Modeling environmental inhalant exposure in rheumatoid arthritis

Abstract

The mucosal origins hypothesis posits that environmental inhalant exposures, including cigarette smoke (CS) and crystalline silica (c-silica), trigger immune responses in the lung mucosa, an extra-articular site, which precede initiating events of rheumatoid arthritis (RA) pathogenesis in distant joints. Epidemiological data strongly associates these exposures with RA risk, especially in genetically susceptible individuals carrying HLA-DRB1 alleles, and with the production of autoantibodies such as anti-citrullinated peptide antibodies (ACPA) and rheumatoid factor (RF). However, establishing causality remains challenging due to unsynchronized exposure and disease onset and the lack of suitable animal models to study early disease events. This review synthesizes evidence linking inhalant exposures to RA, focusing on CS and c-silica, and evaluates experimental animal models used to investigate disease initiation and progression in the context of inhalant exposures. While models like collagen-induced arthritis (CIA) replicate joint pathology, they often fail to capture the lung-joint axis and gene-environment interactions critical for RA onset. We highlight the need for refined models with genetic susceptibility to subclinical autoimmunity to better mimic human RA, emphasizing the importance of standardized exposure protocols to address variability in outcomes. These advancements are crucial for elucidating mechanisms of inhalant exposure-induced RA and developing preventive strategies.

Background

Rheumatoid arthritis (RA) is a systemic autoimmune disease affecting 0.5–1% of the population, marked by symmetric polyarticular arthritis in the small joints of the hands and feet (Firestein 2003; Deane et al. 2017; Smolen et al. 2018). Affected joints present with immune cell infiltration into synovial tissues and progressive cartilage damage. Although the etiology of RA is complex, it is now understood as a multifactorial condition resulting from the interplay of genetic factors—most notably HLA-DRB1 alleles—and environmental exposures (Deane et al. 2017). The complexity of RA is undercut by the diversity of symptoms, presentations, comorbidities, and extra-articular manifestations. This interplay is further emphasized by the mucosal origins hypothesis (Holers et al. 2018), which proposes that inflammation at mucosal surfaces, such as those in the lungs and gut, play a central role in the initiation of RA development (Holers et al. 2018). Inhaled agents, such as cigarette smoke (CS) and crystalline silica (c-silica), are thought to trigger immune responses in the lung mucosa, setting the stage for disease onset (McDermott and Sparks 2023).

Evidence for this hypothesis comes from observations of local immune activation in the lung and the early appearance of autoantibodies in bronchoalveolar lavage fluid (BALF) and sputum, such as anti-citrullinated peptide antibodies (ACPA) and rheumatoid factor (RF), which can be detected years before clinical symptoms appear (Deane et al. 2010; Reynisdottir et al. 2016; Sparks and Karlson 2016). These autoantibodies are considered biomarkers of disease for their role in distinguishing seropositive from seronegative patients (Scherer et al. 2020), and are important prodromes for RA diagnosis. Strong associations with several pulmonary exposures, particularly CS (Willis et al. 2013; Demoruelle et al. 2018) and occupational exposure to c-silica (Mehri et al. 2020), provide additional evidence to support the lung and respiratory mucosa as the initiation site of autoreactivity in the pathogenesis of RA (Sparks and Karlson 2016; Moore et al. 2024). Moreover, the association of identified pulmonary exposures with heightened disease risk echo the clinical recognition that lung diseases are over-represented in RA patients (Demoruelle et al. 2014; Kadura and Raghu 2021; Akiyama and Kaneko 2022). Pulmonary complications, encompassing interstitial lung diseases (ILD), chronic obstructive pulmonary disease (COPD) (Ford et al. 2024), and pleural disorders, have long been recognized as potential consequences of RA (Stack and Grant 1965), and respiratory-related deaths are the single most overrepresented cause of premature mortality among RA patients (Hyldgaard et al. 2017; 2019; Juge et al. 2023). More recently, ILD has also been described before RA onset (Bendstrup et al. 2019; McDermott et al. 2021). Additionally, the generation of characteristic RA autoantibodies in the respiratory mucosa of individuals with chronic lung disease, even in the absence of joint disease, has been observed (Fischer et al. 2012). Thus, these clinical findings, in addition to the extensive epidemiological data associating pulmonary exposures with RA, amount to a growing body of evidence suggesting that the pathogenic precursors of RA originate outside the confines of the joint tissue, as exemplified by pulmonary inflammation following exposure to CS and c-silica.

Despite these strong associations, proving a causal relationship remains challenging. Epidemiological studies are limited by the fact that exposure and disease onset are often not clearly synchronized, and behavioral, occupational, and environmental risk factors are rarely observed in isolation. In addition, a comprehensive understanding of the biological mechanisms that link pulmonary exposures, mucosal inflammation, and subsequent development of RA has yet to be achieved (Pollard et al. 2021). This knowledge gap highlights the need for suitable experimental systems, as the lack of appropriate animal models to study disease initiation has hindered our understanding of the earliest pathogenic events. Existing models of RA, such as the collagen-induced arthritis (CIA) model or the K/BxN model, commonly used to study therapeutics, circumvent early articular events and are therefore ill-suited for studying initiation events (Zhao et al. 2022).

This is a scoping review that consists of a narrative section summarizing the existing evidence on inhalant exposures linked to the development of RA, and a systematic section that evaluates the experimental models used to study mechanisms of disease initiation and development in relation to existing epidemiological data that associates exposures with disease. Particular attention is given to the ability of animal models to reproduce human disease and support a mechanism that links inhalant exposures to disease onset and progression.

Inhalant exposures associated with rheumatoid arthritis biomarkers and disease

While many inhalant exposures have been proposed as potential triggers for RA, establishing a conclusive association has been challenged by the inadequacy of evidence. Consequently, only c-silica (Parks et al. 1999; Barnes et al. 2019) and CS (Silman et al. 1996; Di Giuseppe et al. 2014; Sparks and Karlson 2016) have been confidently associated with the development of RA (Miller et al. 2012). Although the evidence for other pulmonary exposures remains limited, inhalants such as air pollutants are of growing importance for RA pathogenesis (McDermott and Sparks 2023; Bade et al. 2024).

Cigarette smoke

CS is a heterogeneous mixture of solids—including particulate substances such as tar — and gases — including N₂ and CO₂ (Benner et al. 1989; United States Public Health Service 2010). Although CS is a risk factor for both seropositive and seronegative RA, there is a stronger link to RF-positive and ACPA-positive RA (Silman et al. 1996; Di Giuseppe et al. 2014; Sokolove et al. 2016; Hedström et al. 2018). A strong gene-environment interaction between HLA-DR shared epitope (SE) genes and CS has been shown, particularly in ACPA-positive RA (Mattey et al. 2002; Padyukov et al. 2004; Karlson et al. 2010). While both the SE and CS have been shown to confer risk independently, the combination increases risk beyond what can be accounted for by each risk factor. One study found individuals with a long-term history of smoking that were homozygous for the SE to be at a 21-fold increased relative risk for developing ACPA-positive RA (Klareskog et al. 2006). A similar pattern has been observed for the association of CS and SE genes with RF, as one study reported that SE carriers with a history of smoking were 3.7 times more likely to produce RF (Mattey et al. 2002). Not only is there an association between CS and serum autoantibodies in the context of disease, CS has also been associated with RF and ACPA in the absence of clinical disease. A large study found current and past smokers to be twice as likely to be RF-positive when compared to nonsmokers, and within those who had received some amount of CS exposure, the amount, duration, and current smoking status corresponded to the RF titer (Tuomi et al. 1990; Mattey et al. 2002). In a cohort of patients with confirmed serum ACPA and lung disease but without RA, nearly two-thirds had received some amount of CS exposure prior to the study (Fischer et al. 2012). These findings underscore the significant role of CS in promoting RA-associated autoantibody production, both in individuals with and without clinical disease.

Crystalline silica

C-silica, the most abundant component of the earth’s crust, present in rock, soil, and sand (Leung et al. 2012), is aerosolized as a fine, respirable dust during occupational activities such as mining, tunneling, sandblasting, and artificial stone fabrication (Leung et al. 2012). Deposition of these particles in the lung causes severe lung inflammation and can lead to the pneumoconiosis silicosis, characterized by fibrotic nodules (Leung et al. 2012). Since Caplan et al. (Caplan 1953) first noted RA symptoms in silica-exposed workers, numerous reports and studies have associated c-silica exposure and silicosis with autoimmune diseases, including RA, but also systemic lupus erythematosus (SLE), and systemic sclerosis (SSc) (Miller et al. 2012; Janssen et al. 2025). While these studies only include men, due to the nature of occupations associated with c-silica exposure, RA outbreaks have also been reported in women exposed to silica-based scouring powders (Ronsmans and Blanc 2023). Epidemiological data and several meta-analyses indicate a threefold increased RA risk with c-silica exposure, independent of smoking (Mehri et al. 2020; Stolt et al. 2005). Like CS, c-silica exposure poses a particularly high risk for ACPA-positive RA (Stolt et al. 2010). For the association between c-silica exposure and RF, findings are inconsistent (Doll et al. 1981; Aminian et al. 2009; Zaghi et al. 2010). Although c-silica-RA associations described were independent of smoking status, there appears to be an interaction between CS and c-silica exposure. Current smokers with previousc-silica exposure have a 7.4-fold greater risk for developing ACPA-positive RA than their nonsmoking counterparts without c-silica exposure, and 60% of this risk has been attributed to the interaction of c-silica with CS in the lung (Stolt et al. 2010). These findings highlight c-silica as a significant environmental risk factor for RA, with a particularly strong association with ACPA-positive RA and potential interaction with CS.

Other inhalant exposures associated with RA and RA biomarkers

Existing data on the relationship between various other inhalant exposures and RA is inconclusive. Tang et al. investigated the effects of 32 occupational inhalable agents, including silica, asbestos, and engine exhaust, and found all agents to be associated with an increased risk for ACPA-positive RA in a Swedish dataset of RA patients and matched controls (Tang et al. 2023). Some studies suggest a correlation between multiple air pollutants and RA incidence (Zhang et al. 2023) (reviewed in (Bade et al. 2024)), while others, like Gan et al. (Gan et al. 2013), found no link between ambient particulate matter and RA-related autoantibodies in healthy individuals. Asbestos, a group of naturally occurring silicate minerals, is associated with autoimmunity and certain autoimmune diseases (Sporn 2011), but evidence is inconsistent across the six types of asbestos fibers (Bunderson-Schelvan et al. 2011; Pfau et al. 2014). Occupational asbestos exposure often coincides with c-silica exposure, and areas with high asbestos levels, such as Libby, Montana (Marchand et al. 2012), and Wittenoom, Western Australia (Reid et al. 2018), report increased RA prevalence. A case-controlled study of women exposed to textile dust found a notable gene–environment interaction between HLA-DRB1 SE and textile dust exposure, conferring a specifically high risk for ACPA-positive RA (Too et al. 2016). Limited evidence also links organic dust exposure to RA, including military-related burn pits, waste disposal, and chemical pesticides (Parks et al. 2011; 2016; Ebel et al. 2021). Overall, while several inhalant exposures have been associated with RA, inconsistent findings highlight the need for further research to clarify their individual and combined contribution to RA development.

Mechanisms of inhalant exposure-induced autoimmunity and rheumatoid arthritis: observations and hypotheses

The mechanisms that link inhalant exposures, such as CS and c-silica to the development of RA have not been fully defined. Both xenobiotics enter the lungs and respiratory mucosa through inhalation, yet they differ in composition and particle behavior. Size is a defining property; particles smaller than 4 μm can penetrate deep into the alveolar regions, while larger particles tend to deposit in the upper airways. C-silica often forms respirable dust that reaches the alveoli, whereas CS consists of both respirable fractions and larger particles that affect the upper respiratory tract. Once deposited in the lung or respiratory tract, CS and c-silica initiate a cascade of cellular and immune responses (Figure 1).

Figure 1.

Schematic overview of proposed hypotheses that connect inhalant exposures (crystalline silica, c-silica and cigarette smoke, CS) to the development of rheumatoid arthritis (RA). (1) The current schematic overview aims to show the current hypotheses of how an inhalant exposure in the lung could induce RA in distant joints, focused on evidence from CS and c-silica. (2) Inhaled CS and c-silica enter the respiratory tract and deposit according to particle size–larger CS particles in the upper airways, smaller CS and c-silica particles in the alveolar regions. (3) Once deposited, particulate matter induces oxidative stress and the release of damage-associated molecular patterns (DAMPs), activating TLR4 on alveolar macrophages and epithelial cells, triggering release of pro-inflammatory cytokines (IL-1β, IL-6, IL-8, TNF-α). Particles are phagocytosed by alveolar macrophages, leading to lysosomal rupture, activation of the NLRP3 inflammasome, and release of IL-1β and IL-18. (4) Alongside the inflammatory response, stress-related events such as NETosis and the release of PAD enzymes occur, resulting in posttranslational modifications of self-proteins–particularly citrullination–and the generation of neoantigens. (5) Chronic inflammation drives fibrosis (not shown), and formation of inducible bronchus-associated lymphoid tissue (iBALT) containing B cells, T cells, plasma cells, dendritic cells, follicular dendritic cells, and high endothelial venules (HEVs). Within iBALT, antigen-presenting cells (APCs) present self-antigens, including citrullinated self-antigens (generated by PAD2/PAD4 activity during inflammation) via HLA-DR shared epitope alleles to CD4⁺ T cells, leading to activation of autoreactive B cells. (6) These autoreactive B cells differentiate into plasma cells that produce autoantibodies, including anti-citrullinated protein antibodies (ACPA), rheumatoid factor (RF), and antinuclear antibodies (ANA). (7) autoantibodies enter systemic circulation, where they contribute to RA pathogenesis. ACPA bind to citrullinated proteins such as vimentin on osteoclasts, promoting bone resorption. RF forms immune complexes that activate the complement cascade. (8) These events result in synovial inflammation, pannus formation and bone and cartilage erosion.

Tissue damage, inflammation, and fibrosis

The initiating events in the lung upon CS or c-silica exposure have been discussed extensively in multiple review papers (Pollard et al. 2021; Cha et al. 2023; Kawasaki 2025). In short, once deposited in the lung, these particles trigger a range of cellular responses leading to oxidative stress and cell damage, resulting in the release of reactive oxygen species and damage-associated molecular patterns (DAMPs), including the alarmin IL-1α (Carta et al. 2009; Rabolli et al. 2014; Caliri et al. 2021), that activate innate immune receptors like TLRs (Stämpfli and Anderson 2009). This activation triggers a cascade of pro-inflammatory cytokines, including IL-1β, IL-6, IL-8, and TNF-α, which contribute to inflammation and tissue remodeling (Mio et al. 1997). Chronic CS exposure is also associated with fibrosis, as seen in conditions like COPD and interstitial pulmonary fibrosis (Baumgartner et al. 1997; Morse and Rosas 2014), where dysregulated repair processes result in excessive fibroblast activity and extracellular matrix deposition (Chapman 2011). Similarly, c-silica particles are engulfed by alveolar macrophages through scavenger receptors such as SRAI/II (Beamer and Holian 2005) and MARCO (Palecanda and Kobzik 2001; Hamilton et al. 2006; Thakur et al. 2009). However, the particles resist degradation within lysosomes, causing membrane destabilization and the release of lysosomal enzymes (Pavan et al. 2022; 2022). This damage activates the NLRP3 inflammasome (Cassel et al. 2008), leading to the maturation and release of IL-1β and IL-18, which further promote inflammation (Thakur et al. 2008). Over time, repeated c-silica exposure can result in chronic lung injury, fibrosis, and the formation of granulomas characteristic of silicosis (Castranova and Vallyathan 2000; Leung et al. 2012; Mossman and Glenn 2013). Chronic inflammation from c-silica (Bates et al. 2015; Chauhan et al. 2021) and CS (Escolar Castellón et al. 1992; Richmond et al. 1993; D’hulst et al. 2005; van der Strate et al. 2006; Kuroda et al. 2016) may also lead to the development of ectopic lymphoid structures, or even more organized inducible bronchus-associated lymphoid tissue (iBALT), potentially providing a niche for autoreactive B and T cells (Sato et al. 1996; Hogg et al. 2004; Demoruelle et al. 2014; Corsiero et al. 2016; Bombardieri et al. 2017). The formation of follicle-containing lymphoid aggregates in the lung is thought to contribute to the production of autoantibodies that are implicated in the pathogenesis of RA (Wallace et al. 1996; Rangel-Moreno et al. 2006; Manzo and Pitzalis 2007; Duarte et al. 2019). These findings support the idea that inhaled exposures such as CS and c-silica may contribute to RA pathogenesis through a mechanism that goes beyond generic immune activation. Rather than acting solely as adjuvants that amplify inflammation, these exposures appear to initiate a compartment-specific immune response in the lung. The development of organized lymphoid structures, such as iBALT, may create a permissive environment for the activation and expansion of autoreactive B and T cells, as well as for local autoantibody production. This process aligns with the mucosal origins hypothesis, which posits that lung-specific immune events are a critical early step in breaking tolerance and triggering systemic autoimmunity. The selective association of these inhalant exposures with systemic autoimmune diseases such as seropositive RA, rather than with a broad range of inflammatory disorders, further supports a model in which mucosal autoimmunity is central to disease initiation. The mucosal origins hypothesis proposes that chronic inflammation weakens the mucosal barrier, initiating the transition from a local to systemic autoantibody response (Holers et al. 2018).

Autoantibodies

CS and occupational inhalants, including c-silica, are established risk factors for the development of autoantibody-positive RA (Stolt et al. 2010; Di Giuseppe et al. 2014; Sokolove et al. 2016; Hedström et al. 2018; Kronzer and Sparks 2023; Tang et al. 2023). Therefore, the main hypotheses that connect inhalant exposures to the development of RA are based on their potential to induce the production of autoantibodies. iBALT has been identified in the lungs of RA patients and is implicated in local RA pathogenesis through tissue damage and autoantibody production (Rangel-Moreno et al. 2006). Moreover, characteristic RA autoantibodies have been found in the sputum of both seronegative and seropositive patients in preclinical and early stages of RA (Willis et al. 2013; Demoruelle et al. 2017; 2018). The identification of IgA autoantibodies in the serum of RA patients further evidences the lung mucosa as a site of local autoantibody production (Sieghart et al. 2018; Sokolova et al. 2022; Heutz et al. 2024).

Citrullination of self-antigens and anti-citrullinated protein antibodies

Inflammation and oxidative stress can induce post-translational modifications, such as citrullination, homocitrullination, and carbamylation. Citrullination is mediated by peptidyl arginine deiminase (PAD) enzymes, that convert positively charged arginine residues to neutral citrulline, potentially making them immunogenic (Yu and Proost 2022). This process produces the target antigens of ACPA, which are strongly linked to the development and severity of RA (Rantapää-Dahlqvist et al. 2003; Willemze et al. 2012; Khidir et al. 2024). ACPA can be detected years before the onset of clinically apparent arthritis and are associated with experimental arthritis in mouse models (Lundberg et al. 2005; Kuhn et al. 2006). Despite the well-established association of ACPA with RA, their pathogenic role has not been confirmed (He et al. 2024; van der Woude and Toes 2024).

CS promotes protein citrullination in the airways and lungs (Klareskog et al. 2006). The significance of CS in ACPA development is underscored by the strong association between CS, ACPA, and HLA-DR SE alleles. Citrullination increases protein immunogenicity and enhances the binding of modified peptides to SE-containing HLA-DR receptors on the surface of antigen-presenting cells (Hill et al. 2003). Modified peptides are then presented to T cells, eliciting an immune response to citrullinated proteins in SE gene carriers (Hill et al. 2003). Similarly, c-silica exposure has a stronger association with ACPA-positive RA than ACPA-negative RA (Stolt et al. 2010). Silica nanoparticles induce citrullination in vitro (Mohamed et al. 2012), and c-silica exposure induces ACPA in BXD2 (Janssen et al. 2025) and Collaborative Cross mice (Janssen et al. 2025). Combined exposure to CS and c-silica compounds the risk of developing ACPA-positive RA beyond what can be accounted for by each individual exposure, indicating an interaction within the lungs (Stolt et al. 2010). Elevated levels of PAD enzymes, specifically PAD2 and PAD4, are a response to both CS and c-silica. Neutrophils, recruited to the lung upon c-silica (Mohamed et al. 2012) and CS (Vassallo et al. 2014) exposure, are major sources of PAD enzymes, released during NETosis (Thiam et al. 2020). The association between inhalant exposures such as CS and c-silica and protein citrullination is further supported by the link between chronic airway inflammation and ACPA development, even in the absence of exposure. For example, there is a strong association between asthma and elevated circulating ACPA levels prior to RA diagnosis, independent of smoking status (Zaccardelli et al. 2019). This suggests chronic mucosal airway inflammation is central to the development of ACPA. Additionally, smoking is associated with stronger ACPA responses in patients with chronic airway diseases like COPD (Sparks and Karlson 2016). Thus, chronic inflammatory events induced by inhalant exposures are likely precursors of excessive protein citrullination in the lungs. Although the pathogenic role of ACPA in RA is still debated, mechanistic hypotheses have been proposed. ACPA may contribute to joint inflammation by forming immune complexes (IC) and engaging Fc receptors (van Venrooij et al. 2011). Citrullination could also contribute to RA pathogenesis through mechanisms involving ACPA binding to their citrullinated targets. For instance, ACPA binding to citrullinated vimentin on osteoclast surfaces can activate these cells, promoting bone resorption (Harre et al. 2012). In contrast, studies of other inhalant exposures have yet to present clear and consistent findings that mirror the hypothesis developed for CS-induced RA.

Rheumatoid factor

RF represent a group of autoantibodies, mainly IgM (Mattey et al. 2002), against the Fc portion of immunoglobulin G, that are common in RA. RF can be detected in serum several years prior to the onset of disease symptoms, and is used as a biomarker of early disease and a diagnostic tool (Rantapää-Dahlqvist et al. 2003; Nielen et al. 2004; Tan and Smolen 2016). The pathogenic role of RF in RA is subject to continued investigation, but studies suggest their role in targeting existing IC, leading to the formation of RF-IC-complexes that stimulate the complement cascade and eventually produce inflammation (Tan and Smolen 2016). Moreover, sustained and elevated titers of circulating RF correspond to heightened disease risk and severity (Delpuente et al. 1988; van Zeben et al. 1992). Despite the association of CS, and in a lesser extent c-silica exposure, with the development of RF, existing data is unable to confirm a causal relationship between inhalant exposure, the development of RF, and the subsequent development of RA as a mechanism for inhalant exposure-mediated RA.

Unlike ACPA, the association between CS and RF is independent of HLA-DR SE alleles, suggesting that distinct mechanisms control the production of these two autoantibodies (Ishikawa et al. 2019). There is limited evidence to suggest that RF production originates in the lung. Plasma cells that bind IgG have been identified in iBALT in the the lungs of RA patients (Rangel-Moreno et al. 2006), and RF has been detected in the sputum of at-risk and early RA patients (Willis et al. 2013). We propose that the recruitment of excessive immunoglobulin to the lung mucosa following exposure to CS and c-silica initiates the production of RF by creating a reservoir of antigenic targets.

Other autoantibodies

Antinuclear antibodies (ANA), a hallmark of systemic autoimmune diseases, have also been observed in RA (Aitcheson et al. 1980). In the absence of disease, nuclear components are sequestered. Only in conditions of chronic inflammation, like pulmonary inflammation resulting from inhalant exposures, do nuclear components become available in high concentrations. This can overwhelm the immune system’s scavenging capacity, triggering an immune reaction in susceptible individuals and resulting in the production of ANA (Aitcheson et al. 1980; Arnson et al. 2010). However, the contribution of ANA to the pathogenesis of RA has yet to be confirmed. Although ANA tests are often used to screen for rheumatic disease, ANA is a relatively non-specific disease marker. ANA refer to a class of autoantibodies directed against nuclear components and are detected in a significant portion of the general population. Thresholds for what is considered a clinically relevant titer vary. Higher titers confer greater risk for connective tissue disease. CS may cause the release of intracellular antigens through tissue hypoxia or toxin-induced cellular necrosis (Arnson et al. 2010). ANA have also been observed in patients with COPD, which is the most common pathological outcome associated with CS (Núñez et al. 2011). Although c-silica’s role in the development of RA is unclear apart from the limited evidence for anti-citrulline-immunity (Mohamed et al. 2012), c-silica exposure has been investigated extensively for its association with the development of SLE in animal models (Bates et al. 2015; Janssen et al. 2022). Evidence suggests that c-silica’s role in the pathogenesis of SLE is mediated by the formation of ANA, as c-silica exposure has been associated with the development of ANA in both human cohorts (Doll et al. 1981) and animal models (Bates et al. 2015; Mayeux et al. 2018).

Alternate hypotheses

Alternate hypotheses offered to explain development of RA suggest mechanisms such as complement activation, direct effects on synovial fibroblasts, and adjuvant-like properties of inhaled substances may also drive RA pathogenesis (Gravallese and Firestein 2023). For example, components of CS can directly stimulate synoviocytes, altering gene expression (Ospelt et al. 2014), while c-silica can directly promote lymphocyte proliferation (Eleftheriadis et al. 2019). Moreover, the interaction of CS toxins with DNA and the activation of the aryl hydrocarbon receptor (AhR) are emerging areas of interest in understanding the link between inhalant exposures and RA (Fu et al. 2018). Overall, inhalant exposures appear to initiate a cascade of cellular responses leading to lung injury, inflammation, and autoantibody production that appear to be closely linked to the development and severity of RA. Further research is essential to fully unravel these complex pathways and their contributions to disease onset and progression.

Animal models of inhalant exposure-mediated autoimmune arthritis

Experimental animal models have become an essential approach for testing hypotheses arguing for a linkage between environmental exposure and the development of autoimmune diseases such as RA (Miller et al. 2012; Pollard et al. 2018). We used a literature search strategy with inclusion/exclusion criteria (Supplementary File 1) to identify animal models that have been used to examine the role of inhalant exposures in the development of RA-like inflammatory arthritis. We identified a total of 19 studies that describe CS ([ 2) or c-silica (Table 3) exposures to model human RA as well as other inhalant exposures with more limited relevance to RA, such as diesel exhaust and particulate matter (PM) (Table 4) and organic dust (Table 5).

Table 3.

Overview of mouse model studies investigating the effect of crystalline silica and nanoparticles on arthritis development and severity.

Exposure	Treatment schedule	Strain	Model	Effect on arthritis	Mechanistic findings	Ref.
c-silica	30 μL silica solution (40 mg/mL) inhaled intranasally 1 week before CII immunization	Male F1 of female DBA/1J and male B10.q cross	CIA	No augmentation of arthritis	• Increased IgG ACPA • No change in PAD2 or PAD4 expression in lung tissue • No change in anti-CII or anti-MCV antibodies	(Engelmann and Müller-Hilke 2017)
FeNPs	Treated by nebulization 5 hr/day, 5 days/week with aqueous suspension of 0.4 mg/mL FeNPs	DBA/1J	Mild CIA	Accelerated and augmented arthritis	• Reduced anti-CII IgG (not statistically significant) • Increased TNF-α • Increased 4-HNE modified proteins (oxidative stress marker)	(Nowak et al. 2022)
SiNPs	Treated by nebulization 5 hr/day, 5 days/week with aqueous suspension of 0.6 mg/mL SiNPs			Accelerated and augmented arthritis	• Reduced anti-CII IgG (not statistically significant) • Increased TNF-α • Increased 4-HNE modified proteins	(Nowak et al. 2022)

ACPA, anti-citrullinated peptide antibodies; c-silica, crystalline silica; CIA, collagen-induced arthritis; FeNPs, ferric nanoparticles; SiNPs, silica nanoparticles. All comparisons describe a significant difference between the exposure group and corresponding control unless otherwise stated. All antibodies and inflammatory markers were measured in the serum unless otherwise stated.

Table 4.

Overview of mouse model studies investigating the effect of particulate matter on arthritis development and severity.

Exposure	Treatment schedule	Strain	Model	Effect on arthritis	Mechanistic findings	Ref.
Crude PM	Treated by nebulization 5 hr/day, 5 days/week with aqueous suspension of 1.2 mg/mL urban particulate matter (PM) with mean particle diameter 5.85 μm	DBA/1J	Mild CIA	Accelerated and augmented arthritis	• Increased anti-CII IgG1 and ACPA (not statistically significant) • Increased IL-6 (not statistically significant)	(Nowak et al. 2022)
PMΔC	Treated by nebulization 5 hr/day, 5 days/week with aqueous suspension of 1.3 mg/mL plasma-treated PM with 2% carbon content			Accelerated and augmented arthritis	• Increased anti-CII IgG1 and ACPA (not statistically significant) • Increased IL-6 (not statistically significant)	(Nowak et al. 2022)
DEP	DEP with mean diameter 0.4 μm (0.1, 0.3, and 1 mg/mL) suspended in 50 μL PBS administered intranasally every 2 days for 20 days beginning on the day of immunization or daily from days 31–45 after immunization with CII	DBA/1J	CIA	Dose-dependent incidence and augmentation of arthritis	• Increased anti-CII IgG and IgG2a [isotype] • Increased proliferation of spleen cells in response to CII • Increased secretion of IFN-γ, IL-2, and IL-4 from splenocytes	(Yoshino and Sagai 1999)
DEP DIC-DEP	DEP and extracts (5 mg/mL) suspended in 50 μL PBS administered intranasally every 2 days for 20 days beginning on the day of immunization with CII	DBA/1J	CIA	Accelerated arthritis; augmented incidence and severity of arthritis	• Increased anti-CII IgG, IgG2a, and IgG1 antibodies • Increased IFN-γ and IL-4 secretions from splenocytes • Increased inflammatory cells in joint • Increased proliferation of splenocytes in response to CII	(Yu and Proost 2022)
UNE-DEP					Increased anti-CII IgG and IgG2a antibodies • Increased IFN-γ secretions from splenocytes • Increased inflammatory cells in joint • Increased proliferation of spleen cells in response to CII
AMM-DEP				No augmentation of arthritis	• Increased anti-CII IgG1 • Increased IL-4 secretions from splenocytes • Increased proliferation of spleen cells in response to CII
HEX-DEP					NA
BEN-DEP
MET-DEP

ACPA, anti-citrullinated peptide antibodies; AMM-DEP, DEP extracted with ammonia; BEN-DEP, DEP extracted with benzene; CIA, collagen-induced arthritis; DEP, diesel exhaust particles; DIC-DEP, DEP extracted with dichloromethane; HEX-DEP, DEP extracted with hexane; MET-DEP, DEP extracted with methanol; PM, particulate matter; PMΔC, particulate matter with reduced organic content; UNE-DEP, unextracted residues. All comparisons describe a significant difference between the exposure group and corresponding control unless otherwise stated. All antibodies and inflammatory markers were measured in the serum unless otherwise stated.

Table 5.

Overview of mouse model studies investigating the effect of organics on arthritis development and severity.● Increased TNF-α, IL-6, CXCL1, and CXCL2 in BALF● Increased fibronectin and IL-33 in lung tissue● Increased IgG albumin-MAA, CII-MAA, and vimentin-MAA, and anti-MAA in BALF● ACPA not detected(Yoshino

Exposure	Treatment schedule	Strain	Model	Effect on arthritis	Mechanistic findings	Ref.
ODE	Daily intranasal exposure to 50 μL 12.5% ODE in PBS 5 days/week for 5 weeks	Male DBA/1J	CIA	Augmented arthritis (not statistically significant)	• Increased TNF-α, IL-6, CXCL1, and CXCL2 in BALF • Increased fibronectin and IL-33 in lung tissue • Increased pentraxin-2, IgG, IgM, and IgA • Increased IgG ACPA, but not anti-CII antibodies	(Poole et al. 2019)
ODE	Daily treatment with 50 uL intranasal 12.5% ODE in saline for 4–5 weeks	Male DR4-IE	CIA	Augmented minimal arthritis	• Increased TNF-α, IL-6, CXCL1, and CXCL2 in BALF • Increased fibronectin and IL-33 in lung tissue • Increased IgG albumin-MAA, CII-MAA, and vimentin-MAA, and anti-MAA in BALF • ACPA not detected	(Poole et al. 2022)
LPS	Daily intranasal exposure to 100 ng LPS in 50 μL PBS 5 days/week for 5 weeks	Male DBA/1J	CIA	Augmented arthritis (not statistically significant)	• Increased TNF-α and IL-6, but not CXCL1 or CXCL2 in BALF • Increased fibronectin and IL-33 in lung • Increased pentraxin-2, ACPA, anti-MAA, and anti-CIT/MAA IgG	(Mikuls et al. 2021)

ACPA, anti-citrullinated peptide antibodies; BALF, bronchoalveolar lavage fluid; CIA, collagen-induced arthritis; ODE, organic dust extract; LPS, lipopolysaccharide; MAA, malondialdehyde. All comparisons describe a significant difference between the exposure group and corresponding control unless otherwise stated. All antibodies and inflammatory markers were measured in the serum unless otherwise stated.

Rodent models of rheumatoid arthritis

The majority of the studies we identified used established RA animal models (Table 1). The degree to which the experimental arthritis models resemble RA pathogenesis and pathology varies. This is due, in part, to the choice of rodent strains and their genetic background. Despite these differences, many of the established experimental models share pathological and immunological features with human RA, suggesting shared mechanisms between the experimental models and human RA. The differences and similarities in pathology and immunology between animal models of RA and human RA has been extensively reviewed elsewhere (Kannan et al. 2005; Caplazi et al. 2015).

Table 1.

Overview of rodent models used to study rheumatoid arthritis.

Model	Rodent strain	Method	Mechanism	Histological features	Ref.
Collagen-induced arthritis (CIA)	DBA/1; DBA/1xB10.q; Wistar rats	Intradermal injection of native type II collagen (CII) emulsified with Freund’s adjuvant	• Anti-CII antibodies that target endogenous cartilage • Susceptibility to anti-collagen immunity and arthritis controlled by MHC • CII is free of adjuvanticity • Cannot be amplified with mycobacteria	• Synovial hyperplasia • Inflammatory cell infiltration of the synovium joint and cartilage erosion • Periostitis (not present in RA) • No vasculitis	(Brand et al. 2007) (Stuart et al. 1984) (Trentham et al. 1977)
Antigen-induced arthritis (AIA)	C57BL/6	Preimmunization with an antigen followed by direct injection of the antigen into the intraarticular space	• Localized immune response resulting in a unilateral T-cell driven arthritis • Susceptibility determined by T cell ability to respond to antigen, while chronicity requires sustained antigen retention • Th1 anti-antigen response can be boosted or rechallenged by the concomitant or subsequent addition of an adjuvant	• Severe synovitis • Severe articular cartilage erosion • Immune infiltrates include a large population of macrophages and lymphoid cells do not tend to form follicle-like clusters	(van den Berg et al. 2007) (Brackertz et al. 1977)
DR4β/Eα	DR4β(NT)/Eα DRB1*0401.Aβ° C57BL/6 background	Introduction of transgenes onto the B10.RFB3 strain results in a chimeric HLA-DR4 molecule: human DR4β chains paired with endogenous Eα chains	• Does not develop spontaneous arthritis	NA	(Pan et al. 1998)
DR4.AEo	EαDRB1*0401.Ae°	DR4.Aβ° crossed with Ae° results in a completely human HLA-DR4 molecule	• Does not develop spontaneous arthritis	NA	(Taneja et al. 2007)
DQ8	DQA1*0301/DQB1*0302.Ae°	DQ8.Aβ° crossed with Ae° results in a completely human HLA-DQ8 molecule	• Does not develop spontaneous arthritis	NA	(Pan et al. 1998)
TNF ΔARE	TNFΔARE/+ C57BL/6 background	Heterozygous for the TNF ΔAU-rich element	• Overexpression of TNF-α • Spontaneous development of progressive arthritis by 8 weeks of age • RF (IgG and IgM)	• Immune cell infiltration of the synovioentheseal complex in the tarsal joints • Synovial hyperplasia • Cartilage destruction and subchondral bone erosion	(Kontoyiannis et al. 1999) (Denis et al. 2017)
SKG	SKG derived from BALB/c	Point mutation in the ZAP-70 gene	• Hypomorphic • Reduced T-cell receptor signaling causes positive selection of autoimmune T cells • Arthritis develops spontaneously in conventional environments but requires induction in pathogen-free conditions • Autoantibodies: anti-CII antibodies, rheumatoid factor, antibodies to heat shock protein (HSP)-70	• Severe synovitis with subsynovial infiltration of predominantly CD4+ T cells • Villus proliferation of synoviocytes • Pannus formation with cartilage erosion	(Sakaguchi et al. 2003)

Collagen-induced arthritis and antigen-induced arthritis

The collagen-induced arthritis (CIA) model is the most widely used experimental model of autoimmune arthritis (Brand et al. 2007), induced by injecting type II collagen (CII) with adjuvant in genetically susceptible mouse strains like DBA/1 (Trentham et al. 1977; Stuart et al. 1984). The passive transfer of immunoglobulins from a CIA donor to a naiverecipient can induce arthritis in mice that are not susceptible to conventional CIA (Stuart and Dixon 1983), demonstrating that antibodies drive arthritis in this model (Terato et al. 1992). CIA shares many histologic features with RA but differs in the presence of periostitis and the absence of synovial vasculitis. Furthermore, the characterization of CII as an autoantigen in RA has been debated (Stuart et al. 1984). The antigen-induced arthritis (AIA) model involves preimmunization with an antigen followed by the injection of antigen into the articular space (van den Berg et al. 2007), inducing a localized, T cell-driven arthritis with severe synovitis and cartilage erosion. Unlike RA, AIA shows abundant macrophages without follicle-like lymphoid clusters, and its chronicity depends on sustained antigen retention (Brackertz et al. 1977).

Transgenic and genetically mutated

Humanized mice carrying HLA transgenes were developed to study the HLA subtypes in the context of autoimmune diseases like RA (reviewed (Schinnerling et al. 2019)). Specifically, HLA-DR Tg mice are used to identify peptides that bind to MHC class II molecules and study the interaction with CD4 receptors (Schinnerling et al. 2019). To mitigate low positive selection of CD4⁺ T cells and to ensure efficient interactions between MHC-II molecules and CD4 receptors, the DR4β(NT) transgene was introduced onto the B10.RFB3 strain, generating chimeric mice that express the human HLA-DR4 β chain and mouse Eα chain (DR4β/Eα) (Pan et al. 1998). To generate a mouse with entirely human MHC-II, DR4β/Eα mice were crossed with AE° mice (Madsen et al. 1999) to produceDR4.AE° mice that express DR4 molecules (EαDRB1*0401) and lack all four conventional murine chains (Taneja et al. 2007). Similarly, AE° mice were crossed with DQ8/Eα mice to generate DQA1*0301/DQB1*0302 (AEoDQ8) mice (Cheng et al. 2003).

The TNF^ΔARE mouse is heterozygous for a deletion in the AU-rich elements (ARE) of the TNF gene, leading to increased TNF mRNA stability and overproduction of TNF-α (Kontoyiannis et al. 1999). This leads to the spontaneous development of chronic inflammatory arthritis, characterized by synovial hyperplasia, immune cell infiltration, cartilage destruction, and bone erosion, closely resembling human RA (Kontoyiannis et al. 1999). This model develops other TNF-mediated inflammatory diseases, including Crohn’s-like ileitis, and therefore, it is more appropriately characterized as a model of spondyloarthropathies (Denis et al. 2017).

The SKG mouse strain has a spontaneous point mutation in the ZAP-70 gene which renders it hypomorphic and reduces T-cell receptor signaling (Sakaguchi et al. 2003; Keith et al. 2012). The mutation causes positive selection of autoreactive T cells in the thymus, leading to the infiltration of joint spaces by CD4⁺ T cells and resulting arthritis when mice are housed in conventional environments (Sakaguchi et al. 2003).

Inhalant exposures and experimental arthritis models

At present it is unclear whether inhalant exposures can induce RA ‘de novo’ in the absence of genetic susceptibility, whether inhalant exposures can trigger RA development only in the context of genetic susceptibility, or whether they can only exacerbate or accelerate preexisting disease. The following sections evaluate existing research that has attempted to address these questions in the context of different exposures through the use of established models of RA.

Cigarette smoke and its components (Table 2)

Table 2.

Overview of mouse model studies investigating the effect of cigarette smoke and components on arthritis development and severity.

Exposure	Treatment schedule	Strain	Model	Effect on arthritis	Mechanistic findings	Ref.
CS	4 cigarettes/day for 6 days/week for 16 weeks prior to CII immunization (22 weeks total CS exposure)	DBA/1	CIA	Delayed arthritis	• Reduced IgG anti-CII and ACPA • No change in IL-6	(Lindblad et al. 2009)
CSC	Intranasal administration of 10, 50 or 100 μg CSC in solution seven and one days before CII immunization	DBA/1J	CIA	Augmented arthritis (not statistically significant over CIA only)	• Increased IgG anti-CII	(Okamoto et al. 2011)
CS (unrefined)	Nose-only exposure chamber (150 & 600 μg/L) for one hour/day for 5 days/week for 4 weeks following CII immunization	DBA/1J	CIA	Accelerated arthritis by 2 weeks; dose-dependent augmentation of symptoms	• Increased ACPA • Increased citrullinated vimentin and enolase in the tarsal joint and trachea	(Kang et al. 2020)
CS	Exposed to unfiltered tobacco smoke in an inhalation chamber connected to smoking apparatus 4 times/day, 5 days/week for 2 or 4 weeks	(WT) C57BL/6 and TNF-α overexpression (TNFΔARE)		Did not aggravate arthritis	• No change in TNF-α, CXCL1, CXCL2, KC, IL-1β, Atg7, or CCR6 • Slight increase in TNF-induced CCL20	(Allais et al. 2015)
CS	Exposed to cigarette smoke using a Teague TE10 smoking system (3R4F cigarettes), for 5 hours per day, 5 days per week for up to 20 weeks	SKG (point mutation in ZAP70 rendering it hypomorphic with decreased signaling through the T-cell receptor (TCR) complex)		No induction of arthritis	NA	(Keith et al. 2012)
CS	Smoke from 2 cigarettes/45 minutes 10 times/day 5 days/week using a Teague smoking system (3R4F Kentucky research cigarettes) beginning 2 weeks before CII immunization and continued until bleomycin administration (40 days after CII)	DQ8	CIA+ bleomycin-induced lung injury	Augmented arthritis	• Bleomycin did not aggravate arthritis beyond CS alone • CS alone did not induce fibrosis in CIA mice	(Lin et al. 2021)
CS	Exposed to cigarette smoke using a Teague smoking system (3R4F Kentucky research cigarettes) on a daily basis, beginning 2 weeks before CII immunization and continued for up to 10 weeks after immunization	DR4.AE°	CIA	No augmentation of arthritis	• Increased anti-CII and RF • Reduced IL10, IL-13, and IL17	(Vassallo et al. 2014)
		DQ8		Augmented arthritis	• Increased anti-CII and RF • Increased Th17 cells • Increased IL10, IL13, IL17, and TSLP • Higher response to citrullinated CII than native
CS	Exposed to 1, 2, or 3 cigarettes per day (4.5 minutes per cigarette, 5 mice per chamber) twice on days 12 and 17 following the first immunization	Male DBA/1J	CIA	Augmented arthritis	NA	(Talbot et al. 2018)
		Male (WT) C57BL/6, AhR−/−, IL-17R−/−	AIA	Augmented arthritis in WT, but not in AhR−/− or IL-17−/−	• Increased IL-17 mRNA, AhR, and Cyp1a1 expression and Th17 cells in DLN [WT CS-AIA only] • CS-induced aggravation of arthritis is dependent on IL-17RA, AhR activation and elevation of Th17 • Direct activation on CD4+ T cells via Th17 development in vivo is sufficient for the AhR-dependent aggravation of articular inflammatory disease
CS	2 hr/day exposure in exposure machine for 7 days following immunization	Male C57BL/6	AIA	Augmented arthritis	• Increased NETs and MCP-1 in synovial fluid [over HNBT and AIA only control] • Reduced cellularity of DLN and spleen [versus AIA only control] • Reduced IL-2 and increased IL-10 secretions from splenocytes • α7AchR activation impairs splenocyte proliferation and IL-2 secretion	(Heluany et al. 2022)
HNBT vapor				No augmentation of arthritis	• Reduced cellularity of DLN and spleen [versus AIA only control] • Reduced IL-2 and increased IL-10 secretions from splenocytes • α7AchR activation impairs splenocyte proliferation and IL-2 secretion from splenocytes
HQ	Placed in exposure box with 25 ppm HQ aerosol for 1 hr/day during days 21–27 (or 1–14) after CII immunization	Male Wistar rats	CIA	Augmented arthritis	• Increased IL-17+ and AhR+ cells in synovia • Increased IL-6 in synovial fluid • AhR mediates HQ-induced ROS generation and production of IL-17 by neutrophils	(Heluany et al. 2018)
HQ	Placed in exposure box with 25 ppm HQ aerosol for 1 hr/day during days 21–27 (or 1–14) after CII immunization	Male Wistar rats	CIA	Augmented arthritis	• Increased TNF-α and ACPA • Increased neutrophils, collagen deposition, cellularity, and AhR+ cells in synovium • Increased TNF-α and IL-1β secretions from synoviocytes	(Heluany et al. 2018)
HQ	Placed in exposure box with 25 ppm HQ aerosol for 1 hr/day during days 15–21 after CII immunization	Male (WT) C57BL/6, AhR−/−, IL-17R−/−	AIA	Augmented arthritis in WT only	• Increased IL-6 and IL-17 [WT HQ-AIA only], and TNF-α [WT and IL-17−/− HQ-AIA] in femur-tibial joints • Increased H3 citrullination (indicating NET formation) in femur-tibial joints [WT HQ-AIA only]	(Heluany et al. 2021)

ACPA, anti-citrullinated peptide antibodies; AIA, antigen-induced arthritis; CIA, collagen-induced arthritis; CS, cigarette smoke; CSC, cigarette smoke condensate; DLN, draining lymph node; HNBT, heat-not-burn tobacco; HQ, hydroquinone. All comparisons describe a significant difference between the exposure group and corresponding control unless otherwise stated. All antibodies and inflammatory markers were measured in the serum unless otherwise stated.

Of the four studies examining the effects of CS exposure in a CIA model using DBA/1 mice, two found a statistically significant aggravation of arthritis (Talbot et al. 2018; Kang et al. 2020), one found an aggravation of arthritis that failed statistical significance (Okamoto et al. 2011), and one found a significant delay in the development of CIA (Lindblad et al. 2009) (Table 2). Anti-collagen immunity is a major component in the pathogenesis of CIA; however, collagen immunity is required but not sufficient for arthritis (Stuart et al. 1984). Therefore, the delayed development could be explained in part by a reduction in serum anti-CII antibodies and ACPA (Lindblad et al. 2009). Contrastingly, elevated anti-CII antibodies have been observed in the context of arthritis aggravated by CS condensate (Okamoto et al. 2011) and exposure to unrefined CS led to citrullination of the airways and lung as well as increased citrullinated proteins and ACPA in the serum in an aggravated CIA model (Kang et al. 2020).

Exposure to CS did not augment CIA incidence in male DR4.AE° mice and led to a reduction in the incidence of CIA in female DR4.AE° mice; however, CS did lead to significantly greater levels of anti-CII antibodies and augmented citrulline immunity (Vassallo et al. 2014). AEoDQ8 mice carrying the HLA-DQ8 transgene, which occurs in linkage with DR4 in humans, demonstrated increased incidence and severity of CIA along with enhanced levels of anti-CII antibodies and RF as well as augmented expression of Th17 in response to CS exposure (Vassallo et al. 2014). In both DR4.AE° and DQ8 mice, CS exposure primed citrulline immunity as indicated by a greater proliferation of cells in response to citrullinated CII over native CII. Furthermore, CS was found to augment the production of autoantibodies and increase lung compliance in DQ8 transgenic mice, suggesting the development of emphysema following CS exposure (Lin et al. 2021). However, CS was not sufficient to induce lung fibrosis without subsequent treatment with intra-tracheal bleomycin to induce lung injury.

While SKG mice are known to develop spontaneous arthritis (Sakaguchi et al. 2003), exposure to chronic CS was not sufficient to induce arthritis or significant lung injury beyond the accumulation of anthracotic macrophages, exhibiting the same changes that were observed in the WT BALB/c mice (Keith et al. 2012). C57BL/6 mice heterozygous for the TNF ΔAU-rich element (TNF^ΔARE) overexpress TNF-α, which leads to the spontaneous development of progressive arthritis by 8 weeks of age (Allais et al. 2015). After 2 and 4 weeks of CS exposure, there were no significant differences in arthritis in TNF^ΔARE or WT mice. Levels of TNF-α measured in the serum of TNF^ΔARE and WT mice showed no difference between CS and control groups, suggesting CS may not be affecting arthritis through this pathway.

Hydroquinone (HQ) is the most pro-oxidative compound in cigarette tar, but despite the considerable evidence for its carcinogenic and mutagenic properties, it has not been classified as an oral or dermal toxin to humans (Eastmond 2012). HQ has also been shown to aggravate CIA and increase ACPA in Wistar rats (Heluany et al. 2018; 2018). Heluany et al. (Heluany et al. 2018) evidenced the role of the aryl hydrocarbon receptor (AhR) in activating Th17 inflammatory pathways leading to ROS generation and IL-17 secretion from rat neutrophils. This mechanism was further supported by the AhR-dependent aggravation of articular inflammation by HQ in an AIA mouse model (Heluany et al. 2021). HQ-related effects on arthritis, including enhanced edema, hypernociception, synovial hyperplasia, cellular infiltration, and elevated citrullination in the joint space, as well as increased levels of IL-6, IL-17, and TNF-α, were absent in AhR −/− mice and significantly reduced in IL-17R−/− mice. Talbot et al. (Talbot et al. 2018) produced the same results using CS, demonstrating that aggravation of AIA by CS is associated with higher frequencies of Th17 cells in the draining lymph nodes and dependent on IL-17R. Not only were CS-related AIA effects absent in IL-17R−/− mice, but treatment with an AhR antagonist significantly inhibited the proliferation of Th17 cells, indicating that AhR activation is also required for CS-induced aggravation of articular inflammation. An additional study comparing the effects of CS and heat-not-burn tobacco (HNBT) found that CS exposure led to more severe arthritis than HNBT when administered prior to the AIA rechallenge (Heluany et al. 2022). This finding was further supported by more severe lung inflammation following CS when compared to HNBT.

Crystalline silica and silica nanoparticles (Table 3)

Of the two studies investigating the effect of silica exposure on CIA, one found silica nanoparticles to accelerate and aggravate arthritis (Nowak et al. 2022) while the other found no effect with exposure to c-silica (Engelmann and Müller-Hilke 2017) (Table 3). In F1 mice generated from a DBA/1 x B10.q cross, the severity of CIA correlated with serum anti-CII levels; however, c-silica exposure did not augment collagen immunity or arthritis even in the presence of pulmonary inflammation and silicosis (Engelmann and Müller-Hilke 2017). While c-silica did not lead to elevated levels of anti-CII, serum ACPA did increase; however, there was no effect on PAD2 or PAD4 levels. In DBA/1 mice, treatment with CII alone (without Freund’s adjuvant) led to mild CIA, and silica and ferric nanoparticles both accelerated and aggravated CIA but did not increase anti-CII antibodies (Nowak et al. 2022). Although neither silica nor ferric nanoparticles augmented a humoral response, aggravation of arthritis may be the result of inflammatory pathways, indicated by enhanced serum levels of TNF-α and signs of oxidative stress. In the same study, exposure to PM (5.85 μm mean diameter) led to a more rapid acceleration of arthritis with raised levels of anti-CII antibodies; however, serum antibodies were not statistically significant over controls.

Diesel exhaust particles and particulate matter (Table 4)

Diesel exhaust particles (DEP) and DEP extracts have also been shown to augment the incidence and severity of CIA in a dose-dependent manner in DBA/1 mice (Yoshino and Sagai 1999; Yoshino et al. 2002) (Table 4). Disease augmentation was supported by enhanced collagen immunity, indicated by elevated serum anti-CII and increased proliferation of spleen cells in response to CII.

Organic dust and lipopolysaccharide (Table 5)

Organic dust extract (ODE) aggravated arthritis in the CIA susceptible DBA/1 mouse; however, significant differences with CIA-only were limited to reductions in bone parameters (Poole et al. 2019) (Table 5). DBA/1 mice showed increased ectopic lymphoid aggregates and collagen staining in the lung following ODE exposure; however, ODE exposure was not associated with increased anti-CII antibodies in the serum. Conversely, while serum ACPA were elevated with ODE, citrullination in the lung was not evaluated. BALF indicated elevated levels of cytokines TNF-α and IL-6 as well as chemoattractants CXCL1 and CXCL2 with ODE exposure, but autoantibody levels were not measured in the BALF. A male sex bias was observed for ODE-induced lung disease and arthritis in a CIA DBA/1 model (Poole et al. 2019). Similar results were obtained from lipopolysaccharide (LPS) exposure of male DBA/1 mice in a CIA model, suggesting that LPS concentrations within ODE are the pathogenic component driving lung inflammation and corresponding arthritis and autoantibodies (Mikuls et al. 2021).

Exposure of male DR4β/Eα transgenic mice, expressing the HLA-DRB1*0401 SE, and wild type (WT) C57BL/6 to ODE led to a slight augmentation of arthritis; however, signs of arthritis were minimal and demonstrated no significant difference between DR4β/Eα and WT mice (Poole et al. 2022). Although CIA or CIA-ODE combination could not induce a meaningful arthritis phenotype, DR4β/Eα mice exhibited greater citrullination of lung proteins and ectopic lymphoid aggregates than WT mice following ODE exposure; however, ACPA were not detected in the serum. Staining of lung tissues also revealed increased vimentin and malondialdehyde acetaldehyde (MAA) modified proteins following ODE exposure of DR4β/Eα mice, which corresponded to elevated levels of serum anti-CII-MAA and anti-vimentin-MAA as well as anti-MAA antibodies in the BALF. While DR4β/Eα and WT mice demonstrated similar levels of MAA-modified and citrullinated proteins in the lung, elevations of corresponding autoantibodies were potentiated in the DR4β/Eα mice. Considering ODE exposure augmented profound lung inflammation and RA-associated autoantigen expression in DR4β/Eα mice, this may be a useful immunogenetic background to study preclinical phases of disease. Notably, a similar autoantigen profile was observed from exposure of DBA/1 mice to LPS, as well as elevated serum anti-MAA antibodies; however, these mice developed serum ACPA (Mikuls et al. 2021).

Insights gained from existing models

In spite of the inherent limitations of induction models, findings derived from the CIA and AIA mouse models provide a basis for understanding the potential role of CS, PM, DEP, and organic dust in accelerating and exacerbating preexisting arthritis. The underlying mechanisms by which this happens are not fully understood; however, elevated autoantibody levels were observed in many of the included studies. While the pathogenic mechanism of antibodies has been confirmed in CIA, it remains unclear whether autoantibodies have a pathogenic role in human RA. CS concentrate extract was shown to increase circulating concentrations of ACPA in a CIA model, potentially contributing to the exacerbation of arthritis (Kang et al. 2020). Additionally, ODE (Poole et al. 2019) also led to increased ACPA production, but not antibodies against collagen type II (anti-CII) in a CIA model, whereas DEP (Yoshino and Sagai 1999) and PM (Nowak et al. 2022) led to increases in both anti-CII antibodies and ACPA. Beyond measuring serum autoantibody levels, none of the studies examined localized autoantibody production occurring in the lung following mucosal exposure.

Of note is the characterization of ectopic lymphoid aggregates in the lung following exposure to ODE (Poole et al. 2019; 2022). These aggregates were associated with RA-associated autoantigens and serum autoantibodies in DR4β/Eα mice and resemble the iBALT that have been reported in RA patients (Poole et al. 2022). These models did not measure autoantibody levels in the lung; however, considering that ODE exposure augmented profound lung inflammation and RA-associated autoantigen expression in DR4β/Eα mice, this may be a useful immunogenetic background to study subclinical phases of disease. Ultimately, while the CIA model has validated the additive effects of CS, PM, DEP, and organic dust on experimental RA, it has done little to identify the causal mechanisms of RA following pulmonary exposure. Notably, c-silica did not demonstrate a reproducible effect in these models. Although there are inconsistencies, and not all the inhalant exposures that have been studied have augmented CIA, effects on autoantigen profiles provide valuable insights for the immune activation occurring in subclinical phases of disease. The elevation of antibodies implicated in CIA and RA in the context of inhalant exposures without signs of arthritis demonstrates that certain aspects of humoral immunity can be uncoupled from arthritis development.

Aside from the development of autoantibodies, the role of AhR in the development of CS-mediated RA has been highlighted and confirmed by two mouse studies (Talbot et al. 2018; Heluany et al. 2021). These studies demonstrated that the exacerbation of CIA by CS is dependent on AhR activation in Th17 cells and IL-17Ra signaling. Unlike observations made about autoantibodies, these studies confirm the pathological relevance of AhR, IL-17Ra and Th17 in CIA.

CS had no effect on TNF-α levels or arthritis in TNF^ΔARE mice. Crossing TNF^ΔARE onto a RAG1−/− background lacking B and T cells demonstrated that the mechanism driving arthritis was independent of adaptive immunity, suggesting that the deletion of ARE causes synoviocytes to produce TNF spontaneously in this model (Kontoyiannis et al. 1999). Because arthritic events are driven by a localized mechanism, this model is not appropriate to study lung-joint connections in the context of pulmonary exposures. Furthermore, a mutation equivalent to the ARE deletion in this model has not been identified in humans (Kontoyiannis et al. 1999) however, the success of anti-TNF antibodies in the treatment of RA suggests that overproduction of TNF could have mechanistic relevance in the pathogenesis of RA (Elliott et al. 1994). Ultimately, this model may have utility for studying therapeutics, but it is ill suited for studying the etiopathogenesis of RA, especially when exploring the causal mechanisms of environmental exposures like CS. This finding underscores the limitations of current models in accurately replicating the complexities of human RA, particularly regarding gene-environment interactions.

Challenges, gaps, and future directions

In evaluating the existing animal models designed to mimic inhalant exposure-induced RA, we found that while these models provide some insights, they are largely limited in their ability to fully replicate the complex pathogenesis of human RA. The models reviewed often focus on the exacerbation of preexisting disease rather than the initiation of RA. This is a significant limitation of current models, not only for their inability to mimic the chronology of events observed in humans, but the inability to identify the initial triggers of autoimmunity, which is integral for disease prevention. Modeling inhalant exposure-induced RA presents significant challenges, particularly due to the differing requirements for models designed to study disease initiation versus those focused on the chronicity of joint disease. No model will ever be able to fully reproduce the complexity and diversity of human disease; therefore, when modeling disease phenotypes, it is crucial to select an appropriate model for the specific phase of disease under investigation. For instance, while studying the etiology of RA may require models that capture the initiating events and the role of environmental exposures, examining the mechanisms within the joint and the progression of established disease might necessitate different approaches. The CIA and AIA models are widely used because they replicate many aspects of human RA, particularly the inflammatory processes and joint pathology, making them invaluable for studying the mechanisms of ongoing disease. These models offer rapid disease onset and a reproducible progression to symptomatic RA, which is essential for evaluating therapeutic interventions and understanding joint-specific pathology. However, these induced models may not fully capture the gene-environment interactions and extra-articular origins that are increasingly recognized as critical factors in the initiation of RA. Therefore, while CIA and AIA models are excellent for studying joint involvement and disease progression, they may be less suitable for exploring the early events in RA development or the impact of environmental factors, such as inhalant exposures, on disease initiation.

While patient cohorts are frequently classified as seropositive, this categorization does not differentiate RF from ACPA. Although these autoantibodies often co-occur, there is limited evidence that they arise through shared mechanisms or exert similar pathogenic effects. Such broad grouping may obscure relevant distinctions between RF and ACPA and complicate comparisons to animal models.

Although RA is more common in women, it is unclear whether inhalant-associated RA shows a similar sex bias, partly because occupational exposures often involve predominantly one sex. This underscores the need for animal models to assess sex-specific effects, as current models do not reflect the female predominance observed in humans. Future work should prioritize developing sex-specific models and examine whether sex differences vary by exposure type. Exposures more common among women, such as textile dust and cleaning products, should also be studied.

An additional challenge presented by the composite of current studies is reflected by the variability in outcomes between studies that used the same inhalant exposure. One source of variability arises from the diversity of exposure protocols that have been used to study the effects of inhalant exposures. The lack of standardized protocols and reagents presents difficulties for identifying the source of inconsistency between the studies we have discussed. This problem is particularly pronounced in the case of CS, as composition varies between cigarette manufacturers and CS makeup is influenced by the process being used to generate smoke.

The prevalence of smoking across demographics has yielded broad datasets that have been used to identify gene-environment interactions that are occurring in the lungs of smokers with RA. Data is far more limited for other exposures; therefore, gene-environment interactions are not known for many of these exposures. It should not be assumed that the gene-environment interactions for all inhalant exposures mirror those for CS, and efforts should be made to define these in humans.

Future research in the field of inhalant exposure-associated RA should prioritize the development and utilization of mouse models that accurately reproduce the complexities of human RA. We propose a shift from the use of induction models (e.g. CIA) to the use of models with genetic susceptibility for the development of subclinical autoimmunity or late-onset autoimmune arthritis. These models provide a crucial period to study the interplay between genetic predisposition and environmental triggers within the initiation and subclinical phases of disease. It is essential to identify and develop models that exhibit subclinical immune activity, ensuring they can appropriately respond to inhalant exposures without being so aggressive that they mask the specific effects of these exposures. Moreover, selecting models that isolate RA-associated autoimmune responses, in the absence of additional autoimmune diseases, will be critical for clarifying the central mechanisms at play.

Concluding remarks

RA is a complex disease comprised of subclinical and symptomatic phases and arising from genetic and environmental causes. Clinical observations and epidemiological data have identified subclinical features and highlighted the importance of gene-environment interactions in the etiology of RA. Insights derived from patients have illuminated risk factors and suggested potential mechanisms of disease; however, retrospective studies demonstrate the critical role of experimental models in unraveling the complex etiology of RA. This is particularly pronounced in the case of inhalant exposures linked to RA where sub-clinical phases of disease are the primary interest.

In conclusion, the ideal model balances sensitivity to environmental triggers with the specificity required to study RA in isolation. Future efforts should focus on refining these models to better mimic the human condition, thereby enhancing our understanding of how inhalant exposures contribute to the development and progression of RA. This approach will ultimately aid in identifying potential therapeutic targets and preventive strategies for individuals at risk due to environmental and occupational exposures.

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/08958378.2025.2542555.

Data availability statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.