A total of 324 MS patients and 175 age/sex-matched HCs were recruited from the Hospital Clínico San Carlos in Madrid, Spain. Recruitment took place from April 2010 to May 2020, with all participants providing informed consent via forms approved by the Clinical Research Ethics Committee. The MS cohort included patients with relapsing-remitting MS (RR-MS) either active (RR-AMS, with relapses and/or evidence of MRI activity within two weeks of symptoms onset) or stable MS (RR-SMS, defined by the absence of disease activity in the 3 months prior to sample collection) primary progressive MS (PP-MS), and secondary progressive MS (SP-MS) (McDonald criteria) (21). Demographic, clinical, and radiological data were collected from medical records or during study inclusion, as summarized in Table 1 . Participants had not received disease-modifying therapy for at least one month before the blood collection. Specifically, RR-AMS samples were collected prior to corticosteroid administration, and some untreated patients were sampled after switching treatments and completing to another and a washout period.

Peripheral blood (20 mL) was collected from each participant: 10 ml into dry tubes for serum collection and 8 ml into CPT™ tubes (Cell Preparation Tube; Becton Dickinson, Franklin Lakes, NJ, USA) for PBMC isolation. Serum was obtained by centrifugation at 900 x g for 15 min at room temperature (RT) and stored at -80°C until analysis. PBMCs were isolated using Ficoll-Paque™ gradient centrifugation at 900 x g for 30 min at RT. After centrifugation, the cells were washed with saline buffer and centrifugated again at 400 x g for 10 min at RT, discarding the supernatant. The cell pellet was then resuspended in a mixture of 10% dimethyl sulfoxide (DMSO) and 1 ml of fetal bovine serum and stored temporarily at -80° in a MrFrostie container before being transferred to liquid nitrogen (-196°C).

Synthetic peptides with a purity > 95% were sourced from LifeTein (South Plainfield, NJ, USA) and SynPeptide Co Ltd (USA). These peptides were dissolved in DMSO at a concentration of 10 mM and stored at -80°C. The peptides included syncytin-1_{env 486–500} (UniProt accession no. Q9UQF0) with the sequence QMEPKMQSKTKIYRR, and pHERV-W_{env 486–504} (UniProt accession no. Q991W9) with the sequence QIVLQMEPQMQSMTKIYRG.

Serum antibodies against retroviral peptides were detected using an indirect ELISA to measure immunoglobulin G (IgG). To perform the peptide-based indirect ELISA, ninety-six-well immune-plates were coated with 50 µL per well of either the syncytin-1_{env 486-500} or pHERV-W_{env 486–504} peptides, both at equimolar concentrations of 10 μM, diluted in ELISA coating buffer at 0.05 M of carbonate–bicarbonate (pH 9.5, Sigma-Aldrich, St. Louis). The plates were incubated overnight at 4°C. The following day, the microplates were incubated for 1 hour at RT (25°C) using a blocking solution of 1% non-fat dried milk (Sigma-Aldrich, St. Louis, MO, USA) in tris-buffered saline (TBS). After blocking, the plates were washed twice with TBS containing 0.05% Tween-20 (TBS-T). Sera samples, diluted 1:10 dilution in the blocking solution, were added in duplicate wells and incubated for 2 hours at RT. The plates were then washed with TBS-T. Subsequently, the plates were incubated with 100 µL per well of alkaline phosphatase-labeled, Fc-specific, anti-human IgG polyclonal antibodies (1:5000, Sigma-Aldrich, St. Louis, MO, USA), for 1 hour at RT. After another wash, the wells were incubated with 200 µL per well of milli-Q water containing p-nitrophenyl phosphate (Sigma-Aldrich, St. Louis, MO, USA) for 30 minutes at RT in the dark. Optical density was measured at 405 nm using a SpectraMax Plus 384 microplate reader (Molecular Devices, Sunnyvale, CA, USA). Negative control wells were included on each plate, and the mean optical density value from these wells was subtracted from all other data points. Results were normalized against a positive control serum included in all experiments.

Serum samples were assessed using commercial tests: anti-EBNA-1 and anti-VCA IgG (Trinity Biotech, USA), anti-CMV IgG (Vircell, USA), and anti-HHV-6A/B IgG and IgM (Vidia, Ltd., Czech Republic). The analysis was performed using an automated ELISA processing system (DS2, Dynex Technologies, USA). Results were reported in artificial units (AU), calculated by multiplying the index value by 10 (where the index value is the sample absorbance divided by the cut-off value) (22, 23). Each sample was tested in duplicate. Values below 11 AU were considered negative, while samples with values between 9 and 11 AU were reanalyzed.

PBMCs were thawed from liquid nitrogen and washed three times (400 x g, 10 min, RT) in complete medium consisting of RPMI-1640 (Roswell Park Memorial Institute medium; Merck, Darmstadt, Germany), 10% fetal bovine serum, 0.5% streptomycin-penicillin, and 1% L-glutamine. Cell viability was assessed using trypan blue exclusion staining, and cell counting was performed with a Neubauer chamber. The cells were then seeded into a 12-well cell culture plate at a density of 2 x 10⁶ PBMCs per ml and cultured overnight in an incubator at 5% CO₂ and 37°C. The following day, PBMC cultures were washed in complete medium (400 x g, 10 min, RT), counted to assess viability and recovery, and resuspended at a density of 2 x 10⁶ per ml.

Cells from 14 MS patients (7 RR-SMS and 7 PP-MS) and 9 HCs were distributed into 5 tubes, each containing 5 x 10⁵ cells. The experimental conditions included negative control (RPMI to 10% FBS), positive controls (phorbol 12-myristate 13-acetate at 50ng/ml and calcium ionophore at 0.75µg/ml (Sigma Aldrich), vehicle control (DMSO at 0.75%), and two antigen stimulations with syncytin-1_{env 486–500} and pHERV-W_{env 486–504} peptides, both at 75µM, with costimulatory antibodies anti-CD28/CD49d (5µg/ml) (BD Biosciences, San Diego, CA, USA). The cells were incubated at 5% CO₂ and 37°C for 2 h. During the last 6 hours of culture, brefeldin A (BFA) (10 µg/ml) and monensin (0.7 µg/ml) (BD Biosciences, San Diego, CA, USA) were added. After incubation, cell cultures were washed with saline buffer (PBS) and centrifugated at 300 x g for 5 min at RT, discarding the supernatant.

Cell pellets were resuspended in 1 mL of culture medium and incubated with a mix of fluorochrome-conjugated antibodies specific for cell surface markers. The staining was performed for 20 min at 4°C in the dark. The antibodies used included: CD45-V500, CD3-BV421, CD8-APC-H7 and CD19-PE-Cy7 (BD) were added. Following incubation, the cells were washed with PBS and centrifugated at 500 g for 5 min at RT. The supernatant was discarded.

Cells were fixed and permeabilized using Cytofix/Cytoperm (BD Biosciences, San Diego, CA, USA) following the manufacturer’s instructions. After fixation, cells were washed with PERM-WASH buffer and centrifugated at 500 x g for 5 min at RT, discarding the supernatant. Intracellular staining was performed with fluorescein isothiocyanate (FITC) anti-IFN-γ, allophycocyanin (APC) anti-IL-17, Peridinin chlorophyll (PercP-Cy5.5) anti-TNF-α, and Phycoerythrin (PE) anti-GM-CSF. Staining was carried out at 4°C for 20 minutes in the dark. After extensive washing with PERM-WASH buffer, cells were resuspended in 300 μL of 1% paraformaldehyde in PBS. Samples were analyzed using FACS-Canto II 8-color flow cytometer (BD) and data were processed with CytExpert software (Beckman Coulter, USA), acquiring 150.000 events for sample. Analysis strategies for lymphocyte subpopulations are detailed in Supplementary Figure 1 .

Categorical variables were expressed as percentages, while normally distributed numerical variables were reported as mean ± standard deviation, and non-normally distributed variables as median (25th-75th percentile). ELISA results between patients and HCs were compared using the non-parametric Mann–Whitney’s U-test. The accuracy of the ELISA assay was assessed using receiver operator characteristic (ROC) curves. The coefficient of variation (CV) for intra-assay was capped at 10%. Outliers were identified by the Interquartile Range (IQR) method, with a no-negative constant of three. Cut-off values for positivity in each test were set at 95% specificity, and sensitivity was calculated accordingly. The Shapiro-Wilk test was performed to evaluate the normality of data distribution. Non-parametric tests, including Mann-Whitney, Kruskal-Wallis and Spearman correlation, were employed for multiple comparisons and correlation analyses. Discriminant analysis was used to study population distribution, followed by multivariate Hotelling T² analysis. This statistical method identifies which variable discriminates between groups based on quantitative and qualitative measures. The method extrapolates n-1 discriminant functions (DF), where n is the number of groups to be discriminated, which are linear combinations of the selected original quantitative variables. These derived functions can be used to calculate a set of discriminant scores that are used to predict the status of a new observation. Therefore, the first discriminator function, known as DF1, maximizes the variance between the variable’s values. On the other hand, the second discriminator function, known as DF2 which is orthogonal to DF1, maximizes the residual differences between values of these variables. The model parameters are the eigenvalues, a measure of the variance in the variable for each function; the Wilks’ lambda, an index of discriminating power within the range of 0 to 1 (the lower the value, the higher the discriminating power) and the canonical correlation a measure of the associations between the groups formed by the sets of variables and the DF, where the higher this value, the stronger the correlation between the groups and the DFs.

Statistical significance set at p-value of 0.05. Data analysis was performed using GraphPad Prism software (versions 8.0/9.0, San Diego, CA, USA) and SPSS software package (version 28.0, SPSS, Inc).

The peptide ratio was calculated to investigate the potential imbalance between the physiological syncytin-1 and the pathological pHERV-W. Antibody production against the same epitope varies among individuals due to factors like genetics. Analyzing the ratio of antibodies against pHERV-W and syncytin-1 two proteins with different roles, one pathogenic and the other physiological, allows us to account for this variability and to correct the individual heterogeneity in antibody production. Therefore, this ratio reflects the immune system’s response to two distinct antigens, and any imbalance between these antibodies could indicate that one antigen is driving a stronger immune reaction. Such shifts in immune response may provide understanding of the mechanisms and their potential role in disease pathogenesis.

The ratio of anti-pHERV-W_{env 486-504}/anti-syncytin-1_env486-500 IgG response in serum was higher across all categories of MS patients compared to HCs. Specifically, the ratio levels were as follows: RR-SMS 0.87, (0.57-1.21); RR-AMS 0.92 (0.72-1.35); PP-MS 1.14 (0.91-1.35); SP-MS 1.17 (0.973-1.33); HCs 0.79 (0.48–1.02) ( Figure 1 ).

In progressive forms of MS (PP-MS, SP-MS), there was a notably higher pHERV-W_{env 486-504}/syncytin-1_env486-500 IgG response compared to RR-MS, including both RR-SMS and RR-AMS phases ( Figure 1A ). Furthermore, RR-SMS patients with an EDSS > 4 showed higher pHERV-W_{env 486-504}/syncytin-1_env486-500 IgG-antibody titers (1.02, 0.70-1.29) than HCs. However, no significant difference was observed in the humoral response between RR-SMS patients with an EDSS < 2 (0.85, 0.61-1.15) and HCs ( Figure 1B ). Although it exists an imbalance with regard to demographic variables ( Table 1 ), no significant correlation was observed between the antibody response to both epitopes and variables such as age or gender.

Qualitative titer-based ELISA showed a higher prevalence of anti-EBV/EBNA-1 IgG in RR-SMS (95%), RR-AMS (100%), and PP-MS (97%) compared to SP-MS (92%) and HCs (86%). There was no difference in the prevalence of anti EBV/VCA IgG among all groups. Conversely, a lower prevalence of anti-CMV IgG was detected in RR-SMS (59%) compared to RR-AMS (69%), PP-MS (69%), SP-MS (68%), and HCs (74%).

Table 2 presents the quantitative analysis of HHV-6A/B IgM and IgG, along with EBV/EBNA-1, EBV/VCA, and CMV IgG antibody titers, highlighting significant differences between MS patients and HCs. RR-MS patients exhibited higher EBNA-1 and HHV6A/B IgG-IgM titers than HCs, while EBV/VCA titers were significantly elevated in RR-AMS, PP, and SP patients. Notably, lower CMV-IgG levels were observed in RR-SMS patients compared with HCs. Furthermore, graphical representation of this distribution is illustrated in Supplementary Figure 2 . Additionally, moderate positive correlations were detected between pHERV-W_{env 486-504/}syncytin-1_env486-500 and HHV-6A/B IgM⁺ (N =18, r = 0.6, p = 0.009) and IgG⁺ (N = 44, r = 0.3, p = 0.05) levels in PP-MS patients (data not shown).

A Multivariate Discriminant Analysis (MDA) was conducted to determine the factors that most effectively distinguish between the RR-SMS, SP-MS, and PP-MS groups. The analysis considered ten categorical and independent variables, encompassing both clinical and immunological factors. Table 3 presents the model’s parameters, including the eigenvalues, variance, canonical correlation, and Wilks’ Lambda value, along with the matrix structure coefficients that indicate the correlations between each categorical variable and the discriminant functions (DFs). The DF1 accounted for 87.3% of the total variance, demonstrating a strong canonical correlation value of 0.737 and a Wilks’ Lambda value of 0.389, while the DF2 accounted for the remaining 12.7% of the variance.

According to MDA, Supplementary Table 1 (Supplementary Data) summarizes discriminant coefficients used to assess the relative importance of the dependent variables. Standardized discriminant coefficients are provided to compare the relative significance of the clinical and immunological variables in predicting the dependent outcomes. Thus, variables with higher absolute values indicate a greater contribution to the discriminating power of the model.

The application of DF1 and DF2 achieved a correct classification rate of 74.2% for the original cases. Cross-validation of these functions resulted in a classification rate of 71.2%. Table 4 shows the classification matrix, illustrating how accurately the model categorized the participants groups in the study. The highest classification accuracy was achieved for RR-SMS patients at 81.5%, while the lowest was for the PP-MS patients at 63.2%. The greatest challenge in classification was observed between the primary and secondary progressive forms. The discriminative power of the two functions is illustrated in the scatter plot of Figure 2 for DF1/2. DF1 effectively distinguished between progressive forms (PP-MS and SP-MS) to RR-SMS, while DF2, with lower discriminative power, differentiate between primary and secondary progressive forms. Multivariate Hotelling’s T2 analysis, which employs the two discriminant functions, assessed the statistical similarity among patient groups. The Hotelling’s T2 test revealed significant differences between RR-SMS (blue) and PP-MS patients (red) (p < 0.001, centroids distance 2.2) as well as between RR-MS and SP-MS (green) (p < 0.001, centroids distance 2.3). Additionally, PP-MS and SP-MS patients exhibited significant differences (p < 0.001, centroids distance 1.3).

The percentage of CD4+ cells secreting IFN-γ, GM-CSF, TNF-α, and IL-17 was measured in MS patients and HCs ( Figure 3 ). In RR-SMS patients, a significant increase in TNF-α expression was observed in CD4+ cells exposed to syncytin-1_{env 486-500} and pHERV-W_{env 486-504} compared to DMSO control. In PP-MS patients, a significant increase in TNF-α expression was observed in CD4+ cells exposed to pHERV-W_{env 486-504} compared to DMSO. In contrast, none of the tested conditions induced cytokines production in CD4+ cells from HCs ( Figure 3 ). No significant difference in cytokines production by CD4+ cells exposed to DMSO (vehicle control), pHERV-W_{env 486-504} or syncytin-1_{env 486-500} were found between HCs, RR-SMS and PP-MS groups ( Figures 4A, D, G ).

The same experimental conditions applied to CD4 T-cells were assessed for CD8+ T cells. PBMCs from MS patients stimulated to pHERV-W _{env 486-504} showed that CD8+ T–cells expressed cytokines at frequencies greater than 0.5% and 0.4% compared to DMSO and syncytin-1 _{env 486-500}, respectively ( Figure 3 ). In RR-SMS patients, exposure to pHERV-W_{env 486-504} resulted in a significant increase in the expression of all studied cytokines in CD8+ cells compared to DMSO and syncytin-1_{env 486-500}. Similarly, PP-MS patients exhibited significantly higher levels of IFN-γ, GTM-CSF and TNF-α cytokines in CD8+ cells exposed to pHERV-W_{env 486-504} compared to DMSO and syncytin-1_{env 486-500} ( Figure 3 ). In HCs, significant differences were observed only in IFN-γ production by CD8+ cells exposed to pHERV-W_{env 486-504} compared to DMSO and syncytin-1_{env 486-500}. Furthermore, cytokines levels produced by CD8+ cells exposed to pHERV-W_{env 486-504} were significantly different between HCs and both RR-SMS and PP-MS patients ( Figure 4E ). Notably, CD8+ T-cells from PP-MS patients exhibited a substantial increase in GM-CSF (+0.42%; p < 0.01) and TNF-α (+0.34%; p = 0.03) levels compared to HCs, while RR-SMS patients had a significant increase in GM-CSF levels (+0.30%; p = 0.02) compared to HCs. No significant difference in cytokines production by CD8+ cells exposed to DMSO ( Figure 4B ) or syncytin-1_{env 486-500} ( Figure 4H ) were found between HCs, RR-SMS and PP-MS groups.

Intracellular cytokines levels in CD19 B-cells exposed to syncytin-1_{env 486-500} and pHERV-W_{env 486-504} were analyzed ( Figure 3 ). In RR-SMS patients, a significant increase in TNF-α and IL-17 expression were observed with pHERV-W_{env 486-504} compared to DMSO and syncytin-1_{env 486-500}, along with an increase in GM-CSF compared to DMSO ( Figure 3 ). PP-MS patients also showed significant increases in TNF-α expression with pHERV-W_{env 486-504} compared to both DMSO and syncytin-1_{env 486-500}. No significant cytokine increases were observed in HCs. Differences in cytokine levels between HCs, RR-SMS and PP-MS were significant with pHERV-W_{env 486-504} ( Figure 4F ). PP-MS patients had notably higher GM-CSF (+1.03%; p < 0.01) and RR-SMS patients had a higher TNF-α (+1.87%; p = 0.05) compared to HCs. No significant difference in cytokines production by CD19+ cells exposed to DMSO ( Figure 4C ) or syncytin-1_{env 486-500} ( Figure 4I ) were found between HCs, RR-SMS and PP-MS groups.

Significant positive correlations were observed between EBV/VCA-IgG antibody titers and pHERV-W_{env 486-504} cytokine-positive CD19+ B-cells in PP-MS patients ( Figure 5A ), particularly for IFN-γ and GM-CSF ( Figure 5A.1 ). Conversely, a strong negative correlation was found between IFN-γ, GM-CSF and TNF-α levels in RR-SMS patients, as well as between GM-CSF, TNF-α and CMV-IgG antibody titers in PP-MS patients ( Figures 5B, B.1 ). No significant correlation was detected with syncytin-1_{env 486-500} cytokine-positive CD8+ and CD19+ cells in either HC or MS populations.