cDNAs of murine full-length MOG, MBP and proteolipid protein (PLP) were obtained by retrotranscription and PCR amplification of total RNA isolated from mouse brain using the following PCR primers: 5’-AAAGAATTCGATGGCCTGTTTGTGGAGCT-3’ and 5’-AAAGCGGCCGCCAGGAAGACACAACCATCACTCA-3’ for MOG; 5’-AAAGAATTCTAGCCTGGATGTGATGGCAT-3’ and 5’-AAAGCGGCCGCCAGGATTCGGGAAGGCTGA-3’ for MBP; 5’-AAAGAATTCAAGTGCCAAAGACATGGGCT-3’ and 5’-AAAGCGGCCGCGCTCAGAACTTGGTGCCT-3’ for PLP. Fragments containing the myelin proteins included EcoRI and NotI restriction sites that were used for cloning. PCR products were inserted into the pCi plasmid, a cytomegalovirus promoter driven mammalian expression vector (Promega, Mannheim, Germany), digested with EcoRI and NotI to create MOG-, MBP- and PLP-DNA, respectively. The empty vector (pCi) was used as control plasmid. Plasmid DNA was purified from the transformed E. coli strain DH5α and sequenced to verify integrity. Large-scale preparation of plasmid DNA was conducted using Qiagen plasmid kits (Qiagen, Hilden, Germany). Each preparation was checked by agarose gel electrophoresis, and DNA concentration was measured by optical density at 260 nm.

Five- to 8-week-old female C57BL6/J mice were purchased from Harlan (Italy). Experiments were done according to the EU regulations and approved by our institutional Ethics Committee on Animal Experimentation.

Mice were allocated to four groups and vaccinated with plasmid DNA vaccines encoding MOG, PLP, or MBP, or with the control plasmid (pCi). For prophylactic treatment, 100 μg DNA/mouse in phosphate-buffered saline (PBS) was injected intramuscularly in the tibialis anterior muscle at 28 and 14 days before EAE induction. For therapeutic treatment, mice received intramuscular (i.m.) injections of DNA (100 μg/mouse) at days 10 and 24 postimmunization (p.i.).

Anesthetized mice were immunized by subcutaneous injections of PBS containing 50 μg of MOG_35-55 (Proteomics Section, Universitat Pompeu Fabra, Barcelona, Spain) emulsified in complete Freund's adjuvant (Sigma Chemical, St Louis, MO, USA) supplemented with 2 mg/ml Mycobacterium tuberculosis H37RA (Difco Laboratories, Detroit, MI, USA). The animals received an additional intravenous injection of 150 ng Pertussis toxin in 100 μl PBS on the day of immunization and again 48 h later. Mice were weighted and examined daily for clinical signs of EAE, which were scored as follows: grade 0, no clinical disease; grade 1, tail weakness or tail paralysis; grade 2, hind leg paraparesis; grade 3, hind leg paralysis; grade 4, paraplegia with forelimb weakness or paralysis; grade 5, moribund state or death.

Animals were killed with carbon dioxide (>70%) at day 32 p.i. Brain and spinal cord were removed, fixed overnight in paraformaldehyde 4% and embedded in paraffin wax. Brain and spinal cord were cut into 4-μm-thick serial sections. For histopathology, sections were stained with hematoxylin and eosin (H&E), and Kluver-barrera (KB). For immunohistochemistry, endogenous peroxidase activity was blocked by incubating tissue sections in hydrogen peroxide (2%), methanol (70%) and PBS for 20 min. Ag unmasking was developed in citrate 10 mM (pH = 6), 10 mM Tris-EDTA (pH = 9), or protease type XIV (Sigma Chemical). Non-specific protein binding was blocked with 2% bovine albumin in PBS (blocking solution) at room temperature for 1 h. Sections were incubated overnight at 4 °C with the following primary antibodies diluted in blocking solution: rabbit anti-CD3 (DakoCytomation, Glostrup, Denmark) (T lymphocytes) dilution 1:100, Lycopersicon esculentum agglutinin (LEA) (Sigma Chemical) (macrophages/microglia) dilution 1:100, rabbit anti- glial fibrillary acidic protein (GFAP) (DakoCytomation) (astrocytes) dilution 1:500, rabbit anti-Olig2 (Millipore, Billerica, MA, USA) (oligodendrocytes) dilution 1:100, and mouse anti-200kD neurofilament heavy (SMI-32) (Abcam, Cambridge, UK) (axonal damage) dilution 1:100. All samples were incubated at room temperature for 1 h with the following secondary antibodies: biotinylated anti-rabbit or anti-mouse IgG (DAKO) (1:200 dilution in blocking solution); the avidin–biotin-peroxidase complex (ImmunoPure ABC Peroxidase Staining Kits, Pierce, IL USA) (1:100 dilution in PBS) was finally added for 1 h at room temperature. The peroxidase reaction was visualized with 2.5 mg/ml of 3,30-diaminobenzidine and 0.05% hydrogen peroxide. As a background control, the primary antibody (Ab) incubation was omitted. No signal was observed in any of the control slides.

Cell infiltration (H&E) was evaluated according to the following criteria: 0 - no lesion; 1 - cellular infiltration only in the meninges; 2, very discrete and superficial infiltrates in parenchyma; 3, moderate infiltrate (less than 25%) in the white matter; 4, severe infiltrates (less than 50%) in the white matter; 5, more severe infiltrates (more than 50%) in the white matter. Demyelination (KB staining) was scored as follows: 0 - no demyelination; 1, little demyelination, only around infiltrates and involving less than 25% of the white matter; 2, demyelination involving less than 50% of the white matter; 3, diffuse and widespread demyelination involving more than 50% of the white matter.

Three randomly chosen areas (1 mm²) along the spinal cord were analyzed in a blind manner. CD3-, GFAP- and LEA-positive cells were counted in infiltrates manually. SMI-32 and Olig2 quantification was performed with the ImageJ analysis software.

Mice were killed 14 days p.i. with CO₂ and spleens removed. Splenocytes were prepared by grinding the spleens through a wire mesh and cultured in 96-well plates at 2 × 10⁵ cells/well in a total volume of 200 μl of Iscoves modified Dulbecco's medium (PAA Laboratories GmbH, Pasching, Austria) supplemented with 10% HyClone® Fetal Clone I (Thermo Fisher Scientific, Waltham, MA, USA), 50 μmol/ml of 2-mercaptoethanol (Sigma Chemical), 2 mmol/ml of glutamine, 50 U/ml of penicillin and 50 mg/ml of streptomycin, all obtained from Gibco BRL (Paisley, UK). Cultures (three replicas) were stimulated with 10 μg/ml of MOG_35-55. Cells cultured without any stimulus were used as baseline controls. Cultures were incubated in a humidified atmosphere at 5% CO₂ and 37 °C for 72 h. After stimulation, supernatants were collected, and cytokine levels for IFN-γ, IL-10, IL-4, and IL-17 were determined with the FlowCytomix kit (Bender MedSystems, Burlingame, CA, USA) according to the manufacturer’s recommendations. GM-CSF (granulocyte-macrophage colony-stimulating factor) levels were measured using a commercially available ELISA (R&D, Minneapolis, MN, USA).

Nunc-Immuno Plates (Nalgene Nunc International, Roskilde, Denmark) were coated overnight with 0.1 μg/well of MOG_35–55. Serum samples were added in duplicate and after extensive washes were incubated with a secondary horseradish peroxidase-conjugated goat anti-mouse IgG (H + L). Upon addition of TMB Substrate Reagent Set (BD Pharmingen, San Jose, CA, USA), plates were read at 450 nm in an ELISA reader.

Freshly isolated splenocytes (5 × 10⁵) were washed with PBS-azide and incubated with PerCP-labelled anti-CD45, APC-Cy7-labeled anti-CD4, and PE-Cy7-labeled anti-CD25 for 30 min at 4 °C. Cells were then washed, and the intracellular detection of FoxP3 was conducted using a commercial FoxP3 staining protocol (eBioscience, San Diego, CA, USA). Data acquisition was performed on a FACSCanto^TM and analyzed with the DIVA software (BD Pharmingen).

Total mRNA was extracted using TRI reagent (Sigma Chemical) from CNS tissue. Expression was analyzed with the Affymetrix Mouse Gene 1.0 Array using the Ambion WT Expression kit (Applied Biosystems, Foster City, CA, USA) for target amplification and WT Terminal Labeling kit (Affymetrix, Santa Clara, CA, USA) for target labeling. The gene-level log-scaled robust multiarray analysis (RMA) was performed with the Affymetrix Expression Console software. Linear models for microarray data (LIMMA) R package 11 were used to identify differentially expressed genes between the MOG-DNA-treated and control plasmid-treated groups. Genes that were observed as differentially expressed by two-sample t-test with p-value <0.05 were considered significant. Gene Ontology (GO) term enrichment was performed by using Ontologizer 2.0 12. Data were analyzed with this program using a model-based gene set analysis (MGSA) 13. This method significantly reduces the number of redundant categories returned by the classical gene-category analysis. Pathway enrichment analysis was carried out with Ingenuity Pathway Analysis (IPA). Microarray data are stored in the NCBI Gene Expression Omnibus repository and are available at http://www.ncbi.nlm.nih.gov/geo/ with the entry number GSE30482.

Total mRNA was extracted from CNS tissue and retrotranscribed as described above. RT-PCR was performed using the ABI-Applied Biosystems 7900 HT Thermal Cycler (Applied Biosystems) in 384 optical PCR plates. Each reaction contained 2 μl standard/cDNA template, 10 μl of TaqMan Universal PCR Master Mix, No AmpErase®, UNG, 1 μl of corresponding TaqMan® Gene expression assay (Applied Biosystems), and 7 μl of RNase free water following the standard PCR program suggested by the manufacturer. mRNA expression levels of glyceraldehyde-3-phosphate dehydrogenase (Gapdh), brain-derived neurotrophic factor (Bdnf), neurotrophin-5 (Ntf5), platelet-derived growth factor alpha (Pdgfa)FoxP3, glial cell line-derived neurotrophic factor (GDNF) family receptor alpha like (Gfral), GDNF family receptor alpha 2 (Gfra2), GDNF family receptor alpha 4 (Gfra4), and semaphorin-3 F (Sema3F) were determined by RT-PCR relative quantification in MOG-DNA-treated and control plasmid-treated mice. Briefly, the threshold cycle (C_T) value for each reaction, and the relative level of gene expression for each sample were calculated using the 2^-ΔΔCT method 14. To correct for loading differences, the values were normalized according to the level of expression of the housekeeping gene, Gapdh, within each sample. Its C_T value was subtracted from that of the specific genes to obtain a ΔCT value. Differences (ΔΔCT) between the ΔCT values obtained for the control plasmid (calibrators) and the ΔCT values for the MOG-DNA groups were determined. The relative quantitative value was then expressed as 2^-ΔΔCT, representing the fold change in gene expression normalized to the endogenous control and relative to the calibrators. Samples were determined in triplicates. Analysis was performed with the software SDS 2.3 (Applied Biosystems).

Statistical analysis was performed by using the SPSS 17.0 package (SPSS Inc, Chicago, USA) for MS-Windows. Depending on the applicability conditions, Mann–Whitney test or Student’s t-test were used for comparisons of mean values between groups. Differences were considered statistically significant when p-values were below 0.05. Descriptive data are presented as mean values (standard error of the mean - SEM) unless otherwise stated.

To evaluate whether the injection of DNA encoding MOG cDNA can be effective in protecting mice from EAE induction, female C57BL6/J mice were first vaccinated with MOG-DNA or control plasmid (pCi) as described in Methods and depicted in Figure 1A. Mean clinical disease scores were significantly lower in mice vaccinated with the MOG-DNA plasmid vector compared with the control plasmid group (Figure 1B). Prophylactic MOG-DNA treatment also delayed disease onset [mean (SEM): 10.6 days (0.7) in MOG-DNA-treated mice vs. 7.5 days (0.3) in control plasmid-treated mice; p = 0.006], and reduced clinical severity [mean cumulative clinical scores: 1.4 (0.4) vs. 3.8 (0.3) in MOG-DNA-treated and control plasmid-treated mice respectively; p = 0.003] (Figure 1C).

In order to investigate whether these findings were specific for MOG-DNA vaccination or were also observed with other myelin autoantigens, mice were vaccinated with plasmid DNA vaccines encoding MBP or PLP (antigenic controls) using the same above-mentioned prophylactic protocol. As shown in Additional file 1, vaccination with DNA encoding MBP was also associated with a significant reduction in disease severity, although to a lesser degree compared with MOG-encoding DNA vaccines [mean cumulative clinical scores: 2.5 (0.1) vs. 3.8 (0.3) in MBP-treated and control plasmid-treated mice respectively; p = 0.003]. Finally, EAE disease course was similar between PLP-DNA-treated and control plasmid-treated mice (Additional file 1).

We next addressed the question of whether MOG-DNA vaccination could reverse clinically established EAE. To this end, mice were left untreated until disease onset and subsequently vaccinated with MOG-DNA or vector DNA alone, as depicted in Figure 1D. Interestingly, therapeutic MOG-DNA treatment improved ongoing EAE, as reflected by the significant reduction in the mean cumulative EAE clinical score observed in MOG-DNA-treated mice compared with control plasmid-treated mice [2.2 (0.1) vs. 3.0 (0.3); p = 0.031] (Figure 1E and F).

These findings were specific for MOG-DNA treated mice, as mean EAE disease scores did not significantly differ between mice receiving therapeutic treatment with antigenic controls (MBP or PLP) and control plasmid (Additional file 1).

To assess whether EAE clinical improvement was accompanied by decreased neuropathology, histopathological studies were performed in CNS tissue from pCi and MOG-DNA treated mice. H&E and KB staining showed that mice treated with pCi had perivascular cuffs and extensive inflammatory infiltration in the white matter of the spinal cord (Figure 2A), as well as demyelination in areas with moderate to severe inflammatory infiltration (Figure 2B). Mice treated with MOG-DNA, on the other hand, were overall characterized by less severe CNS inflammation (Figure 2A) and significant reductions in microglia/macrophage activation (Figure 2D), astrogliosis (Figure 2E) and axonal damage (Figure 2F). Although demyelination and T cell infiltration were also reduced in MOG-DNA treated mice, differences did not reach statistical significance (Figures 2B and C respectively).

Based on the findings of a reduction in both EAE disease course and CNS pathology in MOG-DNA treated mice, we performed additional experiments in order to characterize the mechanisms by which MOG-encoding DNA vaccines produce their beneficial effects in EAE. All these experiments were conducted in mice vaccinated according to the prophylactic protocol described in Methods. We first measured MOG_35-55-induced T cell immune responses by determining ex vivo the levels of Th1, Th2, and Th17 cytokines using flow cytometry. As shown in Figure 3A, levels of the proinflammatory cytokines IFN-γ and IL17 were significantly reduced in mice vaccinated with MOG-encoding DNA compared to mice treated with control plasmid (p = 0.016 for IFN-γ; p = 0.013 for IL-17). However, levels of the anti-inflammatory cytokine IL-4 did not significantly differ between MOG-DNA treated and pCi-treated mice (p = 0.320; Figure 3A). Finally, levels of IL-10 in supernatants from stimulated mice splenocytes were below the detection limit of the technique.

Based on the IFN-γ and IL17 findings we also measured the levels of GM-CSF, which is known to play important roles in Th17 and Th1 cell function 1516. Supporting the IFN-γ and IL17 data, MOG-DNA treatment was associated with a trend towards decreased GM-CSF levels compared with plasmid-control mice [mean (SEM): 179.5 pg/ml (121.3) vs. 337.6 pg/ml (44.2) respectively; p = 0.065, Mann-Whitney U test)].

Altogether, these observations suggest that MOG-DNA vaccines exert some of their protective effects in EAE by inhibiting the secretion of proinflammatory cytokines such as IFN-γ and IL-17 rather than inducing Th2-type immune responses.

In order to characterize the autoreactive B-cell responses in vaccinated EAE mice, the presence of MOG-specific antibodies was determined in serum samples by ELISA. Anti-MOG antibodies were detected in all animals; however, Ab titers in MOG-DNA vaccinated mice were similar to those observed in control plasmid-treated mice (Figure 3B).

We next investigated whether treatment with MOG-DNA vaccines promoted the development of Treg in EAE animals. As shown in Figure 4A, the percentage of CD4⁺CD25⁺FoxP3⁺ Treg determined by flow cytometry was significantly higher in splenocytes from MOG-DNA-treated mice compared with animals treated with the empty plasmid [20.5% (3.9) vs. 9.3% (1.7), p = 0.025]. Given the potential of Treg to migrate into the CNS in response to inflammation, we also determined the expression of the Treg-specific transcription factor FoxP3 in CNS tissue from EAE mice by RT-PCR. Interestingly, MOG-DNA vaccination was associated with a significant increase in CNS FoxP3 mRNA expression levels compared with the control plasmid-treated condition (p = 0.014; Figure 4B). Taken together, these findings suggest that DNA vaccination with MOG may produce its beneficial effects through an expansion of the Treg population in the periphery and subsequent accumulation of Treg in the CNS.

In order to characterize the mechanisms of action associated with the efficacy of MOG-DNA vaccines, we determined the gene expression profiles induced by DNA vaccination in the CNS using cDNA microarrays. A total of 2,462 genes were differentially expressed between MOG-DNA-treated and control plasmid-treated mice (p < 0.05). Of these, 1,382 genes were downregulated and 1,080 up-regulated in the MOG-DNA-treated group. In agreement with the pathological findings, gene enrichment analysis using the GO terms revealed the inflammatory response category to be the most represented among downregulated genes in mice vaccinated with MOG-DNA (Table 1). A detailed list of differentially expressed genes involved in the inflammatory process is provided in Additional file 2. In order to identify the cellular pathways induced by the effect of MOG-DNA vaccination, a more in depth functional enrichment analysis was performed with the Ingenuity software. Of note, one of the top enriched pathways that contained a high percentage of upregulated genes in the MOG-DNA-treated group was the axonal guidance signaling pathway (Additional file 3), which included genes coding for neurotrophic factors and proteins involved in the remyelination process. These findings pointed to a potential mechanism of action of MOG-DNA vaccines upregulating genes with neuroprotective roles. A search for additional neuroprotective genes (summarized in Table 2) was performed in order to validate microarray findings by an alternative technique.

GO-oriented analysis with differentially expressed genes obtained with microarrays from brain and spinal cord tissue was conducted in order to classify biologically related genes. Ontologizer 2.0 software was used for GO term enrichment analysis.

^aMGSA: score obtained using a model-based gene set analysis. Only categories with a score >0.5 are represented.

^bEnrichment values indicate how many times genes from the corresponding GO categories are overrepresented in the list of differentially expressed genes compared to the list of random genes.

^cExpression: refers to the direction in gene expression observed in MOG-DNA treated mice (n = 5) versus plasmid control mice (n = 5). GO ID: GO identification.

Differentially expressed genes obtained with microarrays between MOG-DNA-treated mice (n = 5) and control plasmid-treated mice (n = 5) were estimated using Student’s t test, as described in Methods.

^aRefers to Entrez gene ID.

^bFold change expression in MOG-DNA-treated mice versus control plasmid-treated mice. Data represent the mean ratio of five arrays. GDNF: glial cell line derived neurotrophic factor.

As depicted in Figure 5, gene expression levels for neurotrophic factors determined by RT-PCR were overall found to be increased in mice vaccinated with MOG-encoding DNA compared with mice receiving control plasmid, and microarray differences were validated for Bdnf (p = 0.029); Gfra2 (p = 0.039) and Gfra4 (p = 0.003). A trend towards higher expression levels was also observed for Ntf5 (p = 0.093), whereas differences for Gfral did not reach statistical significance (p = 0.258).

In microarray studies, two of the neuroprotective genes that were up-regulated in the MOG-DNA treated group, Pdgfa and Sema3f, are known to be involved in the remyelination process (Table 2) and hence were also selected for validation. As shown in Figure 6A, Pdgfa was found to be significantly overexpressed by RT-PCR in MOG-DNA vaccinated mice (p = 0.017), whereas the increased expression levels observed for Sema3f did not reach statistical significance (p = 0.178).

In order to provide histopathological evidence for a potential positive effect of DNA vaccines in remyelination, we also determined the expression of the transcription factor Olig2, a marker strongly expressed in oligodendrocyte progenitor cells (OPC) 17, in CNS tissue of vaccinated mice. Interestingly, a trend towards higher numbers of Olig2⁺ progenitor cells was observed in MOG-DNA-treated mice compared with plasmid control mice (p = 0.07; Figure 6B), suggesting that OPC are augmented in demyelinated CNS lesions from mice vaccinated with MOG-DNA.

Altogether these findings point to a potential neuroprotective effect of MOG-DNA vaccines favouring the remyelination process.