Immunology of MS

1. MS as an Autoimmune Disease

The autoimmune model of MS rests on five empirical pillars:

1. The CSF and lesion infiltrate are dominated by T cells, B cells, and macrophages.
2. Intrathecal IgG synthesis with oligoclonal bands is present in ~95% of patients.
3. The strongest genetic association is with the MHC class II allele HLA-DRB1*15:01.
4. Animal models (EAE) reproduce demyelinating disease by autoreactive T cells alone.
5. Immunotherapies that target lymphocytes (anti-CD20, anti-α4-integrin, anti-CD52, S1P modulators) consistently and powerfully reduce relapse activity.

Unlike classical antibody-mediated autoimmunity (myasthenia gravis, NMOSD), MS is a T-cell-driven, B-cell-supported disease in which the dominant autoantigen has never been identified with certainty — arguing for a bystander or molecular-mimicry mechanism rather than a single dominant epitope.

2. Experimental Autoimmune Encephalomyelitis

EAE, induced in 1933 by Thomas Rivers and refined by Kabat, Wolf and Bering (1947), is the foundational animal model of CNS autoimmunity:

• Induction — immunisation with myelin protein (MBP, PLP1, or MOG) emulsified in complete Freund’s adjuvant, with pertussis toxin to open the BBB.
• Adoptive transfer — CD4⁺ T cells reactive to myelin can transfer disease to naive recipients (Paterson 1960) — cardinal evidence of T-cell mediation.
• Strain dependence — SJL mice get a relapsing-remitting course on PLP139-151; C57BL/6 get a chronic course on MOG35-55. The MOG model produces both T-cell and antibody pathology.
• Therapeutic discovery — glatiramer acetate (originally a synthetic mimic of MBP), natalizumab’s α4-integrin target (Yednock et al., Nature 1992), fingolimod’s S1P, and IFN-β were all validated in EAE.

EAE’s limitations are well known: no model perfectly captures human MS, and several therapies that worked in EAE (anti-TNF, altered peptide ligands) failed or worsened disease in humans. EAE is best read as a model of CNS autoimmunity rather than of MS specifically.

3. The T-Cell Story — Th1, Th17, Tissue-Resident

For forty years MS was modelled as a Th1 disease: CD4⁺ T cells secreting IFN-γ and TNF, driven by IL-12. This picture was revised after 2005 with the discovery of the Th17 lineage (Park et al.; Langrish et al.), which is driven by IL-23 and produces IL-17, IL-22, GM-CSF. Both axes are active in MS:

• Th1 — IFN-γ; reactivates microglia, induces MHC class II, drives macrophage activation.
• Th17 — IL-17, IL-22, GM-CSF; opens the BBB, recruits neutrophils, supports B-cell follicle formation in meninges.
• CD8⁺ T cells — outnumber CD4⁺ in chronic plaques. Clonally expanded; suspected effectors of axonal injury via direct cytotoxicity and IFN-γ.
• Tissue-resident memory T cells (T_RM) — CD69⁺ CD103⁺ populations persist in CNS parenchyma, beyond the reach of peripherally acting drugs.

4. Treg Dysfunction

Functional defects in CD4⁺CD25^highFoxP3⁺ regulatory T cells are a consistent finding in MS (Viglietta et al., J Exp Med 2004): Tregs from MS patients suppress autologous effector T-cell proliferation less efficiently than Tregs from controls. The defect appears thymic in origin (reduced recent thymic emigrant Tregs) and is partially reversed by IFN-β and by daclizumab (anti-CD25, withdrawn after hepatic and encephalitic adverse events). The low-dose IL-2 approach to expanding Tregs is in clinical trials.

5. The B-Cell Resurgence

For decades B cells were considered minor in MS, given the lack of a pathogenic autoantibody. Three lines of evidence rehabilitated them:

1. Persistent intrathecal IgG synthesis (oligoclonal bands) implies a CNS-resident, clonally restricted B-cell population.
2. Meningeal B-cell follicle-like aggregates are present in many SPMS brains and correlate with cortical demyelination (Magliozzi et al., Brain 2007).
3. B cells are highly efficient antigen-presenting cells, secrete cytokines (LT-α, GM-CSF, IL-6) that drive Th17 responses, and prime autoreactive T cells.

The decisive experiment was therapeutic: B-cell depletion with anti-CD20 monoclonals reduces relapse activity by ~50–75% within months — faster than antibody titres can fall, implying that the relevant B-cell function is antigen presentation / cytokine secretion, not antibody production.

6. Anti-CD20 — the Decisive Experiment

The HERMES trial (Hauser et al., NEJM 2008) showed that two infusions of rituximab depleted CD20⁺ B cells from blood and reduced gadolinium-enhancing lesions by 91% over 24 weeks. This single result reorganised MS immunology around B cells. Subsequent trials:

• OPERA-I and OPERA-II (Hauser et al., NEJM 2017) — ocrelizumab (humanised anti-CD20) reduced ARR by ~46% versus IFN-β-1a in RRMS.
• ORATORIO (Montalban et al., NEJM 2017) — ocrelizumab reduced 12-week confirmed disability progression by 24% in PPMS — the first DMT effective in primary progressive disease.
• ASCLEPIOS-I and II (Hauser et al., NEJM 2020) — ofatumumab, fully human anti-CD20, subcutaneous monthly dosing, ~50% reduction in ARR vs teriflunomide.
• ULTIMATE-I and II (Steinman et al., NEJM 2022) — ublituximab, glycoengineered anti-CD20.

CD20 is expressed on pre-B through memory B cells but absent on plasma cells, sparing humoral immunity (in principle). In practice, prolonged CD20 depletion produces hypogammaglobulinaemia in a subset of patients, with attendant infection risk — covered in Part VII.

7. Epstein-Barr Virus as the Causal Trigger

EBV has been suspected as an MS trigger for decades on the basis of repeated epidemiologic associations: MS patients are virtually universally EBV-seropositive, infectious mononucleosis (delayed primary EBV) raises MS risk ~2–4×, and elevated anti-EBNA-1 antibody titres precede MS onset.

The 2022 study by Bjornevik et al. (Science) used a unique resource — serial serum samples from over 10 million US active-duty military recruits — to test causality prospectively:

• 955 individuals developed MS during follow-up.
• Of those, only 1 was EBV-seronegative at the most recent pre-onset sample — effectively all MS cases were preceded by EBV infection.
• The hazard ratio for MS after EBV seroconversion was 32.4 (95% CI 4.3–245); seroconversion to other viruses (CMV) carried no increased risk.
• Serum neurofilament light (a neuroaxonal damage marker) rose only after EBV seroconversion, not before — arguing against reverse causation.

The simplest reading: EBV is a near-necessary but insufficient cause of MS. Mechanistic candidates include direct CNS infection by latently-infected B cells (controversial in autopsy studies), molecular mimicry between EBV proteins and CNS autoantigens, and bystander activation of autoreactive lymphocytes during chronic EBV reactivation.

8. Molecular Mimicry — EBNA-1 and GlialCAM

The molecular-mimicry hypothesis proposes that T- or B-cell receptors raised against a microbial antigen cross-react with a structurally similar self-antigen. The 2022 paper by Lanz et al. (Nature) identified a candidate mimic for MS:

• An MS-derived monoclonal antibody (cloned from intrathecal IgG) was shown to bind a conserved domain of EBV nuclear antigen 1 (EBNA-1, residues 386-405).
• The same antibody bound the CNS adhesion molecule GlialCAM (HEPACAM), expressed on astrocytes and oligodendrocytes.
• Vaccination of mice with EBNA-1 peptides exacerbated EAE in a GlialCAM-dependent manner.
• Anti-GlialCAM reactivity was substantially more common in MS than control CSF.

GlialCAM is unlikely to be the only mimicked antigen — mimicry has been proposed previously between EBNA-1 and αB-crystallin, MBP, MOG — but it is the first that ties EBV directly to a quantitative, mechanistically-coherent CNS autoantigen.

9. MOG, MBP & the Autoantigen Question

The classical candidate autoantigens in MS are surface- or near-surface myelin proteins:

• MBP (myelin basic protein) — cytoplasmic; the immunogen used by Rivers in 1933 EAE; immunodominant peptides MBP85-99 and MBP83-99 bind HLA-DR2 and stimulate CSF-derived T-cell clones.
• MOG (myelin oligodendrocyte glycoprotein) — surface-exposed Ig-superfamily protein; the immunogen of choice for B-cell-driven EAE. MOG-IgG defines the distinct disease entity MOGAD (myelin oligodendrocyte glycoprotein antibody-associated disease), not classical MS.
• PLP1, MAG, CNP, αB-crystallin, neurofascin-155 — subsidiary candidates; T-cell reactivity demonstrable but not dominant.
• GlialCAM, KIR4.1, RASGRP2 — modern candidates raised in the EBV/mimicry era.

Critically, no single autoantigen is universally recognised by MS patient T-cell or B-cell repertoires — a picture better described as epitope spreading over time from an initial trigger to multiple myelin-resident self-proteins.

10. Compartmentalised CNS Inflammation — Why DMTs Stop Working

Modern DMTs largely act on peripheral immune compartments — depleting circulating B cells, blocking lymphocyte egress from lymph nodes (S1P), preventing transmigration across the BBB (anti-α4-integrin). They powerfully suppress relapses but are far less effective in halting progressive, neurodegenerative MS. The prevailing model:

• Once peripheral autoreactive cells have seeded the CNS, a self-sustaining, «trapped» population of T_RM, plasma cells, microglia and meningeal B-cell follicles drives chronic, smouldering damage.
• This compartment is largely inaccessible to peripherally acting DMTs.
• Therapeutic implication: small molecules that cross the blood-brain barrier — especially BTK inhibitors (evobrutinib, tolebrutinib, fenebrutinib) targeting B cells and microglia within the CNS — may be needed for progressive disease.

This concept reframes the relapsing/progressive distinction not as two diseases but as two phases of the same process, in which inflammation is initially blood-borne and later compartmentalised — with profound implications for therapy in Part VII.