Review papers with raw data transparency

Top-level appraisal (what the paper does well)

• Comprehensive synthesis of epidemiology (serology, prospective cohorts), immunology (B/T cell phenotypes), neuropathology, genetics (HLA interactions), and therapeutic efforts—integrates recent high-impact data including longitudinal cohort evidence that established temporal precedence of EBV before MS diagnosis Wahbeh & Sabatino review.
• Clear framing of two mechanistic models: (A) "driver" (ongoing EBV reactivation in B cells drives pathology) and (B) "hit-and-run" (primary EBV infection creates lasting immune dysregulation/molecular mimicry) with pros/cons for each (Figure 2 in the paper).
• Balanced and cautious about limits: explicitly highlights inconsistent CNS detection of EBV, measurement heterogeneity, and need for orthogonal sensitive assays and comparative controls.

Key claims and supporting evidence (selected, with critical notes)

1. EBV almost universal in MS. The review reports >99% EBV seropositivity in MS cohorts and argues near‑universality after assay sensitivity adjustments — consistent with multiple seroprevalence and longitudinal studies Wahbeh & Sabatino on EBV seroprevalence and the longitudinal US Army study showing EBV seroconversion precedes MS diagnosis by years (Bjornevik et al., Science 2022) Bjornevik 2022.
2. Anti-EBV humoral/T-cell signatures are altered in MS. The review compiles consistent observations: elevated anti‑EBNA1 IgG titers in MS (serum and some CSF signals), HLA-DRB1*15:01 interaction with anti‑EBNA1 levels, and mixed CD8+ T cell data (some expansion, functional exhaustion in certain compartments). These immune associations are robust but do not by themselves prove causation Immune signatures summary.
3. EBV detection in CNS is inconsistent across studies. The authors correctly highlight the replication problem: some histologic/molecular studies find EBV transcripts/proteins in perivascular cuffs and meninges, others do not — differences likely reflect tissue sampling, assay sensitivity (EBER ISH, PCR, RNA-seq, in situ protein detection), and pre-analytical variables. Therefore EBV presence in CNS tissue remains contested and requires standardized, orthogonal detection pipelines CNS EBV detection controversy.
4. Driver vs hit‑and‑run remain plausible, not mutually exclusive. The paper’s balanced comparison is its strength: driver model is supported by EBV‑infected B cells in CNS reports and the observation that B‑cell depletion (anti‑CD20) is highly effective in MS (consistent with removing EBV reservoir hypothesis), yet anti‑EBV targeted therapies have not yet given conclusive efficacy Driver vs hit-and-run. The review properly flags alternative interpretations (EBV reactivation as epiphenomenon of systemic immune dysregulation).
5. Translational implications are appropriately cautious. The review catalogs EBV‑directed strategies (antivirals, adoptive EBV T cells, vaccines) and their early/limited results (e.g., small antiviral trials with non-significant trends; ATA188 EMBOLD negative phase II), noting the urgent need for trials measuring both clinical and virologic endpoints and safety considerations around antigen selection for vaccines Therapeutic strategies summary.

Critical weaknesses, blindspots, and recommendations

• Heterogeneous methods across cited studies: the review notes and correctly emphasizes how variance in tissue processing, EBER ISH vs PCR vs RNA‑seq, and serologic platforms make cross-study synthesis noisy — the review could go further by proposing a standardized minimum assay set (quantitative EBV DNA by ddPCR, targeted viral RNA capture + duplex ISH, proteomic validation) and pre-analytic quality metrics.
• Risk of confirmation & publication bias: the authors acknowledge this; but more quantitative meta-analytic syntheses (where possible) or formal bias assessments (Egger funnel, p-curve) across serologic and CNS‑detection studies would strengthen claims that association climbs above publication bias concerns.
• Causality tests remain absent: the review details candidate falsification paths (e.g., MS in EBV-seronegative people; EBV vaccination altering MS incidence) but cannot substitute for prospective interventional data; recommending specific trial designs (see below) would be actionable.
• Therapeutic translation pitfalls: the review notes antigen selection risk for vaccine-induced cross-reactivity; we emphasize that any EBV vaccine trial must pre-specify autoimmune safety endpoints, monitor cross-reactive epitope antibodies/T cells (e.g., EBNA1-GlialCAM peptides), and include HLA-stratified sub-analyses given HLA‑DRB1*15:01 interactions.

Concrete next experiments the review recommends (and which would best falsify/strengthen claims)

1. Multi-center standardized CNS tissue study: prospectively collect fresh frozen and fixed MS and control brain/meninges with prespecified QC; apply (a) ultrasensitive EBV RNA capture sequencing (targeted hybrid-capture), (b) EBER and EBV mRNA duplex RNAscope ISH, (c) single-cell RNA + BCR/TCR with EBV read mapping, and (d) orthogonal proteomics for EBV proteins. Compare lesion stage, meningeal follicles, and control inflammatory neurologic disease. This directly addresses reproducibility concerns raised in the review.
2. Randomized, HLA-stratified EBV vaccine trial for IM prevention/waning seroconversion in adolescents with long-term neurological follow-up and immunologic substudies measuring cross-reactive antibodies/T cell repertoires (EBNA1/GlialCAM/ANO2 mapping) to detect vaccine-induced undesired autoimmunity early.
3. Mechanistic human ex vivo model: infected PBMC-to-CNS organoid co-culture (with autologous B/T cells) to test whether EBV+ B cells present CNS antigens and activate autoreactive T cells — pairing single-cell multiomic readouts and functional killing/antigen‑presentation assays. This complements review suggestions to pair EBV expression with B cell APC function.

Summary conclusion, confidence, and what would change the interpretation

Conclusion: The review is a careful, current, and useful synthesis that lays out why EBV is probably necessary (near-universal seropositivity + prospective seroconversion evidence) but not yet proven sufficient (mechanistic causal chain) for MS; it fairly catalogs data for both driver and hit‑and‑run models and sensibly calls for standardized, orthogonal mechanistic studies and appropriately designed interventional trials Overall paper conclusion.

What would falsify this review's central implication (EBV→MS causal role): (1) reproducible demonstration of sizeable MS incidence among verified EBV‑seronegative individuals; (2) large RCTs of EBV prevention (vaccine) failing to reduce MS incidence after adequate follow‑up; (3) standardized CNS tissue pipelines repeatedly failing to detect EBV expression in MS lesions while detecting it in positive controls — each would substantially weaken the EBV causal claim.