IL-10 Signaling Is Essential for 1,25-Dihydroxyvitamin D₃-Mediated Inhibition of Experimental Autoimmune Encephalomyelitis¹

Multiple sclerosis (MS)³ shows a striking gradient of increasing prevalence with increasing latitude (1). Worldwide studies have attributed this gradient to a strong link between low sunlight exposure and high MS risk (2, 3, 4, 5). Because sunlight exposure produces vitamin D₃ (6), we hypothesized that the high MS risk with low sunlight exposure might reflect insufficient sunlight-catalyzed vitamin D₃ synthesis to support the immunoregulatory functions of the vitamin D₃ hormone, 1,25-dihydroxyvitamin D₃ (7).

Consistent with this hypothesis, 1,25-dihydroxyvitamin D₃ (1,25-(OH)₂D₃) strongly inhibited experimental autoimmune encephalomyelitis (EAE), an MS model (8, 9). The 1,25-(OH)₂D₃-mediated inhibition of EAE required Rag-1-dependent lymphocytes other than the encephalitogenic CD4⁺ T cells (10). To explain this result, we hypothesized that 1,25-(OH)₂D₃ might strengthen the function of lymphocytes that regulate potentially autoreactive T cell responses (10). The CD4⁺CD25⁺ lymphocytes that suppress EAE do so via an IL-10-dependent mechanism (11, 12, 13, 14). IL-10 is a potent inhibitor of APC function, inflammatory T cell activation, cytokine synthesis, and chemokine synthesis (15). Increases in spinal cord IL-10 correlated with EAE remissions (16). Mice with transgenic IL-10 resisted EAE (17), whereas mice with a disrupted IL-10 gene were highly EAE susceptible (18, 19). These data indicate that the IL-10-IL-10R pathway has very important suppressor functions in EAE.

The present experiments evaluated vitamin D₃- and 1,25-(OH)₂D₃-mediated protection from myelin oligodendrocyte peptide (MOG_35–55)-induced EAE in C57BL/6 (B6) mice with targeted disruptions of the IL-10 or IL-10R genes to test the hypothesis that the inhibition mechanism might be IL-10 dependent. Having established that the IL-10-IL-10R pathway was essential, we also evaluated irradiation bone marrow (BM) chimeras to pinpoint the source of the essential IL-10. Furthermore, we investigated a possible direct effect of the hormone on IL-10 or IL-10R expression. We present a model for a bidirectional, IL-10-dependent, anti-inflammatory loop to explain how vitamin D may implement sunlight’s protective biological effects in MS, and discuss the implications of this model for vitamin D-based preventive and therapeutic strategies in MS.

We purchased the B6, B6.SJL-Ptprc^a Pep3^b/BoyJ (B6.CD45.1), and C57BL/6-Il10^tm1Cgn mice (B6.IL-10^−/−) from The Jackson Laboratory. Genentech donated the CFR2.4 (B6.IL-10Rβ^−/−) breeder mice; they were genotyped for the loss of IL-10Rβ and the acquisition of neo as described (20). Mice were housed in our pathogen-free mouse colony at 23°C with 40–60% humidity, 12-h light-dark cycles, and ad libitum access to water. Before experiments, they were fed Lab Diet no. 5008 (PMI Nutrition International) containing 0.33 μg/day vitamin D₃ and 1% calcium. Experiments used male and female mice aged 6–8 wk (age- and sex-matched within experiments). The Institutional Animal Care and Use Committee approved all of the experimental protocols.

The MOG_35–55 was purchased from BioSynthesis and had the amino acid sequence MEVGWYRSPFSRVVHLYRNGK.

A nutritionally complete synthetic diet was formulated as we described (21), except that vitamin D was omitted. The vitamin D₃ (Sigma-Aldrich) was added to this diet in an amount to provide 1 μg/day. Alternatively, the 1,25-(OH)₂D₃ (Sigma-Aldrich) was added in an amount to provide 50 ng/day to females and 100 ng/day to males. These doses significantly inhibited EAE in previous studies (9, 10, 22, 23, 24, 25, 26). Fresh diet was provided three times per week. For prevention studies, mice were fed the synthetic diets before EAE induction (10). For treatment studies, EAE was induced in chow-fed mice, and at the peak of disease, 200 ng of 1,25-(OH)₂D₃ in oil or oil only as the placebo was injected i.p (27). EAE was induced and clinical EAE severity was assessed daily exactly as described (10). Blood samples were obtained before EAE induction and at the end of the study.

For histopathological evaluation, mice were euthanized and perfused with saline. The spinal cords were removed, divided into six equal segments, frozen in OCT compound (Sakura Finetek). And sectioned transversely (10 μm). The cryosections were fixed in 4% paraformaldehyde, stained with Gill’s no. 3 H&E Y (Sigma Diagnostics), and examined using a Zeiss Axioskop microscope equipped with a Plan-Neofluar ×20/0.5 objective. Bright field images were acquired with AxioVision 3.0 software controlling an Axiocam digital camera. For the histopathology analysis, each of six sections per mouse was divided into quadrants, and each quadrant was scored in a blinded fashion as 0 or 1, based on the absence or presence, respectively, of infiltrating inflammatory cells. The histopathology score was recorded as the percentage of spinal cord quadrants that showed a readily identifiable inflammatory cell infiltrate.

Blood was collected, clotted, and centrifuged (2000 × g for 10 min) at 6°C. The serum was decanted and stored at −20°C. The samples, standards, and buffer blanks (2 μl each) were aliquoted into duplicate wells of a 96-well plate. The calcium detection reagent was prepared according to the manufacturer’s directions (Sigma Diagnostics), and 0.25 ml was added to each well. The absorbance at 570 nm less the blank was measured 10 to 30 min later. The Ca²⁺ mM/L serum was determined from a standard curve.

Real-time PCR was performed as described (28) with minor modifications (29). In brief, total RNA was extracted using TRI Reagent (Molecular Research Center) and reverse transcribed from an oligo(dT) primer using the Reverse Transcription System (Promega). Real-time PCR was performed with SYBRGreen PCR Master Mix (Applied Biosystems) in a GeneAmp 5700 Sequence Detection System. Purified and quantified control cDNA was serially diluted, and amplified in each run; the control cDNA vs the threshold cycle formed a standard curve. The transcript abundance, determined with the aid of the standard curve, is reported relative to GAPDH. Published primers for the GAPDH and IFN-γ (30), and for Foxp3 (31) were used. The IL-10 primers were 5′-CCC TTT GCT ATG GTG TCC TT-3′ and 5′-TGG TTT CTC TTC CCA AGA CC-3′; they were designed with Primer Express (Applied Biosystems).

For flow cytometric studies, PBS perfusion was done and single-cell suspensions of spinal cord cells or splenocytes were prepared as described (10, 32). In brief, dissociated spleen cells in cold staining buffer (5% heat-inactivated FBS and 0.1% sodium azide in PBS, pH 7.3) were depleted of RBC, and stained (10⁶ cells/sample) for 30–40 min on ice with predetermined optimal amounts of FITC-, PE-, or allophycocyanin-conjugated mAb. Spinal cord cells were centrifuged through a Percoll gradient, washed, and stained as previously described (10). Reference samples were stained with fluorochrome-coupled isotype control mAb, or single-color specific mAb stains for compensation. Stained samples were analyzed on a FACSCalibur (BD Biosciences) using CellQuest software. The fluorescent mAbs to IFN-γ, IL-10, IL-10R, CD4, CD11b, CD25, CD45.2, and CD45.1 were purchased from Southern Biotechnology Associates or BD Biosciences.

To enumerate cytokine-producing cells, EAE was induced, and when the placebo-treated mice reached stage 2 EAE (20 days post MOG immunization), all mice were euthanized, and the spinal cord and cervical lymph node cells were collected and analyzed (33). The cells were stimulated overnight with immobilized Abs to CD3, cultured 4 h with monensin, washed, and stained with FITC-coupled Abs to CD4 or isotype control Abs. The stained cells were paraformaldehyde-fixed, permeabilized in buffer with 0.01% Triton X-100, and stained with PE-coupled Abs to IFN-γ and allophycocyanin-coupled Abs to IL-10 or isotype control Abs (BD Pharmingen). Flow cytometric data were acquired on a FACSCalibur using CellQuest software (BD Biosciences).

Irradiation BM chimeras were constructed by a standard protocol (34, 35). In brief, recipient mice (6 wk of age) were given antibiotics in the drinking water for 3 days, irradiated with a total of 1200 rad (2 doses 4 h apart), injected i.v. with 2–5 × 10⁶ BM cells, and maintained in a sterile environment with antibiotics for 2 wk. Thereafter, they were maintained under standard conditions without antibiotics. At 6–7 wk posttransplantation, blood lymphocytes were stained and analyzed for CD45.1 and CD45.2 to quantify donor BM engraftment. At 7 wk posttransplantation, the mice were placed on the 1,25-(OH)₂D₃-supplemented or placebo control diet. At 8 wk posttransplantation, EAE was induced and evaluated as above.

Individual mice were analyzed and the mean and SD were calculated for each group of mice. Experiments were repeated at least once. The group sizes are given in the table and figure legends. The significance of differences between the group means was determined using the Mann-Whitney U test (n ≤16), Student’s t test (n > 16), or χ² test (binomial data) as indicated (36); p < 0.05 was considered significant.

The experiments reported here used MOG_35–55-induced EAE in B6 mice to investigate why Rag-1-dependent lymphocytes other than the neuroantigen-reactive CD4 T cells were needed for 1,25-(OH)₂D₃ to inhibit EAE (10). To test the hypothesis that the vitamin D₃ system might be enhancing the function of IL-10-producing regulatory lymphocytes, we studied vitamin D₃-mediated inhibition of MOG_35–55-induced EAE in mice with a Wt or disrupted IL-10 gene. Mice were fed synthetic diets that provided 0 (−D diet) or 1 μg/day (+D diet) of vitamin D₃ beginning at age 4 wk exactly as described (26). At age 8 wk, the +D mice had 56 ± 3 nM/L serum 25-hydroxyvitamin D₃, whereas the −D mice had 8 ± 3 nM/L. All mice were then primed with MOG_35–55 and evaluated for EAE disease. The +D diet significantly reduced the incidence, peak clinical score, and cumulative disease index for MOG_35–55-induced EAE in female but not male B6 mice (Table I). These results confirm the female-specific protective effect of dietary vitamin D₃ that we reported previously for MBP-induced EAE in B10.PL mice (26). However, the +D diet did not inhibit MOG_35–55-induced EAE in B6.IL-10^−/− mice (Table I). We conclude that vitamin D₃-mediated inhibition of EAE requires IL-10 gene expression.

One explanation for the failure of vitamin D₃ to inhibit EAE in B6.IL-10^−/− mice could be an unanticipated requirement for IL-10 to stimulate the conversion of the biologically inactive vitamin D₃ and 25-hydroxyvitamin D₃ compounds into the biologically active hormone, 1,25-(OH)₂D₃. To bypass a possible vitamin D₃ metabolism problem in B6.IL-10^−/− mice, we repeated the experiment except that biologically active 1,25-(OH)₂D₃ was used in place of vitamin D₃. The 1,25-(OH)₂D₃-fed male and female B6 had a lower EAE incidence (p < 0.001), decreased mortality (p < 0.05), and lower peak severity and cumulative disease index (p < 0.0001) than the placebo-fed controls (Fig. 1,A and Table II). Histopathology showed that the 1,25-(OH)₂D₃-fed B6 mice had lesions with inflammatory cell infiltration in 17% of the microscopic fields examined, whereas the placebo-fed controls had lesions in >60% of the fields (Fig. 2). The 1,25-(OH)₂D₃-treated B6 mice also had 66% fewer spinal cord IFN-γ transcripts than the placebo-treated controls, as determined by real-time PCR (Table III). In contrast, the 1,25-(OH)₂D₃-treated and placebo-treated mice had equivalent numbers of IFN-γ-producing CD4⁺ Th1 cells in the lymph nodes (placebo, 16 ± 5%; 1,25-(OH)₂D₃, 16 ± 6%). Collectively, the clinical, histological, and immunological data establish that 1,25-(OH)₂D₃ inhibits MOG_35–55-induced EAE in B6 mice, as it does MBP-induced EAE in B10.PL mice (9, 10, 29).

In contrast, the 1,25-(OH)₂D₃ did not inhibit MOG_35–55-induced EAE in B6.IL-10^−/− mice (Fig. 1 and Table II). The incidence, onset, mortality, peak disease severity, and cumulative disease index were not significantly different for 1,25-(OH)₂D₃-fed and placebo-fed B6.IL-10^−/− mice. Furthermore, the 1,25-(OH)₂D₃-fed and placebo-fed B6.IL-10^−/− mice had equivalent histopathology (Fig. 2) with >60% of the fields showing inflammatory lesions. Lastly, the two groups had equally high numbers of IFN-γ transcripts in the spinal cord (Table III). The 1,25-(OH)₂D₃ also inhibited MBP-induced EAE in Wt B10.PL mice, but not B10.PL mice lacking IL-10 gene expression, confirming these results (data not shown). We conclude that a functional IL-10 gene is necessary for 1,25-(OH)₂D₃ to inhibit EAE.

The precise mechanism for 1,25-(OH)₂D₃-mediated inhibition of EAE is debated. It was recently suggested that elevated serum calcium may be necessary and sufficient to inhibit EAE in female mice, and that 1,25-(OH)₂D₃ may have no functional role other than to elevate serum calcium (37). To investigate this point, we analyzed serum calcium before and during the study shown in Fig. 1. The serum calcium levels were equivalent in all groups before the study (Table IV). The 1,25-(OH)₂D₃-treated B6 mice had no increase in the serum calcium on day 0, but elevated serum calcium by day 53. Therefore, elevated calcium correlated with decreased EAE in these mice. However, the 1,25-(OH)₂D₃-treated B6.IL-10^−/− mice had a significant increase in serum calcium on day 0, and a further increase by day 53 (Table IV), but no decrease in any measure of EAE disease (Table II). Thus, elevated serum calcium was not sufficient to inhibit EAE in the B6.IL-10^−/− female mice. Moreover, the high vitamin D₃ diet inhibited all measures of EAE disease in females without causing hypercalcemia (Table I and Ref. 26). Thus elevated serum calcium was neither necessary nor sufficient to inhibit EAE.

We reasoned that a fully functional IL-10-IL-10R pathway might be necessary for 1,25-(OH)₂D₃ to inhibit EAE. To test this hypothesis, we repeated the prevention study using B6.IL-10Rβ^−/− mice (with a nonfunctional IL-10Rβ-signaling chain) in place of B6.IL-10^−/− mice (20). The results showed that 1,25-(OH)₂D₃ did not inhibit MOG_35–55-induced EAE in B6.IL-10Rβ^−/− mice (Fig. 1,C). The EAE incidence, onset, mortality, peak disease severity, and cumulative disease index were not significantly different for 1,25-(OH)₂D₃-fed and placebo-fed B6.IL-10Rβ^−/− mice (Table II). Collectively, our data show for the first time that the CNS-specific, 1,25-(OH)₂D₃-activated, neuroprotective mechanism depends on a functional IL-10-IL-10R pathway.

We next looked for possible 1,25-(OH)₂D₃-mediated enhancement of IL-10 gene expression. The experiment shown in Fig. 1 A was repeated, and the IL-10-producing CD4⁺ T cells were analyzed by flow cytometry 28 days postimmunization. The IL-10-producing CD4⁺ T cells in the CNS cell suspension were too infrequent to measure. In the cervical lymph node cell suspension, the two groups had equivalent IL-10-producing CD4⁺ T cell percentages (placebo, 15 ± 8%; 1,25-(OH)₂D₃, 12 ± 4%) and fluorescence intensities (data not shown). Furthermore, when 1,25-(OH)₂D₃ or a placebo was administered to mice with acute EAE, the spinal cord and the lymph nodes had equivalent IL-10 transcripts 6-h posttreatment (placebo 0.4 ± 0.2 and 1,25-(OH)₂D₃ 0.3 ± 0.1 IL-10 transcripts per 10³ GAPDH transcripts). We have shown that direct, 1,25-(OH)₂D₃-mediated changes in gene expression can be detected using the 6 h treatment (29). Thus, neither protein nor transcript studies showed 1,25-(OH)₂D₃-mediated enhancement of IL-10 synthesis. However, the data do not rule out 1,25-(OH)₂D₃-mediated effects on minor cell populations like regulatory lymphocytes.

We also looked for 1,25-(OH)₂D₃-mediated enhancement of IL-10R gene expression. The 1,25-(OH)₂D₃ or a placebo was administered to mice with acute EAE, and flow cytometry was performed on spinal cord cells collected 6 h later. We have established that no hormone-mediated changes in the CNS cell populations occur within this time period (29). The nonimmunized control mice had microglial cells expressing an intermediate level of CD11b and the IL-10R (Fig. 3,A) (38, 39, 40). The placebo-treated mice with EAE had the CD11b^intIL-10R⁺ microglial cells, plus a CD11b⁺IL-10R⁻ cell population not found in the nonimmunized control mice (Fig. 3,B). These CD11b⁺IL-10R⁻ myeloid lineage cells appear to be infiltrating, activated macrophages. To our knowledge, this is the first report that CNS-infiltrating, activated macrophages do not express the IL-10R. We found no differences in the IL-10R⁺ or IL-10R⁻ cell percentages or the IL-10R density between the two treatment groups (Fig. 3, B and C, and data not shown). Thus, we found no evidence for a direct effect of 1,25-(OH)₂D₃ on IL-10R expression. However, we cannot rule out effects on minor cell populations like regulatory lymphocytes, astrocytes, or neurons that are not represented in the cell suspension.

Regulatory CD4⁺CD25⁺ T cells that express Foxp3 and suppress EAE by an IL-10-dependent mechanism have been described (11, 12, 13, 14, 41, 42, 43). To see whether 1,25-(OH)₂D₃ increased Foxp3, the prevention study was repeated, and spinal cord Foxp3 transcripts/10³ GAPDH transcripts were quantified 4 wk postimmunization. The placebo (24 ± 6) and 1,25-(OH)₂D₃-fed mice (29 ± 17) had equivalent Foxp3 transcripts, ruling out direct effects on Foxp3.

Having established that a functional IL-10-IL-10R pathway is necessary for the 1,25-(OH)₂D₃-activated protective mechanism, we next sought to determine the source of the essential IL-10. The T and B lymphocytes, infiltrating macrophages, microglial cells, and astrocytes all have the capacity to produce IL-10 (15, 44). Reciprocal irradiation BM chimeras were constructed by transferring Wt BM into lethally irradiated, IL-10^−/− recipients (B6.CD45.1→B6.IL-10^−/−) and vice versa (B6.IL-10^−/−→B6.CD45.1). Cells with a Wt IL-10 gene carried the CD45.1 allotype, whereas cells with an IL-10^−/− genotype carried the CD45.2 allotype as described (34, 35). Control B6.CD45.1→B6 and B6.IL-10^−/−→B6.IL-10^−/− chimeras were also constructed. At 6 wk post-BM transplantation, PBLs were >85% donor origin. At the end of each study, the splenocytes and splenic CD3⁺ T cells were 85–91% donor origin (Table V and data not shown). These CD45 allotype data indicated that all animals were chimeras.

Chimeric mice were fed synthetic diets with or without 1,25-(OH)₂D₃, immunized with MOG_35–55, and evaluated for EAEdisease. As expected, the 1,25-(OH)₂D₃-treated B6.CD45.1→B6 mice had a lower incidence of EAE, a later onset, a lower peak clinical score, and a lower cumulative disease index than the placebo-treated B6.CD45.1→B6 mice (Fig. 4,A and Table V). Also as expected, the 1,25-(OH)₂D₃-treated and placebo-treated B6.IL-10^−/−→B6.IL-10^−/− mice did not differ in the incidence of EAE, the day of onset, the peak clinical score, or the cumulative disease index (Fig. 4,B and Table V). These data confirm that a functional IL-10-IL-10R pathway is necessary for 1,25-(OH)₂D₃-mediated protection from EAE.

Most of the chimeric mice with limited IL-10 gene expression developed such severe and rapidly progressive EAE disease that the experiment was terminated 15 days post immunization. The placebo-treated B6.IL-10^−/−→B6.CD45.1 chimeric mice had significantly more severe clinical EAE disease on days 13 and 14, and a higher cumulative EAE disease index on day 15, than the placebo-treated B6.CD45.1→B6.IL-10^−/− chimeric mice. These data suggest that hemopoietic cell-produced IL-10 is especially critical for protection from EAE.

Surprisingly, the 1,25-(OH)₂D₃ did not inhibit EAE in B6.CD45.1→B6.IL-10^−/− or B6.IL-10^−/−→B6.CD45.1 mice (Fig. 4, C and D; Table V). Thus, limiting IL-10 synthesis to either the radio-sensitive or the radio-resistant cells abrogated 1,25-(OH)₂D₃-mediated protection. These data suggest that IL-10 synthesis from both cell types is essential for optimal 1,25-(OH)₂D₃-mediated protection. In fact, there may be an anti-inflammatory amplification loop involving bidirectional IL-10 signaling.

The present experiments investigated how vitamin D₃ and 1,25-(OH)₂D₃ inhibit EAE induction, in an effort to understand how this sunlight-derived hormone might mediate sunlight’s protective effects in MS. The results supported the hypothesis that the protective mechanism requires a functional IL-10-IL-10R pathway. The vitamin D₃ and 1,25-(OH)₂D₃ inhibited MOG_35–55-induced EAE in B6 female mice, and 1,25-(OH)₂D₃ also inhibited EAE in male mice, consistent with our published data on MBP_Ac1–11-induced EAE in B10.PL female and male mice (9, 26). Importantly, our data show for the first time that vitamin D₃ and 1,25-(OH)₂D₃ failed to inhibit EAE in B6 mice with targeted disruptions of either the IL-10 or the IL-10R genes. Extending this finding, we also show that 1,25-(OH)₂D₃ failed to inhibit EAE in reciprocal BM chimeric mice with restricted IL-10 gene expression. These studies suggest that both the radio-sensitive hemopoietic cells and the radio-resistant brain parenchymal cells must contribute IL-10 for optimal 1,25-(OH)₂D₃-mediated protection. Others reported that adding 1,25-(OH)₂D₃ and dexamethasone to cultures of D011.10 transgenic T cells with APC and peptide increased the frequency of IL-10-producing, regulatory T cells (12). However, our data showed no 1,25-(OH)₂D₃-induced increases in Foxp3 transcripts as a marker of regulatory T cells in the CNS. It is possible that the 1,25-(OH)₂D₃ stimulated IL-10-producing regulatory cells that do not express Foxp3. Our data also did not show an effect of 1,25-(OH)₂D₃ on IL-10 or IL-10R expression in freshly explanted spinal cord cells. However, these data do not rule out an effect on regulatory lymphocytes, astrocytes, or neurons that would have comprised <5% of the cell suspension.

Several caveats apply to our interpretation of the present data. First, we may have missed minor IL-10-independent protective effects of the 1,25-(OH)₂D₃, if the major protective activities were IL-10 dependent and determined the experimental results. We may also have missed IL-10-independent protective effects occurring at a later stage of EAE than was studied here. With respect to the chimera studies, because chimerism was 85–91%, a small percentage of the hemopoietic cells in the B6.IL-10^−/−→B6.CD45.1 chimeric mice had the Wt IL-10 gene. Furthermore, because a minor percentage of microglial cells in irradiation BM chimeras derived from hemopoietic stem cells (45, 46), some brain parenchymal cells in the B6.CD45.1→B6.IL-10^−/− chimeric mice would likely have had the Wt IL-10 gene. Nevertheless, in both cases, the IL-10 producers were insufficient to enable 1,25-(OH)₂D₃-mediated protection, further supporting the concept of a bidirectional IL-10 amplification loop.

To envision how an 1,25-(OH)₂D₃-activated, IL-10- and IL-10R-dependent, Rag-1-dependent protective mechanism could limit encephalitogenic CD4⁺ Th1 cells in the CNS, it is important to consider which CNS cells produce IL-10, and how IL-10 protects the CNS during inflammation. The major IL-10-producing brain parenchymal cells are the microglial cells and astrocytes (44, 47). The IL-10 promotes neuron and glial cell survival by blocking apoptotic signaling. More specifically, IL-10R stimulation regulates the Jak1/Stat3, PI3-kinase, MAP-kinase, suppressor of cytokine signaling (SOCS), and NF-κB life- and death-signaling pathways, ultimately inhibiting both proapoptotic cytokine-induced and mitochondrial-induced apoptosis in neurons and glial cells. The IL-10 also blocks proinflammatory cytokine induction and signaling. Specifically, IL-10R-mediated activation of SOCS reduces proinflammatory cytokine synthesis and cytokine receptor expression and activation. Finally, IL-10 inhibits costimulation of brain-infiltrating T cells via the CD28-CD80/86 pathway, thereby favoring anergy induction.

A model explaining how a 1,25-(OH)₂D₃-activated, IL-10- and IL-10R-dependent, Rag-1-dependent protective mechanism could limit encephalitogenic CD4⁺ Th1 cell activation in the CNS must also take into account which cells express the vitamin D receptor (VDR) (48). Myeloid cells, lymphoid cells, astrocytes, microglial cells, neurons, and oligodendrocytes all express the VDR (reviewed in (49). The T cells and type 2 astrocytes increase their VDR expression when activated, whereas the myeloid lineage cells diminish their VDR expression when activated (49). Our present and previous data rule out a direct effect of 1,25-(OH)₂D₃ on APCs and on naive encephalitogenic CD4⁺ T cells in the absence of additional Rag-1-dependent lymphocytes in the EAE model (10). Therefore, it is most likely that the 1,25-(OH)₂D₃ acted directly on regulatory lymphocytes and/or on CNS-resident cells like astrocytes, neurons, and/or microglia. The requirement for IL-10 gene expression in both hemopoietic cells and brain parenchymal cells suggests that 1,25-(OH)₂D₃ may be acting on both types of cells. The model that is most consistent with our data is bi-directional IL-10 signaling. The 1,25-(OH)₂D₃ could act directly on regulatory lymphocytes to increase their IL-10-dependent suppressive functions, and the regulatory lymphocyte-produced IL-10 could act on the brain parenchymal cells, altering their function such that they stimulate potentially encephalitogenic, autoreactive T cells to apoptose or to become anergic. To complete the bi-directional IL-10-signaling, the 1,25-(OH)₂D₃ could also be acting directly on the brain parenchymal cells to increase their IL-10-dependent suppressive functions. Further experimentation will be required to test this model. In particular, it will be important to demonstrate that there are direct effects of the 1,25-(OH)₂D₃ on regulatory lymphocytes and on brain parenchymal cells like neurons, astrocytes, and microglia that result in an immunosuppressive environment in the CNS.

Given the importance of IL-10 for the protection of the CNS during inflammation (44), it is not surprising that IL-10 deficiency correlates strongly with susceptibility to EAE and MS. The IL-10-deficient mice were highly susceptible to EAE (18, 19). Similarly, IL-10 deficiency appears to be an MS risk factor in humans. The IL-10-secreting T cell frequency was lower in MS patients than controls (50). Also, MS patients had less serum IL-10 protein (51), blood cell IL-10 mRNA (52, 53, 54), and blood cell IL-10 protein (55) than controls. Finally, the IL-10 allele with the low expresser promoter polymorphism was more common in MS patients than controls (56, 57). These correlations suggest a possible cause and effect relationship between IL-10 deficiency and EAE in mice and MS in humans.

Our data showing that IL-10 deficiency in mice eliminated the protective effects of 1,25-(OH)₂D₃ in EAE has very important implications for MS. Specifically, our data suggest that IL-10 deficiency in humans may diminish the protective effects of sunlight and vitamin D₃ in MS-susceptible individuals. There may be synergy between a genetic MS risk factor, IL-10 deficiency, and an environmental MS risk factor, insufficient vitamin D₃ synthesis due to inadequate sunlight exposure, increasing the risk of MS by compromising regulatory lymphocyte functions. If this hypothesis is correct, then it will be important to correct the IL-10 deficiency in MS-susceptible individuals for sunlight and vitamin D₃ prevention and treatment strategies to be beneficial.

We acknowledge Sir E. Donald Acheson (International Centre for Health and Society University College London) for extended and fruitful discussions on the vitamin D-MS hypothesis, and how a genetic IL-10-IL-10R pathway defect could interact with an environmental risk factor, vitamin D₃ insufficiency, to increase MS risk or severity in the human population. We are also indebted to Dr. Julie Olson (Department of Neurological Surgery, University of Wisconsin Medical School) for critically reading the manuscript. We thank Anna M. Rothert for technical assistance with the irradiation BM chimera studies, and Wayne Nehls for animal husbandry.

The authors have no financial conflict of interest.