To compare the efficacy of MOG + VitD or MOG + PARI delivered by different routes, mice were allocated into the following groups: group 1: EAE control group; groups 2 and 3: EAE mice treated with MOG + PARI or MOG + VitD by i.p. route (eight PARI or VitD doses in alternate days and two MOG doses on days 3 and 11 after EAE induction); groups 4 and 5: EAE mice treated with MOG + PARI or MOG + VitD by epicutaneous route (groups treated twice with an epicutaneous patch containing MOG + PARI or MOG + VitD on days 3 and 11); groups 6 and 7: EAE mice treated with PARI or VitD by i.p. route (eight doses in alternate days) and MOG by epicutaneous route (two doses on days 3 and 11). Clinical EAE parameters and serum calcium levels were measured during disease acute phase to compare the efficacy and a possible hypercalcemia of epicutaneous MOG associated with VitD or PARI. This first experimental design is depicted below by a timeline scheme.

To investigate the efficacy of MOG + PARI to control EAE, mice were allocated into five experimental groups: normal mice, non-treated EAE group, and three treated EAE groups: MOG/EAE, PARI/EAE, and MOG + PARI/EAE. Three and 11 days after EAE induction, the animals were treated by epicutaneous application of MOG, PARI, or MOG + PARI in their backs. Clinical parameters were followed for 30 days. Immunological and histopathological evaluations were performed during the acute disease phase, i.e., 19 days after EAE induction. A sham control group, i.e., an EAE group that was treated with a patch containing only the MOG + PARI diluting agent, was also included in the last experiments. The adopted experimental design is illustrated below by a timeline scheme.

Six-week-old female C57BL/6 mice were purchased from the University of São Paulo (USP – Ribeirão Preto, SP, Brazil) and maintained in specific pathogen-free facilities. Mice were allocated in cages (maximum six animals per cage) with sterilized food and water ad libitum.

Eight-week-old animals were subcutaneously injected, into the lower back, with 100 μg MOG_35–55 peptide (MEVGWYRSPFSRVVHLYRNGK) (Genemed Synthesis Inc., San Antonio, TX, USA) emulsified with Complete Freund’s Adjuvant (Sigma Aldrich, St. Louis, MO, USA) containing 4 mg/mL of Mycobacterium tuberculosis (Difco, Detroit, MI, USA). Mice also received two i.p. doses, 0 and 48 h after MOG immunization, of 200 ng Bordetella pertussis toxin (Sigma Aldrich). Clinical score and body weight were assessed daily during 19 or 30 days according to the following criteria: 0, no symptoms; 1, limp tail; 2, hind legs weakness; 3, partially paralyzed hind legs; 4, complete hind legs paralysis; and 5, complete paralysis. The body weight variation was calculated considering the difference between the weight at the moment of EAE induction (day 0) and the one at the acute disease phase (day 19).

Mice were treated by i.p. route with 0.1 µg of 1α,25-dihydroxyvitamin D3 from Sigma Aldrich or 0.1 µg of PARI (Zemplar^®, Abbott Laboratories, Milan, Italy), every other day during 15 days (on days 1, 3, 5, 7, 9, 11, 13, and 15) beginning 24 h after EAE induction. MOG (150 µg) was co-administered by i.p. or epicutaneous route on days 3 and 11. Some experimental protocols required concomitant application of MOG and PARI by epicutaneous route on the 3rd and 11th days after EAE induction. To perform epicutaneous immunization, mice were anesthetized with ketamine/xylazine and then shaved on their backs with a depil machine and a hair removal cream (Veet^®, Reckitt Benckiser Healthcare, Hull, North Humberside, UK). Test substances were applied in a gauze in the center of an occlusive patch (2 cm × 2 cm) (DuoDerm Extra Thin, ConvaTec, Princeton, NJ, USA), which was then affixed to the shaved mice back. A surrounding fringe of adhesive medical tape (Micropore™, 3M United States) was used to assure the proper position and the permanence of the patch at animal’s back for 4 days.

Mononuclear cells infiltrated in the CNS were obtained as previously described (38). Nineteen days after EAE induction, mice were anesthetized with ketamine/xylazine and perfused with 10 mL saline solution. Brain and spinal cord were collected and digested with 2.5 mg/mL of collagenase D (Roche Applied Science, Indianapolis, IN, USA) at 37°C for 45 min. After maceration, suspensions were washed in RPMI medium (Sigma Aldrich) and centrifuged at 450 × g at 4°C for 10 min. Cells were resuspended in Percoll (GE Healthcare, Uppsala, Sweden) 37% and gently laid over Percoll 70%. After centrifugation at 950 × g for 20 min with centrifuge breaks turned off, the ring containing mononuclear cells was collected, washed in RPMI, and centrifuged again at 450 × g for 10 min. Cells were then resuspended in RPMI medium, counted, and analyzed. Pools from two animals were used to get enough cell number to perform the experiments.

Draining lymph nodes (axillary, brachial and inguinal) were collected and adjusted to 5 × 10⁶ cells/mL. Cells plated in supplemented RPMI medium (10% of fetal calf serum and 2 mM of glutamine) were restimulated in vitro with MOG (20 µg/mL). Cytokine levels were evaluated 48 h later by enzyme-linked immunosorbent assay in culture supernatants using IL-2, IFN-γ, and IL-10 OptEIA Sets (Cat#555148; Cat#555138; and Cat#555252) (BD Biosciences, Franklin, San Diego, CA, USA) and IL-6, IL-17, and TGF-β Duosets (Cat#DY406; Cat#DY421; and Cat#DY1679) (R&D Systems, Minneapolis, MN, USA) according to the manufacturer’s instructions.

The BV-2 cell line (BCRJ code 0356) was used to determine the cytokine production and the activation phenotype of microglial cells in vitro. This cell line was generated from a C57BL/6 mice brain tissue transformed by a retrovirus and retains most of the morphological, phenotypical, and functional properties of freshly isolated microglial cells (39). These cells were propagated and tested according to the study by Dulla et al. (40). Briefly, BV-2 cells were cultivated into 24-well plates at 4.0 × 10⁵ cells/well in supplemented DMEM High Glucose medium (10% of fetal calf serum and 1% of sodium pyruvate). After 2.5 h, cells were treated with PARI (1 µM) and MOG (50 µg/mL). 30 min after initial treatment, LPS (100 ng/mL) was added to cultures. Cells were incubated for 24 h in 5% CO₂ atmosphere at 37°C. The supernatants were collected, and the levels of IFN-γ, IL-6, IL-10, and nitric oxide (NO) were determined.

Nitric oxide levels in BV-2 supernatants were measured by quantification of nitrite levels (a stable end product of NO) according to the method described by Green et al. (41). Briefly, 0.1 mL of cell-free supernatant was incubated with an equal volume of Griess reagent containing 1% sulfanilamide (Synth, Diadema, SP, Brazil), 0.1% naphthalene diamine dihydrochloride (Sigma Aldrich), and 2.5% H₃PO₄, at room temperature for 10 min. Nitrite accumulation was then quantified using a chemiluminescence microreader (ELx 800; BioTek Instruments Inc., Winooski, VE, USA). The concentration of nitrite was determined using sodium nitrite (Sigma Aldrich) diluted in distilled water as a standard.

Blood samples were collected after euthanasia by cardiac puncture and serum was obtained after centrifugation at 3,800 × g at room temperature for 15 min. Briefly, 25 µL serum samples were incubated with beads conjugated with fluorochromes from Cytometric Bead Array mouse Th1/Th2/Th17 cytokine kit (BD Biosciences). The acquisition of the beads was performed using a FACS Canto II flow cytometer (BD Biosciences) from Institute of Biosciences (UNESP, Botucatu, SP, Brazil) with FACS Diva software, and the data were analyzed with FCAP Array 3.0 (Soft Flow Inc., St. Louis Park, MN, USA).

Lymph nodes (axillary, brachial and inguinal) and CNS-isolated cells were collected, and the red blood cells were lysed with buffer containing NH₄Cl. To analyze Tregs, samples were incubated with 0.2 µg PerCP-Cy5.5-labeled anti-mouse CD3 (145-2C11), 0.25 µg FITC-labeled anti-mouse CD4 (GK1.5), and 0.25 µg APC-labeled anti-mouse CD25 (PC61.5) for 30 min at 4°C. Intracellular Foxp3 transcription factor was detected using Foxp3 Staining Set (RRID: AB_465935) (eBiosciences, San Diego, CA, USA) according to manufacturer’s instructions. For myeloid DCs analysis, lymph node cells were incubated with 0.1 µg PerCP-labeled anti-mouse F4/80 (BM8), 0.25 µg FITC-labeled anti-mouse CD11c (N418), 0.03 µg APC-labeled anti-mouse MHCII (MS/114.15.2), and 0.125 µg PE-labeled anti-mouse CD86 (GL1) for 30 min at 4°C. For microglial cells, mononuclear cells isolated from CNS were incubated with 0.5 µg FITC-labeled anti-mouse CD45 (30-F11), 0.2 µg PerCP-Cy5.5-labeled anti-mouse CD11b (M1/70), 0.03 µg APC-labeled anti-mouse MHCII (MS/114.15.2), 0.15 µg PE-labeled anti-mouse CD40 (1C10), and 0.1 µg PE-Cy7-labeled anti-mouse PD-L1 (MIH5) for 30 min at 4°C. After staining, the cells were washed, resuspended in flow cytometry buffer, and fixed in 1% paraformaldehyde. Flow cytometry was performed using a FACS Canto II (BD) from Institute of Biosciences (UNESP, Botucatu, SP, Brazil), and the data were analyzed with FlowJo software (TreeStar, Ashland, OR, USA).

RNA was extracted from CNS-isolated cells with TRIzol™ reagent (Life Technologies, Austin, TX, USA). 200 ng of RNA was converted to cDNA using High Capacity cDNA Reverse Transcription kit (Life Technologies) according to the manufacturer’s instructions. Expression of Foxp3 (Mm00475162_m1) and RORc (Mm01261022_m1) was analyzed in comparison to GAPDH (Mm99999915_g1, housekeeping gene) levels. Real time-PCR reactions were performed using TaqMan™ Gene Expression Assays according to manufacturer’s recommendations (Applied Biosystems, Carlsbad, CA, USA). Levels of gene expression were represented as relative fold difference by using the method of delta threshold (2^−ΔΔCt).

The histopathological analysis was performed 19 days after EAE induction, during the acute phase of the disease. After euthanasia, lumbar spinal cord samples were removed and fixed in 10% neutral buffered formalin. Tissues were dehydrated in graded ethanol and embedded in Paraplast Plus (McCormick, St. Louis, MO, USA). Serial sections with 5-µm thickness were cut and stained with hematoxylin and eosin. Five photomicrographs were obtained from each animal with a Nikon microscope. A semiquantitative analysis of CNS inflammation was performed according to the following criteria, as previously described (42): 0, no infiltrates; 1, partial meningeal infiltration; 2, pronounced meningeal infiltration, and 3, pronounced meningeal and some parenchymal infiltration.

Myelin sheath thickness was assessed in lumbar spinal cord samples during the EAE acute phase by transmission electron microscopy. Mice were anesthetized with ketamine/xylazine and perfused intracardially with 10 mL Karnovsky solution. After removing spinal cords from the vertebral canal, tissues were fixed during 24 h in Karnovsky solution at 4°C. Specimens were rinsed in phosphate buffer and treated with 1% osmium tetroxide for 2 h. After another rinse in phosphate buffer, tissues were treated for 2 h with 0.5% uranyl acetate followed by dehydration in graded acetone series and finally embedded in Araldite resin. To assure a representative comparison in the myelin sheath among the experimental groups, a corresponding area of lumbar spinal cord was used to obtain 1-µm thick section. From this area located in the more infiltrated region, ultrathin sections with 90 nm were mounted on 150 mesh formvar-coated copper grids (Electron Microscopy Sciences, Hatfield, PA, USA) and contrasted with 1% uranyl acetate for 20 min and lead citrate solution for 10 min. Sections were examined on a Tecnai Spirit Transmission Electron Microscope (FEI Company, OR, USA) at the Electron Microscopy Center of the Institute of Biosciences (UNESP, Botucatu, SP, Brazil). Ten images per mouse were taken at 2,900× magnification by using a digital camera connected to electron microscope.

To perform the spinal cord barrier permeability assay, we used the Naflu test based on the methodology described by Christy et al. (43). Twelve days after EAE induction, mice were i.p. inoculated with 100 µL fluorescein sodium (NaFlu, Sigma Aldrich). After 20 min, the animals were anesthetized with ketamine/xylazine, and blood samples were collected by cardiac puncture with heparinized syringe. After that, mice were perfused with 10 mL saline solution, and the whole spinal cord sections were collected in microtubes and homogenized with 400 µL 0.9% saline solution. After centrifugation at 9,000 rpm for 10 min at 22°C, the supernatant of spinal cord and the plasma were added in black 96-well immuno plates. The fluorescence was measured in BioTek Synergy™ 4 hybrid microplate reader (emission 485 nm/excitement 528 nm). Results were expressed in relative fluorescence units (RFUs). The following equation was used to evaluate NaFlu incorporation in the spinal cord of each animal: NaFlu incorporation = (spinal cord RFU/spinal cord weight)/(plasma RFU/blood volume).

Data were expressed as mean ± SD or with median and interquartile (25–75%) ranges. To test normality of data, results were analyzed by Shapiro–Wilk’s test. Comparisons between experimental groups were made by one-way ANOVA followed by Tukey’s test for parametric variables and by Kruskal–Wallis test followed by Dunn’s test for non-parametric variables. Chi-square and Fisher exact test were performed for EAE incidence. Statistical analysis was accomplished with SigmaPlot Software Version 12.0 (Systat Software Inc., San Jose, CA, USA), and p < 0.05 was considered significant.

We recently described that MOG + VitD association inhibits EAE development in both prophylactic (44) and therapeutic (37) approaches. In both investigations, MOG and VitD were administered by i.p. route. Considering that VitD and PARI share immunomodulatory activities in vitro (34), that PARI does not trigger hypercalcemia, and that i.p. route is not suitable for human use, it was necessary to evaluate the efficacy of other routes and also the suitability of PARI as a VitD substitute. EAE mice were treated every other day during 15 days with VitD or PARI, and MOG was co-administered on days 3 and 11. This combination of MOG with VitD or PARI was then tested by three different protocols whose objective was to compare the efficacy of different routes. As depicted in Figure 1A, the administration of MOG and VitD or PARI by the same route, i.e., intraperitoneal or epicutaneous, was more efficient to control EAE development than their application by distinct routes. When epicutaneous MOG was associated with VitD or PARI by i.p. route, the disease incidence was still very elevated (around 50%). These data also indicated that PARI does not trigger body weight loss (Figure 1A) neither hypercalcemia (Figure 1D) as occurs with VitD. Comparison of clinical scores, showed in Figure 1B, reveals that MOG + VitD or MOG + PARI delivered by intraperitoneal or epicutaneous route control the disease in a very similar way with no statistical differences (Figure 1C).

As i.p. MOG + PARI was as efficient as epi MOG + PARI and considering that epicutaneous route could be easily translated to human application, this procedure was chosen to be investigated in more details. Initially, to certify that administration of PARI would improve MOG efficacy when the two substances were administered concomitantly by the same epicutaneous route, mice were submitted to EAE induction, and after 3 and 11 days, they received an occlusive patch containing MOG, PARI, or MOG plus PARI (Figure 2A). Animals were daily monitored for disease development during 30 days until EAE chronic phase. Co-administration of MOG + PARI, both by epicutaneous route, determined a much milder disease in comparison with MOG or PARI administration alone (Figure 2B). The protective effect of co-administration was supported by significant lower clinical scores (Figures 2C,D), no body weight loss (Figure 2F) and a clear decreased disease incidence (Figure 2E). As predicted, serum calcium levels in PARI and MOG + PARI groups were not altered (Figure 2G). Sham treatment did not protect the animals (data not shown).

As a recent work from our group indicated downregulation of encephalitogenic cytokine production in the spleen of EAE mice submitted to therapy with VitD + MOG association via i.p. route (37), we expected that this alternative therapeutic route would have the same effect. Unexpectedly, epicutaneous therapy with MOG, PARI, or MOG + PARI similarly increased pro-inflammatory cytokine production by lymph node cell cultures re-stimulated in vitro with MOG, when compared to the EAE control group (Figures 3A–D). Nonetheless, a regulatory immune profile was equally present in the lymph nodes from treated EAE mice, characterized by higher levels of TGF-β (Figure 3F), elevated percentages of Foxp3⁺ Tregs (Figure 3G), and lower MHCII fluorescence intensity in DCs (Figure 3H). Higher amounts of IL-10 were produced by the three treated EAE groups (Figure 3E). The levels of serum cytokines did not follow the same pattern found in lymph node cell cultures. MOG increased IFN-γ (Figure 4A) and TNF-α (Figure 4B), but did not alter IL-6 (Figure 4C) and IL-17 (Figure 4D). Nonetheless, PARI treatment slightly decreased IFN-γ (Figure 4A), TNF-α (Figure 4B), and IL-6 (Figure 4C) serum levels. When PARI was associated with MOG, the reduction of encephalitogenic cytokines in the serum was more pronounced, being the reduction in IL-6 production statistically significant (Figure 4C).

Bearing in mind that EAE is characterized by infiltration of immune cells in the CNS and predominant demyelinated axons in the acute phase of the disease (45), the therapeutic effect of MOG + PARI association could be related to a milder CNS pathology. Stained lumbar spinal cord sections were semiquantitatively evaluated concerning the degree of inflammation during the EAE acute phase (Figure 5A). Although the EAE group presented a pronounced inflammatory process in the meninges and in the spinal cord parenchyma (Figure 5C), MOG and MOG + PARI treatments significantly reduced the extension of the inflammatory infiltrate as illustrated at Figures 5D,F. Nonetheless, the degree of inflammation in the lumbar spinal cord of mice treated with PARI was similar to the one found in non-treated EAE animals (Figure 5E). By using electron microscopy, we found that MOG (Figure 5I) and MOG + PARI (Figure 5K) association conserved nerve fibers with myelin sheath, whereas PARI treatment (Figure 5J) showed areas with demyelinated axons (complete loss of the myelin sheath) similar to EAE control group (Figure 5H). Spinal cord sections from normal mice (control) stained with hematoxylin and eosin or prepared for electron microscopy are represented in Figures 5B,G respectively.

Reduced inflammation revealed by histopathological analysis was further confirmed by determination of the CD3⁺CD4⁺ percentage, by cytometric analysis, in cells eluted from the CNS (Figure 6A). A NaFlu test made during the preclinical phase indicated that reduced inflammation was not due to a decreased blood–spinal cord barrier permeability (Figure 6B). Considering that immunomodulatory properties of vitamin D and its analogs are usually correlated with the suppressive activities of Tregs (34, 46), mRNA expression for Foxp3 and RORc transcription factors and the frequency of CD3⁺CD4⁺CD25⁺Foxp3⁺ cells were evaluated in mononuclear cells recovered from the CNS during the acute phase of the disease. The more efficient therapeutic effect of MOG + PARI treatment in EAE development coincided with a higher ratio between the percentages of CD25⁺Foxp3⁺ and CD3⁺CD4⁺ cells (Figure 6C). Besides, MOG + PARI treatment determined a significantly augmented ratio between relative mRNA expression for Foxp3 and RORc (Figure 6D).

As VitD is able to reach the CNS and to interact with microglial cells (32), a similar characteristic was expected from the analog. This possibility was investigated in microglial cells eluted from treated EAE mice and also in BV-2 cell line. As demonstrated in Figure 7B, microglial cells eluted from the CNS of mice treated with PARI or MOG + PARI showed a similar and significant reduction in the mean fluorescence intensity (MFI) of MHCII. CD40 (Figure 7A) and PD-L1 (Figure 7C) MFI expression were not affected. This downmodulatory effect was confirmed by showing that PARI and MOG + PARI addition to BV-2 cells previously stimulated with LPS significantly decreased MFI MHCII expression (Figure 7E). In this experimental in vitro procedure, a significant reduction in CD40 (Figure 7D) and PD-L1 (Figure 7F) were also observed. The in vitro activity of MOG + PARI revealed other possible regulatory pathways involving cytokines and NO. In this case, MOG + PARI association decreased IL-6 (Figure 7H) and NO (Figure 7J) and increased IFN-γ (Figure 7G) and IL-10 (Figure 7I).