72 five-week-old female C57Bl/6 mice (Tacomic, Tornbjerg, Denmark), with a mean weight of 20.4 g±1.1 g were used for the experiment. They were housed six together in Macrolon IVC-II cages (Scanbur, Karlslunde, Denmark) in standard laboratory conditions; light/dark cycles of 12/12 hours, cage temperature of 22.5±1°C, relative humidity of 52±5% and 75 air changes per hour. Cage maintenance was performed once weekly, and the animals were handled by the same individuals at all times. The animals were weighed twice weekly. The experiment was conducted in strict accordance with the Federation of European Laboratory Animal Science Associations recommendations, and the protocol was approved by the Norwegian Animal Research Authority (permit #2009-1767).

Four diets, only differing in the content of vitamin D₃ were custom made by a commercial manufacturer (Altromin GmbH, Lage, Germany). The vitamin D₃ content was 1) <50 IU/kg (e.g. deficient), 2) 500 IU/kg, 3) 6200 IU/kg and 4) 12500 IU/kg, respectively. The content was verified by HPLC in an independent laboratory (Norwegian Institute for Nutrition and Seafood Research, Bergen, Norway). The different vitamin D₃ doses were chosen to reflect relevant human dietary doses, ranging from deficiency to the doses used in intervention studies in MS. Assuming a daily intake of 5 g of chow per mouse per day, allometric conversion as described earlier ^[31] gives estimated human equivalent doses of 1) <76 IU/day, 2) 760 IU/day, 3) 9425 IU/day and 4) 19003 IU/day for the different diets.

After one week acclimatization, the mice (n = 72) were randomised to one of the four experimental diets. Two weeks after randomisation, when the mice were eight weeks of age, 2/3 of the animals in each diet group (n = 12) were randomised to cuprizone exposure, while the remaining six animals in each diet group served as healthy controls. To induce demyelination, cuprizone (bis-cyclohexanone-oxaldihydrazone, Sigma-Aldrich, St. Louis, MO, USA) 0.2% (w/w) was added to the milled mouse chow. The mice had ad libitum access to chow and tap water during the whole experimental period.

Serum was collected at four time points during the study: 1) before randomisation to the experimental diets, 2) After the two-week wash-in period, before cuprizone exposure, 3) after six weeks of cuprizone exposure, and 4) after two weeks recovery after ending cuprizone exposure.The 25(OH)-vitamin D₃ analysis was performed according to a modified version of a method described previously ^[32]. Briefly, 25 µl serum samples were spiked with 26,27-dexadeuterium-25-OH-Vitamin D₃ (Synthetica AS, Oslo, Norway) as internal standard and extracted with methanol and n-hexane. The n-hexane phase was collected, evaporated to dryness and ejected into a reverse-phase high performance liquid chromatography system. Elution of 25-OH-vit D₃ was performed with methanol/water (88∶12, v/v, with 0.1% formic acid) and the eluate was monitored by a LC/MS-detector (LC/MSD SL, Agilent Technology INC, CA 95051, USA) equipped with a multimode ion-source. 25(OH)-vit. D₃ and internal standard were monitored at 395.0 and 407.3 m/z , respectively, in the APCI positive mode. The mean recovery of 25(OH)-vitamin D₃ was 77.2% (SD 3.9%) and the interassay variation was 4.9%, with a detection limit <4 nmol/l. Serum calcium was analysed using Calsium AS FS (DiaSys Diagnostic Systems GmbH, Holtzheim, Germany) ^[33].

Mice were asphyxiated with CO₂, the brains removed and fixated in 4% paraformaldehyde for at least 7 days, then paraffin embedded. All analyses were performed on 7 µm sections ±1 mm from the bregma ^[34]. Sections were stained with Luxol Fast Blue (LFB) and hematoxyline and eosine (HE). For immunohistochemistry, the sections were dewaxed and rehydrated before antigen retrieval in either citrate buffer (pH 6.2) or Tris-EDTA buffer (pH 9.1). Primary antibodies and their incubation times and temperatures are given in Table 1. Sections were blocked with peroxidase blocking solution (Dako, Glostrup, Denmark), and visualized with EnVision 3.3 - diaminobenzidine (1∶50; 10 min at room temperature (RT); (Dako, Glostrup, Denmark). The tissue sections were counterstained with hematoxylin. For each antibody, omission of the primary antibody served as negative control. Appropriate positive controls were included in each staining procedure. For proteolipid protein (PLP), Neurite outgrowth inhibitor A (NOGO-A), and Mac-3, normal brain tissue from the healthy controls served as positive controls. For the T-lymphocyte marker CD3, sections from tonsils served as positive controls.

All sections were evaluated by two blinded observers, using light microscopy (Leica DMLe, Wetzlar, Germany). For quantification of myelin, sections stained with LFB and PLP were scored semiquantitatively for myelin loss in the midline of corpus callosum using a scale ranging from 0 (complete loss of myelin) to 3 (complete myelination), as described by Lindner and colleagues ^[35]. In tissue sections immunostained for PLP, the midline of the corpus callosum was photographed with identical exposure settings at 40× magnification (Leica DMLe with Leica DC300 camera). Greyscale images were thresholded in order to avoid quantitative registration of low-intensity background staining. The area of PLP immunopositivity in each image was determined using Image processing and analysis in Java (U. S. National Institutes of Health; Bethesda 2009), and expressed as the percentage of pixels, or relative area, in each image with an intensity within the threshold values. The number of oligodendrocytes (NOGO-A immunopositive cells) ^[36], microglia, macrophages (Mac-3 immunopositive cells), and T-lymphocytes (CD3-positive cells) was counted in an area of 0,0625 mm² in the midline of the corpus callosum ^[34], using an ocular morphometric grid.

One-way analysis of variance (ANOVA) was used to analyse diet effects. Fishers' least significant difference method was used as post hoc analysis where applicable. Six cuprizone exposed mice and 6 control mice were included in the data analysis for each diet group. All data are presented as arithmetic mean +/−1 standard deviation. Statistical analyses were performed using PASW Statistics 18.0 (IBM; Chicago, 2010).

The analysis of the vitamin D₃ metabolite 25-OH-D₃ in serum samples obtained at different time points during the study revealed that the vitamin D₃ content in the diet was reflected in the serum of the mice (Figure 1). Prior to the experimental period, all animals were fed a standard rodent diet (Rat and mouse No. I″ maintenance diet from Scanbur, Special Diets Services, Karlslunde, Denmark), containing 900 IU/kg of vitamin D₃. The mean baseline s-25-OH-vit. D₃ level was 54.5±4.2 nmol/L. After two weeks on the experimental diet, the serum levels raised or fell significantly in accordance with the vitamin D₃ content in the diet (p<0.0005). There were no significant differences in the 25-OH-D₃ serum levels between the cuprizone-exposed mice and control mice (p = 0.732). An increase in serum calcium was observed in all mice during the study period, less in the cuprizone exposed mice than in the controls (p = 0.001). There was no difference in the calcium levels of cuprizone exposed mice between the different diet groups (Figure S1). There was no significant difference between the diet groups with regards to baseline body weight or weight changes (Figure S2).

Significant myelin loss was observed in the corpus callosum after six weeks of cuprizone exposure in all diet groups compared to the control mice (p<0.0005). There were no differences in myelin status between the control mice in the different diet groups. The two cuprizone-exposed groups receiving the highest vitamin D₃ content, 6200 and 12500 IU/kg, had significantly less demyelination in the corpus callosum than the groups receiving <50 or 500 IU/kg (p = 0.004) as determined in LFB-stained sections (Figures 2, 3). There were no differences in the extent of demyelination between the groups receiving 6200 IU/kg and 12500 IU/kg, or between the group receiving <50 IU/kg and 500 IU/kg. In tissue sections immunostained for PLP, a similar pattern was observed, as the mice receiving 12500 IU/kg vitamin D₃ in the diet had significantly less myelin loss compared to the vitamin D3 deficient diet (p = 0.02).

In all diet groups, an extensive loss of NOGO-A immunopositive cells was observed after six weeks of cuprizone exposure (Table 2). There were no significant differences in the degree of oligodendrocyte loss (p = 0.18) between the diet groups. Two weeks after discontinuation of the cuprizone exposure, the number of NOGO-A immunopositive cells were similar to the control mice in all diet groups.

Two weeks after discontinuation of the cuprizone exposure, there was no significant difference in myelin status between the cuprizone exposed diet groups (p = 0.74). In the groups fed a vitamin D3 -low (500 IU/kg) or -deficient (<50 IU/kg) diet the myelin status, as evaluated by LFB staining, had improved. This was however not confirmed in tissue sections stained for PLP (Figures 3, 4).

After six weeks of cuprizone exposure, microglia activation and macrophage infiltration were observed in all diet groups compared to the control groups, as measured by immunohistochemical staining for Mac-3. However, in the groups fed a high vitamin D₃-diet (6200 or 12500 IU/kg), the number of Mac-3 immunopositive cells were significantly lower than in the groups fed a diet deficient of vitamin D₃ (<50 IU/kg), or with a low (500 IU/kg) vitamin D₃ content (p = 0.001) (Table 2, Figure 2). There were no significant differences in the Mac-3 immunopositive cell counts between the mice fed 6200 IU/kg and 12 500 IU/kg (p = 0.795), or between the mice fed <50 IU/kg and 500 IU/kg (p = 0.682). Two weeks after discontinuation of the cuprizone treatment, there was a significant reduction in the Mac-3 immunopositive cell density in the groups fed <50 IU/kg and 500 IU/kg (p = 0.018 and p = 0.014, respectively). A reduction was also observed in the groups fed 6200 IU/kg and 12500 IU/kg, this reduction was not statistically significant (p = 0.054 and 0.344) (Table 2, Figure 4).

The extent of T-lymphocyte infiltration in the corpus callosum was studied by immunohistochemical staining for the pan-T-lymphocyte marker CD3. There was no significantly increased infiltration of CD3 immunopositive cells in the cuprizone exposed mice compared to in the controls (data not shown). Only single immunopositive cells contained within blood vessel walls were observed in both groups.