Diet Influences Expression of Autoimmune Associated Genes and Disease Severity By Epigenetic Mechanisms in a Transgenic Lupus Model

Abstract

Objective

Lupus flares when genetically predisposed people encounter appropriate environmental agents. Current evidence indicates that the environment contributes by inhibiting T cell DNA methylation, causing overexpression of normally silenced genes. DNA methylation depends on both dietary transmethylation micronutrients and Erk-regulated DNA methyltransferase 1 (Dnmt1) levels. We used transgenic mice to study interactions between diet, Dnmt1 levels and genetic predisposition on the development and severity of lupus.

Methods

A doxycycline-inducible Erk defect was bred into lupus-resistant (C57BL/6) or lupus-susceptible (C57BL/6xSJL) mouse strains. Doxycycline treated mice were fed a standard commercial diet for eighteen weeks then switched to diets supplemented(MS) or restricted(MR) intransmethylation micronutrients. Disease severity was assessed by anti-dsDNA antibodies, proteinuria, hematuria and histopathology of kidney tissues. Pyrosequencing was used to determine micronutrient effects on DNA methylation.

Results

Doxycycline induced modest levels of anti-dsDNA antibodies in C57BL/6 mice and higher levels in C57BL/6xSJL mice. Doxycycline-treated C57BL/6xSJL mice developed hematuria and glomerulonephritis on the MR and standard but not the MS diet. In contrast C57BL/6 mice developed kidney disease only on the MR diet. Decreasing Erk signaling and methyl donors also caused demethylation and overexpression of the CD40lg gene in female mice, consistent with demethylation of the second X chromosome. Both the dietary methyl donor content and duration of treatment influenced methylation and expression of the CD40lg gene.

Conclusions

Dietary micronutrients that affect DNA methylation can exacerbate or ameliorate SLE disease in this transgenic murine lupus model, and contribute to lupus susceptibility and severity through genetic/epigenetic interactions.

Introduction

Lupus afflicts approximately 1.5 million Americans, 90% of whom are women [1]. Systemic lupus erythematosus (SLE) affects many organs including joints, skin, kidneys, heart, lungs, blood vessels and the brain. Disease ensues when abnormally functioning B and T lymphocytes form auto-antibodies to DNA and nuclear proteins, resulting in immune complex deposition that causes inflammation and tissue damage. While the cause(s) of SLE are unknown, its etiology involves genes that confer susceptibility, as well as hormones and environmental factors (1, 2). Evidence for a genetic contribution comes from familial clustering of lupus casesin which siblings of lupus patients have a 10-20 fold higher risk than the general population in developing SLE, a higher concordance rate in monozygotic twins (20%) compared to dizygotic twins (2%), and known lupus-associated polymorphisms in genes among which are those encoding HLA molecules, complement components, cytokines, and programmed cell death proteins (3).

The discordance of SLE between monozygous twins suggests that non-genetic factors may influence gene expression, triggering lupus(4). However, what these agents are and how they interact with the various predisposing genetic loci to induce lupus are unclear. DNA methylation and histone modifications regulate gene expression through epigenetic mechanisms(5). Drugs such as 5-azacytidine, procainamide andhydralazine as well as UV light trigger lupus-like autoimmunity through their effects on DNA methylation, resulting in autoreactive T cells that promote autoimmunity (2). Reduced Dnmt1 activity causes hypomethylation and overexpression of the immune genes including ITGAL (CD11a), TNFSF7 (CD70), KIR genes and CD40LG in T lymphocytes (2, 6).

Erk pathway signaling is an important Dnmt1 regulator and Erk signaling is inhibited in T cells by hydralazine and in T cells from patients with idiopathic lupus (2, 7, 8). Therefore, environmental agents that inhibit Erk signaling, its upstream regulator PKC-δ, or other conditions such as diet and aging that can decreaseDnmt1enzymatic activity may increase methylation-sensitive gene expression through epigenetic mechanisms to cause a lupus-like disease in genetically predisposed individuals (2, 9, 10).

Diet is an important environmental component and influences gene expression in vivo. Diets rich in methyl donors, administered to pregnant mice, can alter DNA methylation patterns and gene expression in developing embryos (11, 12). Furthermore, dietary methyl donor supplementation can increase total genomic d^mC content in leukocyte DNA(13) while dietary restriction of methyl donors leads to DNA hypomethylation in vivo (14). Lupus patients have significantly reduced levels of methylation-associated micronutrients.(15, 16). We therefore tested the hypothesis that dietary micronutrients necessary for transmethylation would influence lupus disease severity. We have previously developed a transgenic mouse model with an inducible T cell Erk pathway signaling defect that results in demethylation and overexpression of methylation-sensitive genes, causing the development of lupus-like autoimmunity in the female mice (17). The present study uses this model to study the interaction of genes and micronutrients as a potential environmental influence on SLE disease activity and severity. We examined the effect of methyl donor-restricted (MR) and methyl donor-supplemented (MS) diets on the expression of methylation-sensitive T cell genes and lupus disease using mice with the inducible T cell DNA methylation defect on a lupus resistant (C57BL/6), or lupus susceptible (C57BL/6×SJL) hybrid genetic background.

MATERIALS and METHODS

Animals

SJL/J mice were purchased from Jackson Laboratories (Bar Harbor, ME). C57BL/6 mice bearing the TRE-containing dominant-negative MEK (dnMEK) transgene were bred to C57BL/6 mice containing the reverse tetracycline transactivator under the control of the CD2 promoter (CD2-rtTA). Double transgenic (dnMEK+/CD2rtTA+) mice inducibly express a dominant-negative MEK selectively in T lymphocytes in the presence of doxycycline (DOX), leading to ~60% reduction in Erk phosphorylation(17). In the absence of either transgene, DOX administration fails to reduce Erk phosphorylation.

Double transgenic female mice with the following genetic backgrounds and characteristics were generated for the present study:

• P0: C57BL/6(dnMEK+/CD2rtTA+); anti-dsDNA+, lupus nephritis negative (17).

• F1: (C57BL/6×SJL)F1 (dnMEK+/CD2rtTA+); anti-dsDNA+, lupus nephritis positive(17, 18).

• F2: (F1×SJL)F2 (dnMEK+/CD2rtTA+);this study.

The animals were housed in filter-protected cages and provided with standard, irradiated 5053 (Lab Diet, PMI Nutrition International, Brentwood, MO), and water ad libitum. Four mg/ml DOX (Sigma, St. Louis, MO)/5% glucose was administered in the drinking water of selected groups of mice. Protein and hemoglobin in mouse urine were measured by Chemstrip 7 dipstick (Roche, Madison, WI). All mice were bred and maintained in a specific pathogen-free facility by the Unit for Laboratory Animal Medicine at the University of Michigan in accordance with National Institutes of Health and American Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC) International Guidelines. All procedures were approved by the University of Michigan Institutional Animal Care and Use Committee.

Diets

Diets were selected to represent a range of DNA transmethylation micronutrient concentrations. The concentrations of methyl donors and co-factors were based on the micronutrient content of the diets used by Hollingsworth et al. and Delaney et al. (19, 20). Amino acid defined MR (TD.06688) or MS(TD.06690) in the transmethylation micronutrients and co-factors listed in Table 1 were provided by Harlan Laboratories Inc., Madison, WI. Mineral and vitamin premixes were AIN-93M (mineral mix) and AIN-93. The MR diet has low methionine (0.15%) and moderate cysteine (0.25%) and methyl-related nutrients (choline, folate, B12, B6) that are within ranges typically found in other purified or standard diets, including 5053. The MS diet has higher methionine (1.18%) than standard diets, moderate cysteine (0.25%) andis supplemented with the methyl donors choline and betaine. It also contains specific increases in vitamins B12, folic acid, and zinc compared to the standard 5053 and MR diets. The MS diet was approximately eight fold higher in methionine, 14 fold higher in choline, five times higher in folic acid, twenty-five times higher in vitamin B12 compared to MR diet. The standard natural ingredient rodent diet5053, had intermediate levels of methionine but was otherwise similar to the MR diet for other methylation-associated micronutrients.

Table 1.

Transmethylation micronutrient concentrations in mouse diets

	MS Diet #06690	MR Diet #06688	Standard Diet 5053
Methyl Donorsa
Betaine	15	0	Unknown
Methionine	11.8	1.5	7
Choline	16.5	1.15	2
Methyl Cofactorsb
Zn	200	36	87
Folic Acid	16.5	3	3
Vitamin B2	9	9	8
Vitamin B6	8.6	8.6	9.6
Vitamin B12	1.5625	0.0625	0.051

^ag/kg

^bmg/kg

Flow Cytometry

Spleen cells were washed twice in Standard Buffer (PBS containing 1% horse serum and 1mg/ml sodium azide), 4°C. Non-specific binding was blocked by incubating the cells 1 hr on ice in Standard buffer containing 10% horse serum. The cells were then stained in the dark for 1 hr with PE-Hamster anti-mouse CD154 (CD40L), PE-Cy5-rat anti-mouse CD4, or anti-CD11a (BD Pharmingen, Fullerton, CA), washed, then fixed in 2% paraformaldehyde and stored in the dark at 4°C. The cells were analyzed using a FACSCalibur flow cytometer (BD Biosciences, Franklin Lakes, NJ) as previously described (21).

ELISA

Mouse anti-dsDNA IgG antibodies were measured by ELISA as previously described (18). Briefly, Costar (Corning, NY) 96 well flat bottom microtiter plates were coated overnight at 4°C with 10µg/ml dsDNA in PBS, pH 7.2. Various dilutions of mouse sera or murine monoclonal IgG anti-dsDNA antibody (Clone BV16-13, Millipore, Billerica, MA) standard were added and incubated overnight at 4° C. Bound anti-dsDNA antibody was detected using HRP-goat anti-mouse IgG-Fc-specific (Bethyl Labs, Montgomery, TX) antibody and One Step Ultra TMB substrate (Thermo, Rockford, IL) and measured at 450nm.

Bisulfite Conversion and Pyrosequencing

Genomic DNA was isolated from CD4⁺ T cells using the DNeasy blood and tissue kit (QIAGEN), and then bisulfite treated using the EZ DNA Methylation-Gold kit (ZYMO Research, Irvine, CA) according to the manufacturer’s instructions. Pyrosequencing primers for murine CD40lg were designed using PSQ Assay Design software (Biotage, Uppsala, Sweden). We have previously described their sequences and the PCR running conditions for this gene (18).

Statistical Analysis

Student’s t-testChi-square two-tail Fisher Exact test, and linear regression were used as appropriate to determine the significance of differences between groups using SYSTAT software on a Dell PC Optiplex 745 microcomputer. Statistical analyses were also performed by Yebin Tao, Dept. of Biostatistics School of Public Health, and University of Michigan.

Results

Influence of genetic background on anti-dsDNA antibody

Inducing dnMEK expression by DOX treatment in the double transgenic mice decreases Erk phosphorylation and Dnmt1 mRNA, resulting in increased expression of the methylation-sensitive genes Cd11a, Cd40lg and Cd70, and induced anti-dsDNA antibodies(17, 18). The genetic background of mice confers additional factors that influence lupus disease susceptibility and severity (22). We therefore compared the contribution of C57BL/6 (H2^b) and SJL (H2^s) genetic backgrounds for their influence on anti-dsDNA IgG antibody titers in mice with the dnMEK and CD2rtTA transgenes. Female C576BL/6, (B6 X SJL)F1, and (F1 X SJL)F2 mice hemizygous for the dnMEK and CD2rtTA transgenes were given drinking water containing 4 mg/ml DOX/5% sucrose or 5% sucrose alone. Their IgG anti-dsDNA antibody responses after 18 wks of treatment are shown in Figure 1. Transgenic (B6×SJL)F1 mice had significantly (p=0.03, Student’s t-test) higher anti-dsDNA antibody levels than mice with the transgenes on the pure C57BL/6 background. Further increasing the SJL contribution to the genetic background by a second backcross onto SJL significantly (p=0.04 F1 vs F2; P0 vs F2 p=0.001) increased the amount of anti-dsDNA IgG antibody produced. In the absence of DOX treatment, no anti-dsDNA antibody was observed.

Figure 1.

Effect of Genetic Background on anti-dsDNA IgG Antibody. Mice bearing the dnMEK and CD2rt-TA transgenes on the indicated genetic backgrounds were treated for 18 weeks with DOX in their drinking water. Effect of increasing SJL genetic background contribution, P0 vs F1, p=0.03; F1 vs F2 p=0.042; P0 vs F2 p<0.001 (Student’s T-test). No significant differences (p>0.05) in anti-dsDNA antibody were observed between groups in the absence of DOX treatment. Values within the columns are numbers of mice per group.

Although DOX-treated C57BL/6 transgenic mice produce anti-dsDNA antibody, they fail to develop lupus-like organ damage and disease (17). In contrast, SJL mice possess genes that contribute to lupus-like disease when Erk activity is impaired, but they do not spontaneously develop lupus in the absence of DOX activation of the transgene(17, 18). We therefore investigated the effect of dietary transmethylation micronutrients on the epigenetic regulation of lupus susceptibility genes and disease in these strains. Double transgenic C57BL/6 and (F1 X SJL)F2 mice were fed the standard natural ingredient rodent diet 5053 and given DOX in their drinking water, as previously described (18). After 18 weeks, half the mice were fed a MR diet while the remaining animals were fed MS and DOX treatment continued. Anti-dsDNA antibody levels significantly (p=0.006 by linear regression) declined in (F1 X SJL)F2 mice switched to the MS diet and were near background levels eight weeks later (Figure 2). In contrast (F1 X SJL)F2 mice fed the MR diet showed a rise in anti-dsDNA antibody but the change was not statistically significant (p>0.05).

Figure 2.

Anti-dsDNA IgG Antibody Levels Decline in DOX-Treated Transgenic Mice Fed the MS Diet. Standard diet (5053), wks 14–18 vs MS diet wks 20–26, p=0.006 (linear regression); Standard diet vs MR diet p=0.368 (linear regression). MS vs MR diet, wk 26, p=0.01(Student’s t-test). Values are Mean +/− SEM of groups of 5–10 (F1×SJL)F2 female mice.

Diet and Hematuria

The effect of diet on the development of kidney disease was investigated. We previously observed glomerulonephritis and hematuria in this transgenic mouse lupus model (18). As expected, DOX-treated transgenic C57BL/6 mice failed to develop hematuria when fed Standard diet 5053(Fig. 3). They also failed to develop hematuria when fed the MS diet. However, six of eighteen DOX-treated, transgenic C57BL/6 developed hematuria when maintained on the MR diet. Five of the six mice that developed hematuria had 250 erythrocytes/µl. Four of the five animals that developed hematuria also had 30–50mg/dL protein and one had a trace amount of protein in their urine. One mouse had 100 erythrocytes/µl (minimum detectable levels <50 erythrocytes/µl) and a trace amount of protein in their urine. The effect of DOX treatment and the MR diet on hematuria in the C57BL/6 transgenic mice was statistically significant (p=0.01). Two of fourteen DOX-treated transgenic (F1 X SJL)F2 mice fed standard diet developed hematuria, with 50 erythrocytes/µl and 500 erythrocytes/µl with trace and 30mg/dL amounts of proteinuria respectively. However, none of the (F1 X SJL)F2 mice fed the MS diet developed hematuria. Two out of five of the (F1 X SJL)F2 transgenic, DOX-treated mice fed MR diet had 100 erythrocytes/µl 30mg/dL protein and 500 erythrocytes/µl, with 100mg/dL protein respectively. In the absence of DOX no hematuria developed in either strain with any of the diets used (not shown). Increasing the dietary methyl donor content reduced the development of hematuria significantly (p=0.01 by Chi-Square Trends in Proportions) in transgenic mice with SJL genes.

Figure 3.

Influence of Diet on Kidney Function. C57BL/6 or (F1×SJL)F2 mice bearing the dnMEK/CD2rtTA transgenes were treated with DOX in their drinking water and the diets indicated for up to 18 wks. The effect of diet on hematuria was significant (p=0.01, Chi-Square for Trends in Proportions statistical test) in both strains of mice.

The semiquantitative scoring system described by Austin et al. (23)to measure renal disease in lupus patients was used to assess kidney damage in our mice. Light microscopic examination of paraffin sections of kidneys from DOX-treated mice confirmed that the transgenic C57BL6 mice did not develop kidney disease when fed the Standard diet (Fig. 4A), score 0 +/− 0 The glomeruli of these animals were normal and exhibited open capillary loops (arrow) and no increase in cellularity. C57BL/6 animals fed MR diet had focal glomerular hypercellularity with an increase in mesangial matrix (Fig. 4B, arrow average score 4.5 +/− 0.5 range 4–5). The SJL background contains lupus susceptibility genes which contributes to both the development and severity of glomerular inflammation. In (F1xSJL)F2 animals fed the Standard diet, there was mild focal glomerulonephritis with hypercellularity and increased in mesangial matrix (Fig. 4C, arrow average score 3.6 +/− 1.5, range 1–7). (F1 xSJL)F2 animals fed the MR diet had more severe glomerulonephritis than the C57BL/6 mice, with a marked increase in diffuse glomerular hypercellularity and matrix deposition with karyorrhectic nuclear debris (Fig. 4D, dark arrow), and thickening of the glomerular capillary loops consistent with subendothelial deposits (light arrow, average score 5.8 +/− 2.0, range 2–12). In the absence of DOX treatment no animals developed kidney disease (not shown). Taken together, our data indicate that transmethylation micronutrients, particularly methyl donors such as methionine and betaine, can influence lupus-like disease symptoms such as anti-dsDNA antibody and hematuria in our murine transgenic model of SLE.

Figure 4.

Diet and Glomerulonephritis. Formalin-fixed kidneys from transgenic DOX-treated C57BL/6 (P0) and (F1×SJL)F2 mice were routinely processed, stained with H&E and examined by light microscopy for evidence of disease. (A), C57BL/6 mice fed a Standard commercial diet (5053) show normal glomeruli with no increase in cellularity and open capillary loops (arrow) 0 +/− 0, lupus scoring system. (B), C57BL/6 mice fed MR diet. Increased mesangial matrix (arrow). (Example shown is lupus score 4, group average 4.5 +/− 0.5 range 4–5). (C), (F1×SJL)F2 mice fed a Standard diet (5053). Increased mesangial matrix (arrow). (Example shown is lupus score 4, group average 3.6 +/− 1.5, range 1–7). (D), (F1×SJL)F2 mice fed MR diet. Karyorrhectic nuclear debris (dark arrow) and thickening of the glomerular capillary loops consistent with subendothelial deposits (light arrow) are seen, and are more severe than in the C57BL/6 mice (Example shown is lupus score 12, group average 5.8 +/− 2.0, range 2–12).

CD40lg gene expression and methylation

CD40L protein on T cells is elevated in women with lupus and in our transgenic lupus mouse model, and contributes to disease pathogenesis by stimulating B cell antibody production (24, 25). The CD40L gene (CD40LG) is on the X chromosome in both humans and mice and its aberrant demethylation on the inactive X may thus be a determinant of lupus predominance in females compared to males. Increased CD40LG expression with decreased methylation in females but not males has been demonstrated in both species (1, 18, 25). Therefore, the effect of diet on murine CD40lgexpression and methylation was examined. CD4+ T lymphocytes from female transgenic, DOX-treated C57BL/6 mice fed the MR diet overexpressed CD40L protein relative to mice fed the MS diet (MFI 3089+/− 466 vs 864+/− 16, MR vs MS, p=0.017), Fig. 5A. Impaired Erk signaling or inhibition of Dnmt1 contributed to increased CD40L protein on female CD4+ T cells by inhibiting the methylation of CG pairs in the promoter region near the transcription start site (TSS)(18, 25). Therefore the methylation of CG pairs in this region was measured. After 12 weeks of DOX treatment, C57BL/6 transgenic mice fed MR diet had reduced methylation of CG pairs located -46, -43 and -35 5’ to the TSS of the CD40lg promoter compared with mice fed the MS diet (Fig. 5B), and the methylation level continued to decline in mice fed the MR diet. This decline was statistically significant (P=0.01 by linear regression). The CG pairs located -35 and -43 bp 5’ relative to the murine TSS are homologous to regulatory CG pairs in the human CD40LG promoter, and these are demethylated in T cells from female lupus patients and 5-azaC treated T cells from healthy women (25). Demethylation required DOX treatment as no decline in methylation levels in residues -76 through -35 were observed in the absence of DOX (data not shown). This experiment confirms that diet and decreased Dnmt1 synergize to inhibit DNA methylation in these mice.

Figure 5.

Effect of Transmethylation Micronutrient Levels on CD40L Protein Expression and DNA Methylation in Female CD4+ T Cells. (A). Spleen cells from dnMEK/CD2rtTA transgenic C57BL/6 female mice treated with DOX and the diets indicated for 18 wks were stained for CD4+ T cells and CD40L and analyzed by flow cytometry. MR, n=8, Mean +/− S.E.M. MS, n=2, Mean +/− range. PChi-square. (B). Methylation of CG Pairs in the Cd40lg Promoter. Groups of 5–9 C57BL/6 dnMEK+/CD2rtTA+ transgenic mice each. Mean +/− SEM. (p), Student’s t-Test.

Discussion

The development of SLE involves genes that confer disease susceptibility, hormones and environmental factors (1, 2). The present study investigated the gene-environment interaction in lupus development and severity using a double transgenic lupus model with an inducible Erk pathway signaling defect bred onto the ‘lupus resistant’ C57BL/6 or a ‘lupus susceptible’ C57BL/6×SJL hybrid genetic background(17). These mice express a dominant-negative Mek uniquely in CD2+ cells when DOX is administered in their drinking water. Activation of the transgenes suppresses Erk signaling and subsequent Dnmt1 expression, leading to DNA hypomethylation and overexpression of methylation-sensitive genes(17, 18). This is similar to the Erk signaling defect that contributes to human lupus(26). Our present results confirmprevious findings in this model that activation of the transgenes on the ‘lupus resistant’ C57BL/6 strain, fed a standard commercial laboratory diet, induced low levels of anti-dsDNA antibody and DNA hypomethylation, and that higher anti-dsDNA antibody levels and active kidney disease required the presence of other ‘lupus susceptibility’ genes present in the SJL mouse strain(17). Increasing the SJL genetic contribution by an additional backcross further increased the levels of anti-dsDNA antibody and disease severity, thus supporting the findings reported by Sawalha et al. (1) that correlated lupus disease severity and the number of lupus-associated single nucleotide gene polymorphisms in men and women with lupus.

The incomplete concordance of lupus between genetically identical twins strongly supports the involvement of non-genetic factors in the etiology of SLE (4, 27). Drugs such as hydralazine and procainamide can trigger a lupus-like disease in genetically susceptible people through their effects on mechanisms such as DNA methylation, histone modification and signal transduction that control gene expression (8, 28). DNA methylation depends on Dnmt1 activity and S-adenosylmethionine (SAM) levels, the latter of which is regulated by methionine, choline, Zn, B2, B6, B12 vitamins and folate from dietary sources. Therefore, we tested the hypothesis that dietary micronutrients involved in transmethylation reactions can epigenetically modify lupus gene expression and disease severity by continuing to treat our transgenic mice with DOX following the induction of SLE, but changing the diets to a low methionine formulation or a diet rich in methyl donors and cofactors. Our results showed that reducing the methionine and choline content of the diet increased lupus disease severity in genetically susceptible hybrid mice and also caused a milder kidney disease in mice with a lupus-resistant C57BL/6 genetic background. In contrast, a diet enriched in methyl donors and the cofactors zinc, folic acid and vitamin B12 ameliorated both the anti-DNA antibody response and kidney disease. Taken together our results demonstrate that the appearance and severity of lupus disease can be influenced by both lupus susceptibility genes and non-genetic factors that affected DNA methylation.

It is tempting to speculate that human lupus induction and/or severity may be modulated by dietary intervention. Dietary modification has been used successfully in rodent and human studies to influence disease outcome and epigenetically alter the heritable gene expression profile. For example, dietary folate levels modulate hepatocyte DNA methylation in rats(29). A low folate diet caused DNA hypomethylation in lymphocytes of healthy postmenopausal women which could be reversed with a folate-supplemented diet (14). There is limited data from SLE patients that vitamin B6, B12 and folate supplementation ameliorates lupus symptoms (30). However, Wu et al. reported that more than 100 metabolites, many of which contribute to energy metabolism, are significantly altered in SLE patients(16). They further found that methionine and other methyl donors including cysteine, choline, and cofactors such as vitamin B6 were significantly reduced in SLE patients compared to healthy matched controls. Folate depletion increases homocysteine (Hcy) levels, which decreases SAM production resulting in DNA hypomethylation(10, 35). Maintenance of T cell DNA methylation patterns is more sensitive to low folate and methionine levels in older individuals due to decliningDnmt1 levels with age (10). Dnmt1 levels are also inversely proportional to lupus disease activity (31) potentially rendering people with active SLE more sensitive to low levels of micronutrients required for DNA methylation and potentially exacerbating immune disregulation and contributing to disease activity. Similarly, elevated levels of folate or B6 suppressed expression of the methylation-sensitive perforin gene in T cells from lupus patients in vitro15), supporting the possibility that dietary supplementation may have beneficial effects on SLE. Thus, attention to proper nutrition may be particularly important in the elderly and in lupus patients.

Results from the current study show that diets with low levels of methyl donors and cofactors, together with impaired Erk signaling caused progressive hypomethylation in CD40lg regulatory regions of CD4+ T cells, and that the methyl supplemented diet prevented DNA demethylation of this region. CD40L protein on T cells plays an important role in stimulating autoantibody production and is over-expressed in women but not men with SLE due to activation of the CD40L gene on the inactive X chromosome (18, 25, 32). Thus, dietary approaches that target a specific metabolic pathway such as transmethylation, key to lupus pathogenesis is a rational and potentially useful approach to prevention and perhaps therapy.

Successful dietary approaches to treat SLE will probably require simultaneously targeting multiple elements of 1-carbon metabolism as well as other factors that impact methylation pathways. Methionine participates as a methyl donor in SAM biosynthesis. The key reaction in DNA methylation is the DNMT-catalyzed transfer of methyl groups from SAM to dC bases at CpG pairs in DNA, producing methylated (d^mC) DNA and S-adenosylhomocysteine. Intracellular SAM pools are crucially dependent on a number of dietary factors, including folic acid, methionine, choline, betaine, Zn, and vitamins B₆ and B₁₂. While B vitamins and enzyme co-factors are essential for methionine metabolism, they cannot substitute for the required methyl group substrates. The metabolism of methionine to Hcy and remethylation of Hcy back to methionine is tightly controlled such that intracellular SAM levels tend to be maintained during normal variations in dietary methionine content(33). Betaine provides methyl groups for SAM by the action of betaine-homocysteine methyltransferases and one-carbon units via the folate system (34) and was included in our MS diet. Betaine supplementation lowers Hcy and elevates methionine and has been used as a safe and effective treatment for three different forms of homocysteinuria(33), and has been proposed as a therapy for neural tube defects by stimulating cellular methylation reactions(33).

Selection of the methyl donor and cofactor concentrations used in the present study was based on several studies by others in which diet was used to modify DNA methylation and biological responses(12, 36–38). The methionine concentration of our MR diet represented the lowest methionine concentration that could be tolerated over an extended period of time (35). Our MS diet was based on the formulation used by Hollingsworth (19) and the amino acid defined MNS diet used by Delaney et al.(20) to study epigenetic changes acquired in utero. Betaine along with increased dietary methionine, choline and folate were used to boost the concentration of methyl donors and cofactors and resulted in hypermethylation of genes in embryos(12, 19, 20). The MS diet used in the present study reduced anti-dsDNA antibody to near background levels in transgenic mice with defective Erk-triggered lupus-like disease suggesting that micronutrients that enhance transmethylation reactions may ameliorate lupus disease via epigenetic mechanisms. Similarly the reduced methionine content of the MR diet, together with reduced Erk pathway activity could have exacerbated lupus disease via epigenetic mechanisms by causing DNA hypomethylation and enhanced immune gene expression.

The terms ‘low’ and ‘high’ used for the methyl donor and cofactor content of our diets are relative only to one another and not to ‘natural’ diets. Compared to the MR diet the MS diet had ~8X methionine; ~14X choline; ~5X folic acid; ~25X vitamin B12 and 15 g/kg betaine. The methyl donor levels of the MS diet are greater than typically found in commercial rodent diets. The MR diet has lower methionine levels than the common purified diet AIN93G, 5053, and NIH-31used in other studies of DNA methylation in vivo12, 19, 20, 36). Levels of the other micronutrients such as folate, vitamin B6, and vitamin B12 are similar to standard rodent diets.

While our murine study suggests that diets rich in methyl donors may be beneficial for lupus patients, due to size, metabolic rates and differences in nutrient requirements, extrapolating micronutrient levels from rodents to humans is not straightforward. By expressing dietary nutrient levels based on energy, which takes into account the size and metabolic rates of the two species, a rough comparison of dietary micronutrient levels may be performed. Normal Western diets supply approximately 2–4 g sulfur amino acids (methionine + cysteine) per day, 147µg/kcal choline, 0.3 µg/kcal folate and 2.5ng/kcal vitamin B12(37, 38). Men 31–50 years old have an average intake of 2.3g methionine/day while women have 1.6g/day. In a 2000 kcal diet methionine represents 0.8-1.15 mg/kcal(38). Supplementation with five grams of methionine per day is the maximum tolerated dose in people (39). There was 3.1 mg methionine/kcal, approximately 3X that of typical human intake in the MS diet. The MS diet represented 29X choline, 14X folate and 160X the levels of vitamin B12 found in a Western diet. Betaine in our MS diet is 3.9mg/kcal, which is 37.5X that of found in the human diet (40). The amount of methionine in the MR diet is 0.375 mg/kcal which is less than half the amount of methionine consumed by adults fed a Western diet. The MR diet contains twice the amount of choline and folate and six times the amount of vitamin B12 as found in the Western diet.

In conclusion, our results indicate that in a mouse model of SLE it is possible to modulate autoantibody levels and kidney disease severity by dietary manipulation. These data suggest that dietary modification might be a candidate therapeutic approach for future studies in lupus patients.