Erionite and asbestos differently cause transformation of human mesothelial cells



Malignant mesothelioma (MM) is an aggressive tumor associated with environmental or occupational exposure to asbestos fibers. Erionite is a fibrous zeolite, morphologically similar to asbestos and it is assumed to be even more carcinogenic. Onset and progression of MM has been suggested as the result of the cooperation between asbestos and other cofactors, such as SV40 virus infection. Nevertheless, several cases of MM were associated with environmental exposure to erionite in Turkey, where SV40 was never isolated in MM specimens. We show here that erionite is poorly cytotoxic, induces proliferating signals and high growth rate in human mesothelial cells (HMC). Long term exposure to erionite, but not to asbestos fibers, transforms HMC in vitro, regardless of the presence of SV40 sequences, leading to foci formation in cultured monolayers. Cells derived from foci display constitutive activation of Akt, NF-κB and Erk1/2, show prolonged survival and a deregulated cell cycle, involving cyclin D1 and E overexpression. Our results reveal that erionite is able per se to turn HMC into transformed highly proliferating cells and disclose the carcinogenic properties of erionite, prompting for a careful evaluation of environmental exposure to these fibers. The genetic predisposition to the effect of erionite is a separate subject for investigation. © 2007 Wiley-Liss, Inc.

Environmental or occupational exposure to asbestos fibers may cause chronic respiratory diseases, including interstitial lung fibrosis and pleural malignant mesothelioma (MM).1, 2 Pedigree and mineralogical studies indicate that the high incidence of MM in a number of Cappadocia villages in Turkey is caused, in genetically predisposed individuals, by exposure to fibrous erionite. This is a highly pathogenic form of the naturally occurring zeolite, similar in appearance and properties to asbestos.3, 4 Erionite is a strong mutagen,5 considered more carcinogenic than asbestos fibers in man and rodents,6 possibly due to the peculiar property of accumulating iron on its surface, despite its very small content of this element.7 It is well documented that erionite from Oregon has genotoxic properties8 and induces high incidence of MM in rats after both intrapleural inoculation or inhalation,9 although there are no epidemiological data correlating the presence of erionite to MM in Oregon.

Similarly to what occurs with asbestos, exposure to erionite fibers leads to generation of reactive oxygen metabolites from macrophages10 and to increased mRNA levels of the early response c-fos and c-jun proto-oncogenes in mesothelial cells.11, 12 MAPK signaling significantly contributes to pivotal cell functions as cell proliferation and death in response to toxic and inflammatory agents.13 Similarly, PI3K/Akt and NF-κB signaling play a crucial role in determining cell fate after exposure to toxic agents14, 15 and are activated by asbestos fibers as well.16, 17 Interestingly, TNF-α inhibits asbestos-induced cytotoxicity via a NF-κB-dependent pathway.18 These results underscore the key role played by NF-κB in asbestos-induced oncogenesis of human mesothelial cells (HMC). Epidermal growth factor receptor (EGFR) overexpression and activation have been linked to asbestos-induced proliferation.19 Autocrine loops for both EGFR and platelet derived growth factor receptor (PDGFR) have been also reported in MM cells.20 Moreover, we previously demonstrated that an autocrine circuit involving hepatocyte growth factor (HGF) and its receptor Met contributes to HMC transformation21 and others showed that Met is overexpressed in MM cells.22

In vitro studies demonstrated that SV40 and asbestos fibers cooperate to determine HMC transformation,23via PI3K/Akt signaling.24 Therefore, here we compared the proliferating and transforming efficacy of short- and long-term exposures of HMC to erionite, amosite, chrysotile or glass fibers, in absence of SV40 infection. Moreover, to achieve a better understanding on the mechanism by which these fibers determine cell transformation, we further evaluated tyrosine kinase receptor expression and signal transduction in HMC transformed by the different fibers.

Our data reveal that erionite fibers have higher intrinsic transformation ability than asbestos fibers and give a possible explanation for the high incidence of MM in 3 small villages in Turkey, where SV40 virus infection of MM cells has never been detected.3, 4, 25, 26

Material and methods


The sample of erionite fibers from Karain, provided by co-author Dr. Umran Dogan (Ankara, Turkey), was associated with considerable amounts of nonfibrous particles and bundles, with diameter greater than 3 μm and length-to-width ratio (L/W) less than 3:1. The fraction of fibrous material with L/W of 3:1 or greater in this sample is less than 10% by weight, because erionite in Karain occurs as a minor constituent of a volcanic amorphous rock, from which the erionite mineral originates.4, 9, 27, 28, 29 The erionite sample from Oregon, kindly provided by Dr. J. W. Skidmore (Glamorgan, UK), was almost totally made of fibrous particles.

No milling, crushing or ultrasound processing was performed prior to size analysis. Representative portions of each erionite sample was weighted (5 ± 0.5 mg) and dispersed in 20 ml of filtered, deionized, distilled water. About 0.45 ml of the dispersion was filtered on a policarbonate filter (0.1 μm), dried, placed on aluminum stub and plated with gold in a sputter coater.

Each sample was examined using a Philips XL30 scanning electron microscope (SEM) at different magnifications, to determine dispersion and visibility of the fibers. Central fields with adequate fibers dispersion were selected for each sample and length (L) and width (W) of fibers were measured directly on the SEM screen with the ruler of the microscope. Fiber bundles with L/W of 3:1 or greater were considered as single fibers. L and W measurements of about 500 fibers were recorded for each fibrous sample. The size analysis of glass fibers was similarly performed.

Size distribution of UICC asbestos samples (Chrysotile B and Amosite) were acquired from previous data obtained by transmission electron microscopy.30

Erionite (Karain-Turkey and Oregon-USA), amosite, chrysotile (UICC asbestos samples) and glass fibers were dispersed in PBS at 2.0 mg/ml, before autoclaving. Then, amosite, chrysotile and glass fibers, but not erionite fibers, were triturated 8 times through a 22-gauge needle.

Cell cultures

Two primary HMC cell cultures, obtained from patients with heart failure were cultured in Ham's F-12 medium, supplemented with 10% fetal bovine serum (FBS–GIBCO, Rockville, MD) at 37°C in a 5% CO2-humidified atmosphere. HMC were used between the second and the sixth passage. Both cultures gave similar results in all assays performed, as previously characterized.24


HMC cells were exposed to medium containing 2% FBS supplemented with fibers in presence of 10 μM BrdU (Bromo-deoxy-Uridine). Incorporation in neosynthesized DNA was evaluated after 24 hr by the cell proliferation kit (Roche, Basel, Switzerland). Data are expressed as mean increase of DNA-neosynthesis over untreated controls.

Cytotoxicity and DNA adducts

Cells were seeded on multiwell plates and 5 × 103 cell/well were exposed for 24 or 48 hr to fibers at densities ranging from 2 to 10 μg/cm2 in presence of 2% FBS. Experiments performed on foci derived cells were conducted in presence of 100 μM VP16 (Etoposide, Sigma-Aldrich, St. Louis, MO), for 24 hr. Cytotoxicity was assessed by MTT assay, performed in quadruplicate, as previously described.31 Normalized cytotoxicity percentages were obtained according to the ratio: [1-(A570 mean values of extracts from exposed samples/A570 mean values of extracts from control cell samples)] × 100.

DNA adducts were evaluated by high-performance liquid chromatography (HPLC), in the presence or absence of 10 mM N-Acetyl-L-cysteine (L-NAC) (Sigma-Aldrich, St. Louis, MO), on extracts from cells exposed for 5 hr at 10 μg/cm2 to the indicated fibers and are expressed as amount of 8-OHdG per 105 dG, as previously described.32

Exposure to fibers

Short term exposure

Stimulation of sub-confluent cells for 24 hr with medium containing 20% FBS supplemented with fibers at concentrations ranging from 0.1 to 10 μg/cm2.

Long term exposure

Two cycles of treatment, 72 hr each, with low concentrations of fibers. In details, 1 day after plating at 80% confluence in 25 ml flasks, cells were exposed to 2.5 μg/cm2 of each fiber for 72 hr. Then, cells were washed twice with PBS and cells were transferred in larger (75-and 150-cm2) flasks and let growing in medium containing 10% FBS for additional 4 days. Afterwards, the same fiber treatment was repeated as above and cells were grown up to 2 months, by 1:4 periodical passages. Treatments were made in triplicate.


Sub-confluent cells were exposed to Ham's medium supplemented with 2% FBS, containing fibers (10 μg/cm2) or 100 μM VP16 (Etoposide, Sigma-Aldrich, St. Louis, MO) for 24 hr. Nuclei fragmentation was evaluated by staining cells grown on cover slips with 8 μg/ml Hoechst solution (Calbiochem, San Diego, CA) in dark conditions for 1 hr. Samples were fixed in –20°C cold acetone:methanol for 15 min, washed in PBS, mounted in 50% glycerol/PBS and observed with a Leica immunofluorescence microscope. Cells displaying nuclear fragmentation were counted on 10 fields out of at least 50 cells in the same slide. Values are expressed as percentages of Hoechst staining positive HMC over total counted cells. Caspase activity was evaluated by staining cells with CaspACE FITC-VAD-FMK in situ marker (Promega, Madison, WI), followed by flow cytometry.

Signal transduction

Immunoblotting was performed by loading 50 μg of cell lysates in reducing conditions. After separation on SDS-PAGE and transfer to nitrocellulose (Hybond, Amersham, Buckinghamshire, UK), filters were probed with phospho-p38 (Thr180-Tyr182), phospho-NF-κB p65 (Ser536), phospho-JNK (Thr183-Tyr185), phospho-Akt (Ser473), phospho-Erk1/2 (Thr202-Tyr204), Akt, Erk1/2 antibodies from Cell Signalling Technology, Beverly, MA, α-Tubulin antibodies from Sigma-Aldrich as loading controls, Met, EGFR, PDGFRβ, NF-κB p65, JNK, p38, Cyclin D1 and E antibodies from Santa Cruz Biotechnology, Santa Cruz, CA. Detection was performed by the enhanced chemiluminescence system (ECL, Amersham).

Focus forming assay

Cells that survived the “long term” exposure to fibers, were followed 8 weeks, then a focus forming assay was performed in triplicate in 6-well dishes. Cells were plated at a density of 3 × 104/well and grown in 10% FBS Ham's medium. The number of foci per number of seeded cells is expressed as mean number ± standard deviation (SD). Foci arisen from confluent cells were taken and successfully established in cultures as single clones. Biochemical and biological characterization of foci were performed on a pool of clones, with a representative mixture of cells from each original focus.

Cell proliferation

Cells were grown on 24-well plates at a density of 3 × 104/well in Ham's medium supplemented with 2% FBS and containing the indicated fibers (1.25 μg/cm2). Cells were fixed in 11% glutaraldehyde after 0, 24, 48 and 72 hr exposure and stained in crystal violet staining was eluted in 10% acetic acid and absorbance at 595 nm (A595) was measured in an ELISA plate reader.33

Immunochemical staining

Sub-confluent cells plated on glass slide flaskets (NUNC, Rochester, NY) were exposed to Ham's medium supplemented with 2% FBS and containing fibers (10 μg/cm2) for 24 hr and subsequently fixed in 10% formalin. After 1 hr incubation with Ki67 antibodies (Neomarkers, Freemont, CA) at room temperature, biotin-streptavidin immunostaining was performed with UltraVision detection system, according to the manufacturer's instructions. Ki67 positive cells were counted on 10 fields with at least 50 cells in the same slide. Values are expressed as percentages of Ki67 positive cells over total counted HMC.

Cell cycle

Cells were synchronized by 0.1 μg/ml Colcemyd (Sigma-Aldrich) treatment for 24 hr, and then kept in normal medium for 4 days before analysis. They were washed in PBS, fixed in 50% ethanol and stained for 30 min at room temperature with 50 μg/ml propidium iodide (PI-Sigma-Aldrich) in 0.1 M PBS pH 7.2 containing 0.5 mg/ml RNAse. 10,000 events per sample were analyzed by flow cytometry.


Data from cytotoxicity, DNA adducts, cell cycle, DNA neosynthesis, proliferation assays and caspase activity were expressed as mean ± standard deviation (SD) of at least 3 independent experiments. Statistical differences were evaluated by analysis of variance (ANOVA), followed by Tukey's HSD. Statistical analysis of Ki67 immunostaining and Hoechst staining assays was performed by Fisher's exact test among different groups, as indicated in the text. In all statistical evaluation the significance threshold was specified in the text. All statistical tests were two-sided.


Most of erionite fiber samples fall in a mineralogical category with lower cytotoxicity than asbestos fibers

Karain erionite is a minor constituent of a volcanic rock, which is mainly Si-rich glass.4, 9, 27, 28, 29 In addition, montmorillonite, and traces of quartz, feldspar, opal, clay (illite), carbonates were found associated with rock and soil samples from Karain.27, 28 For these reasons, the preparation of an adequately fiber-enriched sample for our tests was not possible. Nevertheless, in accordance with previous studies,28 most of the matrix particulate material associated to the erionite used here is amorphous glass and few clay, carbonate and feldspar minerals. The erionite sample was characterized as having acceptable balance error, to verify that the data obtained fit with chemical composition and structure of the tested mineral.

From SEM examination the fraction of Karain fibrous material with L/W of 3:1 or greater resulted to be in the range 5–10%. In the erionite sample from Oregon this fraction resulted about 75%, with nonfibrous particles being mostly of amorphous nature (Fig. 1), in accordance with previous studies.34

Figure 1.

SEM analysis of erionite fiber samples. SEM analysis of Karain (a) and Oregon (b) fiber samples at ×500 magnification. Central fields with adequate dispersion of fibers were selected for each sample. Bar: 50 μm.

The distribution of fiber sizes of the erionite samples are shown in the Table I. The fraction of fibers (L > 8 μm and W ≤ 1.5 μm) with the dimensional category suggested as having the highest carcinogenic potential in rats (L > 8 μm and W ≤ 1.5 μm)35 was 26.0 and 30.1% in the Karain and Oregon samples, respectively. As shown in Table II, the size distribution of UICC asbestos samples (chrysotile B and amosite) reveals that all particles were asbestiform fibers with diameters smaller than 1 μm. However, the fraction of fibers with L > 8 μm was 88 and 70% for chrysotile and amosite, respectively.

Table I. Number of Fibres in Different Dimensional Categories by SEM
Fibre diameter (μm)Fibre length, μm (%)
  • 1

    Proportion of fibres falling into the dimensional category with highest neoplastic response.

Oregon erionite (n = 495)
>1.51 (0.2)3 (0.6)41 (8.2)
>0.5–1.536 (7.2)55 (11.1)69 (14.0)1
≤0.5117 (23.6)94 (19.0)79 (16.1)1
Karain erionite (n = 500)
>1.51 (0.2)8 (1.6)82 (16.4)
>0.5–1.538 (7.6)69 (13.8)97 (19.4)1
≤0.587 (17.4)85 (17.0)33 (6.6)1
Table II. Size Distribution (%) of UICC Asbestos Samples23
UICC sampleL ≥ 8 μmD ≤ 1.5 μm
Chrysotile B88100

Calculations made on the basis of size distributions of Karain and Oregon erionite fibers revealed that the dust in size range considered more biologically active was constituted of ∼19 and 180 fibers per microgram (F/μg) respectively, whereas the UICC chrysotile sample had about 1.5 × 105 and UICC amosite 1.2 × 105 F/μg. Noteworthy, the number of Oregon erionite fibers expressed in F/μg was very close to the value (150 F/μg) previously reported.8

Therefore, the fraction of more biologically active fibers was considerably lower in the case of erionite fibers, as compared to UICC asbestos samples, particularly in the case of erionite from Karain. In the glass fiber sample only a very small percentage (about 1%) of fibers were included in the more pathogenic size range.

Erionite induces low cell death but high DNA neosynthesis

We compared cytotoxicity induced by Karain (Turkey) and Oregon (USA) erionite fibers with that of asbestos fibers (amosite and chrysotile) and of glass fibers, used as relatively inert control. HMC were exposed for short term (24 hr) to fiber suspensions ranging from 0.1 up to 10 μg/cm2, and cytotoxicity was evaluated by MTT assay. Our data show that dose-dependent cytotoxicity occurred in HMC exposed to erionite fibers, although by far lower than that induced by amosite. The differences evaluated at 5 and 10 μg/cm2 at 24 hr are statistically significant (p ≤ 0.001). Conversely, chrysotile-induced cytotoxicity was comparable with that of erionite. As expected, no detectable cytotoxicity was induced by glass fibers. After 48 hr the cytotoxic response was uniformly increased and differences among amosite and erionite fibers in inducing cell death were less evident, albeit erionite fibers at 5 μg/cm2 still significantly displayed lower cytotoxicity levels than amosite (p ≤ 0.005) and even than glass (p ≤ 0.05) (Fig. 2a). These results were confirmed when apoptosis was evaluated by counting apoptotic cells after Hoechst staining. A higher number of nuclei showing typical chromatin condensation was observed upon cell exposure to pro-apoptotic agent VP16 (11.3%) and to amosite (12.4%), than in cells exposed to Karain erionite (3.9%), Oregon erionite (5.6%) or to glass beads (2.5%). These differences were statistically significant (p ≤ 0.001). The possible effects of oxidative stress by fibers were examined by 8-hydroxy-2′-deoxyguanosine (8-OHdG) evaluation with HPLC and UV/amperometric detection. Interestingly, amosite and erionite fibers significantly induced DNA damage as compared to unexposed controls or to glass (p ≤ 0.001). However, the amount of adducts induced by both erionite fibers was by far higher than that induced by amosite (p ≤ 0.001). When the exposure to fibers was conducted in presence of the anti-oxidant agent N-Acetyl-L-Cysteine (L-NAC), the amount of DNA adducts significantly decreased for amosite and erionite fibers (p ≤ 0.001), (Fig. 2b). This strongly indicates that oxygen reactive species play a role in DNA damage induced by exposure to fibers. Moreover, cell exposure to amosite or erionite fibers at 10 μg/cm2 induced significantly higher BrdU incorporation in comparison with exposure to glass beads (p ≤ 0.005) or chrysotile (p ≤ 0.05), (Fig. 2c).

Figure 2.

Short term exposure to erionite and asbestos induces cytotoxicity, DNA adducts and DNA neosynthesis. (a) Cytotoxicity evaluated by MTT assay at different doses of the indicated fiber types, after short term exposure of 24 and 48 hr. (b) DNA content of 8-OHdG in HMC exposed to the indicated fibers, in presence or absence of L-NAC, evaluated as number of adducts formed per 105 dG. (c) BrdU incorporation in neosynthesized DNA of HMC cells, after 24 hr exposure to different doses of the indicated fiber types, expressed as percentage over untreated control.

Only erionite-induced DNA neosynthesis leads to cell proliferation

Cell proliferation was monitored in HMC cultures to verify whether the considerable DNA neosynthesis observed upon erionite exposure was due either to compensatory synthesis of DNA, to repair fiber-induced chromosome damage, or to actual cell proliferation.

We observed an evident, although not statistically significant, increase in cell proliferation rate after 24 hr upon exposure to both types of erionite fibers, as compared to amosite, chrysotile and glass fibers exposure. After 72 hr, these differences were still evident, even though cells were sub-confluent and attained the growth plateau (Fig. 3a). Also Ki67 staining, a known cell proliferation marker, was determined upon 24 hr exposure to amosite, glass and both types of erionite fibers. Higher percentages of Ki67-positive cells were found in HMC cultures exposed to Karain erionite (6.2%) and to Oregon erionite (5.0%), as compared to amosite (2.6%) or glass (2.1%) fibers, as well as to untreated cells (3.2%). The differences in Ki67 staining induced by both erionite fibers compared to other fiber types are statistically significant (p ≤ 0.005). Moreover, transient Jnk phosphorylation was observed 4 hr after exposure to any fiber type (10 μg/cm2), suggesting a nonspecific stress response (Fig. 3b). Jnk phosphorylation decreased after 8 hr, except in the case of amosite, which evoked instead a sustained Jnk activity, as expected from its high level of cytotoxicity. Erk1/2 and p38 phosphorylation was not detectable after exposure to glass fibers, whereas Erk1/2 phosphorylation was sustained up to 8 hr after exposure to amosite. In cells exposed to erionite fibers Erk1/2 signaling was more transient. The activity of p38 was by far more evident in cells exposed to erionite than amosite, especially in consideration of a reduced p38 expression in erionite treated cells (Fig. 3b). In cells exposed to amosite fibers we observed: reduced proliferation rate (Fig. 3a), increased cell death (Fig. 2a) and significantly lower staining of Ki67, but also elevated BrdU incorporation (Fig. 2c) and sustained Erk1/2 activity (Fig. 3b). One possible interpretation of these data is that DNA neosyntesis occur in cells nevertheless dying, as a consequence of amosite damage.

Figure 3.

Erionite-induced DNA neosynthesis is not only compensation for DNA damage. (a) Cell growth curve of HMC exposed to the indicated fibers and monitored at 24, 48 and 72 hr. Cell number was estimated by Abs595 nm values of the adsorbed crystal violet. (b) Immunoblotting with phospho-JNK, phospho-ERK1/2 and phospho-p38 antibodies on total lysates of HMC exposed to the indicated fibers for 2, 4, 8 and 24 hr. The expression levels of JNK, ERK1/2 and p38 were also examined by immunoblotting and α-tubulin content is also reported as loading control.

As regards erionite fibers, we conclude that DNA-neosynthesis observed after short term exposure to fibers is not only compensatory in consequence of erionite-induced DNA damage, but also allows HMC proliferation. Moreover, the exposure to erionite fibers, displaying low cytotoxicity and high production of reactive oxygen species, exerts on HMC the highest transforming potential.

Erionite fibers “long-term” exposure promotes HMC transformation per se

It has been demonstrated that prolonged exposure of SV40 positive-HMC to asbestos fibers may induce transformation.23, 24 We aimed to verify whether exposure to low doses of erionite fibers could induce mesothelial transformation and if SV40 is required, as for asbestos fibers. We treated HMC according to a “long term” exposure protocol (see Methods). Two months after the “long term” exposure, only cells exposed to erionite fibers underwent loss of cell contact inhibition and several foci arose from the culture. No statistically significant differences were observed between the 2 types of erionite fibers in focus formation. No foci resulted after exposure to amosite, chrysotile and glass fibers (Table III). These cells acquired a clear-cut novel morphology, similar to what we previously described24 for cells exposed to asbestos (Fig. 4a left). Cells obtained from foci were cultured up to 29 passages for HMC-Karain erionite and 48 passages for HMC-Oregon erionite. These cells maintained the phenotype described earlier, grew in low serum, displayed anchorage-independent soft-agar growth and showed a significantly higher proliferation rate than HMC (at 48 hr, p ≤ 0.05 for HMC-Karain erionite and p ≤ 0.005 for HMC-Oregon erionite; Fig. 4a right).

Figure 4.

Erionite fibers “long-term” exposure per se promotes HMC transformation. (a) Left: representative pictures of HMC 7, 30 and 60 days after the long-term exposure to Karain erionite fibers, showing the different acquired morphology. Right: proliferation assay of pooled cell clones from foci obtained after long term exposure to both types of erionite fibers. Cell number was estimated by Abs595 nm values of the adsorbed crystal violet. (b) Cell cycle analysis of synchronized HMC and cells derived from foci elicited by exposure to the two types of erionite. The percentages of cells in the different phases of the cell cycle are given in details below the cytograms. (c) Immunoblotting on total lysates of HMC and of foci derived cells probed with cyclin D1 and cyclin E antibodies. Same filter and loading controls of Figure 5, determined by α-tubulin immunoblotting.

Table III. Foci Formed after 60 Days “Long Term” Exposure to Different Fibre Types
TreatmentFrequency of focus formation1
  • 1

    Number of transformed foci per treated cell.

  • 2

    No foci developed from 3 × 105 cells exposed to fibres.

Karain erionite1.0 × 10−4 ± 0.65 × 10−4
Oregon erionite0.3 × 10−4 ± 0.46 × 10−4

Cell cycle analysis revealed that cells obtained from foci, which escaped serum-dependency for growth, exhibited disappearance of subG1-S hypoploid phase, displayed a strong increase of S-phase entry (Fig. 4b). The differences between percentages of foci cells and HMC in subG1-phase and in S-phase are statistically significant (p ≤ 0.001). Moreover, we observed over expression of cyclins D1 and E, involved in G1/S cell cycle transition (Fig. 4c). In foci derived cells HGFR/Met, EGFR and PDGFRβ were expressed at higher extent than in HMC, as determined by immunoblotting. Analysis of cell signaling revealed that these cells displayed Akt, Erk1/2 and NF-κB activities at higher levels than in untransformed HMC (Fig. 5). Moreover, cells derived from foci became significantly more resistant (p ≤ 0.001 for HMC-Karain Erionite and p ≤ 0.005 for HMC-Oregon Erionite) than HMC to the pro-apoptotic agent VP16 (Etoposide), as determined by MTT assay (Fig. 6a). As well, induction of Caspase activity, evaluated by flow cytometry, was significantly reduced in foci (p ≤ 0.001) as compared with untransformed HMC (Fig. 6b).

Figure 5.

Characterization of cells derived from foci. Immunoblotting on total lysates of HMC performed using either antibodies directed against growth factor receptors (Met/HGFR, EGFR, PDGFRβ) and phospho-specific antibodies recognizing the active form of intracellular effectors (P-p65 NF-κB, P-Akt, P-Erk1/2). Levels of p65, Akt and Erk1/2 were also determined by immunoblotting. The content of cellular proteins has been normalized by α-tubulin immunoblotting.

Figure 6.

Cells from foci become resistant to toxics. (a) Cytotoxicity of the indicated cell types (HMC and cells derived from foci), after exposure to 100 μM VP16, evaluated by MTT assay. (b) Cytofluorimetric analysis of caspase activity, by using the CaspACE FITC-VAD-FMK marker, of HMC and cells derived from foci, after exposure to 100 μM VP16. The percentage of caspase activity increase is indicated on the right.

We conclude that even relatively low concentrations of erionite fibers can cause transformation of HMC.


Our results show that erionite has clear intrinsic transforming properties, associated with low cytotoxicity and high oxygen reactive species formation, combined with high DNA neo-synthesis and high cell proliferation rate. Cell transformation is demonstrated by foci formation in cultures exposed to erionite but not to amosite fibers. Transformed cells obtained from erionite induced foci are highly proliferating and resistant to apoptosis. HGFR/Met, EGFR and PDGFRβ are highly expressed as well, and the intracellular signalling effectors NF-κB, Akt and Erk are also activated.

The characterization by SEM analysis of the fibers used, reveals that the proportion of fibers with L > 8 μm and W ≤ 1.5 μm, close to the dimensional category associated with the highest biological activity,35 was considerably lower for erionite compared to UICC asbestos samples, very rich of fibers in the “biologically active” size range. The nonfibrous material present in our erionite samples should not have a substantial pathogenic influence.36 However, we cannot exclude a possible potentiating effect by some components of this material to the action of the fibrous fraction.37

It has been demonstrated that the number of erionite fibers required to develop peritoneal mesothelioma in rats was orders of magnitude lower than that observed for asbestos fibers.38 This explained the higher carcinogenic potential observed for erionite, even at considerably lower concentration than asbestos.6

In the present work, the comparison of asbestos with erionite fibers indicate that amosite induces relevant HMC cytotoxicity, possibly because of their longer fibers or high iron content.39 Conversely, only erionite fibers promote HMC transformation, possibly because these fibers are less cytotoxic than amosite, having a lower content of iron.7 Chrysotile fibers, in spite of their reduced cytotoxicity, are unable to cause cell proliferation, as previously reported.40 Even short term exposure for 24 hr to erionite fibers induced low cytotoxicity, allowing a high proliferation rate, associated with formation of DNA adducts.

Upon fiber exposure several intracellular signals are generated. Jnk was increased by all fiber types, suggesting a nonspecific stress response and its activity was sustained only upon amosite exposure, probably because of the high cytotoxic property of these fibers. The transient patterns of Erk1/2 activity observed in HMC exposed to amosite fibers were similar to those previously reported for crocidolite exposure,16 which has the same cytotoxicity pattern of amosite.24 Erionite- and amosite-dependent Erk1/2 activity was consistent with cell proliferation evoked by these fibers.12 Higher p38 activity in HMC upon exposure to erionite, can be explained with the highest oxidative stress caused by these fibers41 and is consistent with highest levels of 8OH-dG observed here in the same conditions.

In view of the high level of BrdU incorporation in cells otherwise displaying high percentage of cell death, we interpret DNA incorporation as caused by neo-synthesis of DNA in cells that eventually are dying, as a consequence of amosite induced damage, rather than as enhancement of mitogenesis. Conversely, the 2 types of erionite display the lowest cytotoxicity values, combined with DNA neosynthesis and cell proliferation, evaluated both by growth curve and Ki67 marker. Moreover, genotoxic properties of erionite fibers have been reported8 and are confirmed by our data on 8-OHdG adduct levels, which are significantly higher in cells exposed to erionite fibers of both types. Altogether, these results account for high transforming activity of the few erionite fibers in the pathogenic size range.

Cells from foci induced by erionite displayed the properties of in vitro full transformation, i.e. changes in morphology, growth in soft agar, high number of passages in culture and increased proliferation rate. Moreover, these cells had autonomous NF-κB, Erk1/2 and Akt activities. Whereas MAPK activity is likely associated to increased cell proliferation rate, NF-κB and Akt are responsible of the increased resistance to cytotoxic agents, as suggested by the rescue of sensitivity upon treatment with specific PI3K and NF-κB inhibitors (not shown). Noteworthy, recently it has been demonstrated a key role for NF-κB in asbestos carcinogenesis in HMC.18 The presence of activated p65 NF-κB subunit in erionite-induced foci described here is in agreement with these results. Similarly, these cells expressed high levels of HGFR/Met, EGFR and PDGFRβ, suggesting their possible role in sustaining cell survival.

Asbestos is certainly a universally accepted causative agent of MM,2 although several evidences suggest that SV40 virus may cooperate with asbestos in HMC transformation23, 24 and in the onset of MM.42 We show here that erionite fibers are able to induce HMC transformation per se, regardless SV40 infection. This may explain the high incidence of MM in Cappadocia, a region of Turkey where SV40 virus was never isolated.3, 4, 25, 26 Erionite is a ubiquitous zeolite, also present in Oregon, USA, where SV40 has been identified in several human specimens.26 Although amosite is considered a potent human carcinogen, we show here that, according to other similar findings,23 amosite is not clearly transforming by itself. However, the high level of DNA neosynthesis, as that induced by amosite in these in vitro experiments, can induce infidelity of DNA replication and frequent mismatches. Furthermore other agents (such as SV40) can act as critical co-factors, allowing these damaged cells to survive and to progressively acquire biological transformation.

In conclusion, the mineralogical characteristics of erionite fibers (size, surface area and structure) associated with their high transforming capability, highlight the importance of a careful evaluation of the hazard represented by exposure to low concentrations of other similar fibers, to which human beings could be exposed.


We thank Dr. J.W. Skidmore (MRC Pneucomoniosis Unit, Llandough Hospital, Penarth, Wales) for providing erionite samples. We also thank Dr. M. Rinaldi for advice on the statistical analysis. This work is part of a G.I. Me. (Gruppo Italiano per lo studio e la terapia del Mesotelioma) network program.