Multiple sclerosis (MS) is a demyelinating inflammatory disease of the CNS treated by diverse disease-modifying therapies that suppress the immune system. Severe acute respiratory syndrome coronavirus 2 mRNA vaccines have been very effective in immunocompetent individuals, but whether MS patients treated with modifying therapies are afforded the same protection is not known. This study determined that dimethyl fumarate caused a momentary reduction in anti-Spike (S)-specific Abs and CD8 T cell response. MS patients treated with B cell–depleting (anti-CD20) or sphingosine 1-phosphate receptor agonist (fingolimod) therapies lack significant S-specific Ab response. Whereas S-specific CD4 and CD8 T cell responses were largely compromised by fingolimod treatment, T cell responses were robustly generated in anti-CD20–treated MS patients, but with a reduced proportion of CD4+CXCR5+ circulating follicular Th cells. These data provide novel information regarding vaccine immune response in patients with autoimmunity useful to help improve vaccine effectiveness in these populations.
Multiple sclerosis (MS) is an autoimmune and demyelinating disease of the CNS characterized by the destruction of myelin sheaths (1). Disease-modifying therapies (DMTs) limit the occurrence of lesions and relapses in MS patients, but they also suppress key immunological processes that may be essential for optimal vaccine responses (2).
In the United States, the two most widely used vaccines against the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Spike (S) protein (BNT162b2 from Pfizer and mRNA-1273 from Moderna) have been instrumental in decreasing severity, magnitude, and spread of infection in healthy recipients (3, 4). However, it is not currently fully understood whether MS patients treated with DMTs can mount efficient immune responses after SARS-Co-V2 vaccination. Recent data show low Ab response to SARS-CoV-2 S protein in patients treated with anti-CD20 and fingolimod a month after the second vaccine dose (5–9). However, there are limited data comparing humoral and cellular S-specific response in MS patients treated with different DMTs.
In our study, we assessed SARS-CoV-2 S–specific humoral and cellular immune responses in a cohort of healthy control subjects or MS patients treated with B cell–depleting agents (anti-CD20), sphingosine 1-phosphate receptor agonist (fingolimod) or dimethyl fumarate (DMF), 2 wk and 3 mo after the completion of their mRNA-based vaccine regimen. Our results demonstrate moderate to extensive alterations in immune response to SARS-CoV-2 vaccines depending on the type of DMT and further suggest that future adaptation of vaccines, boosters, and/or DMT administration might be necessary to improve protection in immunocompromised patients.
Materials and Methods
Blood samples were obtained from the Benaroya Research Institute’s Immune Mediated Disease Registry and Repository, under an Institutional Review Board–approved protocol (IRB08108). All study subjects were >18 y of age, and written informed consent was obtained from all enrolled participants prior to participation. Characteristics of study subjects and their use are summarized in Supplemental Table I. Study exclusion criteria included inability or unwillingness of subject or legal guardian/representative to give written informed consent; infection with SARS-CoV-2; and for control subjects, the presence of autoimmune disease was also exclusionary.
Circulating IgG Abs against SARS-CoV-2 S domain S1 Ag were measured in the serum using the Euroimmun immunoassay (Lubeck, Germany). All testing and analyses were performed according to the manufacturer’s protocols, with the OD ratio calculated using the kit calibrator. To standardize results and facilitate comparisons, we converted OD ratio scores for each sample to z scores as described previously (10). A conservative z score ≥ 8 was considered positive to minimize false-positive results.
SARS-CoV-2 S protein peptide array was obtained from BEI Resources. Peptides that demonstrated potent CD8 and CD4 T cell responses (11) were resuspended in DMSO and pooled into class I– and class II–specific mesopools, respectively, and as previously described (12). CEFX Ultra SuperStim Pool peptide was purchased from Innovative Peptide Solutions.
Ag-induced markers assay
Cryopreserved PBMCs from healthy and MS patients were thawed at 37°C and resuspended in prewarmed xeno-free and serum-free media. PBMCs were split into three wells and stimulated for 20 h at 37°C 5% CO2 with either a mesopool (5 μg/ml) of CD4- and CD8-targeted SARS-CoV-2 S peptides to detect vaccine-induced T cell responses, or with CEFX Ultra SuperStim Pool (1μg/ml) to use as a positive control or with equimolar DMSO as a negative control (12).
Flow cytometry immunophenotyping
After Ag-induced marker (AIM) assay, PBMCs were stained for viability (Zombie L/D NIR; BioLegend) and barcoded using CD45 Abs. Barcoded conditions were pooled together and resuspended in staining buffer containing surface staining markers, such that unstimulated (U) and stimulated conditions (S and C) from healthy and MS donors could be stained and acquired together (Supplemental Fig. 1F). Surface markers consisted of gating: CD3 (UCHT1), HLA-DR (G46-6), CD8a (RPA-T8), pan-gdTCR (B1), CD19 (HIB19), CD4 (SK3); and phenotypic: CD26 (M-A261), CXCR3 (1C6), CCR7 (G043H7), CCR6 (G034E3, CD27 (M-T271), CD137 (4B4-1), CD57 (QA17A04), CXCR5 (RF8B2), CD134 (BerACT35), PDL1 (29E.2A3), CCR4 (L291H4), CD25 (BC96), CD127 (hIL7Rm21), CD103 (BerACT8), CD69 (FN50). All samples were acquired on the Aurora spectral flow cytometer (Cytek) and analyzed with FlowJo X (BD).
Analysis and statistics
Cells were enumerated and plotted using FlowJo X and GraphPad Prism9, respectively. GraphPad Prism9 was used to perform statistical analysis with statistical methods described in the figure legends. The p values of statistical significance are depicted by asterisks per star guide scale (nsp > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001).
Changes in B cells and humoral immunity
To determine how different DMTs modulate the immune responses generated by mRNA-based coronavirus disease 2019 (COVID-19) vaccines, we collected serum and PBMCs from MS patients treated with B cell–depleting agent (anti-CD20; Ocrevus), lymphocyte egress modulator (fingolimod; Gilenya), and fumaric acid ester (DMF or Tecfidera), and from healthy control individuals without autoimmunity who received the two-dose regimen of Pfizer or Moderna vaccines. We first enumerated B cell frequencies in the blood of MS patients receiving these DMTs and in healthy control subjects 2 wk after the second dose of vaccination. Anti-CD20–treated patients displayed more than a 50-fold decrease in B cell frequencies and numbers in PBMCs compared with healthy control subjects (Fig. 1A). The few detected B cells had a significantly reduced CD27 and CXCR5 expression (Fig. 1B). Because the CXCL13-CXCR5 axis plays a central role in organizing and attracting cells to the follicles and germinal centers, this could reflect a reduced germinal center formation in these patients. In contrast, patients treated with DMF showed no alteration in B cell frequency and numbers compared with healthy control subjects (Fig. 1A). Fingolimod-treated patients had nearly 10-fold fewer B cells in the PBMCs compared with healthy control subjects (Fig. 1A). However, these differences were not the result of vaccination but of the DMTs themselves, because similar B cell numbers and phenotype were observed prior to vaccine administration (Supplemental Fig. 1A, 1B), including a significant reduction in the proportion of CD27+ B cells, a distinguishing marker of mature, class-switched B cells (Fig. 1B, Supplemental Fig. 1B).
Ocrelizumab (Ocrevus) targets CD20-expressing B cells for Ab-mediated depletion, but a small fraction of B cells remains (Fig. 1A). A longitudinal analysis of serum Abs against the S protein of SARS-CoV-2 was performed to determine whether these B cells were sufficient to form Abs in MS patients treated with anti-CD20. The majority of healthy control subjects showed peak Ab production after the second dose of mRNA vaccine immunization and maintained Ab production 3 mo postimmunization (Fig. 1C, Supplemental Fig. 1C). Only one healthy control subject failed to develop anti-S–specific Abs. In contrast, most anti-CD20–treated patients displayed nearly no detectable S-specific Abs (z score < 8) after the second vaccine dose (Fig. 1C, Supplemental Fig. 1C). Only a few anti-CD20–treated MS patients developed a very low and transient anti–SARS-CoV-2 Ab response (z score of 8–10.5) after the second dose of vaccine, which was not detectable 3 mo postimmunization (Fig. 1C, Supplemental Fig. 1C). This was not correlated with the time since last injection of anti-CD20 (Supplemental Fig. 1D). Despite DMF-treated MS patients showing comparable B cell numbers at 2 wk after second vaccine dose, they had a slight but significant reduction in S-specific Ab responses (Fig. 1C) and associated fewer B cells with a class-switched phenotype (Fig. 1B). The decrease in B cell frequencies observed after the second vaccine dose in the few fingolimod-treated patients we could test correlated with a profound lack of serum S-specific Abs in comparison with healthy control subjects (Fig. 1C).
Effect of DMTs on the magnitude of S-specific T cell responses
During SARS-CoV-2 infection, cytotoxic CD8 T cells are necessary to clear infected cells and can respond well to new variant strains of the virus (11), whereas CD4 Th cells are necessary for optimal B cell and cytotoxic T cell responses and are highly correlated with reduced disease severity of convalescent SARS-CoV-2–infected patients (13). Thus, we next analyzed the representation of these two T cell populations in PBMCs of different patient groups. Fingolimod-treated MS patients had a significant reduction in number of total CD4 and CD8 T cells, although the effect on CD8 T cells was less profound (Supplemental Fig. 1E). Anti-CD20 treatment had no significant effect on overall T cell numbers (Supplemental Fig. 1E). In contrast, DMF-treated patients had a significant reduction in CD8 T cells compared with healthy control subjects (Supplemental Fig. 1E). Next, we investigated the S-specific CD8 and CD4 T cell responses generated by vaccination in DMT-treated MS patients and in healthy control subjects 2 weeks after the second vaccine dose. We used an activation-induced markers assay (12), which allows for the enumeration of S-specific T cells by comparing the upregulation and coexpression of activation markers CD137, CD134, CD25, PD-L1, CD69, and ICOS in PBMCs stimulated with S-derived peptides or vehicle control (Supplemental Fig. 1F, 1G). This assay is more sensitive than IFN-γ detection to identify S-specific CD8 T cell responses (Supplemental Fig. 2A). The two-dose regimen of the Pfizer and Moderna vaccines caused robust expansion of S-specific CD8 and CD4 T cells in healthy patients, indicated by upregulation of CD137, along with CD25, in CD8 T cells and CD134 in CD4 T cells after in vitro stimulation (Fig. 2A–C). Despitehaving a low Ab response to SARS-CoV-2 S protein, anti-CD20–treated patients demonstrated S-specific cytotoxic and Th cell responses comparable in frequencies to those observed in healthy control subjects (Fig. 2A–C). Furthermore, the AIM-specific T cell responses did not correlate with the timing of the anti-CD20 infusion (Supplemental Fig. 1H). A subset of healthy subjects and anti-CD20–treated MS patients tested at 2 wk postvaccination was observed longitudinally for maintenance of S-specific T cell responses at 3 mo postvaccination, because this is a hallmark feature of functional memory T cells. No significant decline was measured in either S-specific CD4 or CD8 T cells compared with healthy control subjects (Fig. 2D). Patients receiving DMF treatment mounted robust S-specific CD4 T cell responses but had a significant decrease in number of S-specific CD8 T cells (Fig. 2A–C). Finally, we could not detect significant CD4 and CD8 S-specific responses in fingolimod-treated patients (Fig. 2B).
Modulation of cytotoxic CD8 T cell diversity by DMTs
DMTs can significantly impact the immune environment, and thus not only the activation but also the differentiation and specialization of vaccine-specific memory CD8 and CD4 T cell responses (2). To characterize the phenotype of the vaccine-specific memory CD8 response, we used a multistep clustering method called DISCOV-R (distribution analysis across clusters of a parent population overlaid with a rare subpopulation) (14) (Supplemental Information). Total CD8 T cells of healthy control subjects and DMF- or anti-CD20–treated MS patients identified in the AIM assay at 2 wk after the second vaccine dose were divided into 12 phenotypic clusters (Fig. 3A, Supplemental Fig. 2B–E). Density overlay showed that healthy control subjects, DMF-treated MS patients, and anti-CD20–treated MS patients had a similar T cell distribution within the uniform manifold approximation and projection (UMAP) space (Supplemental Fig. 2B). No significant influence of age was observed among treatment groups, and cells from older patients colocalized to similar areas on the plot in all treatment groups (Supplemental Fig. 2C). Vaccine type was also compared, but we did not observe differential clustering of T cells within the UMAP space in subjects given the Moderna compared with Pfizer vaccine (Supplemental Fig. 2D). S-specific (ICOS+CD137+) CD8 T cells identified in the AIM assay were mainly distributed into clusters C1–4 (Fig. 3A, 3B, Supplemental Fig. 2E). S-specific T cells from healthy control subjects and anti-CD20–treated MS patients were phenotypically concentrated into similar clusters (especially C1) expressing the markers CD127, CD27, and CXCR3 suggestive of a protective memory-like Tc1 Ag-specific population. Interestingly, this cluster was moderately reduced in DMF-treated patients (Fig. 3A, 3B). Instead, cells from DMF patients were located in C3 (and C4), characterized by expression of CCR6 suggesting a switch of CD8 T cells from a Tc1 to a Tc17 phenotype (Fig. 3A, 3B).
Alteration of S-specific CD4 T cell response by DMTs
Next, S-specific CD4+ T cells (identified based on their coexpression of CD137 and CD134) were characterized using the same DISCOV-R platform. Total CD4 T cells from healthy control subjects and MS patients were used to build a phenotypic landscape of CD4+ T cells (Fig. 4A, Supplemental Fig. 2F–H). Nearest-neighbor clustering was used to divide the landscape into 15 distinct phenotypic clusters, with most clusters present in all parent CD4+ T cell populations (Fig. 4A, Supplemental Fig. 2F–H). CD137+CD134+ S-specific CD4+ T cells were overlaid on these clusters, and their distribution was enumerated to determine their phenotype (Fig. 4A, Supplemental Fig. 2H, 2I). The majority of Ag-specific CD4 T cells resided in clusters C1–5 of the PhenoGraph (Fig. 4B, Supplemental Fig. 2H, 2I). C1 and C4 were characterized by the expression of CCR4, which is important for the trafficking of T cells to the lung (15). C1 contained 60% of all S-specific CD4 T cells in healthy and anti-CD20–treated MS patients and correlated with the most highly activated T cells coexpressing ICOS and CD25 (Fig. 4A, 4B). In DMF-treated patients, S-specific CD4 T cells were significantly less represented in this cluster and more numerous in C4 with low expression of CD25 (Fig. 4A, 4B). In contrast, anti-CD20–treated patients had a significant decrease in Ag-specific CD4 T cells present in C5, marked by high CD25 expression, and in C3, corresponding to the highest expression of CXCR5 and CD27, indicative of circulating T follicular helper cells (cTFH) (Fig. 4B). Furthermore, the longitudinal analysis of CD27- and CXCR5-expressing S-specific CD4 T cells using manual flow cytometry gating showed a significant decline in the proportion of CXCR5+ S-specific CD4 T cells from anti-CD20 patients (Fig. 4C), which was maintained up to 3 mo postimmunization (Fig. 4C).
In summary, our study demonstrates moderate to extensive alterations in MS patients’ immune responses to the SARS-CoV-2 vaccine depending on the type of administered DMTs. B cell–depleting therapies eliminate most B cells from the circulation, causing moderate to significant effects on vaccine-elicited Abs, depending on vaccine type and timing of vaccine administration (16). In the case of SARS-CoV-2 vaccine, we observed an absence of S-specific Abs in MS patients treated with anti-CD20 (Fig. 1C). The impaired Ab response in anti-CD20–treated patients corroborates recently published studies (5–9), but our data extend these observations to 3 mo postvaccination. Conversely, despite having no detectible S-specific Abs, anti-CD20–treated patients mounted similar frequencies of S-specific T cell responses comparable with healthy control subjects. Unlike previous studies, we observed only a mild increase in Ag-specific CD8 T cell responses. This difference could reflect a heterogeneity in patient cohorts between the two studies. Interestingly, ocrelizumab-treated MS patients had a significant decrease in CD4+ T cells expressing high levels of CXCR5, a marker of cTFH, and in those expressing high levels of CD25. Our observations are largely in agreement with a recent study describing that anti-CD20 treatment compromises cTFH responses in SARS-CoV-2–vaccinated subjects (8). Our data extend these findings by demonstrating this phenotype from an unbiased clustering analysis and show that this population remains low 3 mo postvaccination. It will be of interest to determine in future studies whether the limited pool of undepleted B cells could give rise to S-specific Ab-producing B cells, with or without additional TFH after a SARS-CoV-2 vaccine booster shot.
Previous studies have shown that vaccine-mediated immune responses are maintained in patients treated with DMF (17). In this study, the most notable change observed in DMF-treated MS patients was the significant reduction in S-specific CD8 T cell numbers and the change in CD4 and CD8 S-specific T cell phenotypes. S-specific CD8 T cells switched from a Tc1 to a Tc17 phenotype in DMF-treated individuals (Fig. 3A, 3B). The emergence of S-specific CXCR3+CD8+ has been correlated with good immunity in SARS-CoV-2–infected individuals, while an enrichment in CCR6+CD8+ T cells has been preferentially detected in COVID-19–infected individuals who were ventilated, suggesting that the latter subset did not provide an effective cytotoxic antiviral response (18). In addition, a highly activated CD4+ T cell population coexpressing the chemokine receptor CCR4 and other markers (ICOS, CD25, and HLA-DR) was significantly decreased, while a phenotypically similar population but lacking CD25 expression was significantly increased in DMF-treated MS patients compared with healthy control subjects. Because the expression of CCR4 is important for the trafficking of T cells to the lung (15), it is possible that DMF limits the activation of S-specific CD4+ T cells that have the capacity to migrate to the lungs through expression of CCR4. Further studies are required to determine whether the phenotypic changes in these T cell populations can significantly impact protection against SARS-CoV-2 infection and severe disease development.
Among the DMTs analyzed in this study, sphingosine 1-phosphate receptor agonist had the most profound effect on the generation of protective SARS-CoV-2–specific Ab and cellular immune responses. We could not detect significant circulating CD4 and CD8 S-specific responses in fingolimod-treated patients, an effect that is consistent with fingolimod’s mode of action, which blocks lymphocyte egress from the lymph nodes. In addition, we and others (5, 9) did not detect significant anti-S–specific Abs in fingolimod-treated MS patients. Based on experiments performed in mice and reduced humoral response observed following other vaccines (2), it is possible that fingolimod inhibits adequate interactions between T and B cells in the lymph nodes or interferes with the migration of plasma cells in the bone marrow, where they can contribute to long-term Ab production (19) and therefore inhibit the production of high-affinity, class-switched T-dependent Abs (20). One limitation of our study is that our cohort size of fingolimod-treated patients was very limited. Future studies with more fingolimod-treated MS patients will help determine whether the patients analyzed in this study are representative of all patients in this group or whether they represent a subgroup of patients whose response to vaccine is particularly poor.
The presence of cellular immunity in DMF- and anti-CD20–treated and SARS-CoV-2–vaccinated MS patients provides the encouraging possibility that memory T cell populations could provide sufficient protection to avoid severe illness on infection. However, additional studies are required to establish whether and how the change in the phenotype of some CD8 and CD4 T cells subsets can impact protection. Our data on the few fingolimod-treated MS patients show that at least a fraction of this group experiences a significant lack of humoral and cellular responses to SARS-CoV-2 vaccination and warrants additional analysis to determine whether their immune response could be enhanced with a vaccine booster dose, or if DMT administration course could be altered to provide a sufficient window for the generation of protective cellular immunity. Collectively, our study helps elucidate how DMTs modulate vaccine immune response in vulnerable MS patients and offer information to improve vaccine effectiveness in patients with autoimmunity.
We thank Rachel Hartley and Thien-Son Nguyen for assisting with sample and clinical data tracking; Andrew Pickles, Heather White, Kassidy Benoscek, and Kimberly Varner for work in recruitment and enrollment, and in conducting study visits; Carla Greenbaum for providing advice and consulting on the study; and Alice Wiedeman and Kirsten Diggins for helping with the DISCOV-R platform and Colin O’Rourke for help with statistical analysis.
This work was supported by the Benaroya Family Foundation, the Leonard and Norma Klorfine Foundation, Glenn and Mary Lynn Mounger, and Monolithic Power Systems.
E.B. designed and supervised the study. Y.Y. performed the Ag-induced markers assays and the analysis of the data. E.B. and Y.Y. wrote the manuscript. D.J.C. assisted in the study and assay design. P.M. and M.F. participated in the Ag-induced markers assay optimization and preparation of the peptide pools. C.S. and S.L., together with M.K. and L.M., designed the clinical research study. C.S. and S.L. oversaw clinical study design and human sample collection. J.H.B. assisted in study design. A.C., C.M., and M.H.W. were involved in the measurement of Abs. All authors gave feedback and corrections on the manuscript.
The online version of this article contains supplemental material.
Abbreviations used in this article:
coronavirus disease 2019
circulating T follicular helper cell
distribution analysis across clusters of a parent population overlaid with a rare subpopulation
severe acute respiratory syndrome coronavirus 2
T follicular helper cell
uniform manifold approximation and projection
The authors have no financial conflicts of interest.