Asbestos and erionite prime and activate the NLRP3 inflammasome that stimulates autocrine cytokine release in human mesothelial cells
© Hillegass et al.; licensee BioMed Central Ltd. 2013
Received: 11 March 2013
Accepted: 9 August 2013
Published: 13 August 2013
Pleural fibrosis and malignant mesotheliomas (MM) occur after exposures to pathogenic fibers, yet the mechanisms initiating these diseases are unclear.
We document priming and activation of the NLRP3 inflammasome in human mesothelial cells by asbestos and erionite that is causally related to release of IL-1β, IL-6, IL-8, and Vascular Endothelial Growth Factor (VEGF). Transcription and release of these proteins are inhibited in vitro using Anakinra, an IL-1 receptor antagonist that reduces these cytokines in a human peritoneal MM mouse xenograft model.
These novel data show that asbestos-induced priming and activation of the NLRP3 inflammasome triggers an autocrine feedback loop modulated via the IL-1 receptor in mesothelial cell type targeted in pleural infection, fibrosis, and carcinogenesis.
KeywordsAsbestos Mesothelioma Mesothelium Inflammasomes NLRP3
Vascular endothelial growth factor
Human mesothelial cells
- NLRP3 NALP3:
Nod-like receptor-family protein 3
- THP-1 line:
Human macrophage-like cells
Endogenous and exogenous stress-associated danger signals
IL-1 receptor antagonist
Peritoneal lavage fluid
12-O-tetraoctodecanol phorbol-3 acetate
- Act D:
Human primary peritoneal mesothelial cells
Immortalized human mesothelial cells
Tumor Necrosis Factor-α
Scrambled control siRNA
Monosodium urate crystals
Granulocyte-colony stimulating factor
Severe combined immunodeficient
Scanning electron microscopy
Hanks’ Balanced Salt Solution
Quantitative Real-Time PCR
Hypoxanthine phosphoribosyl transferase
Analysis of variance
- Act D:
Human mesothelial cells (HMC) possessing phenotypic and functional features of both epithelial cells and fibroblasts are unique cell types lining the pleural (lung), peritoneal and pericardial cavities. They are target cells of asbestos-associated malignant mesothelioma (MM), unique and devastating tumor types with a poor prognosis. For example, the average life span of patients with MM is approximately 12 months after diagnosis, despite a number of therapeutic strategies (reviewed in [1–3]). The incidence of MM is increasing worldwide, especially in third world countries where use of asbestos is largely unregulated; thus MM remains a global health problem. In addition, abnormal function of mesothelial cells is intrinsic to the development of pleural effusions after infection or pleural injury and pleural fibrosis, an asbestos-associated disease with no effective treatment options .
Mechanisms of pulmonary injury that contribute to the development of pulmonary fibrosis (asbestosis) by asbestos fibers have been studied for decades in clinical and experimental studies (reviewed in [5, 6]). The hallmarks of asbestos inhalation in humans and rodents include early and sustained inflammation causally attributed to initial accumulation of alveolar macrophages that attempt to engulf asbestos fibers, participate in mucociliary clearance of fibers from the lung, and transport fibers to lymph nodes or the lung interstitium. A number of studies have focused on chemotactic factors and cytokine/chemokine release by alveolar macrophages that are essential to neutrophil influx, macrophage-epithelial or macrophage-fibroblast signaling, and the development of epithelial cell hyperplasia and fibrosis. We initially reported mechanisms of activation of the Nod-like receptor-family protein 3 (NLRP3, NALP3) inflammasome and IL-1β production by asbestos fibers in human macrophage-like cells (THP-1 line) and monocytes in vitro. These data support a model in which long asbestos fibers, known to be more inflammatory, carcinogenic and fibrogenic than shorter fibers in rodents (reviewed in [5, 6, 8]), elicited reactive oxygen species through iron-driven redox reactions on the fiber surface as well as by prolonged NADPH oxidase stimulation during frustrated phagocytosis of fibers. Use of NLRP3 knock-out mice showed that the inflammasome was linked functionally to pulmonary inflammation and elevations in IL-1β and KC, a potent neutrophil chemoattractant, in bronchoalveolar lavage fluids. Moreover, NLRP3 knockout mice were more resistant to asbestos-induced lung injury. Subsequently, inflammasome activation by a myriad of endogenous and exogenous stress-associated danger signals (DAMPs), including infectious and toxic agents, has been intensely studied in myeloid cells and very little in non-myeloid cells. This work has generally focused on modulation of the immune system, properties and mechanisms of agents affecting inflammasome activation, and signaling pathways/receptors involved (reviewed in [9, 10]).
Although macrophages play critical roles in clearance as well as inflammation and cell signaling in response to pathogenic fibers in the airways and interstitium of the lung, little is known about whether or not they are required for translocation of asbestos fibers to the pleura or are essential for pleural inflammation and genesis of pleural diseases (reviewed in ). For example, accumulation of fibers in localized acellular areas (pleural plaques) is observed on the pleural surface as opposed to fiber-laden alveolar macrophages and asbestos bodies (iron-coated fibers linked to accumulation and death of macrophages) in the lung [1, 5, 6, 8].
Here we tested the hypothesis that activation of the NLRP3 inflammasome occurs in human mesothelial cells after pleural injury and is linked causally to release of critical cytokines associated with the development of inflammation and pleural diseases. First, we show that in vitro exposures to the naturally occurring, mineralogically distinct fibers crocidolite asbestos and erionite (a fiber type linked to epidemic proportions of MM in areas of Turkey) , cause increases in transcription of NLRP3 mRNA and release of mature IL-1β that are inhibited using NLRP3 siRNA. These changes were accompanied by increases in caspase-1 activity. Elevations in steady-state mRNA levels of NLRP3, IL-1β, IL-6 and IL-8, and release of IL-1β, IL-6, IL-8 and VEGF by HMC were causally linked to an autocrine pathway that was inhibited after addition of the IL-1 receptor antagonist (IL-1ra), Anakinra. In addition, we demonstrate that fiber-exposed HMC cells release the alarmin, HMGB1, via a NLRP3-dependent pathway that is abrogated by blocking the IL-1 receptor (IL-1r). Lastly, we used a well-characterized human xenograft model of peritoneal MM  to show early (1 and 4 wks) production of critical cytokines in peritoneal lavage fluid (PLF) by human MM prior to tumor establishment. Cytokines (IL-8, VEGF, IL-6) in PLF were inhibited most markedly at 1 wk after intraperitoneal (IP) injection of Anakinra in the absence of changes in numbers of macrophages, neutrophils or lymphocytes. Our studies highlight the functional importance of inflammasome-mediated cytokine production via an autocrine pathway in HMC that is perpetuated by durable pathogenic fibers in the pleura. Moreover, data reveal that mesothelial cells are pluripotent cells responding to fiber-induced NLRP3 activation by producing inflammasome-associated pro-inflammatory and angiogenic cytokines via an autocrine feedback loop. We did not observe a significant reduction in spheroid/tumor volume after 4-wks of once daily Anakinra treatment. This may be due to the fact that Anakinra has very short half-life in mice. Future experiments might require a constant infusion of Anakinra for it to become effective. Taken together, our in vitro and in vivo results suggest that selective targeting of the NLRP3 inflammasome or IL-1r may be critical in the prevention and therapy of asbestos-induced pleural diseases.
Asbestos causes NLRP3 priming and activation in human mesothelial cells
Asbestos causes other cytokine release from mesothelial cells
To confirm and determine whether IL-1β and other cytokines were released in both LP9 and NYU474 primary pleural HMC, we performed Bio-Plex cytokine assays on medium from these cells (24 h after asbestos exposure). Tumor Necrosis Factor-α (TNFα) was added to some groups to determine whether or not it was a necessary factor for asbestos-induced cytokine release. The rationale for these studies is that TNFα is mitogenic to mesothelial cells , causes MM cell transformation , and may have pleiomorphic effects on inflammation and/or cell defense as it is linked to both upregulation of the mitochondrial antioxidant enzyme, manganese-containing superoxide dismutase (SOD2) , as well as increased chemokine production by lung macrophages and neutrophils after pathogenic particle exposures .
NLRP3 activation by erionite
HMGB1 is a putative biomarker of inflammation in a number of diseases including endotoxemia, cystic fibrosis, and some tumors including MM [18–20]. In the presence of IL-1β or TNFα, this nuclear protein is released into the cell cytoplasm or secreted. Once extracellular, HMGB1 binds to a number of cytokine proteins that confer increased pro-inflammatory activity . To determine whether erionite caused increased intracellular levels of HMGB1 or its immediate release via an NLRP3 inflammasome-dependent process, LP9 cells were exposed to erionite (SA = 300, 1,200). These studies revealed that although increases in intracellular levels of HMGB1 were not observed at 8 or 24 h by erionite (Figure 4F), toxic concentrations of erionite (1,200) caused significantly elevated levels of HMGB1 in medium (Figure 4G). This finding suggests that although erionite does not increase the expression of HMGB1, it does alter its release from cells.
Anakinra alters cytokine transcription
Anakinra alters cytokines in PLF of SCID mice
Most research on the role of inflammation in asbestos-associated diseases has focused on macrophages, as this is the first cell type accumulating in the lung at sites of initial deposition of inhaled asbestos fibers. In addition, the vast majority of published studies with few exceptions  exploring the role of inflammasomes in the development of a number of diseases have been performed in macrophages or monocytes. Erionite, a non-asbestos fiber type of the zeolite mineral group, is morphologically similar to crocidolite asbestos (Additional file 2: Figure S2A, B), but is more potent in the causation of MM in rodents  and is of contemporary concern because of the high prevalence of MM observed in areas of Turkey where erionite fibers exist in soils and homes [11, 27]. The uniqueness of work here is the demonstration that the pathogenic fibers, asbestos and erionite, but not innocuous particles, such as glass beads, cause both priming and activation of the NLRP3 inflammasome in human mesothelial cells, responses that trigger transcription and production of cytokines critical to the initiation of pleural injury and infection, MM and pleural fibrosis. These responses were validated in primary human pleural and peritoneal mesothelial cells. Most importantly, we link NLRP3 activation to the release of several pro-inflammatory cytokines and VEGF by fiber-stimulated human mesothelial cells in vitro. As these factors are identical to those produced from human MM during the initial stages of inflammation in a SCID mouse MM xenograft model, we believe that IL-1β may be responsible for MM tumor progression also in absence of asbestos fibers. Our studies support a model where an autocrine loop in human mesothelial cells is perpetrated by fiber-induced NLRP3 priming and activation and increased transcription of IL-1β, IL-6 and IL-8 (Additional file 2: Figure S2E). Under these circumstances, HMGB1 may bind to these cytokines, thus increasing its pro-inflammatory properties  and be exuded into serum after injury to the pleura. In support of our findings Yang et al.  have previously shown the release of HMGB1 from asbestos-exposed mesothelial cells, however, we are first to link the asbestos-induced HMGB1 release and NLRP3. Findings of Yang et al. link HMGB1 release as a critical first step in the pathogenesis of asbestos-associated diseases. Both NF-κB and mitogen-activated protein kinase signaling have been implicated as mechanisms regulating the transcription of pro-IL-1β , and asbestos activates both of these pathways in mesothelial cells [29–31]. Transcription and production of IL-1β may be key to induction of VEGF gene expression at both transcriptional and post-transcriptional levels  as well as IL-6 and IL-8 expression . IL-6 also induces VEGF production through a JAK2/STAT3 pathway  and can modulate production of IL-1β in a JAK2-dependent manner . Thus, other pathways of synergy and interaction between these factors may exist in HMC in addition to increased expression and release of these cytokines via inflammasome and caspase-1 activation .
The main purpose of our study here is to demonstrate that stimulation of the IL-1 receptor may be essential for cytokine induction (via an autocrine loop) by asbestos- exposed mesothelial cells in initiation of MM (in vitro studies) as well as in MM tumor progression (in vivo studies). Our in vitro studies show that mesothelial cells secrete IL-1β in response to asbestos/erionite which can then affect the same population of mesothelial cells in an autocrine manner, thus transforming them to become mesothelioma cells. Our in vivo model on the other hand demonstrates that MMs generated from mesothelial cells after asbestos exposures can further secrete IL-1β and/or be affected by IL-1β secreted by host cells, thus affecting their growth/progression (in the absence of asbestos). Therefore, blocking the IL-1 receptor in vivo could be helpful in attenuation of MM growth and progression.
An exciting observation in both in vitro studies and our human MM xenograft model of inflammation was that levels of interleukins and VEGF were blocked using the IL-1r antagonist, Anakinra, currently used for clinical treatment of various inflammatory diseases, including Muckle-Wells syndrome , gout , and arthritis . The lack of significant effects of Anakinra on IL-6, IL-8 and VEGF levels at 4-weeks (Figure 6) could be attributed to two reasons: 1) overwhelming increases in levels of VEGF (4 fold) and IL-8 (7 fold) at 4-weeks; and 2) the short half-life of Anakinra in mice (~4 hours). Future experiments will include repeated dosing or constant infusion of Anakinra to produce significant effects. Our results suggest an autocrine feedback loop resulting in cytokine release that can be blocked at the IL-1 receptor in mesothelial cells. In addition to abrogating the inflammatory mechanisms of mature IL-1β that may also orchestrate the expression, production, and effects of IL-6, IL-8 and VEGF, this approach may be valuable in preventing the ramifications of robust production and release of IL-1 from necrotic and neighboring mesothelial cells after infection  or exposure to asbestos fibers. IL-1α is a key danger signal that triggers both CXCL1 secretion and recruitment of neutrophils via interaction with the IL-1 receptor in murine mesothelial cells . The fact that Anakinra is non-toxic after repeated administration to mice or humans  suggests its use as a novel approach in prevention and treatment of inflammation, a common feature of pleural infection and disease.
In summary, data here question the popular paradigm that initiation of pleural diseases by pathogenic fibers requires cells of the immune system. Moreover, they reveal that unlike macrophages and monocytes, and like MSTO-211H human mesothelioma cells , human mesothelial cells do not require exogenous TNFα or lipopolysaccharide to initiate NLRP3-mediated cytokine release. Conversely, exposure to pathogenic fibers both primes and activates the NLRP3 inflammasome in these cell types.
LP9/TERT-1 (LP9), an hTERT-immortalized cell line that functionally resembles normal human peritoneal mesothelial cells  and human primary peritoneal mesothelial cells (HM3) were obtained from Dr. James Rheinwald (Brigham and Women’s Hospital, Harvard University, Boston, MA). Human pleural mesothelial cells (NYU474) were isolated surgically from cancer-free patients by Dr. Harvey I. Pass (New York University, New York, NY), as was the H2373 sarcomatoid MM cell line . The H2373 MM line gives rise to biphasic MM after injection into SCID mice as verified by a board-certified pathologist (Dr. Kelly Butnor at UVM) . LP9 and H2373 cells were cultured as described previously . NYU474 cells were grown to near confluence in DMEM containing 10% FBS and supplemented with penicillin (50 units/mL) and streptomycin (100 μg/mL). HM3 cells were grown in 50:50 M199:MCDB106 medium (Invitrogen, Carlsbad, CA) supplemented with 15% FBS, 10 ng/mL EGF, 0.4 μg/mL hydrocortisone, 50 units/mL penicillin and 100 μg/mL streptomycin. All cells were incubated at 37°C in 5% CO2 and grown to 80-90% confluence before addition of agents.
Sources of crocidolite asbestos, erionite and glass beads
The physical and chemical characterization of the NIEHS reference sample of crocidolite asbestos has been reported previously . Erionite fibers from Pine Valley, NV, were provided and characterized using similar methods by Dr. Ian Steele (University of Chicago, Chicago, IL). The SA of asbestos and erionite fibers and GB (Polysciences Inc., Warrington, PA) was measured using nitrogen gas sorption analysis. Asbestos, erionite fibers and glass beads were exposed to UV light overnight to inactivate any pathogens growing on them before adding to cell culture.
Scanning electron microscopy (SEM)
Cells grown on Thermonox coverslips (Nalge Nunc International, Naperville, IL) were fixed for imaging as previously described , and critical point-dried using liquid CO2 as the transition fluid in a Samdri PVT-3B critical point dryer (Tousimis Research Corporation, Rockville, MD). Cells on coverslips and fibers on carbon tape were then mounted on aluminum specimen stubs and dried before being sputter-coated with gold and palladium in a Polaron sputter coater (Model 5100; Quorum Technologies, Guelph, ON, Canada) and examined using a JSM 6060 scanning electron microscope (JEOL USA, Inc., Peabody, MA).
Following sterilization under ultraviolet light overnight to destroy lipopolysaccharide or microbial contaminants, minerals were suspended in 1 X Hanks’ Balanced Salt Solution (HBSS) at 1 mg/mL, sonicated for 15 min in a water bath sonicator, and triturated 5 X through a 22-gauge needle. This suspension was added to cells in medium to achieve the desired SA-based concentrations. After 24 h intervals, cells were collected by trypsinization, and counted using a hemocytometer .
Quantitative Real-Time PCR (qRT-PCR)
Total RNA (1 μg) was reverse-transcribed with random primers using the Promega AMV Reverse Transcriptase kit (Promega, Madison, WI) according to the recommendations of the manufacturer as described previously . Transcription was evaluated using the ∆∆Ct method. Duplicate or triplicate assays were performed with RNA samples isolated from at least two independent experiments. The values obtained from cDNAs and hypoxanthine phosphoribosyl transferase (hprt) controls provided relative gene expression levels for the gene locus investigated. The primers and probes used were purchased from Applied Biosystems (Foster City, CA).
Western blot analysis for NLRP3, activated Caspase-1 p20 and HMGB1 in whole cell lysates and supernatants (medium) of asbestos-exposed mesothelial cells
Medium was collected, cells were lysed as previously described , and protein content in cell lysates was determined using the RC DC protein assay (Bio-Rad, Hercules, CA). Medium (500 μL) was concentrated using Amcion® ultracentrifugal filters with a 10 K membrane (Millipore, Billerica, MA) by spinning at 14,000 X g for 30 min. Columns were then reversed into new collection tubes and spun for 2 min at 1,000 X g. Sample buffer was added to the concentrated supernatant, and samples were boiled for 15 min. Western blots were performed as previously described  on both cell lysates (NLRP3, HMGB1) and concentrated supernatants (HMGB1, Capase1-p20). Rabbit polyclonal antibodies for HMGB1 (Abcam, Cambridge, MA), NLRP3 (Novus Biologicals, Littleton, CO) and Caspase-1-p20 (Cell Signaling, Danvers, MA) were used at 1:500 dilutions.
Caspase-1 activity assay
Caspase-1 activity was measured using the Caspase-1 Colorimetric Assay (R&D Systems, Minneapolis, MN), according to the manufacturer’s directions. Caspase-1 activity was determined as OD405 (after subtraction of a blank control) relative to total protein. Protein concentrations were determined by the Bio-Rad protein assay  using the remaining lysate.
ELISA assay for IL-1β and IL-18
The Quantikine Human IL-1β/IL-1f2 Immunoassay (R&D Systems, Minneapolis, MN, measures predominantly mature IL-1β) was used on concentrated cell medium, prepared as described earlier, and the assay performed according to the manufacturer’s instructions. A total of 500 μL of cell supernatant was concentrated. Two hundred μL samples with assay diluents were loaded into 96 well plates pre-coated with IL-1β antibody. For IL-1β release studies in erionite-exposed cells, asbestos exposed HM3 cells and TNFα priming of siCon and siNLRP3 LP9 cells, an ELISA MAX™ Human IL-1β (Biolegend, San Diego, CA, measures both pro and mature form) was used. Wells were pre-coated with an IL-1β capture antibody overnight at 4°C. Fifty μL of concentrated samples or standards then were prepared in assay diluent and allowed to attach to plates overnight at 4°C. IL-18 release was measured using the Human IL-18 ELISA kit (MBL International, Woburn, MA, measures predominantly active IL-18, 0.7% pro IL-18). Medium was collected (500 uL), concentrated and processed according to the ELISA protocol provided. Values were expressed as pg/mL of IL-1β or IL-18 from the original supernatant (non-concentrated).
Transfection of cells with NLRP3 siRNA
On-Target plus non-targeting siRNA #1 (scrambled control), and On-Target plus SMART pool human NLRP3 siRNA (100 nM; Dharmacon, Lafayette, CO) were transfected into LP9 cells at near confluence using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) following the manufacturer’s protocol. The efficiency of NLRP3 knockdown was determined by qRT-PCR.
SCID mouse xenograft model, IP injection of human MM, and retrieval of PLF
H2373 cells (5 × 106 cells in 50 μL 0.9% NaCl, pH 7.4) were injected into the peritoneal cavity of 6 wk-old male Fox Chase SCID mice (n = 3 mice/group/time point) as described previously . Mice were euthanized using sodium pentobarbital before the peritoneal cavity of each mouse was instilled with 5 mL of cold sterile PBS using an 18-gauge needle. The abdomen was gently massaged, and PLF was aspirated back into the syringe and placed on ice. PLF was then centrifuged at 1,000 rpm for 5 min at 4°C, and the supernatant removed and stored at -80°C for Bio-Plex cytokine analysis. Cytospins were prepared from cell pellets to determine total and differential cell counts .
Bio-Plex assays on cell medium in vitro and in PLF
To quantify cytokine and chemokine levels in cell supernatants and PLF, a multiplex suspension protein array was performed using a Human Cytokine 27-plex panel (Bio-Rad) as described previously . Concentrations of each cytokine and chemokine were determined using Bio-Plex Manager Version 3.0 software. Data were expressed as pg cytokine/mL medium.
In vitro and in vivo studies using the IL-1-ra, Ana
For in vitro experiments, cells were pre-treated with 100 ng/mL of IL-1ra Ana (Insight Genomics, Falls Church, VA) for 1 h before administration of asbestos . For in vivo experiments, SCID mice either injected with 500 μL sterile 0.9% NaCl (pH 7.4) or human MM cells (H2373) in saline were then injected with 100 mg/kg Ana (Kineret®) (Amgen, Thousand Oaks, CA) in 500 μL sterile 0.9% NaCl (pH 7.4)  or 500 μL sterile 0.9% NaCl (pH 7.4) alone. Mice were injected IP daily (with saline or Ana) for 1 wk observations or 3 X IP weekly for 4 wks before euthanization and collection of PLF for Bio-Plex cytokine analyses and cytospin preparations as described above.
Data were evaluated either by analysis of variance (ANOVA) using the Student Neuman-Keul’s procedure for adjustment of pair-wise comparisons between groups or a Student’s t-test. Differences in gene expression values determined by qRT-PCR were evaluated using a Student’s t-test. Differences with p values ≤ 0.05 were considered statistically significant. All experiments were repeated in duplicate or triplicate. Figures represent individual experiments, unless otherwise noted, with Mean ± SEM presented.
Dr. Pamela M. Vacek, Department of Medical Biostatistics at UVM, assisted with statistical analyses. We acknowledge the recently deceased Dr. Jurg Tshopp, a previous collaborator and pioneer in inflammasome biology, for his enthusiasm, helpful discussions, and advice on approaches used here. The Pine Valley, NV erionite sample was kindly contributed and characterized by Dr. Ian Steele, The University of Chicago, Chicago, IL. We also acknowledge the Vermont Cancer Center (VCC) DNA Analysis Facility for assistance with qRT-PCR studies, and thank Jennifer Díaz for her preparation of the manuscript for submission. This work was supported by National Institutes of Health Grants R01 ES021110 (AS), T32 ES007122 (BTM for JMM, SLM), and J. Walter Juckett Fellowship from the Lake Champlain Cancer Research Organization (LCCRO), Burlington, VT (JMH).
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.