Amphibole Asbestos as an Environmental Trigger for Systemic Autoimmune Diseases

Abstract

A growing body of evidence supports an association between systemic autoimmune disease and exposure to amphibole asbestos, a form of asbestos typically with straight, stiff, needle-like fibers that are easily inhaled. While the bulk of this evidence comes from the population exposed occupationally and environmentally to Libby Amphibole (LA) due to the mining of contaminated vermiculite in Montana, studies from Italy and Australia are broadening the evidence to other sites of amphibole exposures. What these investigations have done, that most historical studies have not, is to evaluate amphibole asbestos separately from chrysotile, the most common commercial asbestos in the United States. Here we review the current and historical evidence summarizing amphibole asbestos exposure as a risk factor for autoimmune disease. In both mice and humans, amphibole asbestos, but not chrysotile, drives production of both antinuclear autoantibodies (ANA) associated with lupus-like pathologies and pathogenic autoantibodies against mesothelial cells that appear to contribute to a severe and progressive pleural fibrosis. A growing public health concern has emerged with revelations that a) unregulated asbestos minerals can be just as pathogenic as commercial (regulated) asbestos, and b) bedrock and soil occurrences of asbestos are far more widespread than previously thought. While occupational exposures may be decreasing, environmental exposures are on the rise for many reasons, including those due to the creation of windborne asbestos-containing dusts from urban development and climate change, making this topic an urgent challenge for public and heath provider education, health screening and environmental regulations.

Graphical abstract

1. Introduction

1.1. Statement of Problem and Purpose

The prevalence of autoimmune diseases is increasing worldwide, with considerable evidence that environmental exposures are driving most of this increase [1]. Because of the public health impact, increasing costs, and challenges of treatment, a critical next step is to identify contributing environmental factors. This effort would enhance exposure mitigation, allow public education, and increase screening of exposed populations, thus helping to reduce the burden of autoimmunity on society. One important environmental exposure that needs to be considered is asbestos, a fibrous silicate mineral, and similar elongated mineral fibers such as erionite and fluoroedenite. It is well documented and widely accepted that exposure to crystalline silica dust increases both antinuclear autoantibody production and the risk of systemic autoimmune diseases (SAID) including systemic lupus erythematosus (SLE), systemic sclerosis (SSc) and rheumatoid arthritis (RA) [2–4]. The mechanisms of silica-induced autoimmunity are related to a complex series of events, many of which have been replicated in experiments using amphibole asbestos [4–6]. Therefore, while on-going research into the interaction between cells and mineral fibers is still needed, these similarities provide robust support and background for understanding induction of autoimmunity by amphibole asbestos.

The purpose of this paper is to review the science regarding asbestos exposure and autoimmunity and to call for the acknowledgement of amphibole asbestos as a driver of autoimmunity and autoimmune diseases in order to encourage more research and effective preventative measures.

2. Methods

2.1. Strategy for Mapping Potential Exposures to Amphibole Asbestos and Erionite

A geographic information system (GIS) approach was used to evaluate the geological potential for fibrous amphibole and erionite by utilizing documented occurrences [7–18] and small-scale geologic mapping compilations [19]. The mapping that was used in this study was designed for use at coarser scales, so more localized interpretations are not appropriate. However, these same methodologies can be used at localized scales to help determine where future studies should be directed. These geologic map data contain attributes that describe the different surface rocks that are present in an area [19].

Mapped rock units were selected that included metamorphic rocks, rocks formed from the intrusion of magmas, and rocks that were subjected to tectonic processes [19]. These rocks are often present in mountainous areas where there has been a history of tectonic uplift, exposing windows and slices of crust that may have been subjected to conditions appropriate for the possible formation of fibrous amphiboles. For erionite, rock units of higher potential include volcanic rocks that contain tuffs. Known erionite localities are in the western United States (US) where geologically more recent volcanic activity resulted in the deposition of volcanic ash combined with hydrothermal or diagenetic alteration [16]. Further details on production of the maps are included below in Section 5.2.

2.2. Search Strategy and Selection Criteria for Epidemiological Studies (Table 1 and Figure 1)

Table 1:

Human studies of Asbestos Exposure and Autoimmunity

As described in Section 2.2, studies ranging from case studies to epidemiologic analyses were compiled in chronological order.

Year	Exposure	Type of asbestos	Autoimmune Outcome	Reference
1958	Occupational	Unknown	Case study: Caplan Syndrome (rheumatoid pneumoconiosis), with asbestos exposure.	Rickards, 1958
1961	Occupational	Amosite, Chry, Crocidolite	Case study: Caplan Syndrome (rheumatoid pneumoconiosis), with asbestos exposure.	Telleson, 1961
1965	Insulation workers	Unknown	Frequency & titers of RF significantly elevated in workers vs matched controls	Pernis, et al. 1965
1970	Asbestos workers, screening	Mixed	Frequency of ANA and RF significantly elevated above expected in normal population	Turner Warwick, 1970
1973	Asbestos workers, screening	Mixed	Frequency of ANA and RF significantly elevated above expected in normal population	Turner Warwick, 1973
1973	Factory Workers	Unknown	Frequency of ANA and RF not elevated above expected, but more frequent in groups with longer exposure	Turner Warwick, 1973
1973	Naval Personnel	Likely mixed	Frequency of ANA and RF not elevated above expected, but more frequent in groups with longer exposure	Turner Warwick, 1973
1974	Asbestosis patients	Mixed	Increased frequency of ANA with exposure, with or without asbestosis	Lange, 1974
1974	Chrysotile mine	Chrysotile	No association between exposure and rheumatic symptoms	White, 1974
1974	Shipyard workers	Mostly Chrysotile	Frequency of RF, not ANA, elevated above expected	Stansfield, 1974
1976	Asbestos Miners	Anthophyllite (Amphibole)	RF & ANA frequency stated to be normal	Toivanen, et al, 1976
1977	Asbestos cement	Mixed	Frequency of ANA and RF significantly elevated above expected in normal population	Kagan, et al. 1977
1978	Asbestosis patients	Mixed	Elevated RF, ANA, serum Ig in patients compared to matched no exposure controls.	Haslam, et al. 1978
1978	Asbestos workers	Mixed	Frequency of ANA and RF higher than healthy blood donors in Finland	Huuskonen, 1978
1979	Asbestos worker	Mixed	17 yr. Follow up on Telleson patient 1961, ANA, polyarthritis, complex autoimmune.	Greaves, 1979
1979	Asbestos workers	Mixed	Elevated ANA in workers compared to expected	Gregor, 1979
1980	Textile workers	Mixed	Elevated frequency and titers of ANA, not RF, in workers compared to matched controls.	Lange, 1980
1982	Shipyard & Textile	Mixed	Case Study: Positive ANA, RA after asbestos exposure	Gaensler, 1982
1983	Asbestos workers	Chrysotile	Elevated serum IgG but not ANA	deShazo, 1983
1983	Asbestos workers	Mixed	Elevated ANA, RF, C3 and immune complexes in workers compared to controls	Lange, 1983
1983	Asbestos cement	Amphibole	Elevated ANA (not RF) in workers compared to expected, especially in those with asbestosis	Doll, 1983
1985	Insulation workers	Chrysotile	No elevation of ANA or RF compared to expected and controls	Zone, 1985
1986	Asbestos workers	Mixed	Frequency of ANA elevated in exposed compared to controls	Lange, 1986
1989	Whitewash	Tremolite (Amphibole)	ANA slightly elevated above expected, especially with longer exposure latency	Zerva, 1989
1992	Asbestos workers	Unknown	ANA, but not RF, frequency elevated in asbestos exposed subjects compared to expected	Al Jarad, 1992
1993	Asbestos workers	Mixed	ANA significantly elevated above matched control group	Tamura 1993, 1996
1993	Asbestos milling	Mixed	ANA frequency & CIC significantly elevated in exposed compared to matched unexposed controls.	Nigam 1993
1994	Asbestos workers	Mixed	dsDNA autoantibodies elevated in workers compared to matched unexposed controls	Marczynski, 1994
2003	Asbestos workers	Mixed	ANCA frequency and titer elevated in exposed workers, compared to matched unexposed controls	Pelclova, 2003
2004	RA Patients	Unknown	Occupational asbestos exposure associated with RA	Olsson, 2004
2005	Vermiculite mining	Libby Amphibole	ANA, not RF, frequency and titer elevated in exposed compared to matched unexposed controls	Pfau, 2005
2006	Occupational	Unknown	ANCA vasculitis significantly associated with exposure to either silica or asbestos	Rihova, 2005
2006	Occup/Environmental	Libby Amphibole	Case/Control study found association between exposure and RA, and trend for SLE and SSc	Noonan, 2006
2007	Occupational	Unknown	Occupational Asbestos as significant contributor to Systemic Sclerosis	Gold, 2007
2012	Vermiculite mining	Libby Amphibole	Increased frequency of ANA & MCAA associated with exposure .	Marchand, 2012
2015	Vermiculite vs Steamfitters	Amphibole vs Chrysotile	Increased frequency of ANA with amphibole exposure, not chrysotile	Pfau, 2015
2016	Textile workers	Unknown	Exposure significantly associated with increased risk of RA	Too, 2016
2918	Case series	Libby Amphibole	SAID prevalence higher than US expected, incl RA, SLE, SSc, Sarcoidosis, other syndromes	Diegel, 2018
2018	Mining & Environmental	Crocidolite	Increased frequency & titer of ANA with exposure	Reid, 2018
2018	Environmental	Fluoroedenite (Amphibole)	Increased frequency of ANA in exposed compared to unexposed matched controls	Ledda, 2018
2018	Mining & Environmental	Libby Amphibole	Elevated frequency and titers of ANA with Amphibole compared to Chrysotile	Pfau, 2018
2018	Vermiculite vs Steamfitters	Fluoroedenite (Amphibole)	ANA frequency increased in exposed compared to matched controls	Rapisarda, 2018
2019	Textile workers	Unknown	Higher risk of both sero-pos and sero-neg RA compared to unexposed controls.	Ilar, 2019
2020	Case series	Chrysotile	Frequency of ANA not elevated for chrysotile exposure	Lee 2020
2020	Mining & Environmental	Libby Amphibole	SAID mortality elevated with exposure	Larson, 2020
2021	Environmental	Crocidolite	Frequency of ANA elevated among exposed compared to Australia expected prevalence.	Carey, 2021
2022	RA Patients	Unknown	Asbestos exposure associated with increased risk of RA	Tang, 2022
2023	Asbestosis patients	Unknown	Prevalence of RA elevated among asbestosis patients compared to expected in Finland	Keskitalo, 2023

Figure 1:

Associations Between Asbestos Exposure and ANA in Humans.

All studies included provided the odds ratio (OR) and 95% Confidence Interval (CI) for the presence of positive ANA tests among those exposed compared to a control group, or provided the data for those calculations, as described in the Methods, Section 2.5. Type of Asbestos was either provided in the paper or determined based on information provided about the exposures.

The MEDLINE repository was searched using PubMed for eligible articles without a specific time frame, using MeSH major search terms “asbestos”, OR “asbestos, amphibole”, OR “asbestos, chrysotile” OR “asbestosis” AND “autoimmune” OR “rheumatoid”. Searches including “ANA” OR “antinuclear autoantibodies” AND “autoimmune” OR “rheumatoid” were also performed. Since MEDLINE started in 1966, bibliographies from papers in the 1960’s and 1970’s were perused to find earlier papers in the late 1950’s. These searches gathered 164 studies which were read and evaluated. Papers were excluded if the article referenced the search terms but did not actually study the association or did not report data. Table 1 includes the remaining 48 studies that evaluated autoimmune effects of asbestos in humans, including case studies, epidemiological studies, and serological studies. Studies were excluded from ANA frequency analysis presented in the Forest plot (Figure 1) if they were a) simple case studies, b) review articles that did not include extensive literature review or lacked statistical analysis, or c) the study was performed in rodents (see Table 2). All remaining articles were used in Figure 1 analysis if any of the following was true: a) age and sex-matched internal controls were used or b) age and sex distributions were provided and the study mentioned either the ANA testing method, or the source of normal expected prevalence values used. Although smoking, duration of exposure and type of occupation and/or exposure were usually mentioned, lack of these details did not exclude them from our analysis but are mentioned in the text. With those criteria applied, 31 studies are included in Figure 1.

Table 2:

Rodent studies of Asbestos Exposure and Autoimmunity

As described in Section 2.3, studies using animal models with asbestos exposure to test for autoimmune responses were compiled in chronological order.

Year	Strain	Sex	Treatment/route/duration	Comments	Citation
1965	Rabbits	Not given	Chrysotile, i.v. & i.t.	RF increased with exposure to quartz dust, not chrysotile	Pernis, 1965
2008	C57BL/6 mice	Female	LA, I.t., 7 mo.	LA increased ANA, anti-Ro52, anti-dsDNA, immune complexes	Pfau, 2008
2008	C57BL/6 mice	Female	LA, I.t., 7 mo.	LA induced anti-SSA/Ro52 autoantibodies that bind to LA-induced apoptotic blebs	Blake, 2008
2011	C57BL/6 mice	Female	LA, tremolite, I.t., 7 mo.	Both amphiboles induced anti-fibroblast antibodies	Pfau, 2011
2012	Lewis rat	Female	LA, amosite, I.t., 13 weeks	Both fibers increase ANA, no exacerbated RA	Salazar, 2012
2013	Lewis rat	Female	LA, amosite, I.t., 13 weeks	Both fibers increase ANA, anti-Jo-1, No immune complex or anti-dsDNA	Salazar, 2013
2013	C57BL/6 mice	Female	LA, Chry, I.t., 7 mo.	LA (not Chry) increased ANA and IL-17	Ferro, 2013
2014	C57BL/6 mice	Female	Erionite, tremolite, LA, chrysotile; i.t., 7 mo.	Erionite and amphiboles induced ANA and Th-17 cytokines; chrysotile did not	Zebedeo, 2014
2014	C57BL/6 mice	Both	LA, Wollastonite, i.t., i.p., 1-7days	B1a trafficking is induced by LA from pleural and peritoneal cavities	Pfau, 2014
2016	C57BL/6 mice	Both	LA, I.t., 7 mo.	LA induces MCAA in mice, and MCAA induced peritoneal collagen production	Gilmer, 2016
2017	C57BL/6 mice	Both	LA, AZA, o.p., 7 mo.	Low dose (3ug/mouse) of both fibers induced ANA, Th-17 cytokines, B cell activation, proteinuria, lung & pleural fibrosis.	Pfau, 2017
2019	C57BL/6 mice	Both	LA, i.p. 3 days	Synthetic flaxseed diglucoside inhibits LA-induced acute inflammation and WBC trafficking	Christofidou-Solomidou_2019
2021	C57BL/6 mice	Both	LA, I.t., 14 days.	Synthetic flaxseed diglucoside inhibits LA-induced late inflammation, class switching, and WBC trafficking	Badger, 2021
2022	Various	Both	Review: silica & asbestos	Both silica and amphibole asbestos are relevant to systemic autoimmune disease	Janssen, 2022
2023	C57BL/6 mice	Female	Amosite, o.p., 4 mo.	anti-dsDNA and efferocytosis increased	Lescoat, 2023

Abbreviations:

i.v. Intravenous

i.t. Intratracheal

i.p. intraperitoneal

o.p. Oropharyngeal

RF Rheumatoid Factor

LA Libby Amphibole

Chry Chrysotile asbestos

AZA Arizona Amphibole

RA Rheumatoid Arthritis

AFA Anti-Fibroblast Antibodies

MCAA Mesothelial Cell Autoantibodies

WBC White Blood Cells

2.3. Search Strategy and Selection Criteria for Rodent Studies (Table 2)

In order to create Table 2, similar searches were performed to identify studies in rodents (in vivo) that tested hypotheses about asbestos-related autoimmunity or inflammatory and immune responses that are associated with autoimmune mechanisms. We also used a recent metanalysis of rodent studies related to asbestos and silica exposures that evaluated study quality during their selection [5] to select studies that authors determined had used appropriate controls and had adequate animal numbers. There have been many rodent studies that evaluated inflammatory responses related to fibrosis and carcinogenesis, but those were excluded.

2.4. Search Strategy and Selection Criteria for Studies Comparing Asbestos and Silica (Table 3)

Table 3:

Comparison of the Efects of Crystalline Silica, Amphibole Asbestos, Chrysotile Asbestos

	Amphibole Asbestos	Silica	Chrysotile Asbestos	References
Chronic inflammation	Yes	Yes	Yes	Mossman (1998)
Mesothelioma	Yes	No	Yes	Attanoos (2018); Steenland (1997)
ROS Production	Yes	Yes	Yes	Blake (2007); Hu (2006); Kussainova (2023)
Thiol depletion	Yes	Yes	Yes	Blake et al. (2007); Hu (2006); Afaq (2000)
NALP3 Inflammasome	Yes	Yes	Yes	Li (2012); Dostert (2008)
Profibroticb	Yes	Yes	Yes	Mossman (1998)
ANA, esp. Lupus-associated	Yes	Yes	No	Pfau (2008a); Mayeux (2018); Ferro (2013)
Glomerular damage	Yes	Yes	DNFa	Pfau et al. (2008b): Mayeux et al. (2018)
Kidney Immune Complexes	Yes	Yes	DNF	Pfau et al. (2008b): Mayeux et al. (2018)
Proteinuria	Yes	Yes	DNF	Pfau et al. (2017): Mayeux et al. (2018)
Decreased Treg cells	Yes	Yes	No	Pfau (2018aza); Brown (2004), Otsuki (2012)
Innate B cell trafficking	Yes	Yes	Yes	Pfau (2014); Brown (2004); Ferro (2013)
Increased regulatory B cells	No	Yes	Yes	Ferro (2013); Liu (2016)
Ectopic Lymphoid Tissues	Yes	Yes	No	Pollard (2020); Chauhan (2021); Rom (1992)
Th-17 response (IL17, IL6, TNF)	Yes	Yes	No	Ferro (2013); Chen (2014)
Apoptosis of macrophages	Yes	Yes	Yes	Blake (2007); Pfau (2004); Hamilton (1996)
Antibodies to apoptotic blebs	Yes	Yes	DNF	Blake (2007); Pfau (2004)
Impaired Efferocytosis	Yes	Yes	DNF	Lescoat (2020, 2023)
PAD2/4 enzymes	Yes	Yes	DNF	Desai (2017); Mohamed, 2018c
NET formation	Yes	Yes	DNF	Desai (2017); Rada (2018)

^aDNF = Data not found

^bAsbestosis = diffuse fibrosis, lower lung zones; silicosis = silicotic nodules, upper zones

^cMesothelioma patients, environmental asbestos exposure, type unknown but likely mixed

PubMed searches were performed to identify studies that compared the risk of “autoimmune” responses by “silica” AND “asbestos” in vitro or in vivo. Since a number of those studies did not differentiate between chrysotile and amphibole, further searches were done to find studies on those same outcomes (inflammatory or autoimmune) in which specifically chrysotile or amphibole asbestos was used. That is why for most outcomes, multiple studies are cited for each outcome in Table 3.

2.5. Statistical Analyses for Data in Figure 1

Odds ratios (OR) and 95% confidence intervals (95% CI) comparing exposure groups with controls were often provided in the papers themselves. If they were not provided, these values were calculated using an on-line calculator (https://www.medcalc.org/calc/odds-ratio.php) using values provided in the papers in terms of the frequency of measured outcomes in their experimental group compared to their control group. If no control group was used, peer-reviewed papers were selected that provided the frequency of the outcome (such as positive ANA tests) in a normal, healthy U.S. population during a specific period of time and using a particular ANA or RF testing protocol. Autoantibody testing protocols have evolved from the 1960’s to current, becoming increasingly sensitive. Therefore, we used normal control values consistent with the testing method or time period used in the study being evaluated. As much as possible, the control/comparison values were matched for age and sex to the study population.

The comparison papers are as follows:

For comparisons with studies performed in the 1960’s, Alexander, et al., was used [20].

For comparisons with studies performed in the 1970’s and 1980’s, Shu, et al., was used [21], except for papers comparing at 1:40 serum dilution, in which case Tan, et al., was used [22].

For comparisons with studies performed in the 1990’s –2020’s, Dinse, et al., was used [23].

3. The Case for Amphibole Asbestos Induction of Autoimmunity

3.1. Amphibole Versus Chrysotile Asbestos Mineralogy and Health Effects

‘Asbestos’ is a commodity and regulatory term describing minerals historically treated as a single commercial entity. However, in the US, the regulatory term encompasses six minerals that come from two different mineral groups. Chrysotile [Mg₃Si₂O₅(OH)₄] is a sheet silicate mineral from the serpentine mineral group, whereas tremolite [Ca₂ Mg₅ Si₈ O₂₂ (OH)₂], actinolite [Ca₂ (Fe, Mg)₅ Si₈ O₂₂ (OH)₂], riebeckite (crocidolite) [Na₂ (Fe³⁺₂, Fe²⁺₃) Si₈ O₂₂ (OH)₂], anthophyllite [Mg₂ (Mg)₅ Si₈ O₂₂ (OH)₂], and cummingtonite-grunerite (amosite) [(Fe, Mg)₂ (Fe, Mg)₅ Si₈ O₂₂ (OH)₂] are from the amphibole group and are double chain silicate minerals. The physical properties of each mineral are controlled by their chemical composition and their atomic structure. Thus, although asbestos is treated as a single commodity based on a set of marketable properties, these properties can vary extensively among these six different minerals. In general, the amphibole asbestos minerals as a group are better than chrysotile for withstanding high temperatures and resistance to acids, whereas chrysotile is generally more flexible and thus better for weaving and spinning as compared to the amphibole minerals [24, 25]. All of these asbestos minerals are naturally-occurring, and form as a result of metamorphism, hydrothermal alteration, and metasomatism [9, 26]. These geological processes are common over large areas of the US, especially in and around numerous mountain ranges in both the eastern and western portions of the US (Figure 3). Therefore, people can be exposed through either or both the use of asbestos containing products or through environmental exposures [7, 26, 27].

Figure 3:

Potential Extent of Exposure Risk for Amphibole Asbestos and Erionite in the U.S.

GIS was used to map the potential for exposure to fibrous amphibole and erionite by utilizing documented occurrences and small-scale geologic mapping compilation, as described in Sections 2.1 and 5.2. Figure 3A shows sites of known or predicted occurrence of asbestos. Figure 3B shows these occurrences with an overlay of population density according to the 2020 census data.

While crystalline silica dust is well known to cause SAID, designation of asbestos as a related causative factor has been hampered by a failure of the historic literature to clearly define whether exposures were chrysotile or amphibole. Research over the last 20 years suggests that only the amphibole forms of asbestos drive autoimmunity, while chrysotile asbestos does not [28, 29]. Because chrysotile is the most common commercial asbestos used in the US, most occupational studies have evaluated primarily chrysotile exposure, and these have not supported an association between this form of asbestos and autoimmunity [30, 31] (See Table 1 and Figure 1). However, with newer studies focusing on amphibole exposures, evidence for a link with autoimmunity is being revealed [32–34].

3.2. Lessons from Libby, Montana

Large deposits of vermiculite were discovered in the Cabinet Mountains of northwestern Montana in the early 1900’s. Due to its ability to be expanded (“popped”) by heating, vermiculite was found to have excellent insulation and fire-proofing qualities. The Zonolite Company began mining and marketing the material for insulation, plaster, and a garden soil amendment. Libby, Montana, a growing lumber and mining community at that time, lies at the base of Zonolite Mountain, and a highly productive expansion and bagging plant opened and shipped large quantities of the material throughout the country by railroad. Over 200 such processing facilities opened up along those rail routes to further the production capacity and distribution of products.

The presence of asbestos in the ore was known as early as the 1960’s, but this was not made public until high mortality rates among workers at the mine and expansion plant were revealed late in the 1990’s [35], and subsequently asbestos-related diseases were described among workers at other Libby vermiculite expansion plants [36]. In the late 1990’s, the State Medical Officer was concerned about anecdotal reports of high rates of SAID, especially RA and SLE, in the county. Researchers from the University of Montana were asked in 2002 to assess the veracity of this observation and to evaluate the possibility that this could be related to asbestos exposure.

Data supporting high rates of positive tests for anti-nuclear autoantibodies (ANA) among Libby miners and residents were published in 2005 [37], and a nested-case control study was published in 2006 supporting an elevated risk of SAID among people exposed to LA [38]. Other studies documenting the autoantibody specificities (ANA and antibodies to extractable nuclear antigens, ENA) were published in 2008 [39] and 2018 [40] revealing a high frequency of lupus autoantibodies such as anti-histone, anti-chromatin, and anti-Ro52. Data providing strong evidence of a lupus-like disease in mice from LA exposure, including autoantibody production and autoimmune glomerular damage, were published in 2008 [41]. A review of the prevalence of autoimmune diseases in Libby, along with a case series, was published in 2018 [32].

The presence of pathogenic autoantibodies targeting mesothelial cells in people exposed to LA was reported in 2012 [42], and this has been documented in LA-exposed mice as well [43]. These mesothelial cell autoantibodies (MCAA) have been shown to drive fibrosis in mice [44] and are associated with progressive pleural disease in humans [45]. These data demonstrate the ability of MCAA to drive progressive autoimmune induced pleural fibrosis, which is a severe and deadly disease seen among those exposed to LA [46–48].

Libby Amphibole is spread by wind and water and was used in millions of homes and other buildings throughout the US. The World Trade Center buildings in New York City released dust containing significant amounts of asbestos. This included Monokote spray-on insulation which contained LA and insulated the first 40 floors of the North Tower [49]. Now, due to disease latency, doctors are beginning to see increases in mesothelioma [50], sarcoidosis [51] and autoimmune diseases [52] in those exposed during rescue/recovery efforts, but no increase so far in autoimmune diseases in the broader General Responder Cohort (GRC), many of whom were only exposed later [53].

3.3. Epidemiological Studies

The earliest paper linking asbestos exposure with rheumatic disease was a case of Caplan Syndrome in a man exposed to asbestos in the navy and an asbestos factory [54]. Caplan Syndrome is the concurrence of coal dust or silica pneumoconiosis with RA [55], but in this case, because his silica exposure was minimal and his lung pathology was asbestosis rather than silicosis, the cause was determined to be asbestos. A second such case was described by Tellesson, et al. in 1961 [56]. Since then, over 40 studies have evaluated the association between asbestos exposure and autoimmunity (Table 1). The early work focused on rheumatoid factors, with elevated risk occurring with amphibole or mixed exposures [57–62]. No association was found between exposure and RF in papers where the exposure was known to be chrysotile [30, 31, 63, 64]. An elevated risk of RA with asbestos exposure has been reported several times, and all showed increasing risk with increasing years of exposure [38, 65–67]. Asbestos was second only to silica in strength as a risk factor for both sero-positive and sero-negative RA in a large case-control study in Sweden [68]. Recently, the prevalence of RA among asbestosis patients was found to be significantly elevated over expected prevalence in Finland [69]. A nested case-control study found an increasing risk for RA, SLE and SSc with increasing exposure pathways to LA [38]. Gold et al., found occupational asbestos exposure to be a significant contributor to systemic sclerosis [70]. Diegel, et al., reported higher than expected rates of RA, SLE, SSc and sarcoidosis in the screening population of Libby Montana, but not of organ-specific autoimmune diseases such as multiple sclerosis, type 1 diabetes, or Sjogren’s Syndrome [32]. Interestingly, the detailed evaluation of the Libby population revealed a high frequency of patients with several severe symptoms that are associated with SAID, but who did not meet specific diagnostic criteria for any of the known SAID. Symptoms included primarily joint swelling, Raynaud’s phenomenon, severe fatigue, unexplained rashes or sores, and bleeding problems (clotting, bruising, bleeding). The authors’ conclusion was that it is possible that exposure triggers autoimmune processes that lead to symptoms in a given individual which may represent a complex overlap between several autoimmune diseases [32].

Over the last 15 years, reviews have been published regarding asbestos exposures and autoimmunity [28, 29, 71]. Those reviews introduced the hypothesis that the reason for only a weak overall epidemiological association between asbestos and autoimmunity was because studies did not clearly distinguish the types of asbestos in those exposures. Figure 1 is a forest plot of studies that evaluated associations between asbestos and ANA in humans, illustrating the strength of that association for individual studies on ANA specifically and for all of the ANA studies combined (OR= 2.78, 95% CI 2.39-3.23). The association between asbestos exposure and ANA is very strong for the overall data, which includes amphibole, chrysotile, and mixed exposures. The association is strongest for amphibole asbestos exposure, and is not statistically significant for chrysotile exposure. These data support the hypothesis that amphibole asbestos, but not chrysotile, is a trigger for non-tissue-specific autoantibodies. Other studies of asbestos with other SAID antibodies, rheumatoid factor (RF), and anti-neutrophil cytoplasmic antibodies (ANCA) are not as conclusive, partly due to varying study designs and sample sizes (Table 1).

3.4. Animal studies

Despite limitations of using animal models of human disease, rodent studies of asbestos exposure over the last 20 years have provided critical information about mechanisms of pathology, pathogenicity determinants, and data to interpret observations in humans and begin to evaluate causation, rather than just association [72].

Table 2 summarizes studies done in rodent models specifically focused on autoantibodies and evidence of SAID. In addition, rodent studies on asbestos and autoimmunity were recently thoroughly reviewed in a metanalysis that evaluated the quality of the study designs as well as the findings [5]. Their overall conclusion was that both crystalline silica and asbestos are highly relevant to the development of autoimmune conditions. Although there are fewer studies with asbestos than silica that directly evaluated autoimmune outcomes, the studies were determined to be of high enough quality and consistency to say that the association between asbestos and autoimmunity was supported [5]. Most of the work has been done with amphibole asbestos including LA, tremolite, and amosite. Three studies tested chrysotile compared against amphiboles, and in all cases chrysotile did not increase ANA, RF or autoimmune signs above control levels [61, 73, 74].

In addition to ANA, seven months after exposure to LA, mice developed immune complexes deposited in kidneys [41, 74], evidence of glomerular damage [41], proteinuria [75], a Th-17 cytokine profile [73–75], autoantibodies targeting apoptotic bleb contents [76], and autoantibodies to mesothelial cells (MCAA) [44, 75] and fibroblasts [77]. Other humoral changes elicited by amphibole asbestos in mice were altered immunoglobulin isotypes, lymphocyte trafficking, splenomegaly, and B1a cell activation [41, 73, 75, 78, 79].

Most of the studies were performed in the common C57BL/6 mice, which are not prone to develop spontaneous autoimmunity until senescence. The only other animals used were rabbits [61] and Lewis rats [80, 81]. Rabbits exposed to pure chrysotile did not develop RF activity, but those exposed to crystalline silica did [61]. Rats exposed to LA or amosite developed high titer ANA and proteinuria [81]. No clear evidence of kidney disease was seen at the 7-month time point, but as rats live nearly twice as long as mice, this time point may have been premature for development of disease. Salazar, et al., also used the collagen-induced arthritis (CIA) and the peptidoglycan-polysaccharide (PG-PS) models of RA in Lewis rats to test whether LA or amosite exposure affected the disease course [80]. Although neither asbestos exposure increased the rate or severity in either arthritis model, both exposures did increase the number of mice with positive ANA. More studies are clearly needed in various rodent strains to tease apart genetic susceptibility, but all studies of amphibole asbestos in rodents have shown evidence of autoimmune mechanisms.

4. Mechanisms of silicate-induced autoimmunity

4.1. Summary of Mechanisms of Induction

This section provides a comprehensive overview of the proposed mechanisms involved in the pathogenesis of amphibole asbestos-induced autoimmunity, including inflammation, immune dysregulation, oxidative stress, genetic susceptibility, and autoantibody production. Understanding these mechanisms is crucial for developing targeted therapeutic strategies and preventive measures.

Amphibole asbestos exposure is associated primarily with systemic autoimmune diseases, including systemic sclerosis, systemic lupus erythematosus, rheumatoid arthritis. As with crystalline silica, the mechanisms underlying the development of autoimmunity in response to amphibole asbestos exposure are complex and multifactorial. Figure 2 illustrates a model of pathogenicity consistent with the current literature described below.

Figure 2:

Illustration of Cellular Mechanisms Linking Amphibole Asbestos Exposure with Autoimmune Responses and Pathogenicity.

Amphibole asbestos induces pathways of inflammation, ROS and cell death that mimic those of crystalline silica, leading to inflammasome activation, release of DAMPs, loss of regulatory T cell activity, and activation of B cells with autoantibody production. Cell death pathways result in altered signaling, antigen modification, and immune cell recruitment. Altered signaling leads to immune dysregulation, TH17 responses and lost tolerance. Amphibole asbestos “checks all the boxes” for activation of autoimmune responses.

4.1.1. Inflammation and Immune Dysregulation

Amphibole asbestos fibers induce chronic inflammation in the lungs and other affected tissues through the activation of immune cells such as macrophages and lymphocytes, as well as several epithelial cells lining the lungs and pleura. This persistent inflammation leads to immune dysregulation, characterized by the production of pro-inflammatory cytokines and chemokines. Dysregulated immune responses contribute to the development of autoimmune diseases by promoting tissue damage and perpetuating inflammation [82]. The inflammatory mechanisms are triggered through a variety of receptors, including scavenger receptors and Toll-like receptors, all of which evoke inflammatory cell recruitment but also trigger various forms of cell death including apoptosis and the release of damage associated molecular patterns (DAMPs) [83–85]. A key cell being exposed and carrying these receptors is the macrophage, which, if undergoing cell death, becomes unavailable for effective clearance of cellular debris. A proposed model implicates asbestos-induced apoptosis of the very cells needed for clearance of cell debris in an environment in which antigen presentation is enhanced due to recruitment of more activated inflammatory cells, thereby increasing the risk of presenting modified self-antigen to the adaptive immune system and lost tolerance [82, 86, 87].

4.1.2. Oxidative Stress

Exposure to amphibole asbestos fibers results in the generation of reactive oxygen species (ROS) and reactive nitrogen species (RNS) within cells. These highly reactive molecules induce oxidative stress, leading to damage to cellular components such as DNA, proteins, and lipids. Oxidative stress plays a crucial role in the pathogenesis of autoimmune diseases by promoting inflammation, activating immune cells, and contributing to tissue damage [82]. There are important differences, however, in the induction of oxidative stress and signaling between chrysotile and amphibole asbestos, in that amphibole asbestos induced more ROS than chrysotile [88]. Asbestos-induced oxidative stress can also impact type of immune response through redox signaling and lipid peroxidation [89–91], leading to immune dysregulation.

4.1.3. Genetic Susceptibility

Genetic factors play a significant role in determining individual susceptibility to autoimmune diseases triggered by amphibole asbestos exposure. Polymorphisms in genes involved in immune regulation, inflammation, and tissue repair have been implicated in the development of autoimmune diseases. These genetic variants may interact with environmental factors, including asbestos fibers, to modulate immune responses and contribute to disease pathogenesis [92, 93].

4.1.4. Epigenetic Modifications

Although little work has been done in this area, amphibole asbestos exposure can lead to epigenetic modifications, primarily DNA hypermethylation of specific promoters in asbestos-induced cancers, but also global methylation changes [94] and production of a variety of microRNAs associated with mesothelioma [95]. These epigenetic alterations can affect the expression of genes involved in immune responses, inflammation, and autoimmunity.

4.1.5. Autoantibody Production

Exposure to amphibole asbestos fibers is associated with the production of autoantibodies. These autoantibodies contribute to the pathogenesis of autoimmune diseases by promoting inflammation, immune complex formation, and tissue damage. The production of autoantibodies may result from the dysregulation of B-cell activation and tolerance mechanisms in response to asbestos exposure [79, 82, 96]. For more mechanistic details specifically related to asbestos, see Section 6.

5. Extent of the Public Health Risk

5.1. Asbestos Definitions and Exposure Routes

Asbestos is a global public health risk that is best known as a cause of pulmonary diseases including mesothelioma, lung cancer, pulmonary fibrosis, and pleural diseases. The link between asbestos exposure and these diseases initially occurred due to outbreaks among occupationally exposed populations early in the 20^th century [97]. Decades later, regulatory agencies such as the Occupational Safety and Health Administration (OSHA) developed exposure standards in an attempt to reduce exposure among workers, and they defined the six previously mentioned fibrous minerals as asbestos due to their commercial value. Recently, the U.S Environmental Protection Agency (EPA) recognized Libby Amphibole (LA), a group of mineral fibers co-existing and mined along with vermiculite in Libby MT, as a seventh in this group [98]. This is important because the last twenty years have revealed that amphiboles that were not previously included in the regulated six fibers (such as Libby Amphibole, winchite, fluoroedenite), also cause the same lung and pleural diseases as the regulated asbestos, making the traditional occupational terminology and definitions obsolete. However, it is a slow process to update regulatory standards that incorporate this new knowledge. Therefore, most environmental asbestos exposures are currently poorly recognized and under-regulated. Also, it is a common myth that asbestos is banned in the US. Although mining for asbestos specifically is banned, the manufacture and use of asbestos-containing products is not. EPA’s recent “asbestos ban” only limited specific uses of chrysotile, but not amphiboles, and does nothing to protect people from exposure to asbestos already in homes and buildings [99]. Also, it is not illegal to mine other important minerals such as talc, iron and vermiculite, all of which commonly co-occur with fibrous amphiboles.

It is now known that asbestos is also associated with several other diseases, including cancers of the larynx and ovaries and possibly cancers of the stomach, pharynx and colorectum [100], all likely due to the ability of mineral fibers to be transported throughout the body by the lymphatics and to interact with and disrupt immune function [28, 101, 102] in similar ways as crystalline silica [4]. These revelations must raise the concern that the association between amphibole asbestos and autoimmune disease is part of the increasing frequency of environmental SAID.

Amphibole asbestos is a common component of environmental dusts created by urban development, erosive winds, off-road vehicle recreation, and mining activities in rock containing asbestos [7, 9, 27, 103–106]. It is also released from use of asbestos-containing products [107, 108]. For example, the desert dusts of southern Nevada contain asbestos fibers, primarily amphibole [7, 9], and residents of the Las Vegas metropolitan area and beyond are exposed to high levels of dust during wind storms. In fact, high rates of mesothelioma have been measured in women/children in this region [109], increasing concerns of asbestos exposure as arid lands expand due to climate change [110]. Minnesota and Wisconsin iron mines and outcroppings contain amosite (commercial asbestos name for cummingtonite-grunerite solid solution amphibole) [104]. Therefore, while amphibole asbestos is not a new exposure, it is a newly revealed environmental exposure that is increasing in intensity and geographic scale due to anthropogenic development and global climate change.

5.2. Geological Basis of Amphibole Asbestos Exposure

5.2.1. Analytical Basis for Mapping Exposure Risk

To better assess the risk of exposure in the US, GIS maps were developed that include locations of previously reported naturally-occurring amphibole asbestos [7, 8, 10–15, 17, 18], and also the locations of sites where Libby, Montana vermiculite was processed [111] (Figure 3). The Libby vermiculite contains amphibole asbestos including winchite, richterite and tremolite (see Sections 3.1, 3.2). For the amphibole asbestos occurrences, we reviewed the datasets and only compiled locations where specifically amphibole asbestos was documented and do not include chrysotile occurrences. We also documented erionite, a fibrous zeolite [16], because it also elicits autoimmune responses [71, 74]. In total, naturally-occurring amphibole asbestos was reported at 468 locations, Libby amphibole/vermiculite processing centers at 104 locations, and naturally-occurring erionite at 95 locations.

5.2.2. Mapping the Potential for Exposure to Naturally-Occurring Amphibole Asbestos and Erionite

To better predict regions that have the potential for human exposure to naturally-occurring amphibole or erionite, we used the known sites of fibrous amphibole or erionite combined with the known geological rock-forming processes that form these minerals to produce a map of areas in the US with the geological potential to contain fibrous amphibole or erionite (Figure 3A). Predicting the distribution of geologic environments and rocks that are more likely to contain amphibole or erionite fibers can be accomplished through analysis of geologic maps [7, 112–115]. Similarly, erionite’s environment of formation is largely a result of diagenesis or hydrothermal alteration, most commonly, although not always, of volcanic rocks [16, 116].

Note that there are several caveats to the data shown in Figure 3A: (1) The data only show the potential for bedrock to contain these fibrous minerals, and therefore do not take into account geological processes such as erosion, transport, and deposition via ice, water, or wind which can spread a single geological occurrence in bedrock into soils and sediments dispersed over a wide geographic area [7, 8, 27, 117]. These processes happen on geological timescales, which further contribute to how hazardous minerals can be transported over great distances. Thus, the geographic distribution of potentially-occurring amphibole asbestos or erionite in Figure 3A may be significantly underestimated; (2) Conversely, the presence of rock types that have the potential for naturally-occurring amphibole asbestos or erionite does not imply that these minerals will be present, only that fiber development is possible in these geological settings. Additionally, due to the resolution of the map units [19], the areas for potential amphibole asbestos or erionite depicted in Figure 3A are likely over-estimating potential bedrock occurrences of these minerals. More detailed geologic research is needed to determine mineral occurrences for specific areas. (3) Some boundaries depicted in Figure 3A are artificial, such as along the Minnesota and North or South Dakota state lines. This is a result of geological map units not yet being reconciled across State boundaries [19]; (4) The sites showing known naturally occurring amphibole asbestos and erionite were largely found as a result of specific exploration for these minerals (or as a result of exploration for other commodity-based minerals). This type of exploration is focused on large, economically-obtainable geologic mineral deposits, which ignores rocks with low concentrations. Thus, many more areas than are currently known may contain low concentrations of amphibole asbestos or erionite, which could pose hazards for health [118, 119]. Much more site-specific research is needed to better define the occurrences. (5) Lastly, the potential for human exposure for naturally-occurring amphibole asbestos and erionite is significantly controlled by additional site-specific factors, some of which include climate, vegetation, wind speed/direction, and perhaps most importantly, human behavior and human disturbance or use of soil and rock. Any actions, natural or anthropogenic, that increase the potential for these minerals to become airborne, also increase the potential human exposures and associated health risks.

5.2.3. Mapping Exposures Relative to Population Density

Finally, we mapped the vermiculite/Libby amphibole, naturally-occurring amphibole asbestos and erionite locations, relative to the 2020 population density (https://data.census.gov/) (Figure 3B). This was performed to better understand potential human interactions with Libby amphibole/vermiculite, and naturally-occurring amphibole asbestos or erionite. These data show that potential human exposures to amphibole asbestos or erionite can occur across the nation and in both rural and densely-populated areas (Figure 3B). The raw vermiculite mined in Libby, Montana contained up to 26% asbestos and the milled vermiculite contained 0.3 to 7% [120]. Approximately 80% of all vermiculite used worldwide from the 1920’s to 1990 may have been from Libby, Montana [35]. This material was mined in Libby and then transported and processed in many areas across the US [121–123].

The EPA estimated that in 1979, in the US alone, 13 million people were exposed to LA by living near vermiculite exfoliation facilities and 106 million people were exposed from consumer products [122]. The 104 Libby amphibole/vermiculite locations shown in Figure 3 do not all have the same level of risk of exposure. Some were large exfoliation facilities, in which the vermiculite was heated and expanded (“popped”) and this processing method likely released more asbestos than other methods [111]. The potential human exposure routes include occupational inhalation by employees, exposure to contaminated clothing to employees and household contacts, and community exposures through airborne dust and/or soil contamination [123]. The timeline and degree of exposure as well as the number of people exposed will vary with each site. Even though these sites are no longer actively being used to transport or process vermiculite, many areas likely have remaining indoor and outdoor soil contamination [111]. Additionally, the potential occurrences of human exposure to LA in vermiculite are more than those shown in Figure 3, because across the nation, these materials still exist as insulation in housing, as soil amendments, and various other products such as plasters, fireproofing, other insulation [111, 123, 124].

6. Criteria for SAID induction by specific exposures

6.1. Application of Criteria to Amphibole Asbestos

An international group of experts in the area of autoimmune diseases developed guidelines for identifying specific environmental agent-associated autoimmune diseases [125]. In order to establish relationships between environmental factors and specific outcomes, the Hill Criteria are often used [126], which include strength of the data and quality of the experimental studies, consistency of the data, specificity of the association, temporal plausibility of the exposure-outcome relationship, and biological plausibility. The literature review above has addressed issues of strength, quality, consistency, and specificity of the data. Below, issues of temporal (Section 6.1.1) and biological plausibility are addressed (Section 6.1.3), as well as genetic issues (Section 6.1.2). With all of those factors addressed, additional guidelines to establish an exposure/outcome association include removal of the exposure and its effects when possible (leading to resolution), then re-challenge if appropriate and assessing if there is a recurrence of the effect. Although such guidelines are consistent with known autoimmune processes, asbestos is unique. Unlike drugs and other chemicals that are removed from the body through metabolic processes, asbestos is a mineral and little metabolism occurs. Fibers remain in the body virtually forever, as evidenced by the presence of both chrysotile and amphibole fibers in lungs of miners after their death in their 70’s and 80’s, 25-35 years after their last exposure [127]. Even though chrysotile is said to be “cleared” from the lung, in reality, this clearance is fragmentation and fraying, simply leading to smaller fibers that are more easily trafficked to other tissues [128]. Because of this persistence in the body, it is not possible to perform “dechallenge” and “rechallenge” to assess the role of asbestos in autoimmunity. The same is true for crystalline silica, which is well-accepted as a trigger exposure for autoimmunity.

6.1.1. Temporal Association

When reviewing the methods for the studies listed in Table 1, seroconversion to ANA positive, or development of a SAID, occurred in a reasonable length of time, at least 18-20 year latency after exposure [129]. In studies that did not have 20 year’s latency, ANA or disease did not develop [130, 131].

6.1.2. Genetic Factors

For most autoimmune diseases, particularly systemic AID, monozygotic twin studies show largely incomplete concordance, and large-scale analysis of polymorphisms show that significant genetic associations applied only to subgroups of patients [93]. Other studies in twins and family members have shown that although genetics clearly play a role in susceptibility to SAID, environmental factors appear to be more significant contributors [132–134]. Genetic contributions appear to be multi-genic, such that risk of disease increases with increasing numbers of genetic risk markers [135].

Several studies have explored genetic contributors to asbestos-induced disease, and beyond the cumulative effect of multiple genetic risk factors, there is very little evidence to link genetics specifically with asbestos diseases. Al Jarad et al. studied the effect of multiple HLA alleles on the development of asbestosis, ANA, or RF after asbestos exposure, and they found no association of any of the alleles with those outcomes [92]. Therefore, while genetic factors can increase the risk of SAID, the exposure can be the “second hit” that tips the scale toward disease.

6.1.3. Mechanistic Biological Plausibility.

Early in the research on asbestos related diseases (ARD), several studies evaluated humoral/immune changes in subjects exposed to asbestos in order to determine whether there were markers that predicted asbestosis, since this interstitial fibrosis did not occur in all asbestos-exposed subjects. Conclusions from a series of studies by Turner-Warwick et al. were that: a) excess ANA and RF occurred in a large proportion of exposed subjects; b) patients with autoantibodies had increased severity of asbestosis, suggesting that the autoantibodies exacerbate asbestos-caused damage, and; c) because not all fibrosing diseases are associated with ANA or RF, and since the auto-antibodies were present in excess in subjects who were exposed but had no radiographic evidence of lung disease, these autoantibodies are likely not caused by fibrotic processes [62, 131]. Other studies have shown increased ANA with asbestos exposure, but particularly elevated among patients with asbestosis [57, 136–138] or with progressive pleural disease [45], consistent with the idea that the autoantibodies contribute to a fibrotic process. Lange et al. measured serum ANA, IgG and IgM, and found all elevated in exposed individuals, but, although immunoglobulin levels were higher in patients with asbestosis, only ANA correlated statistically with asbestosis [138]. The authors hypothesized that the progression of fibrosis in asbestosis is driven by immune complexes. Other studies have shown elevated immune complexes in asbestos-exposed subjects [139, 140]. Kagan et al. measured serum immunoglobulins in asbestosis patients and found elevated levels of IgA, IgG and IgM. They also found elevated levels of cold lymphocytotoxins, defined as autoantibodies against lymphocytes, ANA and RF in the patients, all suggestive of humoral immunological disturbance [141].

In 1980, Lange et al. similarly showed elevated immunoglobulins and a high frequency of ANA in a different group of workers, and postulated that asbestos-induced cellular processes cause modification of self-antigens and reduce T-cell surveillance, resulting in autoantibody production that then drives a fibrotic response [137]. Alteration of self-antigen by asbestos is supported by work showing increased C3 binding to macrophages after amphibole exposure [142], the presence of antibodies to self-antigens of macrophages exposed to amphibole [86], and the presence of antibodies to surface antigens of fibroblasts [77] and mesothelial cells [42] in the blood of amphibole-exposed subjects. The loss of T-cell regulation with asbestos exposure has been demonstrated by many groups, and is nicely reviewed by others [137, 143–145]. In brief, the most consistent findings point to a scenario in which asbestos interacts with alveolar macrophages and epithelial cells, leading to production of chemical mediators that drive the following: a) depressed cellular immune response and anti-tumor immunity [144, 146, 147]; b) increased Th-17 responsiveness, which is implicated in the breakdown of self-tolerance [73, 75, 148]; and c) altered suppressor/regulatory T cell function [144, 149, 150]. The Th-17 response is essential for the development of tertiary lymphoid structures (TLS) which initially develop as granulomas of innate immune cells and then organize into collections of lymphocytes and antigen-presenting cells like macrophages [151]. TLS facilitate interactions between activated B-cells and antigen-presenting cells (APC), creating localized nodes of immune activation. Cell death, antigen modification and, neutrophil extracellular traps (NETs) increase antigen presentation of self-antigens.

6.2. Crystalline Silica and Amphibole Asbestos: Mechanistic Comparison

The mechanisms of silica-induced autoimmunity are related to a complex series of events, including oxidative stress from interaction with cells, cell death processes that release nucleic acids and other damage associated molecular patterns (DAMPs) such as high mobility group box 1 protein (HMGB1), modification of self-antigens and their extrusion in apoptotic blebs, and chronic inflammatory responses with oxidative stress [152–154]. Similarly, amphibole asbestos induces oxidative stress [76, 88], apoptotic mechanisms with extruded modified self-antigen, activation of inflammasomes and DAMPS [86, 155, 156], and leads to production of inflammatory mediators, effects on B and T effector cell populations, and the production of lupus-associated autoantibodies [40, 41, 153]. The effective clearance of apoptotic cells, called efferocytosis, is inhibited by both silica and amphibole asbestos while phagocytosis by antigen presenting cells is left intact [84]. Mice exposed to crystalline silica develop a Th-17 cytokine profile [157] as do mice exposed to amphibole asbestos as described above. Further, the development of TLS from silica has only been described in autoimmune-prone mice [158, 159], but not humans. However, TLS-like structures containing antigen-presenting cells, lymphocytes and even plasma cells have been described in people exposed to mixed amphibole/chrysotile asbestos [160] and in non-autoimmune-prone mice exposed to amphibole asbestos [161]. No similar structures have been described in people or mice exposed only to chrysotile.

Tying the inflammatory and oxidative effects of both crystalline silica and asbestos exposure to autoimmunity, studies have shown that treatments which ameliorate oxidative stress after exposure reduce several of these effects, including inflammatory markers, development of B cell activation, cell trafficking, class switching and autoantibodies [78, 82, 96, 162]. Therefore, the mechanisms that lead to autoimmunity with crystalline silica are mimicked by amphibole asbestos in vitro and in vivo (Reviewed in [5, 82, 145, 163]). In their metanalysis, Janssen, et al., stated that although both amphibole asbestos and crystalline silica can drive lupus-like disease in mice, amphibole asbestos was able to evoke such a response in healthy, non-autoimmune-prone mice, whereas silica was not, making the evidence for amphibole asbestos slightly stronger in mice than for silica [5]. Table 3 reviews effects of silica and asbestos in terms of cellular mechanisms related to autoimmunity.

There are interesting differences between cellular responses to crystalline silica and asbestos, particularly related to cellular management of oxidative stress. Janssen, et al. found that crystalline silica and crocidolite (an amphibole) both cause oxidative stress with production of reactive oxygen species in the lung, but the the patterns of responsive antioxidant enzyme expression differed [164]. Both inhaled minerals induced high levels of mRNA for manganese superoxide dismutase (MnSOD) and glutathione peroxidase within 9 days after exposure, but actual enzyme activity was only sustained with crocidolite. Crocidolite also increased the activity of total SOD and catalase while silica did not. During that same timeline of about 2 weeks, silica-exposed rats had a more robust influx of polymorphonuclear cells (PMN) in bronchoalveolar lavage than asbestos-exposed rats, suggesting a different inflammatory timeline. However, both inhalants induced similar amounts of collagen in the lungs, suggesting that these different patterns did not protect the lung from damage, but may be related to the histological differences between silicosis and asbestosis [164]. The System xc⁻ anti-oxidant mechanism is activated in amphibole-exposed macrophages, but not silica-exposed [90]. Proposed reasons for the differences include a) ability of macrophages to phagocytose silica versus asbestos (although there is good evidence that all but very long fibers > 10 μm in length are phagocytosed), b) the specific location in the lung where deposits of inhaled mineral dusts occur due to shape and size, and c) specific surface properties (including presence of iron, although many pathogenic asbestos fibers do not have surface iron) of the mineral dust particles [90, 145, 164, 165]. Despite these controversial differences, the cellular responses to crystalline silica and asbestos are similar [145] (Table 3), providing further support of the ability of amphibole asbestos to induce autoimmune responses via mechanisms similar to those of crystalline silica. The need for exposure specificity is emphasized by the fact that only crystalline silica, but not amorphous silica, induces autoimmunity. Further research exploring the interaction between cells and silicate particles and fibers is still needed so that targeted therapies for disease and possibly disease susceptibility might be developed.

7. Weight of evidence summary

7.1. Number of Studies

The primary argument (historically) against asbestos as a trigger for autoimmunity has been what appeared to be a relatively limited number of papers evaluating autoimmune disease compared to those for crystalline silica. Occupational asbestos exposure cohorts have been used historically for research on fibrotic diseases (asbestosis and pleural fibrosis) and cancers (mesothelioma, pulmonary carcinoma), with little focus on autoimmune disease, except where evidence showed that the autoantibodies play a role in the progression and severity of other asbestos-related diseases [45, 57–60]. However, when compiled, there is actually quite a robust literature. In addition, the analyses of risk for autoimmune disease with asbestos exposure are very similar to those with crystalline silica exposure [65, 66, 68], and the mechanisms of asbestos-induced autoimmunity are very similar to those of crystalline silica-induced autoimmunity, as discussed above in Section 6.2.

7.2. Distinguishing Types of Asbestos and Mixed Exposures

A secondary limitation in the studies is that few papers distinguished between chrysotile and amphibole asbestos, often just reporting “asbestos workers”. Sometimes occupational details were provided that allowed a presumption: for example, textile workers are primarily exposed to chrysotile, but for specialized acid-resistant fabric, amosite or crocidolite may be added. Brake-pad workers, pipe insulators and steamfitters are primarily exposed to chrysotile [166–169]. In a few studies, little or no association was found between asbestos and autoimmune outcomes even if the exposure was mixed or primarily amphibole. The problem in those studies was that the long latency for autoimmune disease was often not taken into consideration. Disease latency for SAID associated with crystalline silica exposure is said to be 20-25 years [129], and studies showing no association between asbestos and autoimmunity had a relatively short average follow-up of 10-15 years [130, 131].

It has been proposed that a difference between chrysotile and amphibole asbestos effects on the immune response may have to do with types of cell death and self-antigen modification, which affects the titers, specific targets, and pathogenicity of autoantibodies produced [40]. More research is needed regarding the induction of various cell death pathways which variably alter self-antigens, and how those are affected by different asbestos types. A recent excellent literature review discussed mechanisms involved in exposure-induced autoimmunity, including silica and asbestos, and cell death pathways [82]. Their conclusion was that both silica and asbestos invoke similar pathways of cell death, and suggested that differences in specific outcomes are more related to dose, route of exposure, cell types that were studied, and possibly surface properties that determine how particles interact with cells. This latter hypothesis has been supported by another group that has been comparing different asbestos preparations and demonstrating different signaling depending on combinations of receptors that are bound by the fibers [84].

Since 2005, a Japanese group led by Dr. Otsuki has been comparing the cellular effects of crystalline silica and chrysotile asbestos as related to induction of autoimmunity by silica but cancer by chrysotile, especially focused on T cell behavior and activation/suppression after exposure [170]. Overall, their work has emphasized that silica drives cellular events leading to immune stimulation and autoimmunity, while chrysotile causes suppression of anti-cancer immunity by inhibiting T-cell activation [171, 172]. When this group compared chrysotile with crocidolite (an amphibole), however, significant differences were seen in that chrysotile, but not crocidolite, inhibited cytotoxicity in an allogenic peripheral blood mononuclear cell challenge [173]. This supports the hypothesis that chrysotile suppresses cellular immunity, but crystalline silica and amphibole asbestos progress to immune hyperactivity.

Finally, because crystalline silica is known to induce autoimmunity [174], and is a common constituent in many types of rock formations, there is concern that crystalline silica co-exposure is the actual trigger, not the asbestos, during asbestos exposures. However, there is evidence that this is highly unlikely. First, three large occupational exposure studies controlled for both crystalline silica and asbestos exposure by using detailed exposure questionnaires and interviews, and all three showed association with autoimmune disease for crystalline silica and asbestos independently [65, 68, 70]. Importantly, Reid, et al., compared the risk for ANA among miners versus residents of the local town, Wittenoom, Australia, where exposure did not contain respirable crystalline silica, and there was no difference in ANA frequency between the residents and miners, whose exposure likely contained some crystalline silica dust [34]. Second, the commercial applications of crystalline silica and asbestos have different desirable properties such that extraction of both minerals is optimized to avoid contamination by the other. The fibrotic diseases from silica and asbestos are different enough that they have unique names (silicosis vs asbestosis) and the disease characteristics help identify the exposure [175–177]. Finally, in mouse studies, the animals are exposed to preparations of highly purified asbestos fibers with minimal crystalline silica that were collected from veins of nearly pure amphibole [75], In particular, several studies used elutriated (purified and size-fractionated) asbestos and found very similar effects as those seen with raw asbestos preparations [178, 179].

For non-occupational exposures, there are some additional reasons to lessen the concern that co-exposure to crystalline silica is the actual trigger for autoimmunity. Although both crystalline silica (quartz) and the asbestos minerals are naturally-occurring in many different types of rock and soils [100], under natural conditions, it is more difficult for crystalline silica to break into respirable particles, likely because it is harder than most other minerals [180]. Studies routinely find that finer fractions of soils and airborne dust contain significantly less crystalline silica as compared to coarser fractions [181–183]. However, crystalline silica is still an important component of airborne dust [184] and therefore has been the topic of study for non-occupational exposures particularly in arid and semi-arid climates or where human activities such as farming, construction, and demolition activities create airborne dust [185]. However ambient levels of crystalline silica in non-occupational settings are below the levels that cause disease [185] In contrast, in industrial uses, crystalline silica undergoes significant fracturing and grinding, which is able to create fine-grained crystalline silica particles (≤5 μm) and therefore is much more hazardous than naturally-occurring crystalline silica because of the potential for increased human exposure, higher surface area of smaller particles, greater potential to be inhaled more deeply into the lung, and freshly fractured crystalline silica has been reported to be significantly more cytotoxic [182, 185–187]. As already mentioned, the industrial uses of crystalline silica and asbestos minerals are quite different and thus contamination is strongly avoided.

8. Discussion and Conclusions

The tragedy of exposures to Libby Amphibole through mining and use of vermiculite over nearly a century led the Centers for Disease Control and Prevention (CDC) and the Agency for Toxic Substances and Disease Registry (ATSDR) to fund research and screening to improve the understanding of asbestos-related diseases. Over 100 papers have been published from work related to those exposures over the last 25 years. This work has revealed much of the knowledge that now questions the old paradigms of exposure and health effects of mineral fibers, and builds on the work that started in the 1960’s to understand asbestos impacts on the immune system. Research focused on specific mechanisms of induction, including genetic factors affecting susceptibility to autoimmunity, is still needed for both crystalline silica and amphibole asbestos in order to develop targeted therapies.

Key points from this review include the following:

1. Both human and animal studies support an association between amphibole exposure and ANA.

2. Both human and animal studies support an association between amphibole exposure and SAID.

3. The strength of research coming from studies of LA-exposed populations negates weaknesses of previous studies that did not distinguish between chrysotile and amphibole asbestos.

4. There is evidence that amphibole asbestos can increase severity and progression of pleural fibrotic disease through production of autoantibodies targeting collagen-making cells.

5. Widespread environmental exposures to amphibole asbestos, some of which include mineral fibers that do not meet the strict regulatory definitions of asbestos, such as LA, fluoroedenite, and other environmental amphiboles, are increasing the risk for development of autoimmune diseases, creating a current critical public health problem.

6. Recognizing amphibole asbestos as a trigger for SAID would increase awareness and opportunities for screening, and exposure mitigation.

HIGHLIGHTS:

• Asbestos exposure presents a current and global public health risk

• Many exposure pathways exist, including urban development, wind, and recreation.

• Amphibole asbestos (AA) is associated with ANA and systemic autoimmune diseases

• Recognition of AA as a SAID trigger would advance research and lead to reduced risk

Abbreviations:

ANA

Antinuclear Autoantibodies

ANCA

Anti-Neutrophil Cytoplasmic Antibodies

ARD

Asbestos Related Disease

ATSDR

Agency for Toxic Substances and Disease Registry

CARD

Center for Asbestos Related Disease

LA, LAA

Libby Amphibole Asbestos

MCAA

Mesothelial Cell Autoantibodies

RA

Rheumatoid Arthritis

RF

Rheumatoid Factor

SAID

Systemic Autoimmune Disease

SLE

Systemic Lupus Erythematosus

US

United States

Data availability -

All data for this review are published and available in the literature.