Bacterial lipopolysaccharide both renders resistant mice susceptible to mercury-induced autoimmunity and exacerbates such autoimmunity in susceptible mice

Abstract

The initiation and severity of systemic autoimmune diseases are influenced by a variety of genetic and environmental factors, in particular bacterial infections and products. Here, we have employed bacterial lipopolysaccharide (LPS), which non-specifically activates the immune system, to explore the involvement of innate immunity in mercury-induced autoimmunity in mice. Following treatment of mouse strains resistant [DBA/2 (H-2^d)] or susceptible [SJL(H-2^s)] to such autoimmunity with mercuric chloride and/or LPS or with physiological saline alone (control), their immune/autoimmune responses were monitored. Resistant DBA/2 mice were rendered susceptible to mercury-induced autoimmunity by co-administration of LPS, exhibiting pronounced increases in the synthesis of IgG1 and IgE, high titres of IgG1 deposits in the kidneys and elevated circulating levels of IgG1 antibodies of different specificities. Furthermore, the percentages of the T cells isolated from the spleens of DBA/2 mice exposed to both mercury and LPS that produced pro-inflammatory cytokines were markedly increased by in vitro stimulation with phorbol myristate acetate (PMA) and ionomycin, which was not the case for splenic T cells isolated from mice receiving mercuric chloride, LPS or saline alone. In addition, exposure of susceptible SJL mice to mercury in combination with LPS aggravated the characteristic features of mercury-induced autoimmunity, including increased synthesis of IgG1 and IgE, the production of IgG1 anti-nucleolar antibodies (ANolA) and the formation of renal deposits of IgG1. In summary, our findings indicate that activation of the innate immune system plays a key role in both the induction and severity of chemically induced autoimmunity.

Introduction

The underlying aetiologies of many systemic autoimmune diseases, including systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), multiple sclerosis (MS) and systemic sclerosis (SS) or scleroderma, are still unknown. It has been proposed that the aetiology, severity and recurrence of such diseases are multifactorial, involving both exogenous factors, such as infectious organisms and environmental toxins/pollutants, and endogenous hormonal, immunological and genetic factors [1,2]. In agreement with this suggestion is the observation that flare-ups of SLE in patients undergoing remission can often be triggered by infections, UV light, ingestion of certain food-stuffs and/or physical, mental or hormonal stress (reviewed in [3]). Although such phenomena are poorly understood, it is conceivable that infectious organisms induce autoimmune responses by molecular mimicry and disruption of immunoregulation, whereas toxins and/or drugs alter cellular responsiveness and the immunogenicity of self-antigens [3–6].

Among the environmental factors that may contribute to the development and/or exacerbation of autoimmune diseases is the bacterial endotoxin/lipopolysaccharide (LPS), a component of the cell wall of Gram-negative bacteria distributed widely in the environment, as well as inhabiting the digestive tracts of humans and other animals [7–9]. It is now well established that interaction of LPS with the LPS-binding protein (LBP) causes the membrane-bound or soluble form of CD14 and the recently identified Toll-like receptor 4 (TLR4)-MD-2 complex to activate the innate immune system non-specifically, resulting in the production of proinflammatory cytokines together with up-regulation of co-stimulatory molecules [10–12]. Moreover, probably via this same receptor complex, LPS is capable of inducing polyclonal activation of B cells, which leads to the production of antibodies of different specificities, e.g. the production of autoantibodies by murine splenic cells, both in vitro and in vivo[13–16].

It has been proposed that such non-specific activation of innate immunity by LPS enhances the responses of the adaptive immune system to both non-self (i.e. pathogen-related) and self (host-related)-antigens [12,17,18]. Thus, several studies indicate that under certain conditions LPS might initiate and/or accelerate autoimmune processes. For example, immunization of susceptible mouse strains with mouse thyroglobulin in combination with LPS can induce an autoimmune thyroiditis which resembles the analogous syndrome (Hashimoto's thyroiditis) in humans [19]. Moreover, LPS enhances polyclonal activation of B cells and exacerbates nephritis in mouse strains prone to develop lupus, including MRL/lpr/lpr (NZB × NZW)F1 and BXSB mice [20–22].

Mercurial compounds are widely spread environmental pollutants that have been shown to induce a systemic autoimmune reaction in various mammals, and particularly in rodents [23–26]. In susceptible mice, mercury-induced autoimmunity is characterized by CD4⁺ T cell-dependent polyclonal activation of B cells, increased serum levels of immunoglobulin (Ig) G1 and IgE, production of autoantibodies of different specificities (especially anti-nucleolar autoantibodies (ANolA) and formation of deposits of IgG in the kidney [25–27]. Indeed, in most of these respects this autoimmunity resembles the lupus developed by (NZB × NZW)F1 hybrids and BXSB mice [28,29], as well as human systemic lupus erythematosus [30]. In addition, as is the case for the development of lupus in susceptible animals and for human SLE, genetic susceptibility to mercury-induced autoimmunity is determined by both H-2 and non-H-2 background genes [31–35].

Because LPS is known to exacerbate the autoimmune manifestations of lupus in mouse strains prone to develop this condition spontaneously [20–22], the present investigation was designed to test the hypothesis that LPS also influences the frequency and/or severity of mercury-induced autoimmunity in both resistant and susceptible strains of mice. Monitoring of various immunological parameters −including activation of B cells in vivo and T cells in vitro, production of specific autoantibodies, synthesis of various cytokines and formation of renal IgG deposits revealed that LPS is able to render resistant animals susceptible to mercury-induced autoimmunity, as well as to exacerbate such autoimmunity in susceptible mice.

Materials and methods

Animals and treatment with mercuric chloride and/or LPS

Female SJL (H-2^s) and DBA/2 (H-2^d) mice (6–8 weeks old at the beginning of each experiment) were obtained from M&B A/S (M&B A/S, Ry, Denmark) and housed in the animal facilities either at the Department of Immunology, Stockholm University, or at the Department of Medicine, Karolinska University Hospital at Huddinge, with a 12-h dark/12-h light cycle and access to standard laboratory chow and tap water ad libitum. Four groups of each of these strains, each group containing four to five mice, were employed. Groups 1 and 2 were injected subcutaneously (s.c.) with 0·1 ml of a solution containing 0·4 mg HgCl₂ (Merck, Darmstedt, Germany) per ml sterile physiological saline (to give a dose of 1·6 mg/kg body weight) every third day for 4 weeks. Group 2 also received intravenous (i.v.) injections of 50 µg LPS (Escherichia coli, 055:B5; Sigma-Aldrich Sweden, Stockholm) in 0·1 ml sterile physiological saline through their tail veins on the first day of mercury treatment and once again 2 weeks later. The animals in group 3 received LPS alone in the same manner, while the mice in group 4 (the control group) were injected s.c. with 0·1 ml physiological saline alone according to the same schedule employed for mercury treatment. These experiments were approved by the Northern Stockholm Ethical Committee for Animal Experimentation.

Collection of blood, spleens and kidneys and preparation of serum and cell suspensions

Following the 4-week treatment period, the mice were bled by retroorbital puncture under light isofluorane anaesthesia and thereafter killed by cervical dislocation and their spleens and kidneys dissected out aseptically. The blood samples were allowed to clot at 4°C and then centrifuged to obtain the serum, which was stored at −20°C until being assayed for antibody/autoantibody levels. Approximately three-quarters of each spleen was teased apart with forceps in Earle's balanced salt solution (EBSS) in order to obtain single cell suspensions, which were washed three times and resuspended in 5 ml of this same solution for performance of the protein A plaque assay.

The protein A plaque assay

The numbers of splenic cells secreting antibodies belonging to different Ig classes and subclasses were quantified utilizing the protein A plaque assay described by Gronowicz et al. [36], employing rabbit anti-mouse IgM, IgG1, IgG3 (Organon Teknika, Durham, NC, USA) and IgG2b (Nordic Immunological Laboratories, Tillburg, the Netherlands) as the developing reagents.

Quantification of mouse IgE by enzyme-linked immunosorbent assay (ELISA)

Total serum levels of IgE were determined with a sandwich ELISA procedure, as described previously [37]. A rat anti-mouse IgE monoclonal antibody (mAb) (R35-72; Pharmingen, San Diego, CA, USA) was utilized as the ‘capture’ antibody and a biotinylated rat anti-mouse IgE mAb (R35-92; Pharmingen) as the ‘detection’ antibody.

Detection of ANolA by indirect immunofluorescence (IIF)

The levels of IgG1- and IgG2a-type ANolAs in serum samples were determined employing indirect immunofluorescence. For this purpose, HEp-2 cells grown as monolayers on slides (Immuno Concepts, Sacramento, CA, USA) served as the substrate and FITC-conjugated goat anti-mouse IgG1 and IgG2a (Southern Biotechnology, Birmingham, AL, USA) as the visualizing antibodies. The pattern and titres of different ANolAs were subsequently assessed under a Reichard–Jung Polyvar microscope (Vienna, Austria), with serum samples that exhibited no specific green fluorescence at a dilution of 1 : 50 being assigned a value of zero. The highest dilution at which nucleolar fluorescence could still be detected was defined as the titre of the IgG1 and/or IgG2a ANolA.

Activation of T cells in vitro

The remaining quarter of each spleen (see above) was pooled with the corresponding samples from the mice in the same group and single cell suspensions subsequently prepared in complementary RPMI-1640 (cRPMI) (Invitrogen AB, Stockholm, Sweden). These spleen samples had to be pooled in order to obtain adequate numbers of cells for intracellular staining of cytokines and flow cytometry. Erythrocytes in these cell suspensions were lysed with a solution of NH₃Cl (StemCell Technologies Inc. Vancouver, Canada) and the other cells then washed twice in phosphate-buffered saline (PBS)-Dulbecco's medium, re-suspended in 5 ml cRPMI and adjusted to a density of 1 × 10⁶ cells/ml.

To stimulate the T cells, 10⁶ cells from this suspension were plated in duplicates in each well of six-well flat-bottomed plates (BD Labware, NJ, USA) containing 4-ml medium supplemented either with phorbol 12-myristate 13-acetate (PMA) (10 ng/ml, final concentration), ionomycin (1 µg/ml) and brefeldin A (10 µg/ml) (all were purchased from Sigma, Stenheim, Germany) or with brefeldin A alone (control wells). Following incubation for 6 h at 37°C in a humidified atmosphere containing 5% CO₂, the cells were harvested and centrifuged at 510 g for 5 min at 10°C and the pellets thus obtained washed in PBS-Dulbecco's medium (Invitrogen AB, Stockholm, Sweden) by centrifugation. The resulting pellet was fixed by incubation at room temperature for 10–20 min in 1 ml 4% paraformaldehyde. Thereafter, the fixed cells were washed twice in PBS, re-suspended in 1 ml PharMingenStain buffer (FBS) and stored at 4°C until the staining procedure was performed.

Intracellular staining of cytokines and flow cytometry

The mAb employed for detection of surface markers and intracellular cytokines (all purchased from PharMingen, San Diego, CA, USA) are documented in Table 1. Intracellular staining for interleukin (IL)-2, interferon (IFN)-γ and tumour necrosis factor (TNF)-α was performed according to the standard protocols recommended by the BD Pharmingen company. Briefly, the cells were first stained (for 30 min on ice in the dark) for CD4, CD8 and CD19 using the appropriate mAb conjugated with phycoerythrin (PE). Thereafter, the cells were washed twice with staining buffer, permeabilized Perm/Wash™ solution, and stained for intracellular IL-2, IFN-γ and TNF-α using the PE-conjugated mAbs directly specifically against these cytokines (see Table 1).

Table 1.

The monoclonal antibodies used for detection of surface markers and intracellular cytokines.

Raised in	Specificity	Conjugated with	Clone number	Isotype form
Rat	Mouse CD4	FITC1	H129·19	IgG2a
Rat	Mouse CD8	PE2	53–6·7	IgG2a
Rat	Mouse CD4	PerCp3	RM4-5	IgG2a
Rat	Mouse CD3	FITC	17A2	IgG2b
Rat	Mouse CD19	PE	1D3	IgG2a
Rat	Mouse IL-2	PE	S4B6	IgG2a
Rat	Mouse IL-4	PE	BVD4–1D11	IgG2b
Rat	Mouse IDN-γ	PE	XMG1·2	IgG1
Rat	Mouse TNF-α	PE	MP6-XT22	IgG1
Rat	Unknown4	FITC	R3-34	IgG1
Rat	Unknown	PE and FITC	R35-95	IgG2a
Rat	Unknown	PE	R3-34	IgG1
Rat	Unknown	FITC	R35-38	IgG2b

All these monoclonal antibodies were purchased from PharMingen (San Diego, CA).

¹Fluorescein isothiocyanate

²phycoerythrin

³peridinin chlorophyll protein

⁴negative isotype control

All samples were analysed on a FACSCalibur flow cytometer with CellQuest software (BD Immunocytometry Systems). The percentages of cells producing IL-2, IFN-γ and/or TNF-α was calculated as 100 × (the number of cells expressing both the cytokine and CD3/the total number of cells expressing CD3) and the fold-increase caused by in vitro stimulation calculated by dividing this percentage for stimulated cells by the corresponding value for the cells prior stimulation.

Detection of renal deposits of IgG1

The presence of glomerular deposits of IgG1 antibodies was detected by direct immunofluorescence (DIF), as described previously [37]. Briefly, 5 µm-thick cryostat sections of kidney tissue were fixed in acetone and incubated with serial dilutions of fluorescein isothiocyanate (FITC)-conjugated goat anti-mouse IgG1 antibodies (Southern Biotechnology), starting with an initial dilution of 1: 40. When no specific green fluorescence was detected at this initial dilution, the sample was assigned a value of zero. The highest dilution of the antibody at which specific fluorescence could still be seen was defined as the end-point titre for the glomerular deposits.

Enzyme immunoassays

For quantification of serum IgG1 antibodies directed against both self and non-self antigens, micro-ELISA plates (Costar, Cambridge, MA, USA) were first coated with 50 µl single-stranded (ss)-DNA (10 µg/ml in phosphate-buffered saline (PBS); Serva, Heidelberg, Germany), chicken collagen type II (10 µg/ml; Sigma Chemical Co, St Louis, MO, USA) or bovine thyroglobulin (10 µg/ml in 0·035 M carbonate–bicarbonate buffer, pH 9·8; Sigma Chemical Co, St Louis, MO, USA) (self-antigens) or with the non-self antigen trinitrobenzene sulphonic acid coupled to bovine serum albumin (TNP-BSA) (10 µg/ml in 0·035 M carbonate-bicarbonate buffer, pH 9·8; a kind gift from Dr Carmen Fernandez, Department of Immunology, Stockholm University) and then incubated overnight at 4°C. Thereafter, the plates were washed three times with PBS containing 1% Tween-20 (PBS-Tween) and unoccupied binding sites blocked by incubation with 1% BSA in PBS for 2 h at room temperature. Subsequently, the plates were washed again three times with PBS-Tween and 50 µl serially diluted serum samples (in PBS-Tween, starting with a dilution of 1 : 10) added to the wells and the plates again incubated overnight at 4°C. Following three more washes with PBS-Tween, the plates were incubated with alkaline phosphatase-labelled goat anti-mouse IgG1 (diluted 1 : 2000 in PBS-Tween; Southern Biotechnology) for 2 h at 37°C, followed by addition of 50 µl of the phosphatase substrate p-nitrophenyl phosphate (1 mg/ml dissolved in 100 mM diethanolamine-HCl, pH 9·8; Sigma Chemical Co). After a 40-min incubation at room temperature, the OD_{405 nm} was quantified using an ELISA reader (Anthos, Labtec Instrument, Salzburg, Austria).

Expression and statistical analysis of the data

The numbers of cells secreting IgG1 antibody, serum levels of IgE, the titres of IgG1 ANolAs and titres of renal IgG1 deposits are all expressed as means ± standard error (s.e.). Differences between these parameters for the control and treated groups were analysed for statistical significance using analysis of variance (anova), with a P-value of <0·05 being considered statistically significant. All statistical analyses were performed utilizing WinSTAT software (R. Fitch Software, Medina AB, Vänerborg, Sweden). The experiments shown in Figs 1 and 4 are representative of three and those in Fig. 3 and Table 2 are representative of two independent experiments.

Fig. 1.

Administration of bacterial lipopolysaccharide (LPS) during treatment with mercuric chloride renders DBA/2 mice susceptible to mercury-induced autoimmunity. Group 1 (five mice per group) (filled hexagons) received repeated subcutaneous injections of subtoxic doses of mercuric chloride (Hg) for a period of 4 weeks. In addition to similar administration of mercuric chloride, group 2 (filled circles) was injected i.v. with LPS (50 µg per mouse) on the 1st and 15th days of mercury treatment. The two control groups 3 (filled squares) and 4 (open squares) received LPS alone on two occasions or repeated injections of physiological saline, according to the corresponding schedules used for groups 2 and 1, respectively. At the end of the period of treatment (a) the percentages of splenic cells secreting IgG1 antibodies (PFC) were determined employing the protein A plaque assay. In addition (b) the levels of IgE in the serum samples were determined employing a sandwich enzyme-linked immunosorbent assay (ELISA) assay and (c) the kidneys were analysed for the presence of glomerular deposits of IgG1 utilizing a direct immunofluorescence technique, as described in the Materials and methods. The mean ± s.e. values are presented as thin vertical lines. *P < 0·05 for the difference between the two groups indicated. These results are representative of three independent experiments.

Fig. 4.

Exposure to bacterial lipopolysaccharide (LPS) potentiates the development of mercury-induced autoimmunity in susceptible SJL mice. This experiment was the same as that described in the legend to Fig. 1, except that SJL mice were used and serum levels of anti-nucleolar autoantibodies (ANolA) (d) were also determined, employing an indirect immunofluorescence procedure. The mean ± s.e. values are presented as thin vertical lines. *P < 0·05 for the difference between the two groups indicated. These results are representative of three independent experiments.

Fig. 3.

Administration of bacterial lipopolysaccharide (LPS) together with mercuric chloride to DBA/2 mice increases their production of IgG1 antibodies to both exogenous and endogenous antigens. The levels of IgG1 antibodies directed against TNP (a), ssDNA (b), collagen (c) and thyroglobulin (d) in serum samples from the same animals presented in Fig. 1 were determined employing sandwich enzyme-linked immunosorbent assay (ELISA) assays, as described in detail in the Materials and methods. In all cases a serum dilution of 1 : 40 was utilized and the values obtained were within the linear range of the dilution curve. The mean ± s.e. values are presented as thin vertical lines. *P < 0·05 for the difference between the two groups indicated. These results are representative of two independent experiments.

Table 2.

Percentages of the CD3⁺ T cells in the spleens of DBA/2 mice treated with mercuric chloride (Hg) and/or bacterial lipopolysaccharide (LPS) that express interleukin (IL)-2, tumour necrosis factor (TNF)-α and/or interferon (IFN)-γ, with and without in vitro stimulation by phorbol myristate acetate (PMA) and ionomycin.

	Unstimulated splenic T cells*			Stimulated/unstimulated			Stimulated splenic T cells*
Treatment	IL-2	TNF-α	IFN-γ	IL-2	TNF-α	IFN-γ	IL-2	TNF-α	IFN-γ
Saline	3·9 ± 0·7	3·9 ± 1·2	3·5 ± 0·5	14 ± 1·8	50 ± 5	9 ± 0·3	3·6	13	2·6
LPS	7·2 ± 1·5	17 ± 3	16 ± 3·5	15 ± 3·3	37 ± 0·7	14 ± 6	2	2	0·9
Hg	3 ± 0·1	0·7 ± 0·4	1 ± 0·4	13·5 ± 0·6	34 ± 1·3	14 ± 0·4	4·5	49	14
LPS + Hg	0·3 ± 0·2	0·4 ± 0·1	0·4 ± 0·2	16 ± 0·9	58 ± 3·5	13 ± 0·8	53	145	33

^*The percentages of T cells producing IL-2, TNF-α and/or IFN- γ, with and without in vitro stimulation by PMA and ionomycin, in pooled spleen samples from the same groups of animals presented in Fig. 1. These cytokines were detected employing intracellular staining and flow cytometry, as described in Materials and methods. The percentage of cytokine-producing cells was calculated as: 100 × (the number of CD3⁺ cells producing the cytokine)/(the total number of CD3⁺ cells). These results are representative of two independent experiments.

Results

Administration of LPS during treatment with mercuric chloride renders resistant DBA/2 mice susceptible to mercury-induced autoimmunity

We and other investigators have demonstrated that DBA/2 (H-2^d) is the only inbred strain of mice that exhibits no immune/autoimmune response at all upon exposure to mercury [31–34]. This was confirmed here by the observations that the percentages of splenic cells producing IgG1 antibodies, serum levels of IgE and renal deposits of IgG1 (Figs 1 and 2), as well as the production of circulating IgG1 antibodies directed against various antigens (Fig. 3) were not significantly different in DBA/2 mice treated with mercuric chloride or saline alone. Nor did treatment of these same animals with LPS alone exert any statistically significant effect on any of these parameters (Figs 1, 2 and 3).

Fig. 2.

Immunofluorescent detection of glomerular deposits of IgG1 in the kidneys of DBA/2 mice administered mercuric chloride in combination with lipopolysaccharide (LPS). The kidneys from the same animals presented in Fig. 1 were tested for the presence of glomular deposits of IgG1 employing the direct immunofluorescence procedure described in the Materials and methods. Representative kidney sections from animals administered saline (a), LPS (b) or mercury (c) alone show no green fluorescence intensity; whereas the kidney of a DBA/2 mouse treated with mercury in combination with LPS (d) displays intense green fluorescence in the glomeruli.

In sharp contrast, DBA/2 mice treated with mercury in combination with LPS exhibited increased percentages of splenic cells secreting IgG1 antibodies (Fig. 1a), elevated serum levels of IgE (Fig. 1b) and high titres of IgG1 deposits in their renal glomeruli (Figs 1c and 2d). Moreover, this combined treatment also enhanced serum levels of specific IgG1 antibodies in DBA/2 mice exposed to both exogenous (TNP) and endogenous (ssDNA, collagen and thyroglobulin) antigens (Fig. 3a–d). However, administration of mercury and LPS to these mice did not result in a significant increase in the serum levels of ANolA (not shown).

Cytokine production by splenic T cells isolated from DBA/2 mice treated with mercuric chloride and/or LPS

T cells, and in particular CD4⁺ T-cells, together with the cytokines they produce have been shown to play crucial roles in the development of mercury-induced autoimmunity [24–27,38]. Therefore, in attempt to elucidate the possible mechanism(s) by which LPS renders DBA/2 mice susceptible to mercury-induced autoimmunity, a separate experiment was performed to analyse the in vitro cytokine production by splenic T cells (with and without stimulation) isolated from DBA/2 mice subjected to different treatments. As shown in Table 2, the spleens of animals treated with saline or mercury alone contained intermediate and comparable numbers of T cells producing IL-2, TNF-α and/or IFN-γ. In these two cases, in vitro stimulation caused a similar increase in the percentage of cells producing IL-2, but more pronounced increases in the percentages of cells producing TNF-α and/or IFN-γ following treatment with mercuric chloride alone (Table 2). On the other hand, the spleens of animals administered LPS alone contained a higher percentage of T cells producing IL-2, TNF-α and/or IFN-γ than did the spleens of the mice in the other groups (Table 2). However, a 6-h in vitro exposure of these cells to PMA/ionomycin enhanced only the percentages of cells producing IL-2 and/or TNF-α and this effect was slight (Table 2).

In contrast, the spleens of mice treated with Hg in combination with LPS exhibited the lowest percentages of T cells producing all of the cytokines examined here (Table 2). At the same time, in vitro stimulation of these cells led to a more potent elevation in the percentages of cells producing IL-2, TNF-α and/or IFN-γ than was observed in the case of any of the other groups (Table 2).

Treatment with LPS potentiates mercury-induced autoimmunity in susceptible SJL mice

As expected, 4 weeks of treatment with mercury alone induced an immune/autoimmune response in susceptible SJL mice, characterized by increased percentages of splenic cells secreting IgG1 antibodies (Fig. 4a), elevated serum levels of IgE (Fig. 4b), increased titres of renal IgG1 deposits (Fig. 4c) and high serum titres of IgG1 ANolA (Fig. 4d). Treatment with LPS alone did not alter these parameters in comparison with control administration of saline alone (Fig. 4a). However, the SJL mice that received both mercury and LPS exhibited even higher levels of all four of these manifestations of autoimmunity than did animals injected with mercuric chloride alone (Fig. 4a–d). In contrast, LPS had no influence on the elevation in the percentages of splenic cells secreting IgM, IgG2b and/or IgG3 antibodies or the enhancement of serum levels of IgG2a ANolA elicited by administration of mercury alone (data not shown).

Discussion

In the present investigation we have demonstrated that while, as expected, treatment with mercuric chloride or LPS alone did not elicit an immune/autoimmune response in DBA/2 mice, combined treatment renders these animals highly susceptible to mercury-induced autoimmunity. Indeed, our observation that most of the characteristics of mercury-induced autoimmune developed potently in the mice subjected to combined treatment indicates that under certain circumstances two or more environmental factors can act in concert to break self-tolerance and initiate the development of autoimmunity, even in resistant individuals. Our findings also suggest that activation of the innate branch of immune responses by bacteria or their products plays a crucial role with regard to this synergistic effect. Interestingly, two types of observations support this suggestion. First, it has been shown that pretreatment with broad-spectrum anti-microbial drugs including tylosin, ivermectin and metronidazole is able, to a large extent, to prevent mercury-induced necrotizing vasculitis in susceptible Brown Norway rats (39). Secondly, it has been demonstrated that exposure of transgenic B10.S mice (which lack the T cell receptor for the peptide derived from the encephalitogenic myelin proteolipid protein (PLP) and are thus resistant to experimentally provoked autoimmune encephalomyelitis (EAE-resistant) to LPS results in activation of TLR-4, the breakdown of self-tolerance and induction of EAE [40].

Production of ANolA, the hallmark of mercury-induced autoimmunity [25–27], has been shown to be strictly controlled by H-2 genes, i.e. mouse strains carrying H-2^s, H-2^q or H-2^f genes produce high levels of ANolA after treatment with mercury, where strains with H-2^d, H-2^b or H-2^z do not [25–27,31–34]. Here, we found that although exposure of mercury-treated DBA/2 mice (which carry the H-2^d gene) to LPS triggers the production of autoantibodies of several specificities, LPS is not able to elicit the synthesis of ANolA in these animals. Thus, treatment with LPS apparently cannot overcome H-2-associated resistance to mercury-induced autoimmunity in mice.

Certain aspects of the mechanism(s) by which LPS renders DBA/2 mice susceptible to mercury-induced autoimmunity are addressed by our present studies on in vitro cytokine production by both stimulated and unstimulated splenic T cells. We found that the spleens of DBA/2 mice receiving LPS alone contain the highest percentages of T cells which, without stimulation, produce mainly inflammatory (IL-2, IFN-γ, TNF-α) cytokines; but that these percentages are relatively unaltered by in vitro priming with PMA/ionomycin. These findings indicate that in vivo, LPS alone and in the absence of antigen can stimulate murine T cells to produce different types of cytokines and simultaneously reduce their capacity to respond to mitogens. This conclusion is supported by a number of reports that, either directly [41,42] or indirectly via activation of accessory cells [43,44], LPS can activate T cells to proliferate and produce cytokines.

It has been demonstrated that in connection with an antigen-specific T cell response, administration of LPS promotes the migration of antigen-specific T cells into B cell-rich follicles, where they participate in antibody production [45,46]. Moreover, the antigen-experienced T cells that remain in the lymphoid tissues for several weeks following exposure to an antigen and an adjuvant such as LPS exhibit the functional characteristics of memory T cells [47], i.e. upon activation in vitro, these cells rapidly produce high levels of a broad spectrum of both pro- and anti-inflammatory cytokines [47,48]. Similarly, our present observation that splenic T cells isolated from DBA/2 mice treated with mercury and LPS and subsequently primed in vitro produce large amounts of different types of cytokines, including IL-2, TNF-α, and IFN-γ, suggests that these T cells also demonstrate functional characteristics reminiscent of those of memory T cells. Therefore, it appears likely that co-administration of LPS with mercury enhances the generation and activation of T cells specific for self-antigens altered by mercury and that, after aiding autoreactive B cells, some of these cells are transformed into memory T cells. Clearly, further study is required to test this hypothesis.

Furthermore, we observed that in vitro priming of splenic T cells isolated from DBA/2 mice treated with mercury alone causes larger increases in the percentages of these cells expressing TNF-α and/or IFN-γ than was the case for splenic T cells from untreated mice, but that the increase in IL-2 was similar. These results suggest that even in these resistant mice, mercury alone can activate certain T cells, which can develop further into memory-type T cells with a low capacity for production of IL-2. This proposal receives strong support from the findings that exposure of mice to antigen alone, in the absence of adjuvant, leads to the development of a few antigen-specific CD4⁺ T cells that remain in the lymphoid tissue for a long period of time and exhibit a phenotype characteristic of memory cells [49,50]. However, unlike the memory cells produced in response to administration of antigen together with adjuvant, these antigen-experienced T cells do not produce IL-2 or effector lymphokines [49,50].

Thus, it is possible that in the resistant mice and in the absence of non-specific stimuli (here, LPS) alteration of the presentation of self-antigens by mercury is inefficient, because this presentation is performed by dendritic cells that do not express optimal levels of co-stimulatory molecules. This situation, together with a lack of inflammatory cytokines, leads to the activation of specific T cells with low proliferative capacity, pronounced susceptibility to apoptosis and the ability to develop into functionally defective memory cells. In contrast, the presence of LPS converts tolerogenic dendritic cells into autoimmune dendritic cells, which are capable of activating nontolerant CD4⁺ T cells.

Interestingly, it has recently been shown that non-specific stimuli, including LPS, are not able to break cross-tolerance to self-antigens in autoreactive cytotoxic T (CD8⁺) cells, unless help from specific autoreactive CD4⁺ T cells is provided [51]. Although, as the authors have suggested, the exact underlying mechanisms for this co-operation between ‘autoreactive CD8⁺ and CD4⁺ T cells and non-specific stimuli require further investigations’[51], this interesting finding, together with our present observations, suggest that autoreactive CD4⁺ T cells are more prone than autoreactive CD8⁺ T cells to activation by non-specific stimuli.

Finally, we also show here that exposure of SJL mice (which are susceptible to mercury-induced autoimmunity) to LPS twice during the course of mercury treatment potentiates most, if not all the immune/autoimmune responses elicited by mercury, including the formation of IgG1, synthesis of IgE, development of renal deposits of IgG1 and production of IgG1 ANolA. Such LPS-induced exacerbation of mercury-induced autoimmunity in SJL was not strain-dependent, as we have made identical observations in another mercury-susceptible mouse strain, A.SW (H-2^s) (data not shown). Thus, these findings are in agreement with other reports that LPS exacerbates autoimmune manifestations in mouse strains genetically prone to develop a lupus-like autoimmune disease [20–22].

Although the mechanism(s) underlying this potentiation of mercury-induced autoimmunity by LPS remains to be elucidated, several lines of evidence indicate that signals provided by co-stimulatory surface molecules play a key role in this connection. For instance, like any other adaptive immune response, mercury-induced autoimmunity has been shown to be dependent on the expression of co-stimulatory molecules such as CD40, CD80 (B7-1) and CD86 (B7-2) on professional antigen-presenting cells, including macrophages, dendritic cells and activated B cells [52–54]. On the other hand, it is now well established that by interacting with the TLR-4-MD-2 complex and CD14 on professional antigen-presenting cells, LPS initiates TLR signalling, which exerts adjuvant effects on adaptive immune responses [55,56].

This signalling involves various specific adaptor proteins, including MyD88 (myeloid differentiation factor 88), MAL (MyD88 adaptor-like), Trif (Toll receptor-associated activator of interferon) and TEAM (Toll receptor-associated molecule), as well as recruitment of the Tank-binding protein kinase-1 ((TBK-1), which leads eventually to the production of IFN-α/β[55,56]. Synthesis of IFN-α/β would then up-regulate the expression of co-stimulatory molecules on antigen-presenting cells [57], thereby generating robust antigen-specific T and B cells [55,56]. Therefore, it is conceivable that in SJL mice treated with mercury alone, antigen-specific T cells are activated primarily by dendritic cells, which constitutively express both MHC and co-stimulatory molecules; whereas in animals administered mercury in combination with LPS, other antigen-presenting cells such as macrophages and B cells also express co-stimulatory molecules and are thus capable of activating T cells and thereby potentiating the immune/autoimmune responses induced by mercury.

Considered together, our present findings reveal that activation of the innate immune system plays a key role in induction and exacerbation of mercury-induced autoimmunity. This conclusion supports the hypothesis that environmental risk factors, such as infections, can either initiate and/or accelerate autoimmune processes. Accordingly, we suggest that simultaneous exposure to certain environmental factors can cause individuals who are normally genetically or immunologically resistant to become susceptible to autoimmune diseases.