two groups of interest was evaluated by unpaired Student’s t test. Differences were considered significant at P < 0.05.