Silica Triggers Inflammation and Ectopic Lymphoid Neogenesis in the Lungs in Parallel with Accelerated Onset of Systemic Autoimmunity and Glomerulonephritis in the Lupus-Prone NZBWF1 Mouse

Genetic predisposition and environmental factors influence the development of human autoimmune disease. Occupational exposure to crystalline silica (cSiO2) has been etiologically linked to increased incidence of autoimmunity, including systemic lupus erythematosus (SLE), but the underlying mechanisms are poorly understood. The purpose of this study was to test the hypothesis that early repeated short-term cSiO2 exposure will modulate both latency and severity of autoimmunity in the lupus-prone female NZBWF1 mouse. Weekly intranasal exposure to cSiO2 (0.25 and 1.0 mg) for 4 wk beginning at 9 wk of age both reduced latency and increased intensity of glomerulonephritis. cSiO2 elicited robust inflammatory responses in the lungs as evidenced by extensive perivascular and peribronchial lymphoplasmacytic infiltration consisting of IgG-producing plasma cells, and CD45R+ and CD3+ lymphocytes that were highly suggestive of ectopic lymphoid tissue (ELT). In addition, there were elevated concentrations of immunoglobulins and the cytokines MCP-1, TNF-α and IL-6 in bronchoalveolar lavage fluid. cSiO2-associated kidney and lung effects paralleled dose-dependent elevations of autoantibodies and proinflammatory cytokines in plasma. Taken together, cSiO2-induced pulmonary inflammation and ectopic lymphoid neogenesis in the NZBWF1 mouse corresponded closely to systemic inflammatory and autoimmune responses as well as the early initiation of pathological outcomes in the kidney. These findings suggest that following airway exposure to crystalline silica, in mice genetically prone to SLE, the lung serves as a platform for triggering systemic autoimmunity and glomerulonephritis.


Introduction
Development of autoimmunity is widely acknowledged to be impacted not only by genetic predisposition, but by environmental factors that potentially accelerate initiation of autoimmune responses and increase severity of organ-specific pathologies [1]. Systemic lupus erythematosus (SLE) is a prototypical autoimmune disease that affects 300,000 Americans, most of whom are women, and is often associated with glomerulonephritis and kidney failure [2]. In the United States, it is estimated that at least 1.7 million workers are employed in occupations (e.g. mining, construction, manufacturing and custodial industries) with potential exposure to crystalline silica (cSiO 2 ) [3]. Human epidemiological studies have linked occupational exposure to crystalline silica to increased SLE incidence [4][5][6]. For example, a population-based, case-control study conducted with participants from the Carolina Lupus Cohort reported that development of SLE was more prevalent in individuals with occupations associated with cSiO 2 exposure [5]. Similarly, in a cohort of 15,000 uranium miners heavily exposed to cSiO 2 , 28 men were diagnosed with SLE, whereas the predicted SLE prevalence for a group of this size was 1.5 individuals [7]. Finckh and coworkers [6] found that occupational exposure to cSiO 2 , but not organic solvents, contributed to increased SLE in urban women. Other studies report SLE-like symptoms in individuals with silicosis, including lymphocyte activation and elevated antinuclear antibodies, immunoglobulins, and immune complexes [8][9][10][11].
Both in vivo and in vitro studies have yielded several lines of evidence that cSiO 2 -induced cell death of alveolar macrophages is a critical event in cSiO 2 -induced exacerbation of autoimmunity [12][13][14][15][16][17]. Phagocytosis of cSiO 2 particles by alveolar macrophages induces NF-κB activation [18,19]. This results in production of proinflammatory cytokines, including TNF-α [20], which are correlated with disease activity in individuals with SLE [21,22]. Aberrant apoptosis and/or necrosis of cSiO 2 -laden alveolar macrophages not only releases these particles, enabling continued pulmonary re-exposure to cSiO 2 and sustained inflammation, but could also expose normally sequestered nuclear antigens (e.g. dsDNA). Both excessive apoptosis and defective clearance of apoptotic debris contribute to aberrant self-antigen presentation, driving consequent autoimmune sequealae [12,23].
Mouse strains that spontaneously develop SLE have been used to explore how toxicant exposure affects the pathogenesis of autoimmunity [24][25][26][27]. One such strain, NZM2410, bred from the widely used SLE model, NZBWF1 [24], exhibits rapid onset of lupus nephritis with most mice succumbing to glomerulonephritis by 25 wk of age [28]. When intranasally instilled with cSiO 2 , both male and female NZM2410 mice exhibit accelerated onset and increased severity of SLE-related responses including elevated autoreactive antibodies, proteinuria, and glomerulonephritis [29]. It was further determined that concentrations of TNF-α in bronchoalveolar lavage fluid of cSiO 2 -exposed NZM2410 mice were significantly higher than saline-treated control mice [30]. Autoantibodies isolated from sera of cSiO 2 -exposed NZM2410 mice were found to bind to apoptotic debris derived from alveolar macrophages [31]. Importantly, the effects of cSiO 2 on autoimmunity could be attenuated in this strain in vivo by concurrent instillation with rottlerin, an inhibitor of apoptosis [32].
There are several issues that limit the use of the NZM2410 phenotype to uncover mechanisms by which cSiO 2 triggers SLE. Not only does this strain develop rapid onset and progression of lupus nephritis compared to other lupus-prone mouse models [24], but the majority of NZM2410 mice succumb to glomerulonephritis within 2 wk of proteinuria appearance [33]. Furthermore, contrary to SLE prevalence in human population and other murine models [24], NZM2410 mice do not display female sex bias. Finally, while NZM2410 mice have been useful for delineating genetic contributions to SLE, most preclinical studies of putative therapeutic agents employ the NZBWF1 mouse, which displays slower development of autoimmunity [25]. Taken together, investigation of subtle underlying mechanisms and preventative measures to counter toxicant-triggered SLE might be enhanced by employing the less-penetrant NZBWF1 phenotype.
The purpose of this study was to test the hypothesis that short-term repeated intranasal cSiO 2 exposure will both decrease latency to onset and increase severity of autoimmunity in female NZBWF1 mice. The results presented herein suggest that cSiO 2 triggered robust autoimmune and inflammatory responses in the lung, including the formation of ectopic lymphoid tissue, that closely paralleled early onset of glomerulonephritis and exacerbated systemic autoimmune responses.

Materials and Methods Mice
All experiments were approved by the Institutional Animal Care and Use Committee at Michigan State University in accordance with the National Institutes of Health guidelines for animal use. Female 7 wk old lupus-prone NZBWF1 and control C57Bl/6 mice were obtained from Jackson Laboratories (Bar Harbor, ME), randomized into treatment groups, and allowed to acclimate for 2 wk prior to cSiO 2 exposure. Upon arrival and throughout the study, mice were fed semi-purified American Institute of Nutrition (AIN)-93G diet containing high oleic safflower oil [34]. Mice were housed 4 per cage with free access to food and water and maintained at constant temperature and humidity (21°C-24°C and 40-55%, respectively) under a 12-h light/ dark cycle.
The experimental design for cSiO 2 exposure, summarized in Fig 1, was based on previous protocols [29][30][31][32]35]. cSiO 2 (Min-U-Sil-5, average particle diameter 1.5-2.0 μm) was originally obtained from Pennsylvania Sand Glass Corporation (Pittsburgh, PA). Stock suspensions of each dose were prepared fresh in sterile phosphate buffered saline (PBS) just prior to intranasal exposure and sonnicated for 1 min before use. At 9 wk of age, mice were anesthetized with 4% isoflurane and instilled intranasally with 0, 0.25 and 1.0 mg of cSiO 2 in 25 μl PBS. To mimic intense occupational exposure to cSiO 2 over a short time period, mice were instilled once per wk for 4 wk. Urine was collected weekly and evaluated for proteinuria by 2P urine reagent strips (Cortez Diagnostics, Inc., Calabasas, CA). Mice were sacrificed and tissues collected 12 wk after the final exposure to cSiO 2 . At this time point, more than 50% of animals exposed to the high cSiO 2 dose exhibited moderate proteinuria, defined as 300 mg/dl. In addition to NZBWF1, female C57Bl/6 mice were used as a control to discern histopathologic effects of high cSiO 2 exposure in kidney and lung. C57Bl/6 mice were dosed with either 25 μl PBS or 1.0 mg cSiO 2 in 25 μl PBS following the above protocol and terminated 12 wk after the final exposure.

Necropsy and tissue collection
Animals were euthanized by intraperitoneal injection with 56 mg/kg BW sodium pentobarbital and exsanguination via the abdominal aorta. Blood was collected using heparinized syringes and centrifuged at 3500 x g for 10 min at 4°C for separation of plasma. The collected plasma was stored at -80°C until future analysis. Following BALF collection as described previously [36], lungs were intratracheally fixed with 10% (v/v) neutral buffered formalin at constant pressure (30 cm H 2 O) for a minimum 1 h and stored in fixative until further processing for histology. The right kidney was excised and the cranial portion fixed for 24 h in 10% neutral buffered formalin.

Kidney histopathology
Formalin-fixed kidneys were paraffin-embedded, sectioned and stained with either hematoxylin and eosin (H&E) or Periodic acid-Schiff and hematoxylin (PASH). Stained sections were evaluated for lupus nephritis in a blinded fashion by a board-certified veterinary pathologist using a modified International Society of Nephrology/Renal Pathology (ISN/RPS) Lupus Nephritis Classification system [37]. Slide sections were graded as follows: (0) no tubular proteinosis and normal glomeruli; (1) mild tubular proteinosis with multifocal segmental proliferative glomerulonephritis and occasional early glomerular sclerosis and crescent formation; (2) moderate tubular proteinosis with diffuse segmental proliferative glomerulonephritis, early glomerular sclerosis and crescent formation; and (3) marked tubular proteinosis with diffuse global proliferative and sclerosing glomerulonephritis.

Lung histopathology
Randomly orientated, serial sections of the formalin-fixed left lung lobe were processed routinely and embedded in paraffin. Tissue sections (5 μm) were deparaffinized and stained with H&E for histopathology. Tissues were semi-quantitatively scored by a board-certified veterinary pathologist for the following lung lesions: (a) presence of lymphocytic cell infiltration within perivascular and peribronchial regions, (b) alveolitis defined as the presence of alveolar infiltration of vacuolated macrophages, neutrophils, and lymphocytes, granuloma formation in the alveolus, type II epithelial cell hyperplasia, and thickened alveolar wall, and (c) presence of alveolar proteinosis. Individual lungs were graded for these lesions using the following criteria (% of total pulmonary tissue examined): (0) no changes compared to control mice; (1) minimal (<10%); (2) slight (10-25%); moderate (26-50%); (4) severe (51-75%); or (5) very severe (>75%) of total area affected.
Immunohistochemistry for IgG was performed on paraffin-embedded lungs. Tissue sections (5 μm) were deparaffinized and subjected to heat-induced epitope retrieval with citrate buffer (pH 6.0) for 30 min at 100°C. Endogenous peroxidase was blocked with 3% (v/v) H 2 O 2 / methanol for 30 min at room temperature followed by blocking non-specific protein binding with normal horse serum (Vector Laboratories Inc., Burlingame, CA) for 30 min. Tissue endogenous avidin and biotin binding sites were blocked by incubation with Avidin D (Vector Laboratories) followed by d-Biotin (Sigma Aldrich, St. Louis, MO), each for 15 min. Goat antimouse IgG γ-chain specific (Alpha Diagnostic Inc., San Antonio, TX) was diluted 1:4000 in normal antibody diluent (ScyTek Laboratories Inc., Logan, UT) and incubated for 60 min followed by incubation with 11μg/ml biotinylated horse anti-goat IgG H+L (Vector Laboratories, Inc.) in normal antibody diluent for 30 min. Peroxidase enzyme was incubated on slides for 30 min with Ready-to-Use Elite Peroxidase Reagent (Vector Laboratories, Inc.). Finally, slides were incubated for 15 min with Nova Red substrate (Vector Laboratories, Inc.) and counterstained with hematoxylin.
B and T lymphocytes were identified using rat anti-mouse CD45R (BD Biosciences, San Jose, CA) (1:300) and rabbit anti-mouse CD3 (1:100), respectively on deparaffinized tissue sections (4 μm) that were subjected to heat-induced epitope retrieval with citrate buffer (pH 6.0) for 30 min at 100°C (CD45R) or 125°C for 15 min (CD3). Following pretreatments, formation of avidin-biotin complex and primary antibody staining were performed as described above. Bound CD45R and CD3 antibodies were detected by incubation with biotinylated rat and goat anti-rabbit in conjunction with Peroxidase Reagent and Nova Red substrate.

BALF processing and differential staining
Cytological slides from 150 μl BALF were prepared by centrifugation at 400 x g for 10 min using a Shandon Cytospin 3 (Shandon Scientific, Pittsburgh, PA), allowed to air dry, and stained with Diff-Quick (Fisher Scientific). A total of 200 cells were counted and cells identified as monocytes/macrophages, lymphocytes, and polymorphonuclear (PMN) leukocytes using morphological criteria. Remaining BALF was centrifuged at 2400 x g for 15 min and supernatant collected and stored at -80°C for IgG and cytokine analysis.

IgG and autoantibody measurement
Total IgG, IgA and IgM were quantitated in BALF using an ELISA method described previously [38]. Briefly, 96-well plates were coated with 50μL of 10 μg/ml goat anti-mouse IgA, IgM, or IgG, respectively (Alpha Diagnostics, Inc.). The standard curve was generated using mouse reference serum (Bethyl Laboratories, Inc., Montgomery, TX). Plates were read on an ELISA reader (Molecular Devices, Menlo Park, CA) at 650 nm and Ig concentrations calculated from the standard curves using Softmax software (Molecular Devices). Total Ig antibodies for nuclear antigens and double-stranded DNA (dsDNA) in plasma were measured according to manufacturer's instructions using commercial ELISA kits (Alpha Diagnostics, Inc.).

Cytokine analysis
BALF and plasma were diluted two-fold and analyzed for the inflammatory cytokines IL-6, IL-10, MCP-1, IFN-γ, TNF, and IL-12p70 using a Mouse Inflammation Cytometric Bead Array (BD Biosciences). Data was acquired using a FACSCalibur flow cytometer (BD Biosciences) and cytokine concentrations were calculated from standard curves using FCAP Array Software (BD Biosciences).

Statistics
All treatment groups consisted of 8 mice and data presented as group mean ± SEM. Data were plotted and analyzed using SigmaPlot 11.0 for Windows (Jandel Scientific; San Rafael, CA). Grubb's test was performed to identify and exclude outliers from statistical analysis. Differences between cSiO 2 -and vehicle-treated NZBWF1 mice were analyzed by one-way ANOVA on Ranks with Dunn's method. Differences between cSiO 2 -and vehicle-treated C57Bl/6 mice, and NZBWF1 and C57Bl/6 histopathological lung grades were analyzed by Mann-Whitney Rank Sum Test. The Spearman rank-order correlation coefficient was used to correlate cSiO 2 dose to experimental endpoint for NZBWF1. A P value of < 0.05 was considered statistically different for all study outcomes.

Results cSiO 2 elicits early onset of proteinuria and glomerulonephritis
The effects of weekly intranasal exposure to 0, 0.25 mg, or 1.0 mg cSiO 2 from 9 to 12 wk of age on renal function were monitored by measuring urinary protein at weekly intervals. Proteinuria ( 300 mg/dl) was observed in 13% and 62% of NZBWF1 mice treated with 1.0 mg cSiO 2 at 10 and 12 wk post-dosing, respectively (Fig 2). In contrast, proteinuria was not evident in Fig 2. cSiO 2 exposure accelerates development of proteinuria in NZBWF1 mice. Proteinuria was monitored weekly until sacrifice 12 wk after the final cSiO 2 exposure when most mice exposed to 1.0 mg cSiO 2 were over threshold ( 300 mg/dl). Proteinuria was not detected in NZBWF1 mice dosed with 0.25 mg cSiO 2 or vehicle. C57Bl/6 mice exposed to cSiO 2 or vehicle did not develop detectable proteinuria over the course of the experiment. To confirm nephritogenic effects of cSiO 2 , kidney sections from NZBWF1 mice at 12 wk post-treatment were assessed histopathologically following H&E and PASH staining (Fig 3A  and 3B). Kidneys from NZBWF1 mice treated with 1.0 mg cSiO 2 exhibited frequent tubular proteinosis and proliferative glomerulonephritis (Fig 3B). Lymphocytic infiltration at the renal pelvis as well as sclerosis and crescent formation were additional common findings in animals treated at this dose. These responses were largely absent in NZBWF1 mice treated with vehicle ( Fig 3A) or 0.25 mg cSiO 2 as well as control C57Bl/6 mice treated with vehicle or 1.0 mg cSiO 2 (Fig 4A and 4B).
When mice were graded individually for glomerulonephritis severity, only 13% (1/8) of NZBWF1 mice treated with vehicle were ranked 1 due to the occurrence of minimal tubular proteinosis and rare glomeruli with segmental mesangial hypercellularity, whereas the remaining vehicle-treated mice were ranked 0 ( Fig 3C). In contrast, cSiO 2 elicited renal lesions characteristic of lupus nephritis, with 38% (3/8) and 63% (5/8) of NZBWF1 mice treated with 0.25 mg and 1.0 mg of the agent, respectively, being ranked 1. Additionally, 25% (2/8) of NZBWF1 mice treated with 1.0 mg cSiO 2 had renal lesions ranked 2 or 3, whereas no mice treated with vehicle or 0.25 mg cSiO 2 developed such severe lesions. Interestingly, while all C57Bl/6 mice exposed to vehicle were ranked 0, indicating no renal lesions consistent with glomerulonephritis, 38% of mice dosed with 1.0 mg cSiO 2 developed renal lesions ranked 1 or 2 ( Fig 4C). Thus, cSiO 2 appeared to induce modest glomerulonephritis in some of the C57Bl/6 control mice without onset of evident proteinuria.

cSiO 2 induces marked lymphoplasmacytic infiltration in lung parenchyma
Histologically, conspicuous lymphoplasmacytic infiltrates were present in interstitial tissues surrounding bronchioles and associated blood vessels throughout the lung in NZBWF1 mice exposed to 1.0 mg cSiO 2 (Fig 5C). This response was less severe in NZBWF1 mice treated with 0.25 mg cSiO 2 and minimal in NZBWF1 mice treated with vehicle ( Fig 5A and 5B; Table 1). C57Bl/6 mice exposed to 1.0 mg cSiO 2 had minimal perivascular and peribronchial lymphoplasmacytic infiltrates as compared to vehicle controls (Fig 5D and 5E) and NZBWF1 mice treated with cSiO 2 (Table 1).
In addition to the pronounced inflammatory cell infiltration observed in the interstitial tissue of the lung, diffuse alveolar lesions were histologically evident after cSiO 2 exposure in NZBWF1 and C57Bl/6 mice. In contrast to the aforementioned lymphocytic response in the lung tissues, the severity of alveolar lesions in NZBWF1 mice was comparable to C57Bl/6 mice given the same amount of cSiO 2 ( Table 1). Dosing of NZBWF1 mice with 0.25 mg cSiO 2 also elicited the responses observed with 1.0 mg cSiO 2 in this strain, albeit to a more modest extent.

cSiO 2 elicits IgG producing plasma cells in lung
Based on the prominent lymphoplasmacytic response in the lungs of cSiO 2 -treated NZBWF1 mice, the effects of intranasal cSiO 2 on IgG production in these tissues was assessed immunohistochemically. Numerous IgG-laden lymphocytes were found present in the perivascular and peribronchiolar lymphoplasmacytic aggregates in cSiO 2 -treated but not vehicle-treated NZBWF1 mice (Fig 6B and 6D). In addition, H&E staining of lung sections from NZBWF1 mice treated with cSiO 2 revealed occasional plasma cells containing Russell bodies (conspicuous cytoplasmic accumulation of hyaline material) that were immunohistochemically Representative light photomicrographs of PASH stained kidney sections in NZBWF1 mice at 24 wks of age exposed to vehicle (A) and 1.0mg cSiO 2 (B). NZBWF1 mice exposed to 1.0mg cSiO 2 developed extensive glomerulonephritis (black arrow) and tubular proteinosis (asterisk). NZBWF1 mice were individually graded for lupus nephritis following the modified ISN/ RPS classification system as described in the Materials and Methods (C). Animals exposed to cSiO 2 developed more severe lesions characteristic of lupus nephritis than vehicle-exposed mice. cSiO 2 dose significantly correlated with lupus nephritis (Spearman rank-order coefficient = 0.64, p < 0.05).  Representative light photomicrographs of PASH stained kidney sections in C57Bl/6 mice at 24 wks of age exposed to vehicle (A) and 1.0mg cSiO 2 (B). Some renal histopathological lesions were observed in cSiO 2exposed C57Bl/6 mice (B) with notable renal tubular proteinosis (asterisk) and several glomeruli with global mesangial hypercellularity (white arrows). C57Bl/ 6 mice were individually graded for lupus nephritis as described in the Materials and Methods (C). Animals exposed to cSiO 2 developed more severe lesions characteristic of lupus nephritis than vehicle exposed mice.   identified as containing IgG globulin. Taken together, these results suggest that both production and secretion of IgG by infiltrating plasma cells is induced in both lungs and airways after intranasal cSiO 2 exposure in NZBWF1 mice.
CD45R+ and CD3+ infiltrating lymphocytes in lung organize into ectopic lymphoid tissue after cSiO 2 exposure in NZBWF1 mice Infiltrating lymphocytes induced by cSiO 2 in lungs of NZBWF1 mice were further characterized by immunohistochemistry for B and T lymphocytes. Distinct aggregates of CD45R+ cells (B lymphocytes) (Fig 7B, 7F and 7J) interspersed with CD3+ cells (T lymphocytes) (Fig 7C, 7G and 7K) were observed peripheral to both bronchiolar airways and blood vessels in lung interstitial tissue of NZBWF1 mice exposed to 1.0 mg cSiO 2 . The distinct localization of CD45R+ aggregated cells in conjunction with infiltrating CD3+ lymphocytes in the lungs following cSiO 2 exposure in NZBWF1 mice structurally resembled the spatial architecture of ectopic lymphoid tissue (ELT). The identification of ELT in parallel with the observation of IgG-producing plasma cells suggest that these structures functionally contribute to initiation of antigen-specific adaptive immune responses in lungs of NZBWF1 mice following intranasal cSiO 2 exposure. cSiO 2 induces leukocyte infiltration and elevates IgG, IgA, and IgM in BALF Effects of enhanced lymphocytic infiltration following intranasal cSiO 2 exposure were also evident in the airways. In the recovered BALF, treatment with 0.25 mg and 1.0 mg cSiO 2 in NZBWF1 mice elicited increased recruitment of lymphocytes (33% and 43%, respectively) relative to vehicle control where the majority of cells were monocytes/macrophages in both NZBWF1 and C57Bl/6 mice (Fig 8). Relative to NZBWF1 mice, 1.0 mg cSiO 2 in C57Bl/6 mice elicited a modest increase of lymphocytes in BALF (23%). Similar PMN leukocyte responses were observed after 1.0mg cSiO 2 exposure in both NZBWF1 (19%) and C57Bl/6 mice (17%).
Based on the observed increase in lymphocytes following intranasal cSiO 2 exposure, we performed ELISA to quantitate IgG, IgA and IgM in BALF of NZBWF1 and C57Bl/6 mice. IgG in BALF of NZBWF1 was substantially elevated in a dose-dependent manner following cSiO 2 Table 1. cSiO 2 exposure increases severity of lymphocytic cell infiltration in lungs of NZBWF1 mice relative to C57Bl/6 mice. treatment ( Fig 9A). cSiO 2 also increased IgA and IgM in BALF of both NZBWF1 and C57Bl/6 mice, albeit to a lesser extent than IgG (Fig 9B and 9C).

cSiO 2 triggers systemic autoantibody and proinflammatory cytokine responses
Marked elevations of total Ig anti-dsDNA antibodies ( Fig 11A) and anti-nuclear antibodies ( Fig 11B) were detected in plasma of cSiO 2 -dosed NZBWF1 mice compared to vehicle-treated mice. No changes in plasma anti-dsDNA antibodies were detected in C57Bl/6 mice exposed to

Discussion
Human epidemiological findings support the contention that airway exposure to cSiO 2 is etiologically linked to development of autoimmunity [4][5][6][39][40][41][42][43]. However, such studies shed little light on the mechanisms that underlie this putative association. Towards that end, we investigated the effects of short-term repeated cSiO 2 exposure on autoimmune disease development in the NZBWF1 female mouse, a widely used experimental animal model for SLE. Several novel observations regarding this lupus-prone mouse strain were made 12 wk after the final cSiO 2 exposure. First, cSiO 2 triggered the onset of and exacerbated glomerulonephritis. Second, concurrent with kidney effects, cSiO 2 induced marked lymphocyte infiltration in the lungs as well as dramatic increases in total immunoglobulins (IgG, IgA and IgM), and proinflammatory cytokine concentrations. Third, a striking observation in the lungs of NZBWF1 mice was the identification of extensive ELT that likely contributes to early triggering of SLE after cSiO 2 exposure. Finally, in plasma, we found that cSiO 2 exposure elicited marked elevations in autoreactive antibodies and proinflammatory cytokine concentrations. Collectively, these findings  suggest intensive cSiO 2 exposure via the airways over a discrete period evoked long-lasting injurious inflammatory and autoimmune responses that were evident at the site of exposure (lung), systemically, and at a distal tissue site (kidney). Elevated plasma titers of autoreactive antibodies are an immunological hallmark for human SLE. Systemic autoantibodies form immune complexes with circulating self-antigens (e.g. dsDNA and nucleosome fractions). When deposited in the kidney [29,[44][45][46], these immune complexes promote proinflammatory cytokine and chemokine production that mediate infiltration by mononuclear cells and consequent tissue injury [47]. Congruent with these mechanisms, heightened severity of glomerulonephritis was observed here in cSiO 2 -treated NZBWF1 mice concurrently with early onset of proteinuria. These findings are consistent with previous observations in cSiO 2 -exposed NZM2410 mice [29].
Several epidemiological studies have identified an association between occupational crystalline silica exposure and development of end-stage renal disease [39,48,49], particularly glomerulonephritis [50,51]. In a previous study by our laboratory, NZBWF1 mice fed diet identical to that employed here (safflower oil-based AIN-93G) spontaneously developed proteinuria at 30 wk of age [34]. Herein, NZBWF1 dosed with 1.0 mg cSiO 2 exhibited proteinuria beginning at 22 wk of age. Thus, intranasal exposure to cSiO 2 likely decreased the time to onset of lupus nephritis by 8 wk. Interestingly, kidney lesions were also increased in cSiO 2 -treated C57Bl/6 mice, albeit to a more modest extent. The observation of renal pathology consistent with lupus nephritis in C57Bl/6 mice supports the notion of an association between cSiO 2 exposure and kidney disease. Consistent with these findings, autoimmune-like responses in kidneys including IgG deposition have been previously described after intratracheal administration of fibrogenic particles in C57Bl/6 mice [52,53]. While overt proteinuria ( 300 mg/dl) did not develop over the 12 wk duration of our study, it might be speculated that given sufficient time, short-term repeated airway exposure to cSiO 2 might induce autoimmune nephritis in C57Bl/6 and other mice without evident genetic predisposition to SLE.
Although a prior study of cSiO 2 -induced pulmonary inflammation in NZM2410 autoimmune mice qualitatively described inflammatory cell infiltration in the lung [29], this report is the first to semi-quantitatively grade this response. A critical finding in this study was that NZBWF1 mice exposed to cSiO 2 exhibited extensive lymphoplasmacytic infiltration in peribronchial and perivasculature regions of the lungs. Similar but more modest effects were observed in C57Bl/6 mice. This observation is consistent with the findings of Singh et al. [54] who suggested that the magnitude of perivasculature and peribronchial inflammation induced by pulmonary toxicant exposure is dependent on genetic background.
A critical question relates to the relevance of the doses employed in this study to actual human exposure. Using the NIOSH limit for respirable cSiO 2 established at 0.05mg/m 3 /d and the assumption that human ventilation rate is 6.0L/min, we calculate that humans could be exposed to 1433mg silica in 40 years of work (8h/d for 5 d/wk) with respirable cSiO 2 . Based on the assumption that mice ventilation rate is 0.03L/min, we estimate that an equivalent lifetime exposure to crystalline silica would be 8.28 mg in mouse. Therefore, the two cumulative doses of cSiO 2 used in this study, 1 and 4 mg, represent approximately one eighth and one half, respectively, of a human lifetime exposure to cSiO 2 at the recommended NIOSH exposure limit.
Although inhalation exposure of rodent species is the gold standard for modeling human airway exposure to respirable toxicants such as crystalline silica, the need for specialized inhalation exposure chambers, rigorous technical considerations and high operational costs limit these experiments to only a few laboratory facilities. A comparative study of intratracheal, trans-oral, and intranasal instillation in Balb/c mice receiving a single 1.0 mg dose of cSiO 2 revealed that each method was capable of eliciting significant pulmonary inflammation 4 h after exposure [59]. Intranasal instillation resulted in comparatively less inflammation relative to both trans-oral and intratracheal instillation. This could be attributed to trapping of cSiO 2 particles in the nasal cavity, resulting in delayed passage to the lower airways [59]. However, this effect is not likely to be an issue in long-term studies of intranasal exposure to cSiO 2 , as mice treated in this manner eventually develop pulmonary fibrosis, a hallmark of cSiO 2 exposure [35,60,61]. These latter studies support the contention that the high-dose, short-term repeated intranasal exposure employed here is a suitable model for investigations of environmentally triggered onset of autoimmunity in lupus-prone mice. Nevertheless, in the future it will be of interest to compare the effects of other routes of cSiO 2 exposure, especially inhalation, on development of autoimmunity in NZBWF1 mice.
An unexpected observation in this study was extensive ectopic lymphoid neogenesis in the lung after cSiO 2 exposure in NZBWF1 mice. ELT morphologically and functionally resemble secondary lymphoid tissues (e.g. lymph nodes) by supporting maturation and differentiation of B cells into plasma cells that produce class-switched antibodies [62]. ELT have been previously identified in target tissues of several autoimmune diseases, including the kidneys of aged NZWBF1 mice [63], synovial joints in rheumatoid arthritis [64,65], and salivary glands in Sjögren's syndrome [66][67][68], where they contribute to local manifestations of autoimmunity. Although events that drive formation of ELT remain poorly defined, neogenesis of ectopic lymphoid structures is indicative of tissue-specific chronic inflammation [69] and importantly, may influence systemic autoimmunity [70]. Taken together, cSiO 2 -triggered formation of pulmonary ELT likely contributed to the exacerbated systemic autoimmunity observed in NZBWF1 mice following cSiO 2 exposure.
Elevated BALF concentrations of TNF-α, MCP-1, and IL-6 observed here in cSiO 2 -exposed NZBWF1 mice are key observations because these cytokines play critical roles in both initiating innate and amplifying adaptive immune responses. The sustained elevation of these cytokines in BALF up to 12 wk after cSiO 2 exposure suggests that the initial innate immune response was insufficient to contain the inflammation, supporting a previously described model for continuous alveolar macrophage apoptosis after cSiO 2 exposure with ensuing self-antigen presentation [12]. Robust elevation of TNF-α in BALF has been similarly reported in cSiO 2 -exposed NZM2410 mice [30]. TNF-α induces apoptosis in alveolar macrophages and therefore could contribute to increased apoptotic debris containing autoantigens [12,31]. This cytokine can also stimulate production of additional proinflammatory cytokines, which could further amplify the immune response to cSiO 2 [71][72][73][74]. Released predominately by monocytes/macrophages, MCP-1 recruits monocytes from systemic circulation to local sites of inflammation after tissue injury in attempt to contain or resolve further inflammation. The heightened elevation of this chemokine in BALF 12 wks following cSiO 2 exposure further supports the notion that cSiO 2 induces a continuous cycle of alveolar macrophage apoptosis, as sustained recruitment of circulating monocytes to the lungs by MCP-1 is likely an innate attempt to repopulate lost alveolar macrophages and resolve inflammation. We observed robust increases in the percent of lymphocytes in BALF recovered from cSiO 2 -treated NZBWF1 mice. Previous studies in cSiO 2 exposed NZM2410 mice have indicated that both B and T cells contribute to increased lymphocytes in BALF [30]. IL-6 has a well-established role in both CD4+ T cell survival and Bcell differentiation to Ig-secreting plasma cells [75]. Taken together, the increase of lymphocytes in the presence of IL-6 in BALF likely promotes antigen-specific effector responses in the lungs. Indeed, in the lung, cSiO 2 -induced elevation of IL-6 corresponded to significant increases of IgG in BALF of NZBWF1 mice.
Mechanistic studies have indicated that alveolar macrophages are key to not only cSiO 2 -induced pulmonary pathogenesis [12], but are also essential for development of IgG immune complex-mediated inflammation in the lungs [76,77]. Following exposure to chrysotile asbestos, in vitro production of superoxide anion by alveolar macrophages was enhanced by stimulation with IgG [78]. Proinflammatory cytokines in lungs were markedly reduced in C57Bl/6 mice deficient in FcγRIII, which mediates activation of mononuclear phagocytes by binding the Fc region of the IgG molecule [79]. The notion that IgG both initiates and enhances lung inflammatory responses through the alveolar macrophage is therefore of possible significance to cSiO 2 triggering of SLE in autoimmune-prone mice. Extensive deposition/production of IgG in lungs of cSiO 2 -exposed NZBWF1 mice could act synergistically with defective clearance of apoptotic macrophages, further contributing to acceleration of autoimmunity.
Airway exposure to cSiO 2 stimulates alveolar macrophages, epithelial cells, and fibroblasts that mediate recruitment of circulating monocytes, neutrophils, and lymphocytes by releasing an array of inflammatory mediators including cytokines [57,[80][81][82]. The close proximity of these inflammatory cells to both the airways and vasculature in this study suggest that these cells are not only capable of mediating production of proinflammatory mediators that impact the lung, but that may also be secreted into systemic circulation, thereby exacerbating development of systemic autoimmunity. It is notable that cSiO 2 -induced plasma increases of TNF-α and IL-6 mirrored elevations of these cytokines in BALF. Importantly, systemic concentrations of TNF-α and IL-6 correlate with SLE disease activity in humans [83] and treatment with exogenous IL-6 exacerbates glomerulonephritis in NZBWF1 mice [84]. There was also a trend towards elevated MCP-1 in plasma of NZBWF1 mice exposed to cSiO 2 . Urinary MCP-1 concentration has been identified as a biomarker of disease activity in lupus nephritis [85,86], and one study indicated that renal expression of MCP-1 correlates with NF-ΚB activation in kidney [87]. Overall, these results suggest that elevated plasma proinflammatory cytokines induced after cSiO 2 exposure might further contribute to production of plasma autoantibodies as well as exacerbated renal pathology. Interestingly, cytokine array analysis of cSiO 2 -exposed NZM2410 mice failed to reveal any significant difference in plasma cytokines IL-4, IFN-γ, IL-10, IL-12, and TNF-α [30] suggesting some inherent differences in the response to intranasal cSiO 2 might exist between that strain and the NZBWF1 employed here.
To summarize, the results presented here suggest that following airway exposure to cSiO 2 , the lung serves as a platform for the early triggering and exacerbation of systemic autoimmunity and glomerulonephritis in the NZBWF1 mouse. This model can serve as a starting point for further studies to gain insight into toxicant-triggered autoimmunity. First, it will be essential to characterize antigen-presenting cell and lymphocyte subpopulations recruited to and migrating out of the lung after cSiO 2 exposure. These cells have the potential to drive subsequent tissuespecific homing of effector cell populations that mediate pathological outcomes in the lung and kidney. Second, while it is apparent that cSiO 2 induces plasma elevation of proinflammatory cytokines, further studies are warranted to ascertain if these originate from lung ELT, inflamed kidneys, and/or immune tissues such as spleen. Third, this model of cSiO 2 -accelerated lupus can be used to study potential approaches for prevention and intervention in occupationally exposed human populations. A particularly attractive approach is the consumption of n-3 polyunsaturated fatty acids in fish oil which have been shown to delay onset and severity of autoimmune nephritis in NZBWF1 and other models [34,88]. Finally, the strategy described herein could be used to investigate whether other airborne toxicants that cause inflammatory responses in the lung might similarly exacerbate autoimmunity.
Supporting Information S1 Fig. cSiO 2 did not significantly alter anti-dsDNA Ab's in plasma of C57Bl/6 mice. Antibodies in plasma at sacrifice were measured by ELISA. Data are mean ± SEM (n = 7-8/gp) and were analyzed by Mann-Whitney Rank Sum Test. (TIFF)