PwMS and HD, belonging to this prospective single-center study, were recruited at the AOU San Luigi Gonzaga, Orbassano (TO, Italy) according to the following inclusion/exclusion criteria. A diagnosis of MS, according to the most recently revised McDonald criteria (29), and eligibility for anti-SARS-CoV-2 vaccination were considered as inclusion criteria. Any medical condition that does not allow the signing of informed consent and a prior history of symptomatic SARS-CoV-2 infection or breakthrough infection before the third dose were considered as exclusion criteria. All the subjects in the study were vaccinated with two doses of Comirnaty (ex mRNA BNT162b2) mRNA vaccine (Pfizer/BioNTech Inc, BioNTech Manufacturing GmbH) and then with the third dose (booster) of Comirnaty or Spikevax (ex mRNA-1273) vaccine (Moderna, Moderna Biotech Spain S.L.). COVID-19 disease was not reported from any of the subjects before vaccination. COVID-19 infection after vaccination was determined by self-reported positive COVID-19 test during the follow-up and/or presence of nucleocapsid Ab (Anti-N) in collected serum samples.

Blood and sera were collected immediately before the first dose of Comirnaty vaccine (Pfizer/BioNTech Inc, BioNTech Manufacturing GmbH) (P0), 4 weeks ( ± 15 days) (P1) and 6 months ( ± 15 days) (P6) after the booster vaccination. Sera were immediately frozen for further analysis. PBMCs were isolated by density gradient centrifugation using Histopaque-1077 (Sigma-Aldrich, St. Louis, MO, USA) from heparinized venous blood.

Anti-RBD IgG titers were measured with the SARS-CoV-2 RBD IgG ELISA (EIA-6150, DRG Diagnostics, Marburg, Germany, EU; lot number 142K061) following manufacturer instructions. Results are expressed in IU/ml (log10), and the cut-off threshold corresponded to 1.4 IU/ml (log10), according to manufacturer indications. Anti-N IgG were measured with the SARS-CoV-2 (COVID-19) IgG ELISA test (NOVATEC Immunodiagnostica GmBH, Dietzenbach, Germany, EU; lot number COVG-053) according to manufacturer method.

The human embryonic kidney cells (293T, ATCC, CRL-3216), the baby hamster kidney cells (BHK21, ATCC, C-13) and the hepatocyte derived cellular carcinoma cell line (Huh7) (ECACC, Cat num: 01042712) were grown as monolayers in Dulbecco’s modified Eagle’s medium (DMEM) (Sigma-Aldrich St. Louis, MO, USA), supplemented with 10% (v/v) heat inactivated fetal bovine serum (FBS) (Sigma-Aldrich) and a 1% (v/v) antibiotic solution (penicillin–streptomycin, Sigma-Aldrich) at 37˚C in 5% CO2 atmosphere.

SARS-CoV-2 pseudotyped viruses were produced and titrated according to Nie et al. (21) and were analyzed by means of Western Blot analysis to verify the presence of the SARS-CoV-2 S protein on the VSV envelope (30). The detailed protocols are reported in the Supplementary Materials .

The neutralization assay was performed according to Almahboub et al. (31) and Nie et al. (21). Huh7 cells were pre-seeded in a 96 well/plate at a density of 1.2x10⁴ cell/well. The following day, serum samples were heat inactivated at 56°C for 30 min in a water bath. Then, a fixed amount of SARS-CoV-2 pseudovirions (650 TCID₅₀/well) (21) was incubated with serial dilutions of serum samples (from 1:20 to 1:14580 in duplicate) for 1h at 37°C in continuous oscillation. As controls, six wells were incubated with only culture medium (CC wells) and six wells were infected but not treated (VC wells). After 1h incubation, the pre-treated virus was inoculated on Huh7 cells for 24h at 37°C to evaluate the residual viral infectivity. The detection was performed by adding the Britelite plus reporter gene assay system (PerkinElmer) to cells in a 1:1 ratio with the culture medium, for 2 min in the darkness at RT. 150µl of each well were then transferred to a corresponding 96-well chemiluminescence detection plate and the RLU were read in the Infinite F200 luminescence reader (TECAN). Inhibition (%) of luciferase activity from each serum dilution was calculated as follows: 100 - [(mean RLU of each sample - mean RLU of CC)/(mean RLU of VC - mean RLU of CC) x 100]. Inhibition (%) were then plotted against each dilution using four-parameter logistic (4PL) curve, and 50% inhibitory dilution (ID₅₀) values for each sample were calculated using GraphPad Prism software, version 8.0 (GraphPad Software, San Diego, CA, USA). As recommended by the World Health Organization (WHO) (29), the neutralization assay was calibrated and validated with the Working Standard Reagent for anti-SARS-CoV-2 immunoglobulin (National Institute for Biological Standards and Controls –NIBSC-, code: 21/234), that was also employed as positive control at each run of the experiment. As negative control, a serum sample from an uninfected and unvaccinated person was used.

PBMCs were cultured at 10⁷/mL in RPMI-1640 Medium (Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% FBS (Corning, New York, NY, USA) and stimulated or not with PepTivator^® SARS-CoV-2 Prot_S Complete (Miltenyi Biotec, Bergisch Gladbach, Germany, EU) at a final concentration of 0.6 nmol of each peptide/mL. PepTivator SARS-CoV-2 Prot_S Complete is a pool of lyophilized peptides, covering the complete protein coding sequence (aa 5–1273) of the surface or S glycoprotein of SARS-CoV-2 (GenBank MN908947.3, Protein QHD43416.1). Cells were incubated at 37°C for two hours and then 5μg/mL of Brefeldin A (Sigma-Aldrich, St. Louis, MO, USA) was added to cells to allow intracellular cytokine staining. PBMCs were incubated for further 16 hours and then prepared for staining. To detect anti-S specific CD4 and CD8 T cells, stimulated cells were stained for the surface antigen CD4 and CD8 (BioLegend, San Diego, CA, USA); fixed with Cyto-Fast Fix/Perm Buffer Set (BioLegend) and intracellular stained with anti-IFN-γ mAb (BioLegend). Stained PBMCs were acquired on CELESTA FACS (BD Biosciences, San Jose, CA, USA) and analyzed with FlowJo software (Ashland, OR, USA) Version 10. 50.000 CD4+ events were acquired and analyzed. The frequency of Spike-specific IFNγ-CD4+ and CD8+ T cells was obtained by subtracting cytokine background obtained from unstimulated cells.

This study obtained ethics approval from the ethics committee of AOU San Luigi Gonzaga, Orbassano (TO), Italy; Ref. number #117-2021). All the subjects included in the study consented to participate in the study.

Data sets used during this study are available from the corresponding author on reasonable request.

Anti-RBD IgGs titers and ID₅₀ values of the inhibition curves were calculated by a regression analysis using GraphPad Prism software, version 8.0 (GraphPad Software, San Diego, CA, USA) by fitting a quadratic curve and a variable slope-sigmoidal dose-response curve and statistically compared with the F-test, respectively. Ab levels were transformed on a log10 scale, to normalize their distribution and according to previous literature (12, 14).

Statistical analysis was performed using ANOVA Analysis of variance followed by Bonferroni post-test or t-test as reported in the legends to the Figures. Multivariable analysis was performed using a linear regression model computed by R version 3.6.3 (2020-02-29). Model was used to compare log-transformed P1 Anti-RBD IgG titer across patients treated with different DMTs, after adjusting for age, sex, EDSS score, MS type: relapsing remitting MS (RRMS), secondary progressive MS (SPMS), primary progressive MS (PPMS), MS disease duration, booster type (Comirnaty/Spikevax/COVID-19), Ab levels in the P0 samples. P1 Ab titers were included in the model to compare log-transformed P6 Anti-RBD IgG titers across subjects.

11 untreated pwMS, 59 pwMS under different DMTs and 24 HD were recruited and prospectively followed-up from their first shot of anti-COVID-19 mRNA vaccine (Pfizer) to 6 months after the third dose (Pfizer/Moderna). Demographic and clinical characteristics are reported in Table 1 .

The titer of anti-RBD IgGs induced by the full cycle of anti-SARS-CoV-2 vaccination (three doses) was evaluated in serum samples collected immediately before vaccination (P0), one (P1) and six months (P6) after booster. Moreover, the anti-N Ab titration was performed to evaluate a response to the natural infection occurred after vaccination.

Treated (T) pwMS (2.4 ± 1.2; p = 0.001) showed a significant lower level of anti-RBD IgGs compared to HD (3.6 ± 0.2) at P1 while not treated (NT) pwMS (3.3 ± 0.3) showed comparable levels with HD ( Figure 1A ). At P6, no significant difference was observed comparing anti-RBD IgG levels in HD (3.5 ± 0.2) with NT pwMS (3.4 ± 0.5) and T pwMS (2.7 ± 1) ( Figure 1A ). Subsequently, pwMS were divided according to anti-N positivity and their anti-RBD IgGs level were compared ( Figure 1B ). No statistical difference was observed, suggesting that anti-RBD IgGs is not related to a possible natural infection after vaccination. PwMS under interferon were excluded from this analysis because none of these subjects experienced natural COVID-19 infection after vaccination.

To investigate the effect of therapies on anti-RBD IgGs, pwMS were then divided according to DMTs ( Figure 1C ). All T pwMS showed comparable levels of anti-RBD IgGs with exception of T pwMS under ocrelizumab (1.3 ± 1; p<0.0001) and fingolimod (1.6 ± 1.3; p=0.0009) that showed significant lower levels of Ab respect to NT pwMS (3.3 ± 0.3). This difference is maintained at P6 for pwMS under ocrelizumab (1.1 ± 0.7, p=<0.0001) compared to NT pwMS at P6 (3.4 ± 0.45). Interestingly, a significant difference in anti-RBD IgG titers was not observed between P1 and P6 within each group suggesting a long-lasting durability of anti-RBD IgGs.

Finally, the association of factors included in Table 1 to anti-RBD IgG levels at P1 and P6 was explored by a multivariable regression analysis. Results of this analysis are reported in Table 2 . Ab titers at P6 were significantly associated with P1 Ab level (p = 0.0099) and ocrelizumab therapy (p = 0.0012). We confirmed that anti-RBD IgG titers at P1 were associated with treatment with ocrelizumab (p < 0.00005) and fingolimod (p = 0.0005) which both showed significantly reduced anti-RBD Ab levels compared to NT pwMS. Moreover, anti-RBD IgG titers at P1 were significantly increased in subjects that had Spikevax booster with respect to Comirnaty (p = 0.0294). No association with any other considered factor was found.

SARS-CoV-2 pseudovirions were produced according to a previously reported protocol published on Nature Protocols by Nie et al. (21), which is briefly described in the Material and methods section. A viral stock with a titer corresponding to 1.5x10⁵ TCID₅₀/ml was produced and used throughout the study. As reported by Figures 2A, B , a moderate to extensive cytopathic effect was observed in flasks transfected with pcDNA3.1.S2 plasmid and infected with the VSV-G pseudotyped virus after 48h or 72h, respectively. In order to verify the incorporation of SARS-CoV-2 spike glycoprotein on the VSV particles, pseudovirion production was characterized by means of western blot analysis. As reported in Figure 2C , the S protein was efficiently expressed on pseudovirion envelop: specific bands were detected in the lane of SARS-CoV-2 pseudovirions, whilst no specific band was found in the VSV-G pseudovirions (generated with the same procedure as SARS-CoV-2 pseudovirions) in the corresponding position. The monomer of the S protein (S1 + S2) was observed at a position of about 190 kDa and the S2 domain was detected at 90kDa. The SARS-CoV-2 pseudovirions, together with G*ΔG-VSV, were tested against a VSV-M specific Ab, showing a common band at 26kDa. Altogether, our results indicated that we generated a well-defined SARS-CoV-2 pseudovirion production suitable for the neutralization assays.

The serum samples collected at P1 and P6 were subsequently tested by means of the neutralization assay, in order to directly evaluate the Ab function and efficacy against SARS-CoV-2 pseudovirions. Focusing on sera collected at P1, we compared the neutralizing activity of samples from pwMS with that of HD ( Figure 3A , grey dots). As reported, no statistical difference in the neutralizing activity was observed between HD (2.7 ± 0.4) and NT pwMS (2.5 ± 0.6) or pwMS under different therapies, with exception of pwMS under ocrelizumab. As expected, considering the mechanism of action of ocrelizumab and the previously reported low levels of anti-RBD Ab, a significant reduction of the neutralizing activity was observed in pwMS under anti-CD20 therapy (0.8 ± 0.4; p=0.0001) compared to NT pwMS (2.5 ± 0.6). In particular, despite a small production of anti-RBD Ab, the neutralizing activity of sera from ocrelizumab-treated MS patients was always under the limit of detection (i.e NT₅₀ <1.3), indicating an absence of neutralization capacity.

Similar results were observed focusing on sera collected at P6 ( Figure 3A , blue dots). No statistical difference in the neutralizing activity at P6 was observed between HD (3.0 ± 0.4) and untreated pwMS (3.2 ± 0.9) or pwMS under different therapies, with the exception of pwMS treated with ocrelizumab (0.5 ± 0.5; p< 0.001) showing a neutralizing activity always under the limit of detection. Additionally, the potential reduction of the neutralizing ability after several months from the three doses was evaluated. As reported in Figure 3A , we didn’t observe significant differences comparing the neutralizing titers at P1 and P6 within each group, suggesting that, where present, the neutralization activity against SARS-CoV-2 is maintained over time. A difference in neutralizing efficacy is visible at P6 comparing T pwMS in which a natural SARS-CoV-2 infection occurred after vaccination (3 ± 0.6, p = 0.04) with uninfected T pwMS (2.4 ± 0.5) ( Figure 3B ), suggesting that natural infection may increase neutralizing response in these subjects. However, the same difference is not visible in HD and NT pwMS. Similarly to what was previously done for anti-RBD quantification, we excluded pwMS under interferon from this analysis because none of these subjects got COVID-19 after vaccination.

Overall, we observed a robust correlation between the previously reported anti-RBD Ab levels and the Ab neutralizing efficacy in HD (Pearson correlation; R=0.78, p= 6.9e-06) and in pwMS (Pearson correlation; R=0.85, p< 2.2e-16) ( Figure 4 ).

To determine the levels of S-specific T-cell activity, the number of CD4+ and CD8+ cells releasing IFNγ was assessed by cytofluorimetry after exposure of PBMCs to a 15-mer peptide pool covering the S protein of Wuhan wild-type SARS-CoV-2 ( Supplementary Figure 1 ). One and six months after vaccination, all groups of pwMS showed a similar frequency of S-specific IFNγ producing- CD4+ and CD8+ T cells comparable to that of HD ( Figures 5A, C ). Notably, T pwMS in which occurred a natural COVID-19 infection after vaccination display a higher frequency of both CD4 (0.24% ± 0.15; p=0.016) and CD8+(0.19% ± 0.18; p= 0.04) S-specific T cells response compared to T pwMS who remains protected from the infection (0.09% ± 0.03 and 0.05% ± 0.02 respectively, Figures 5B, D ). These results suggest that COVID-19 disease may increase the S-specific T cells repertoire, more than the vaccination alone.