Autoantibodies in outbred Swiss Webster mice following exposure to gold and mercury

in


Introduction
Autoimmunity arises from a combination of genetic, environmental, and stochastic factors (Theofilopoulos et al., 2017).In this study, we have examined the autoimmune effects of the environmental factors gold (Au), and mercury (Hg) using an outbred mouse strain, Swiss Webster (SW).These two metals are tied together due to gold mining (Steckling et al., 2014).
In the 1860-90´s, large amounts of mercury were used for extracting gold in the USA in artisanal and small-scale gold mining (ASGM).In the first decade of the 2000´s, these activities again surged due the high price of gold.Nowadays, 10-15 million people in 70 countries are engaged in ASGM (Seccatore et al., 2014) which provides 15-20% of gold production in the world.ASGM is by far the biggest source of mercury pollution and contributes to one-third of the global mercury emission (Bishop et al., 2020).
The main chemical forms of mercury (inorganic, organic, and elemental) have had an important effect on human health throughout the last three centuries, but the mode of exposure to mercury has varied.Occupational exposure to mercury dominated in the 18 th and 19 th centuries however it is still a major health issue.Studies on gold miners have shown a correlation between mercury exposure and serum antinuclear antibodies (ANA) prevalence and titer (Gardner et al., 2010;Motts et al., 2014).
The catastrophic industrial pollution with methyl-mercury (Me-Hg) in the Minamata Bay in 1956, causing neurologic damages and deaths started a new era of mercury effect of human health (Harada, 1995).In the 1960´s, a vicious circle of contamination by Me-Hg was discovered.Due to the linked food chains, Me-Hg contaminated the wildlife and humans over the globe.In the 1990´s, an example for the food chain contamination by Me-Hg, due to consumption of whale meat, was found in the Faroe Islands with a significant negative impact on the intellectual capacity in children (Grandjean et al., 1998).
Apart from the use in jewelry and electronic components, gold was used for more than a century to treat rheumatic disease (Lansdown, 2018).However, gold treatment causes many side effects, some of them serious, and most of the patients have to stop the treatment in the long run (Bendix and Bjelle, 1996).Adverse effects such as thrombocytopenia, granulocytopenia, proteinuria, and nephrotic syndrome are common among these patients (Adachi et al., 1984;Bensen et al., 1984;Bigazzi, 1999).
Several new drugs, like monoclonal antibodies, and biologics, have practically eliminated gold as an anti-rheumatic substance in the western world (Burke et al., 2006;N.I.H, 2017).
In animal models, the immune response to mercury includes molecular activation, increased cytokine expression, polyclonal B-cell activation, lymphoproliferation, hypergammaglobulinemia, development of autoantibodies, and glomerular and vascular immune complex (IC) deposits (Pollard et al., 2019).Whereas the immune response to gold shows similarity to effects of H-2 s mice exposed to mercury, it differs in some respects: a delay in the immune response, a weaker lymphoproliferation, and the lack of IC deposits in organs (Pietsch et al., 1989;Havarinasab et al., 2007c;Havarinasab et al., 2009).
The susceptibility of mouse models to mercury-and gold-induced autoimmunity is under genetic control, particular by the H-2 major histocompatibility complex (Hultman et al., 1992;Havarinasab et al., 2007c).However, genes outside the H-2 region are shown to play an important role as well (Alkaissi et al., 2018).In genetically susceptible strains of mice (H-2 s , H-2 q , H-2 f ), both mercury and gold induce antinucleolar antibodies (ANoA).The ANoA in some strains target the nucleolar protein fibrillarin (Robinson et al., 1986;Reuter et al., 1989;Robinson et al., 1997).Anti-fibrillarin autoantibodies (AFA) in serum are observed in 10% of systemic sclerosis (SSc) patients (Arnett et al., 2000;Sharif et al., 2011), an autoimmune disease with uncertain etiology but with links to environmental agents.AFA are mainly detected in patients with the sever form of SSc (Mora, 2009).
Studies using genetically defined inbred strains of mice have contributed to understanding how xenobiotics modulate the immune system; however, inbred mice do not represent the genetic heterogeneity in human populations.Recently, the use of outbred mice in experimental research has proved advantageous in studying different immunological issues including inflammation (Barone et al., 2018), infection (Martin et al., 2017), and vaccine development (Sunagar et al., 2018).Outbred mice stocks originating from the Swiss mice (Chia et al., 2005) show a high amount of genetic variations comparable to human populations (Rice and Obrien, 1980;Cui et al., 1993).Outbred mice have been exposed to mercury, silver, and silica and their autoimmune reactions have been recorded (Abedi-Valugerdi, 2009;Arefieva et al., 2016;Mayeux et al., 2018).In this study, we more extensively examined autoimmune responses caused by gold or mercury salts using outbred SW mice and studied new target antigens for autoantibodies induced by the two heavy metals.

Mice
In total 31 female SW mice (Taconic, Ry, Denmark) with 10-18 weeks of age at the start of the study were used (ethical # 75-14).The animals were housed in steel-wire cages in temperatureand humidity-controlled environment under 12-hour light-dark cycles and given sterilized food pellets and water ad libitum.All the experiments were performed according to the guidelines for care and treatment of experimental animals recommended by the EU Directive 2010/63/EU and the animal experiments ethical committee in Linköping.

Sample collection
Through retro-orbital bleeding, blood samples were collected from AuTM-exposed and control mice before (week 0) and after 5, 10, and 15 weeks from the start of the exposure.HgCl 2- exposed mice and the corresponding controls were sampled before (week 0) and after 5 weeks of mercury exposure.Sera were separated from the blood samples by centrifugation at 500 g for 15 minutes and stored at -20 ˚C until analysis.Upon sacrifice, tissue from left kidney and spleen were harvested from AuTM-or HgCl2-exposed mice for IC deposition analysis.

Tissue immune complex deposition
To assess immune complex deposition, sections from kidney and spleen were stained using FITC-conjugated goat anti-mouse IgG and IgM antibodies (Southern Biotech.), as well as anti-C3c antibodies (Organon Technica, West Chester, PA) as described previously (Hultman et al., 1992).The titer for IgG and C3c were determined by serial dilution of the antibodies, and the IgM fluorescence was graded semi-quantitatively.

Detection of serum anti-dsDNA antibodies by Crithidia luciliae assay
As previously described (Aarden et al., 1975), Crithidia luciliae slides (Binding Site) were used as substrate to analyze the antibodies to dsDNA.The slides were incubated with sera diluted 1:10, followed by addition of FITC-conjugated goat anti-mouse IgG (Southern Biotech.).Two independent evaluators examined the slides qualitatively by fluorescence microscopy.
NZB/NZW F1 mouse serum was used as positive control.

Investigation of ANA/ANoA specificity by Western blotting
To assess the specificity of serum autoantibodies, Western blotting was performed on sera as previously described (Warfvinge et al., 1995).Briefly, nuclear extract isolated from mouse liver (Chan and Pollard, 1992) were separated by SDS-PAGE (sodium dodecyl sulfatepolyacrylamide gel electrophoresis) using 12.5% and 18% Tris-HCl gels (Bio-Rad).The separated protein bands were then blotted onto 0.45-mm nitrocellulose membranes (Bio-Rad) at 0.8 mA/cm 2 under water cooling (Criterion Blotter; Bio-Rad) for 30 minutes.The membranes were blocked, overnight at 4˚C, using 5% non-fat dry milk (Bio-Rad).The lanes on the membranes were then cut, using a blade, into separate strips and each incubated with diluted mice sera for 60 minutes.Following repeated washing steps, HRP-conjugated goat anti-mouse IgG (Southern Biotech.) was added followed by chemiluminescent HRP substrate (Immobilon Western, Millipore Corporation; Billerica, MA).High performance chemiluminescence films (Amersham Hyperfilm ECL; Amersham, GE Healthcare, Buckinghamshire, UK) were then exposed to the membrane strips and developed with Kodak D-19 developer (Eastman Kodak; NY).Reference sera of histone H1 (AE-4) mouse monoclonal antibody (Santa Cruz Biotech., CA), histone H2A (L88A6) mouse monoclonal antibody (Cell Signaling Technology, Danvers, MA), human fibrillarin (Binding Site), and anti-PM/Scl-100 produced in rabbit (Sigma) were used.

Statistics
The differences between ANA/ANoA positivity rates induced by AuTM or HgCl2 exposure were examined with Fisher's exact test.Mann-Whitney U test was used to determine the differences in serum antibody levels between AuTM-and HgCl2-exposed animals compared to controls.GraphPad (GraphPad Software Inc.; San Diego, CA) was used for statistical analyses.
A p-value < 0.05 was considered statistically significant.

Gold exposure caused development of serum ANA and polyclonal B-cell activation in Swiss
Webster mice After 5 weeks of exposure to AuTM, 40% of the SW mice developed serum ANA with speckled patterns.The ANA response peaked at week 10 with homogeneous, speckled, and rim patterns (Sur et al., 2018).The ANA in the SW mice exposed to AuTM was primarily of IgG2a (90%), IgG2b (50%), and IgG1 (20%) isotypes.However, after 15 weeks of exposure the grading of total IgG ANA decreased, slightly fewer mice were ANA-positive, the ANA pattern switched to rim, and the ANA of IgG2a and IgG2b isotypes were equally present (70 %).The IgG1 isotype appeared seldom (10%) at week 15 (Table 1).Exposure to AuTM for 15 weeks induced in 50% of the SW mice, renal mesangial IC deposits of IgG with only low titers (40 ± 20 (mean ± SEM)).No IC deposition of IgG, IgM, or C3c were detected in the spleen (data not shown).
Serum total anti-DNP and anti-ssDNA antibodies, markers for polyclonal B-cell activation (Havarinasab et al., 2007b), were both elevated after AuTM exposure (Fig. 1).Anti-DNP antibodies showed significantly higher levels in AuTM-exposed mice compared to the corresponding controls starting from week 5 and remained high until week 15.Significantly higher levels of anti-ssDNA antibodies compared to controls were detected at weeks 10 and 15 compared to the corresponding controls.The response in the AuTM-exposed mice were higher with anti-DNP antibodies rather than with anti-ssDNA antibodies.There was a larger variation in titer in the AuTM-exposed mice compared to the controls.(Fig. 1A, B).No significant increase was detected in serum IgG1 or IgG2a levels following exposure to AuTM compared to the corresponding controls.(data not shown).As the SW mice are genetically heterogeneous, their immune response to gold exposure showed variation as evident by higher reciprocal antibody levels against DNP, ssDNA, and chromatin in a subset of mice compared to rest of the mice in the group (Appendix B: Fig. S1).

Gold exposure induced antibodies against chromatin, histones, and dsDNA in Swiss
Webster mice Exposure to AuTM led to significantly higher levels of serum ACA compared to corresponding controls at week 5.The ACA levels continued to rise with highest levels at week 15 (Fig. 1C).
Serum AHA levels analyzed at week 15 showed significantly higher levels after AuTM exposure compared to corresponding controls (Fig. 1D).Among the AuTM-treated mice, one remained unresponsive.Furthermore, Crithidia luciliae assay showed that 4 out of 10 mice exposed to AuTM for 15 weeks developed anti-dsDNA antibodies.
Western blotting showed that sera from mice exposed to AuTM for 15 weeks (Fig. 2, lanes 2-5) reacted with a 32-33-kDa, and a 14-kDa protein corresponding to mouse histone H1 (lane 6) and histone H2A/H2A.X reference sera (lane 7), respectively.Sera collected from mice at pretreatment time and sera from the control mice at week 15 did not react with nuclear antigens (data not shown).

Mercury exposure caused serum ANoA, tissue IC deposition, and polyclonal B-cell activation in Swiss Webster mice
Five weeks of exposure to 8 mg/L HgCl2 caused serum total IgG ANoA with a dense clumpy nucleolar pattern in all mice (Table 2).ANoA of the IgG isotypes IgG1, IgG2a and IgG2b were present in all mice.None of the control mice showed ANA/ANoA after 5 weeks or at pretreatment time.
HgCl2 exposure for 5 weeks led to IC deposition of IgG, IgM, and C3c in renal mesangium and spleen of SW mice.The IC deposition was more prominent in kidneys than in spleens.
Serum levels of anti-DNP and anti-ssDNA antibodies were both significantly increased in mice exposed to HgCl 2 at week 5 compared to the corresponding controls.However, the anti-DNP antibodies levels in the HgCl2-exposed mice rose to high titers compared to the corresponding controls, whereas the increase in anti-ssDNA antibodies was small yet significant (Fig. 3A, B).
The serum IgG1 and IgG2a levels increased after 5 weeks of HgCl 2 exposure (Fig. 3C, D).No significant changes in serum ACA levels were induced by HgCl2 exposure after 5 weeks (data not shown).

Mercury exposure induced serum antibodies corresponding to anti-fibrillarin and anti-PM/Scl-100 antibodies in Swiss Webster mice
Western blotting showed that sera from SW mice exposed to HgCl2 for 5 weeks (Fig. 4, lanes 2, 3, 5 and 6) reacted with a 34-kDa protein corresponding to human anti-fibrillarin antibody (lane 1), and with a 100-kDa protein corresponding to rabbit anti-PM/Scl-100 reference serum (lane 7).Sera collected from mice at pretreatment time and control mice at week 5 did not react with any nuclear antigens (data not shown).

Discussion
In this study we report that genetically heterogeneous (outbred) SW mice exposed to AuTM or HgCl2 showed immune and autoimmune response evident by polyclonal B-cell activation and serum ANA/ANoA development.AuTM exposure induced autoantibodies to chromatin components including histones H1 and H2A, and HgCl2 triggered development of autoantibodies that correspond to anti-fibrillarin and anti-PM/Scl-100 antibodies in SW mice.
While immune response to mercury has been previously shown in outbred mouse stocks (Abedi-Valugerdi, 2009), our study documents this for gold for the first time, and further suggests SW mice as a potential animal model to study metal-induced immune response.
The ANA in AuTM-exposed mice peaked at week 10 and were primarily of the IgG2a isotype.
The ANA staining patterns following exposure to AuTM were mainly homogenous and speckled; this differs from the response previously observed in genetically susceptible strains of mice with H-2 s (B10.S, A.SW, SJL) or H-2 t2 (A.TH) haplotypes which show autoantibodies against nucleolar antigens (ANoA) targeting fibrillarin, when exposed to gold (Havarinasab et al., 2007c;Havarinasab et al., 2009).In these inbred mouse strains (H-2 s , H-2 t2 ) the susceptibility to AFA is genetically restricted to the H-2A locus in mice with the "A" background (Havarinasab et al., 2007b;Pollard and Hultman, 2014).Because of the heterogeneity of SW mice in the H-2 region, the gold-induced autoantibodies may target more than one nuclear substructure such as histones and nucleosome.Antibodies against different nucleosome structures have been reported to play a crucial role in Th-cell-dependent immune response (Liu and Davidson, 2012).However, in the present study we were not able to pinpoint the specificity of ANA in mice following exposure to AuTM.
We performed line immunoassay (LIA) as an initial step to examine the specificity of the autoantibodies induced following exposure to gold and mercury (Methodology available in Appendix A).LIA is an assay for diagnosis of serum autoantibodies in patients.Since many autoantigens show similarities between humans and mice, we tested the ANA by using mouse detecting antibodies to investigate autoantibodies in mice serum.Our screening using LIA indicated serum autoantibodies against nucleosomes (9 out of 10 mice) and dsDNA (5 out of 10 mice) following 15 weeks of AuTM exposure (Table S1).
The homogenous nuclear staining pattern we observed in this study after AuTM exposure is linked to ANA against chromatin, histones, and dsDNA (Dellavance and Andrade, 2019), and is in humans is observed in drug-induced lupus and systemic lupus erythematosus (SLE) patients (Mehra and Fritzler, 2014).This, in addition to positive reaction for antibodies against nucleosomes detected with LIA, prompted us to investigate whether antibodies against chromatin, histones, or dsDNA are induced by gold exposure as they are closely interconnected in nucleosome and chromatin structure (Cutter and Hayes, 2015).Increased serum ACA and AHA levels shown by ELISA, anti-dsDNA antibodies detected by Crithidia luciliae assay, together with antibodies detected with Western blot reacting with nuclear proteins corresponding to histones H1 and H2A, at 32-kDa and 14-kDa respectively, point to an immune response with autoantibodies against chromatin components in SW mice after exposure to AuTM.No or weak tissue IC deposition in inbred mice after exposure to gold (Havarinasab et al., 2007c;Havarinasab et al., 2009) and silver (Hultman et al., 1995) has been documented previously.Similar to inbred mice, the outbred SW mice showed no, or negligible rates of, IC deposition following AuTM exposure.
Development of IgG-ANoA of different isotypes in SW mice following HgCl2 exposure tallies with the effect of mercury in the genetically susceptible strains of mice like A.SW and B10.S (H-2 s ) (Reuter et al., 1989), as well as other outbred mouse stocks (Robinson et al., 1984).This study, for the first time, documents tissue IC deposition after mercury exposure in outbred SW mice.Tissue IC deposition is one of the features of murine mercury-induced autoimmunity (Pollard et al., 2019).Renal and splenic IC deposits in SW mice exposed to HgCl 2 show further concurrence between metal-induced autoimmunity in inbred and outbred mice.
Mercury is known to induce AFA in genetically susceptible mice (Pollard et al., 1997).Our Western blot analyses also showed a band at 34-kDa corresponding to fibrillarin in sera of outbred SW mice exposed to HgCl2 for 5 weeks.This might be a shared feature in the immune response to mercury between genetically susceptible H-2 s mice and outbred SW mice.
Moreover, Western blotting showed antibodies reacting with a ca 100-kDa nuclear antigen which corresponds to PM/Scl-100, to our knowledge not previously detected after mercury exposure in neither inbred nor outbred mice.Both AFA and anti-PM/Scl-100 antibodies may show nucleolar staining pattern on HEp-2 cells, and are associated with SSc and sclerodermapolymyositis overlap syndrome (SSc/PM) in humans (Chan et al., 2015).LIA showed antibodies against PM/Scl-100 (7 out of 8 mice), and antibodies against fibrillarin (3 out of 8 mice) following mercury exposure (Table S1).While a recent study (Hamaguchi et al., 2019) reported no correlation between LIA, indirect immunofluorescence (IIF) and immunoprecipitation (IP) in detecting anti-PM/Scl antibodies in SSc patients, other human studies show a strong correlation between LIA, IIF and IP for detection of PM/Scl-100 antibodies (Ghirardello et al., 2010;Jaskowski et al., 2011;Bonroy et al., 2012).We also detected other bands in the immunoblots in both gold-and mercury-exposed mice, particularly the latter.SW mice are genetically heterogeneous and less investigated from an immunological aspect.The target antigens for the unidentified bands remain to be investigated.
In contrast to IgG1 and IgG2a levels in AuTM-exposed mice that were not affected compared to controls, HgCl2 exposure led to increased levels of both Th1-associated IgG2a (Finkelman et al., 1988) and Th2-associated IgG1 (Kuhn et al., 1991), suggesting that both Th cells are involved in the immune response to HgCl2 in outbred SW mice.The coexistent Th1 and Th2 cell responses to mercury has been previously detected in both inbred (Hu et al., 1999;Havarinasab et al., 2007b) and outbred (Abedi-Valugerdi, 2009;Arefieva et al., 2016) mice.
Exposure to environmental agents in humans usually occurs at low doses and over long periods of time which differs from experimental studies.The AuTM dose in this study (22.5 mg/kg.bw)has previously induced autoimmune response in inbred mice (Havarinasab et al., 2007c;Havarinasab et al., 2009).This dose corresponds to a 4 mg internal dose of gold (4 mg/kg.bw) which is considerably lower compared to the dose in patients who developed nephritis (Cramer et al., 1983) and thrombocytopenia (Adachi et al., 1987).
Our previous investigation of HgCl2 doses in mice showed that 8 mg/L HgCl2 in drinking water is more efficient in triggering autoantibody development compared to lower or higher doses (Havarinasab et al., 2007a).Eight mg/L HgCl2 in drinking water corresponds to an internal mercury (Hg) dose of 148 µg/kg.bwper day which can be compared with the estimated lowestobserved-adverse-effect level (LOAEL) of mercury in humans (230-630 µg Hg/kg.bw/day)(UNEP, 2008).
The genetic heterogeneity of the SW mice is reflected in the varying response levels by each animal, detected in most serum parameters measured, after AuTM (Fig. S1) and HgCl2 (Fig. S2) exposure, with certain mice showing consistently low or high response to the heavy metals compared to the rest of the animals in the groups.
Our data shed more light on the mercury-induced autoimmune response in outbred strains of mice.Furthermore, we show AuTM to be a trigger for systemic autoimmune response in outbred SW mice inducing ANA against several nuclear components.Recent studies have reported autoimmune response in outbred mice following exposure to environmental agents (Arefieva et al., 2016;Mayeux et al., 2018)

Figure 2 .
Figure 2. Immunoblotting of SW mice sera, exposed to AuTM for 15 weeks, using SDS-

Table 1 .
suggesting outbred animals as potential animal models to study how environmental agents affect immune system.Our findings showing that HgCl2 and AuTM trigger development of autoantibodies in SW mice comparable to observations in susceptible inbred mice propose SW mice as a promising animal model to investigate pathobiology and mechanisms for xenobiotic-induced autoimmunity.This warrants further studies on effect of environmental agents on the immune system using SW mice as a more appropriate representative of human populations.AuTM-induced serum antinuclear antibodies (ANA) in SW mice.Percentages indicate the fraction of mice positive for total IgG or each IgG isotype in each group (positive/total).Significantly different compared to week 0: a = p < 0.001; b = p < 0.01; c = p < 0.05 (Fisher's exact test).n: number of mice; ANA: antinuclear antibodies; Sp.: speckled.Grading scale: 0 for negative; 1 for slight positive; 2 for moderate positive; and 3 for strong positive.

Table 3 .
Renal and splenic immune complex (IC) deposits in and HgCl2-exposed SW mice.