1,25-Dihydroxyvitamin D3 acts directly on the T lymphocyte vitamin D receptor to inhibit experimental autoimmune encephalomyelitis

Authors

  • Christopher G. Mayne,

    1. Division of Rheumatology, Department of Pediatrics, Medical College of Wisconsin, Milwaukee, WI, USA
    Search for more papers by this author
    • These authors contributed equally to the work.

  • Justin A. Spanier,

    1. Department of Biochemistry, College of Agricultural and Life Sciences, University of Wisconsin-Madison, Madison, WI, USA
    Search for more papers by this author
    • These authors contributed equally to the work.

  • Lance M. Relland,

    1. Division of Rheumatology, Department of Pediatrics, Medical College of Wisconsin, Milwaukee, WI, USA
    Search for more papers by this author
  • Calvin B. Williams,

    1. Division of Rheumatology, Department of Pediatrics, Medical College of Wisconsin, Milwaukee, WI, USA
    Search for more papers by this author
  • Colleen E. Hayes

    Corresponding author
    1. Department of Biochemistry, College of Agricultural and Life Sciences, University of Wisconsin-Madison, Madison, WI, USA
    • Department of Biochemistry, College of Agricultural and Life Sciences, University of Wisconsin-Madison, 433 Babcock Drive, Madison, WI 53706, USA Fax: +1-608-262-3453
    Search for more papers by this author

Abstract

Multiple sclerosis (MS) is an incurable autoimmune neurodegenerative disease. Environmental factors may be key to MS prevention and treatment. MS prevalence and severity decrease with increasing sunlight exposure and vitamin D3 supplies, supporting our hypothesis that the sunlight-dependent hormone, 1,25-dihydroxyvitamin D3 (1,25-(OH)2D3), inhibits autoimmune T-cell responses in MS. Moreover, 1,25-(OH)2D3 inhibits and reverses experimental autoimmune encephalomyelitis (EAE), an MS model. Here, we investigated whether 1,25-(OH)2D3 inhibits EAE via the vitamin D receptor (VDR) in T lymphocytes. Using bone marrow chimeric mice with a disrupted VDR only in radio-sensitive hematopoietic cells or radio-resistant non-hematopoietic cells, we found that hematopoietic cell VDR function was necessary for 1,25-(OH)2D3 to inhibit EAE. Furthermore, conditional targeting experiments showed that VDR function in T cells was necessary. Neither 1,25-(OH)2D3 nor T-cell-specific VDR targeting influenced CD4+Foxp3+ T-cell proportions in the periphery or the CNS in these studies. These data support a model wherein 1,25-(OH)2D3 acts directly on pathogenic CD4+ T cells to inhibit EAE.

Introduction

Multiple sclerosis (MS), a genetically and immunologically complex neurodegenerative disease, is considered to be a T-cell-mediated autoimmune disease. The characteristic focal CNS lesions show T-lymphocyte and mononuclear cell infiltration, inflammatory mediators, demyelination, oligodendrocyte loss, reactive astrocyte formation, and axonal injury and loss 1. Current MS therapies seek to halt the T-cell-mediated autoimmune attack, the origins of which are not known.

Environmental factors may determine MS disease in individuals carrying genetic risk factors 2. There is a strong negative correlation (r=−0.9) between UV light exposure and MS prevalence 3. Noting this association, the UV light dependence of vitamin D3 synthesis 4, and vitamin D receptor (VDR) expression in activated T lymphocytes 5, 6, we hypothesized that 1,25-dihydroxyvitamin D3 (1,25-(OH)2D3) might control pathogenic autoimmune T-cell responses in MS 7. Abundant and compelling evidence now supports this hypothesis 8. Serum 25-hydroxyvitamin D3 (25-OH-D3) correlated inversely with MS risk and disease activity. Further, MS disease activity decreased after MS patients were given vitamin D39 or 1,25-(OH)2D3 therapy 10. The least active VDR alleles correlated with the greatest MS risk 11. Moreover, Cyp27b1 mutations that decreased 1,25-(OH)2D3 synthesis increased MS risk 12, 13. Lastly, the strongest known MS susceptibility gene, HLA-DRB1*1501, included a putative VDR responsive element in its promoter 14. Collectively, these studies implicate vitamin D3, 1,25-(OH)2D3, and VDR in determining MS risk and disease activity, and suggest that vitamin D3 and 1,25-(OH)2D3 may be useful in preventing and/or treating T-cell-mediated autoimmune attacks.

To elucidate how vitamin D3 and 1,25-(OH)2D3 may reduce MS risk, we have investigated how they inhibit EAE, an MS model. In EAE, neural antigen-primed, CD4+ T cells penetrate the perivascular space, become re-activated, invade the brain parenchyma, and initiate autoimmune-mediated damage. When vitamin D3-supplemented animals were immunized to induce EAE, highly localized 1,25-(OH)2D3 synthesis occurred in the CNS and correlated with EAE inhibition in females 15. Exogenous 1,25-(OH)2D3 also inhibited EAE development, prevented disease progression, and reversed established disease 16–18. The protective mechanism involved regulatory lymphocytes 17, IL-10 19, sensitization of pathogenic CD4+ T lymphocytes to apoptotic signals, a halt to further inflammatory cell recruitment, and enhanced expression of transcripts associated with CNS cell survival 20, 21.

To understand more precisely how the 1,25-(OH)2D3-mediated protective mechanism limits autoimmunity, the present experiments sought to identify the target cells of 1,25-(OH)2D3 action through selective inactivation of the VDR gene. The VDR gene is expressed in activated T and B lymphocytes, monocytes, macrophages, DC, neurons, astrocytes, microglia, and oligodendrocytes 22. To test whether 1,25-(OH)2D3 limits autoimmunity through direct actions on autoimmune T cells, we evaluated BM chimeric mice with an inactive VDR gene in hematopoietic cells or CNS-resident cells, and in mice with an inactive VDR gene in T lymphocytes. The data firmly establish that 1,25-(OH)2D3 limits autoimmunity through direct actions on T cells. Because CD4+Foxp3+ Tregs had very few VDR transcripts and were unaffected by 1,25-(OH)2D3 or T-cell-specific VDR targeting, our current model suggests an action of 1,25-(OH)2D3 directly on pathogenic T lymphocytes.

Results

The VDR is essential for 1,25-(OH)2D3-mediated EAE resistance

The first experiments tested the requirement for a functional VDR in 1,25-(OH)2D3-mediated EAE inhibition through the study of C57BL/6 (B6) mice with the Tokyo mutant VDR−/− (B6.VDR−/−) (Supporting Information Fig. 1A) 23, 24. This allele encodes a non-functional VDR protein lacking the first of two DNA-binding zinc finger motifs 25 (Supporting Information Fig. 1B). For this and all prevention studies, mice were fed diets without or with an optimal amount of 1,25-(OH)2D3 (females 50 ng/day; males 100 ng/day) continuously beginning 3 days before immunization with myelin oligodendrocyte glycoprotein peptide (MOG35–55) and EAE disease severity was scored daily. The 1,25-(OH)2D3 decreased the incidence 53%, delayed the onset ∼4 days, reduced the peak disease score 77%, and diminished the cumulative disease index (CDI) 87% in B6 mice, but provided no benefits in B6.VDR−/− mice (Fig. 1, Table 1). The VDR−/− mice were only evaluated for 19 days due to their high morbidity. The CD4 transcripts in the CNS correlated with and validated the clinical disease scores (Supporting Information Fig. 2B). Thus, 1,25-(OH)2D3 inhibits EAE through an action on the genomic VDR, not a non-genomic receptor.

Figure 1.

Effect of VDR gene inactivation on 1,25-(OH)2D3-mediated inhibition of EAE. (A) B6 and (B) B6.VDR−/− mice were fed diets that provided 1,25-(OH)2D3 or placebo continuously beginning 3 days before MOG35–55 peptide immunization. EAE disease severity was scored daily thereafter. Data show the composite mean±SD (17–34 B6 mice/group from three independent experiments; 5–6 B6.VDR−/− mice/group from one experiment).

Table 1. Effect of VDR gene inactivation on 1,25-(OH)2D3-mediated inhibition of EAEa)
Strain1,25-(OH)2D3Incidence (%)b)Onset (day)Peak disease scoreCDI (19 days)c)CDI (28 days)
  • a)

    a) Mice were fed diets that provided 0, 50 ng/day (females), or 100 ng/day (males) of 1,25-(OH)2D3 continuously beginning 3 days before MOG35-55 peptide immunization. EAE disease severity was scored daily thereafter. Shown is the composite mean±SD. Incidence data were subjected to a Chi-squared test. Disease onset, peak disease severity and CDI data were subjected to a Mann–Whitney test (n≤12) or a Student's t-test (n>12). ** indicates p<0.01 for placebo compared to 1,25-(OH)2D3-treated mice.

  • b)

    b) Mice with a clinical score ≥1 for two consecutive days were considered to have EAE and the first of the two consecutive days was the day of onset. The number of mice is given parenthetically.

  • c)

    c) The CDI was calculated by summing each animal's daily EAE disease scores for the indicated number of days.

B6100 (34/34)15±33.0±1.310.2 ±7.127.2±14.2
B6+29** (5/17)19±60.7±0.9**1.2±3.0**3.4±6.5**
B6.VDR−/−83 (5/6)11±22.5±1.113.8±8.4n.d.
B6.VDR−/−+100 (6/6)12±32.2±0.812.3±5.3n.d

Hematopoietic cell VDR is essential for 1,25-(OH)2D3-mediated EAE resistance

We next examined BM chimeric mice to determine whether 1,25-(OH)2D3 acts on radio-sensitive hematopoietic cells, CNS-resident cells, or both to prevent EAE. Studies were performed 6–7 wk after lethal irradiation and BM stem cell transfer, when hematopoietic cell reconstitution was complete but before significant microglial cell turnover occurred 26. The host and donor mice differed for the CD45 allotypes. Flow analysis of CD45.1 and CD45.2 showed that donor cell engraftment was >90% in B6 mice reconstituted with B6.CD45.1 stem cells (denoted B6.CD45.1→B6) and B6.VDR−/−→B6.CD45.1 mice, but somewhat lower in the B6.CD45.1→B6.VDR−/− mice (Table 2).

Table 2. Effect of VDR gene inactivation in hematopoietic cells or CNS cells on 1,25-(OH)2D3-mediated inhibition of EAEa)
Chimera1,25-(OH)2D3Chimerism (%)b)Incidence (%)c)Onset (day)c)Peak disease scoreCDIc)
  • a)

    a) The prevention study using BM chimeric mice and the statistical analyses were performed as described in Table 1 footnotes. Shown is the composite mean±SD. * indicates p<0.05 and ** indicates p<0.01 for placebo compared to 1,25-(OH)2D3-treated mice.

  • b)

    b) Flow cytometric analysis of CD45.1 and CD45.2 immunostained splenocytes harvested at the end of the study.

  • c)

    c) Incidence, onset, and CDI were calculated as described in Table 1 footnotes. The B6.CD45.1→B6 and B6.CD45.1→B6.VDR−/− chimeric mice were followed for 28 days. The B6.VDR−/−→B6.CD45.1 chimeric mice were followed for 23 days due to high morbidity.

B6.CD45.1→ B690±286 (12/14)14±42.8±1.226.7±15.7
B6.CD45.1→ B6+91±119 (3/16)**19±11.0±1.7**4.6±9.6**
B6.VDR−/−→ B6.CD45.193±2100 (14/14)13±13.9±0.736.6±5.9
B6.VDR−/−→ B6.CD45.1+93±2100 (16/16)13±14.1±0.934.7±3.3
B6.CD45.1→ B6.VDR−/−79±989 (8/9)10±12.6±0.930.4±11.5
B6.CD45.1→ B6.VDR−/−+82±763 (5/8)11±11.6±0.9*14.5±10.0**

An EAE prevention study was performed. The data showed that 1,25-(OH)2D3 reduced the incidence 78%, delayed the onset ∼5 days, decreased the peak disease score 64%, and diminished the CDI 83% compared to placebo in control B6.CD45.1→B6 mice (Fig. 2, Table 2). In B6.VDR−/−→B6.CD45.1 mice lacking VDR function in hematopoietic cells, the 1,25-(OH)2D3 did not alter the incidence, onset, or peak disease score. These mice had a particularly aggressive EAE disease course such that euthanasia was required 23 days post immunization (their CDI reflects only 23 days). In B6.CD45.1→B6.VDR−/− chimeras lacking VDR function in CNS cells, the 1,25-(OH)2D3 decreased the incidence 29%, reduced the peak disease score 38%, and diminished the CDI 52%. The greater disease reduction in the B6.CD45.1→B6 chimeras compared to the B6.CD45.1→B6.VDR−/− chimeras may reflect the ∼20% VDR−/− hematopoietic cells in the latter due to incomplete donor BM stem cell engraftment. Thus, only hematopoietic cells must express a functional VDR for 1,25-(OH)2D3-mediated EAE resistance.

Figure 2.

Effect of VDR gene inactivation in hematopoietic cells or CNS cells on 1,25-(OH)2D3-mediated inhibition of EAE. (A) Analysis of CD45.1 and CD45.2 on splenocytes from chimeric mice. The B6.CD45.1→B6 chimeras (top), B6.VDR−/−→B6.CD45.1 chimeras (middle), and B6.CD45.1→B6.VDR−/− chimeras (bottom) were produced by transfer of hematopoietic stem cells into lethaly irradiated recipients. The flow cytometric data shown are representative of 8–16 mice/group. (B) Effect of 1,25-(OH)2D3 on EAE induction in chimeric mice. A prevention study was performed as described in Fig. 1 legend. EAE severity was scored daily. These data are the composite mean±SD (8–16 mice/group from two independent experiments).

VDR gene inactivation in T lymphocytes

The next experiments selectively inactivated the VDR gene in T lymphocytes. Mice expressing the CD4-cre transgene 27 were crossed to mice carrying the enhanced green fluorescent protein (EGFP) transgene with a STOP cassette flanked by loxP sites (EGFPfloxed-STOP transgene) 28. CRE-mediated excision of the STOP cassette marks the CRE-expressing cells with EGFP. A flow cytometric analysis showed no EGFP+ splenocytes in Cre mice, and only EGFP+CD4+ and EGFP+CD8+ T cells in Cre+ mice (Supporting Information Fig. 3 and data not shown). These data confirm that CD4-cre excises marked DNA sequences at the CD4+CD8+ double-positive stage of T-cell development 27, but low EGFP expression precluded quantification of excision efficiency.

Next, homozygous B6.VDRfl/fl mice with loxP sites flanking exon 2 of the VDR gene 29 were crossed to CD4-cre transgenic mice and the F1 progeny were backcrossed to B6.VDRfl/fl mice. For experiments, we selected control CreVDRfl/fl mice and littermate Cre+VDRfl/fl mice with CRE-mediated excision of VDR exon 2 generating the non-functional Tokyo VDR−/− allele in T cells (Fig. 3A). Genomic DNA isolated from flow-sorted splenic CD4+ T cells and kidney cells was amplified with VDR allele-specific PCR primers (Fig. 3A). The Cre+VDRfl/fl CD4+ T-cell DNA yielded the 348 bp amplicon indicative of the exon 2-deleted VDR allele (Fig. 3C). The faint 217 bp amplicon indicative of the non-deleted VDRfl allele may indicate a few non-CD4+ T cells or <100% efficient CRE-mediated exon 2 deletion. The CreVDRfl/fl CD4+ T-cell and kidney cell DNA from both strains yielded only the 217 bp amplicon. RNA isolated from flow-sorted splenic CD4+ T cells and kidney cells was reverse transcribed and the cDNA amplified with VDR allele-specific PCR primers (Fig. 3B). The CreVDRfl/fl CD4+ T cells and the kidney samples of both strains yielded the amplicon derived from the non-deleted VDRfl transcript (Fig. 2D). PCR primers were chosen such that the exon 2-deleted VDR transcript did not yield an amplicon. The faint band from Cre+VDRfl/fl CD4+ T-cell cDNA indicated most VDR transcripts lacked exon 2.

Figure 3.

VDR gene inactivation in T cells. (A) The VDR gene structure in CD4-Cre (top) and CD4-Cre+ (bottom) mice. Shown are the first four exons (numbered boxes), the loxP sites (open triangles), the normal and alternate transcriptional start sites (black arrow and white arrow, respectively), and the PCR primer-binding sites (gray arrows). The genomic VDRfl/fl allele was detected as the 217 bp PCR product of the intronic PCR primers flanking the 5′ loxP site. The genomic VDR−/− allele was detected as the 348 bp PCR product of the intronic PCR primer upstream of the 5′ loxP site and the intronic PCR primer 5′ to exon 3. (B) The VDR cDNA structure in CD4-Cre (left) and CD4-Cre+ (right) mice. Shown are the first four exons and the PCR primer-binding sites (gray arrows). The control VDR cDNA was detected as the 231 bp PCR product of the forward PCR primer in exon 2 and the reverse primer in exon 3. The VDR−/− cDNA did not yield a PCR product, because the forward PCR primer in exon 2 cannot anneal to the cDNA. (C) PCR analysis of VDR sequences in the genomic DNA isolated from CD4+ T cells or kidney cells. Genomic DNA was isolated from flow-sorted VDRfl/flCD4-Cre and VDRfl/flCD4-Cre+ splenic CD4+ T cells and analyzed by PCR using the primers depicted in panel A. (D) PCR analysis of VDR cDNA from CD4+ T cells or kidney cells. Flow-sorted CD4+ T-cell RNA was isolated, reverse-transcribed, and analyzed by PCR using the primers depicted in panel B. GAPDH served as the control.

Inactivation of VDR in CD4+CD8+ double-positive thymocytes did not affect T-cell development, as evidenced by comparative analysis of thymocytes and splenocytes from CreVDRfl/fl and Cre+VDRfl/fl mice. No significant differences were detected in percentages of splenic CD4+ and CD8+ T cells (Supporting Information Fig. 4 and data not shown) or CD4+Foxp3+ T cells (CreVDRfl/fl; 9.4±3.4%; Cre+VDRfl/fl, 8.8±1.7%), or the CD4+ T-cell proliferative response to CD3 stimulation with IL-2 addition (Supporting Information Fig. 5). These data are consistent with data from VDR−/− mice 30.

T-cell VDR is essential for 1,25-(OH)2D3-mediated EAE resistance

An EAE prevention study was performed using mice with and without T-cell-selective VDR gene inactivation. In the control B6.CreVDRfl/fl mice, the 1,25-(OH)2D3 decreased the EAE incidence 47%, reduced the peak disease score 39%, and diminished the CDI 63% compared to placebo controls (Fig. 4A, Table 3). However, the 1,25-(OH)2D3 completely failed to inhibit EAE in the B6.Cre+VDRfl/fl mice with T-cell-selective VDR gene inactivation (Fig. 4B). Histopathological analysis of transverse spinal cord sections showed no immune cell infiltration or CNS demyelination in the 1,25-(OH)2D3-treated B6.CreVDRfl/fl mice (Fig. 4C), but abundant immune cell infiltration and CNS demyelination in the 1,25-(OH)2D3-treated B6.Cre+VDRfl/fl mice (Fig. 4D). Since CD8+ T cells were not required for 1,25-(OH)2D3 to inhibit EAE 31, we conclude that a functional VDR is essential in CD4+ T cells.

Figure 4.

Effect of VDR gene inactivation in T cells on 1,25-(OH)2D3-mediated inhibition of EAE. (A) Effect of 1,25-(OH)2D3 on EAE induction in B6.VDRfl/flCD4-Cre mice. (B) Effect of 1,25-(OH)2D3 on EAE induction in B6.VDRfl/flCD4-Cre+ mice. A prevention study was performed as described in the Fig. 1 legend. EAE severity was scored daily. The data are the composite mean±SD (9–16 mice/group from two independent experiments). (C) Histopathology of spinal cord sections from 1,25-(OH)2D3-treated B6.CreVDRfl/fl mice. (D) Histopathology of spinal cord sections from 1,25-(OH)2D3-treated B6.Cre+VDRfl/fl mice. Transverse sections were stained with hemotoxylin and eosin (top panels) or Luxol Fast Blue (bottom panels). The images shown are representative of three mice per group. The magnification was 4× (left panels) or 10× (right panels). The scale bars represent 0.1 mm. The dotted line outlines a region with immune cell infiltratration and corresponding demyelination in the adjacent section.

Table 3. Effect of VDR gene inactivation in T cells on 1,25-(OH)2D3-mediated inhibition of EAEa)
CD4-cre genotype1,25-(OH)2D3Incidence (%)b)Onset (day)Peak disease scoreCDI (28 days)
  • a)

    a) The prevention study using B6.Cre+VDRfl/fl and B6.CreVDRfl/fl mice and the statistical analyses were performed as described in Table 1 footnotes. Shown is the composite mean±SD. * indicates p<0.05 and ** indicates p<0.01 for placebo compared to 1,25-(OH)2D3-treated mice.

  • b)

    b) Incidence, onset, and CDI were calculated as described in Table 1 footnotes.

83 (10/12)11±32.6±0.931.8±15.2
+44 (4/9)15±51.6±0.9*12.1±9.6**
+100 (14/14)13±41.9±0.922.8±7.8
++100 (16/16)15±52.3±0.622.0±8.8

VDR transcripts and 1,25-(OH)2D3 responsiveness of CD4+Foxp3+ T cells

Activated Th1, Th2, and Th17 CD4+ T cells expressed abundant VDR transcripts 17, 32, but VDR gene expression in CD4+Foxp3+ Tregs has not been investigated. To determine VDR expression in Tregs, CD4+EGFP+ and CD4+EGFP T cells were flow sorted from the spleens of Foxp3EGFP reporter mice carrying a bicistronic allele with EGFP in the 3′UTR of the Foxp3 gene 33 (Fig. 5A). The full flow sorting strategy is shown in the Supporting Information Fig. 6. RNA from the sorted cells was subjected to RT-PCR analysis for Foxp3 and VDR transcripts. The EGFP+ T cells (11–14% of CD4+ splenocytes) had many Foxp3 transcripts but few VDR transcripts, whereas the EGFP T cells had few Foxp3 transcripts and many VDR transcripts (Fig. 5B and C and data not shown). This inverse relationship between Foxp3 and VDR gene expression suggests that at 1,25-(OH)2D3 levels ∼0.5 nM, the Kd for VDR binding 34, the 1,25-(OH)2D3 might influence the CD4+Foxp3 T cells more than the CD4+Foxp3+ Tregs.

Figure 5.

VDR transcript expression by CD4+Foxp3 T cells and CD4+Foxp3+ Tregs. (A) Flow sorting of Foxp3-expressing CD4+EGFP+ Tregs and Foxp3-non-expressing CD4+EGFP Tconv populations from Foxp3EGFP mouse splenocytes. (B) The relative abundance of VDR and GAPDH transcripts in flow-sorted CD4+EGFP+ Tregs and CD4+EGFP Tconv cells. Transcript abundance was estimated by RT-PCR on serially diluted samples. The data shown are representative of three independent experiments. (C) The relative abundance of VDR transcripts in flow-sorted CD4+EGFP+ Tregs and CD4+EGFP Tconv cells. Transcript abundance was quantified by RT-qPCR using L32 transcript abundance as the reference. The data shown are the mean± SD for 6 mice/group; * indicates p<0.02 (Mann–Whitney test).

Despite low VDR expression in CD4+Foxp3+ Tregs, it was of interest to determine whether 1,25-(OH)2D3 expanded this population to reduce EAE disease. Accordingly, an EAE prevention study was performed with B6.CreVDRfl/fl and B6.Cre+VDRfl/fl mice. The Foxp3+, IFN-γ+, and IL-17+ CD4+ T cells were quantified 12 days post EAE induction. The proportions (and numbers) of peripheral CD4+IFN-γ+, CD4+IL-17+, and CD4+Foxp3+ T cells did not vary by genotype, treatment group, or EAE disease score (Supporting Information Table 1 and Fig. 7). In the CNS, the Foxp3+ T cells constituted 8–12% of CD4+ T cells, regardless of genotype, treatment group, or EAE disease score. The Th17 cells in the CNS were too variable to draw conclusions.

Possible 1,25-(OH)2D3 effects on CD4+Foxp3+ Tregs were also investigated in vitro. Flow-sorted CD4+EGFP splenocytes from Foxp3EGFP mice were activated and examined for EGFP (Supporting Information Fig. 8); 1,25-(OH)2D3 addition did not affect the EGFP+ cells. Thus, during the early stages of EAE, neither T-cell-specific VDR gene targeting nor 1,25-(OH)2D3 treatment altered the proportions of Foxp3+, IFN-γ+, or IL-17+ CD4+ T cells in the periphery, or the Foxp3+ T cells in the CNS.

Discussion

Compelling evidence has implicated vitamin D3, 1,25-(OH)2D3, and the VDR in determining MS risk and MS disease activity. However, there is incomplete knowledge of the molecular mechanisms that apply. Our new data show that 1,25-(OH)2D3 exerts protective biological effects in the EAE model through the nuclear VDR in hematopoietic cells, and more specifically in T lymphocytes. The 1,25-(OH)2D3 failed to inhibit EAE disease induction in chimeric mice lacking a functional VDR in hematopoietic cells, and in mice lacking a functional VDR specifically in T lymphocytes. Our analysis revealed an inverse relationship between VDR and Foxp3 gene expression, with abundant VDR transcripts only in Foxp3 T cells and few VDR transcripts in Foxp3+ T cells. Neither 1,25-(OH)2D3 nor T-cell-specific VDR inactivation influenced the CD4+Foxp3+ T cells in the periphery or the CNS in the early stages of EAE. To our knowledge, this is the first evidence obtained entirely in vivo that 1,25-(OH)2D3 acts directly on immune system cells and more specifically on T lymphocytes in autoimmune disease control.

It was surprising that 1,25-(OH)2D3 did not act directly on CNS cells to prevent EAE, since VDR transcripts have been reported in neurons, astrocytes, microglia, and oligodendrocytes 22. The 1,25-(OH)2D3 inhibited EAE in chimeric mice lacking a functional VDR in CNS cells, indicating the VDR was dispensable in these cells for EAE prevention. These data do not rule out 1,25-(OH)2D3 effects on VDR-expressing CNS cells for EAE treatment 18, 21. On the contrary, 1,25-(OH)2D3 treatment in EAE rapidly upregulated cell type-specific genes associated with CNS cell survival (calcium/calmodulin-dependent protein kinase II-δ transcripts in reactive astrocytes; neurotrophic tyrosine kinase receptor B in neurons) 20. A minimal model would be 1,25-(OH)2D3 acting directly on pathogenic T lymphocytes to prevent EAE, and on pathogenic T lymphocytes and CNS cells in established EAE to resolve inflammation and prevent or reverse neurological damage.

Another model for autoimmune disease prevention is 1,25-(OH)2D3 action on myeloid DC to establish a tolerogenic phenotype 35. This model raises the question whether VDR targeting in myeloid DC may have contributed to the present results. We did not detect floxed gene deletion in CD4-Cre+EGFPfloxed-STOP splenocytes other than CD4+ and CD8+ T cells. Others did not detect floxed gene deletion in flow-sorted CD11c+ myeloid DC 36, 37. Floxed gene deletion occurred in plasmacytoid DC 38, which have been ruled out as 1,25-(OH)2D3 targets 39. Therefore, it is unlikely that VDR targeting in myeloid DC contributed to our results. Importantly, when T cells lacked VDR function, 1,25-(OH)2D3 responsiveness of other hematopoietic cells like DC was not sufficient to prevent EAE.

Although CD4+Foxp3+ T cells had low VDR gene expression and did not fluctuate in response to 1,25-(OH)2D3 or T-cell-specific VDR targeting, it remains possible that 1,25-(OH)2D3 enhances CD4+Foxp3+ Treg function. CD4+Foxp3+ Tregs capable of preventing EAE immunopathology by an antigen-specific, IL-10-dependent mechanism have been described 40. Consistent with the possibility that 1,25-(OH)2D3 might enhance Treg function, we found that EAE resistance persisted after 1,25-(OH)2D3 treatment was withdrawn, and that 1,25-(OH)2D3 failed to inhibit EAE in TCR-transgenic Rag-1−/− mice lacking Tregs 17 and in IL-10-null or IL-10R-null mice with compromised Treg function 19. Also, 1,25-(OH)2D3 enhanced TGF-β1 gene transcription in vivo in the EAE model 41. Thus, 1,25-(OH)2D3 enhancement of Treg function, in particular TGF-β1 or IL-10 gene expression, has not been ruled out.

Our data indicate that a direct effect of 1,25-(OH)2D3 on Th1 and Th17 cells in the CNS to prevent autoimmunity is a likely possibility. Activated Th1 and Th17 cells expressed abundant VDR transcripts and would be highly responsive to 1,25-(OH)2D317, 32. We previously ruled out inhibition of Th1 cell priming, cytokine production, and access to the CNS as mechanisms by which 1,25-(OH)2D3 prevented EAE 17. The data presented here rule out 1,25-(OH)2D3 inhibition of Th17 cell priming and cytokine production in the periphery in the early stages of EAE, but do not rule out 1,25-(OH)2D3 inhibition of Th17 cell entry or function in the CNS. In the terminal stages of EAE, others found fewer splenic Th17 cells and lower IL-17 production in 1,25-(OH)2D3-treated mice compared to placebo controls 32, 42. Further analysis of 1,25-(OH)2D3 effects on Th17 cells in the CNS at early stages of EAE is needed and is underway.

We previously documented 1,25-(OH)2D3-mediated enhancement of pathogenic T-cell sensitivity to CNS-derived apoptotic signals in EAE treatment studies. In the CNS, there was rapid 1,25-(OH)2D3-mediated induction of pro-apoptotic genes and repression of anti-apoptotic genes 20, and rapid increases in apoptotic CD4+ T cells followed by decreases in CD4+ T-cell numbers and cytokine synthesis 21. Since CD4+ T-cell apoptosis and decreased pro-inflammatory cytokines and chemokines were not found in the periphery, we proposed that the apoptotic signals were delivered within the CNS 21. Low VDR expression in CD4+Foxp3+ Tregs may allow them to escape sensitization to apoptotic signals.

In summary, strong evidence implicates vitamin D3, 1,25-(OH)2D3, and the VDR in determining MS risk and disease activity. EAE studies have suggested at least two possible molecular mechanisms by which the vitamin D endocrine system could reduce MS risk and severity, an action on induced Tregs to increase their function 17, 19, 43, and an action on encephalitogenic T cells enhancing their responsiveness to apoptotic signals. The present experiments provide the first in vivo evidence that 1,25-(OH)2D3 acts directly on the VDR in T lymphocytes to inhibit EAE, a result that is most compatible with sensitization of encephalitogenic T cells to CNS-derived apoptotic signals.

Materials and methods

Mice

All experiments used adult male and female mice (age 6–8 wk); the groups were sex-matched within experiments. The mice were maintained at 23°C with 40–60% humidity and 12-h light–dark cycles. Drinking water was provided ad libidum. The B6 and B6.SJL-PtprcaPepcb/BoyJ (B6.CD45.1) mice were obtained from the Jackson Laboratory (Bar Harbor, ME) at backcross generation 22, and bred in our pathogen-free mouse colony. The B6.VDR−/− mice (backcross generation 10), originally produced in the lab of Dr. Shigeaki Kato 24, were a gift from Dr. Hector DeLuca (Department of Biochemistry, University of Wisconsin). The B6.VDR+/− strain was maintained in the heterozygous state, since homozygous mice are infertile. The B6.VDR+/− heterozygotes were intercrossed and the offspring genotyped for B6.VDR−/− homozygotes to produce experimental mice. The B6 mice with loxP sites flanking exon 2 of the VDR gene (B6.VDRfl/fl), also produced in the Kato lab 29, were a gift from Dr. James Fleet (Department of Foods and Nutrition, Purdue University). The ROSA26 reporter mice expressing the EGFPfloxed-STOP transgene 28 were a gift from Dr. Akihiro Ikeda (Department of Genetics, University of Wisconsin). Mice expressing two copies of the Tg(CD4-Cre)1Cwi transgene on a C57BL/6 background (B6.CD4-cre) were purchased from Taconic (Hudson, NY) 27. The B6.Foxp3EGFP reporter mice have been described 33, 44; these mice were produced by Dr. Talal Chatila (University of California at Los Angeles). The Foxp3EGFP mice were maintained at the Medical College of Wisconsin. All animal experimentation was conducted in accord with accepted standards of humane animal care. The University of Wisconsin's Institutional Animal Care and Use Committee approved the experimental protocols (Protocol No. A00847).

Experimental diets

Before experiments, the B6.VDR−/− mice were fed a high lactose rescue diet 45, while other mice were fed standard chow (Lab Diet ♯5008; PMI Nutrition International, Brentwood, MO) with 1% calcium that provided 0.33 μg/day of vitamin D3. The synthetic diet for experiments was formulated from the high lactose diet with all essential nutrients except vitamin D3 as described 15. The 1,25-(OH)2D3 (Sigma Aldrich, St. Louis, MO) was dissolved in absolute ethanol (1 mg/mL) and stored at 4°C. It was added to the synthetic diet in an amount sufficient to provide 50 ng/day (females) or 100 ng/day (males), based on a measured daily consumption of 4.0 g dry weight of diet per mouse 41.

EAE induction and analysis

Mice were fed the experimental diets continuously beginning 3 days before EAE induction. Fresh synthetic diet was provided three times per wk. The MOG35–55 peptide had the amino acid sequence MEVGWYRSPFSRVVHLYRNGK (BioSynthesis, Lewisville, TX). MOG35–55 was dissolved in 0.1 M acetic acid (4 mg/mL), emulsified in an equal volume of CFA supplemented with Mycobacterium tuberculosis H37 Ra (4 mg/mL; Difco, Detroit, MI). Each anesthetized mouse received a 0.05-mL s.c. injection of the emulsion (200 μg MOG35–55 total) in each hind flank near the base of the tail. On the day of immunization and again two days later, each mouse was injected i.p. with 200 ng of pertussis toxin (List Biological Laboratories, Campbell, CA). EAE severity was assessed daily as follows: 0, no disease; 1.0, limp tail; 2.0, limp tail and mild weakness of hind legs; 3.0, limp tail, moderate weakness of hind legs and markedly wobbly gait; 4.0, paralysis of both hind legs without foreleg weakness; 5.0, both hind and one foreleg paralysis; 6.0, moribund/dead.

BM chimeric mice

Irradiation BM chimeric mice were produced by a standard protocol 26, 46 as we described previously 19. BM cell engraftment was quantified as we described 19. EAE experiments were performed 6–7 wk post-transplantation.

Flow cytometry

Staining of cells for flow cytometric analysis was performed as we described 47. Stained samples were analyzed on a FACScalibur using CELLQuest software. Flow sorting was performed on a FACSVantage cell sorter (BD Biosciences, Franklin Lakes, NJ). The biotin-coupled mouse mAb to mouse CD45.2 (Clone 104), the PE-coupled mouse mAb to mouse CD45.1, and the FITC-coupled rat mAb to mouse CD4 (Clone GK1.5) were obtained from Southern Biotech. The streptavidin-PerCP-Cy5.5 was obtained from BD Biosciences.

Genomic DNA isolation, RNA isolation, reverse transcription, and PCR

Genomic DNA was isolated from flow-sorted CD4+ T cells and kidney tissue using proteinase K digestion and ammonium actetate purification [48]. Genomic VDR PCR was performed using one forward primer (TCTGACTCCCACAAGTGTACCACGG) and two reverse primers, GCCTTCTATGCATATCTGAAAGCT yielding a 217 bp amplicon and CCAGGTGAGTTTACCTACCACTTCCC yielding a 348 bp amplicon.

Total cellular RNA was extracted from spinal cord tissue, kidney tissue, and flow-sorted CD4+ T cells using Tri Reagent (Molecular Research Center, Cincinnati, OH), and flow-sorted CD4+EGFP+, and CD4+EGFP T cells using RNeasy mini kit (Qiagen, Valencia, CA) according to the manufacturer's protocols. For cDNA synthesis, 1 μg of total RNA was reverse transcribed from an oligo(dT) primer using the AMV Reverse Transcription System (Promega, Madison, WI) or the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) according to the manufacturer's protocols. The amplification and detection were accomplished with a 7500 Fast Real-Time PCR System (Applied Biosystems) and SYBR Advantage qPCR Premix (Clontech, Mountainview, CA). The VDR cDNA was amplified using the primer pair forward TGACTTTGACCGGAATGTGCCT and reverse TTCATCATGCCAATGTCCACGCAG. GAPDH cDNA was amplified using primer pairs forward TTCACCACCATGGAGAAGGC and reverse GGCATGGACTGTGGTCATGA or forward AGGTCGGTGTGAACGGATTTG and reverse TGTAGACCATGTAGTTGAGGTCA. Spinal cord CD4 cDNA was amplified using the primer pair forward TCACCTGGAAGTTCTCTGACC and reverse GGAATCAAAACGATCAAACTGCG. L32 cDNA was amplified as a normalization control for VDR qPCR using primer pairs forward GGAAACCCAGAGGCATTGAC and reverse TCAGGATCTGGCCCTTGAAC. VDR qPCR was performed on T-cell cDNA using primer pairs forward ATCTGTGAGTCTTCCCAGGAGAGC and reverse TGACTTTGACCGGAATGTGCCT. Primers were purchased from Integrated DNA Technologies (Coralville, IA) or Invitrogen Life Technologies (Carlsbad, CA).

Data analysis

Individual mice were analyzed and the mean±SD was calculated for each group of mice. The group sizes are given in the table footnotes and figure legends. The significance of differences between the group means was determined using the Mann–Whitney test, Student's t-test, or Chi-squared test (binomial data) as detailed in the footnotes and legends.

Acknowledgements

We thank Drs. Hector DeLuca, Shigeaki Kato, James Fleet, and Akihiro Ikeda for providing B6.VDR−/−, B6.VDRfl/fl, and ROSA26 reporter mouse strains. We are indebted to Faye Nashold and Lauren Brown for technical assistance. This work was supported by the National Multiple Sclerosis Society research grant RG 3107-C-4 and grant MSN119798 from the US Dept. of Agriculture to C. E. H.

Conflict of interest: The authors declare no financial or commercial conflict of interest.

Ancillary