EFB Dengue Literature Review

council final report-Ansari2025

12 min read

Council Final Report: Ansari2025

Paper: Ansari A et al. (2025). A peripheral T helper subset drives the B cell response in dengue. Cell Reports 44:115366.

Review date: 2026-05-14 (updated 2026-05-14 — PDF-based review replacing source-page-based review)

Council composition:

  • Council Head: Claude Opus 4.6 (synthesis and verdict)
  • Methodology Critic: Sonnet agent (experimental design, statistics, controls)
  • Claims Validator: Sonnet agent (evidence-to-claim mapping)
  • Contextual Critic: Sonnet agent (literature fit, contradictions, gaps)
  • Strengths Advocate: Sonnet agent (genuine contributions, novelty)

Input: Full PDF (raw/Ansari2025.pdf) — main text (pp. 1–20), STAR Methods (pp. 21–29), Supplementary Figures S1–S7 (pp. 31–45), Supplementary Tables S1–S4 (pp. 46–53). All council members read the PDF directly.

Note: This report supersedes the earlier council review conducted from the wiki source page only (pdftoppm was unavailable). Key corrections from direct PDF access are marked with [PDF CORRECTION] below.

THE VERDICT

A genuinely important paper that earns its landmark status for two specific contributions — identifying Tph as the dominant CD4⁺ helper in dengue and demonstrating IL-21-dependent B cell help via clean coculture experiments. However, it overclaims in several areas (DN2 identity, “first direct evidence,” concurrent EF+GC), and direct PDF review reveals both a factual correction (Tfh coculture comparison exists) and new concerns (day-of-sampling confounder, canonical Tph identity mismatch, coculture used recall T cells not acute).

CLAIM-BY-CLAIM ASSESSMENT

#ClaimCouncil VerdictStrength
1Tph (CXCR5⁻PD-1⁺) are ~75% of activated CD4⁺ T cells in acute dengueSupported within the ICOS⁺Ki67⁺ activated gate (~75%) and as ~53.4% of CD45RA⁻ CD4⁺; total CD4⁺ contribution unknown. [PDF CORRECTION] Cells have Th1 signature (CXCR3⁺, T-bet, IFN-γ), not canonical Tph (MAF⁺, CXCL13⁺) — “Tph” label is an interpretive choiceMODERATE
2Tph accumulate in severe dengueSupported (p=0.006 Kruskal-Wallis; p=0.01 DF-WS vs severe) but confounded — [PDF CORRECTION] severe patients sampled later (median 8±4 vs 5±2 days post-onset), so expansion may reflect disease duration not severity. No multivariable adjustmentMODERATE
3Helper vs. cytotoxic Tph subclusters (scRNA-seq)Hypothesis-generating only — [PDF CORRECTION] 4,361 cells from 3 patients (not 4 as previously reported). Co-expression of GZMB and IL21 in some cells (Figure S7D) suggests spectrum, not binary. 13 shared TCR clonotypes insufficient for lineage conclusionsWEAK
4Tph drive memory B cell → plasmablast via IL-21Best-evidenced claim — clean coculture + blocking. [PDF CORRECTION] Tfh coculture comparison arm EXISTS for plasmablast generation (Figure 6H: CXCR5⁻PD-1⁺ > CXCR5⁺PD-1⁺, p=0.0006; also Supplementary S6F-G). However, the IL-21 blocking experiment (Figure 6J) was performed ONLY with CXCR5⁻PD-1⁺ T cells — no parallel Tfh blocking arm to show cytokine-dependence is Tph-specific. Caveat: coculture used memory T cells from seropositive donors, not acute patients (acute cells died in culture)STRONG
5CD21⁻CD11c⁺ B cells represent EF-phenotype (DN2-like) expanding in denguePhenotypic expansion supported (p<0.0001 DN vs naive, Friedman test, n=20). But panel lacks T-bet, CXCR5, FCRL5 — cannot distinguish DN2 from DN1 or activated memoryMODERATE
6Neutralizing Ab paradox (non-neutralizing IgG up, FRNT₅₀ not different by severity)Well-supported — but [PDF CORRECTION] FRNT used only DENV-2 (strain S-16803), and severe subgroup for FRNT is only n=10. Single-serotype limitation notedSTRONG
7Concurrent EF + GC pathways (CXCL13 elevation)CXCL13 elevated at 2–5 days (p<0.01 vs HD) but not GC-specific — can be produced by Tph cells themselves, macrophages, inflamed endothelium. No direct GC assessment (no CD10⁺ B cells, no GC histology)WEAK
8”First direct evidence of extrafollicular B cell activation in dengue”Defensible for the Tph-B cell axis specifically; but “extrafollicular” identity of B cells is inferred from 2-marker proxy without tissue-level confirmation. Authors acknowledge “lack of direct evidence of IL-21⁺ T cell interactions with extrafollicular B cells in DENV-infected tissues” in LimitationsMODERATE

TOP STRENGTHS

  1. The IL-21 blocking experiment remains the paper’s crown jewel. IL-21R-Fc fusion protein reduced plasmablast output ~60% in autologous cocultures; anti-IL-10 ~25%; anti-IL-4 no effect. This cytokine hierarchy is mechanistically informative and goes beyond correlation. [PDF CORRECTION] The Tfh coculture comparison arm for plasmablast generation (Figure 6H) shows CXCR5⁻PD-1⁺ generates significantly more plasmablasts than CXCR5⁺PD-1⁺ (p=0.0006 ANOVA), establishing functional superiority. Supplementary Figure S6F-G provides additional head-to-head data. This was incorrectly flagged as missing in the previous review. Nuance: The IL-21 blocking experiment itself (Figure 6J) was performed only with CXCR5⁻PD-1⁺ T cells — there is no parallel Tfh blocking arm to determine whether Tfh-driven plasmablast generation is also IL-21-dependent (or uses a different cytokine axis).

  2. Cohort size (n=170) is strong for dengue T cell immunology — three consecutive dengue seasons (2017–2019), systematic WHO severity classification across three strata (DF without WS n=24–37, DF with WS n=66–116, severe n=26–33; numbers vary by assay). Longitudinal follow-up in 32 convalescent patients with paired acute samples shows 4-fold Tph contraction from acute to convalescence.

  3. Lymph node tissue samples (n=10) are rare in dengue immunology — ENT surgery patients who were DENV-seropositive. CXCR5⁻PD-1⁺ cells at comparable frequencies in blood and lymph node; LN tissue cells respond to DENVpep stimulation. Confirms tissue residency, not blood-only artefact.

  4. Places dengue firmly within the Tph/EF framework with disease-specific data rather than inference from SLE/COVID-19 analogues. The field’s attention is genuinely reoriented from Tfh to Tph.

  5. AIM assay enrichment data (CXCR5⁻PD-1⁺ cells: 2.88% DENV-specific vs. pTfh: 1.35% after background subtraction) provides concrete functional enrichment beyond phenotypic proportions. Uses validated 180-peptide DENV megapool covering all four serotypes.

  6. Technical rigour in scRNA-seq — hashtag antibody demultiplexing, explicit regression of dissociation-induced stress gene signatures, CCA-based batch correction. Methodologically sound despite small n.

TOP CONCERNS

Major Methodological Issues

  1. Missing isotype control for blocking reagents. Anti-IL-10 and anti-IL-4 require matched isotype controls to exclude Fc-mediated effects. IL-21R-Fc (chimeric protein) has no obvious isotype equivalent, but the absence of any non-specific blocking control is a gap. Severity: MAJOR CONCERN. Affects: IL-21 pathway specificity (Figure 6J).

  2. scRNA-seq underpowered — [PDF CORRECTION] 3 patients, not 4. 4,361 cells from 3 acute secondary dengue patients at a single site. Ten clusters from this cell number risks over-splitting (cluster 7/Th17 = ~78 cells). Co-expression of GZMB and IL21 in some cells (Figure S7D) and shared expression of PDCD1, MKI67, CCR2, CCR5 (Figure 7F) suggest a spectrum rather than discrete subsets. Severity: MAJOR CONCERN. Affects: helper/cytotoxic dichotomy, subcluster generalizability.

  3. B cell panel lacks T-bet, CXCR5, FCRL5. Cannot formally distinguish DN1 (CXCR5⁺) from DN2 (CXCR5⁻) within IgD⁻CD27⁻. T-bet (the defining EF transcription factor) and FCRL5 unstained. The CD21⁻CD11c⁺ proxy could include activated memory or other non-EF populations. Severity: MAJOR CONCERN. Affects: extrafollicular B cell identity (Figure 7H-J).

  4. [PDF CORRECTION] Day-of-sampling confounder for severity. Severe dengue patients were sampled later (median 8±4 vs 5±2 days post-symptom-onset for DF without WS, per Table S1). Tph expansion could reflect disease duration rather than severity per se. No adjustment for sampling day in severity analyses. Severity: MAJOR CONCERN. Affects: severity association (Figure 2F-I).

  5. [PDF CORRECTION] “Tph” identity is debatable. These CXCR5⁻PD-1⁺ cells show Th1 transcriptional signatures (CXCR3⁺, T-bet, IFN-γ) per scRNA-seq, not the canonical Tph profile described by Rao et al. 2017 in RA (MAF⁺, IL-21⁺, CXCL13⁺). The authors acknowledge this in the Discussion but still adopt the Tph label in Highlights. May mislead readers expecting the autoimmune Tph phenotype. Severity: MAJOR CONCERN. Affects: framing and field interpretation.

  6. No multiple testing correction across the study. Within-panel corrections applied (Dunn’s, Friedman), but no global FDR across dozens of comparisons. The EF B cell correlation (Figure 7J: r=0.56, p=0.009, n=20) would not survive stringent correction across all correlations tested. Severity: MINOR LIMITATION.

  7. Serotype not documented. Dengue confirmed by NS1 antigen/IgM/IgG serology, but infecting serotype never reported. FRNT tested only DENV-2. Multiple serotypes co-circulate in Delhi. Severity: MAJOR CONCERN. Affects: all immune response characterisation.

  8. Healthy donor sex imbalance. HD cohort 99% male (blood bank donors) vs dengue cohort 63% male. 36-percentage-point gap is unaddressed and could confound all HD vs. dengue comparisons. Severity: MAJOR CONCERN. Affects: Figures 1–2, 5–7.

Contextual Concerns

  1. The naive vs. memory tension is unresolved and more significant than the authors acknowledge. In SLE (Jenks2018, Tipton2015) and COVID-19 (Woodruff2020), EF responses are primarily naive-derived (resting naive → aNAV → DN2 → PB, with germline/low-SHM BCR). Ansari2025 claims dengue EF is primarily memory-derived (Tph preferentially help memory B cells; Figure 6I). This is not a minor variation — it implies a fundamentally different precursor biology. The authors do not engage with:

    • GodoyLozano2016’s paradoxically lower SHM in secondary than primary (cuts against memory recall)
    • Priyamvada2016’s high SHM in sorted PBs (consistent with memory but from a different angle)
    • Singh2026’s DN/atypical MBC accumulation data (the upstream precursor population)
    • [PDF CORRECTION] Whether the coculture bias toward memory reflects the use of seropositive donor memory T cells (acute cells died in culture), not genuine in vivo precursor preference

  2. The SHM paradox remains completely unresolved. Low bulk SHM (GodoyLozano2016) vs. high sorted PB SHM (Priyamvada2016) is the central molecular question for EF vs. GC origin. This paper adds cellular phenotyping but no BCR sequencing — the paradox persists unchanged. The paper does not acknowledge this gap.

  3. Missing citations. Tipton et al. 2015 (aNAV as EF ASC precursors in SLE), Wei et al. 2007 (original DN B cell characterisation), Scharer et al. 2019 (epigenetic DN2 programming), and Sanz lab comprehensive reviews are absent. Engagement with the foundational Sanz lab EF framework is superficial.

WHAT THE PAPER PROVES vs. WHAT IT IMPLIES

Actually demonstrated:

  • Tph cells (CXCR5⁻PD-1⁺) are numerically dominant among activated CD4⁺ T cells in acute dengue (~53.4% of CD45RA⁻ CD4⁺; ~75% of ICOS⁺Ki67⁺)
  • [PDF CORRECTION] Tph provide SUPERIOR B cell help compared to pTfh in autologous coculture (p=0.0006, Figure 6H) — functional dominance, not just numerical
  • Tph help is IL-21-dependent (~60% PB reduction with blocking), preferentially driving memory B cell → plasmablast differentiation
  • CD21⁻CD11c⁺ B cells expand within the IgD⁻CD27⁻ gate during acute dengue and correlate with Tph frequency (r=0.56, p=0.009, n=20)
  • Non-neutralizing anti-NS1/prM IgG correlates with severity; FRNT₅₀ does not (but FRNT is DENV-2 only, severe n=10)
  • Tph frequencies contract from acute to convalescence (n=32 paired)
  • DENV-specific T cells (AIM assay) are more enriched in the CXCR5⁻PD-1⁺ subset than in pTfh

Implied but not demonstrated:

  • That Tph cause severe dengue (vs. being co-elevated; day-of-sampling confounder unresolved)
  • That the CD21⁻CD11c⁺ cells are bona fide DN2 (T-bet⁺, CXCR5⁻, FCRL5⁺)
  • That these cells are truly “Tph” in the Rao2017 sense (Th1 signature, not canonical Tph)
  • That EF and GC pathways genuinely operate concurrently (CXCL13 is not GC-specific; may originate from Tph themselves)
  • That cytotoxic GZMB⁺ Tph are a stable, reproducible subpopulation (n=3, co-expression data suggest spectrum)
  • That the Tph→memory B cell axis produces the ADE-competent antibodies that drive immunopathology
  • That IL-21 blocking would reduce severity-associated (non-neutralizing) IgG while preserving neutralizing titers in vivo
  • That the memory B cell preference in coculture reflects in vivo precursor biology (vs. an artefact of using seropositive donor memory T cells)

REMAINING GAPS

  • No BCR sequencing — the EF vs. GC molecular origin of the plasmablast wave is still phenotypically inferred, not molecularly confirmed
  • No tissue-level visualisation of extrafollicular foci in dengue tissue (authors acknowledge this)
  • No adequately powered primary vs. secondary Tph comparison — key question for whether memory recall vs. naive activation drives the EF response in primary infection
  • No link between Tph axis and ADE-competent antibody output — the central translational question
  • Cytotoxic Tph function completely unknown — GZMB expression without functional killing data
  • No serotype stratification — single-site (New Delhi) with undocumented serotype distribution across three seasons
  • No temporal adjustment for severity — day-of-sampling confounder requires multivariable modelling
  • Canonical Tph identity unresolved — are these Tph (Rao2017 definition) or Th1-like activated effectors with B cell help capacity?

CORRECTIONS FROM PDF-BASED REVIEW

The following errors or gaps in the previous source-page-based council review were identified through direct PDF access:

ItemPrevious ReportPDF-Based Correction
Tfh coculture comparisonListed as missing (Concern #1)Partially exists — head-to-head plasmablast generation comparison EXISTS (Figure 6H, p=0.0006 + S6F-G), but IL-21 blocking (Figure 6J) was done ONLY with Tph — no parallel Tfh blocking arm
scRNA-seq patient count”4 patients”3 patients
Day-of-sampling confounderNot identifiedSevere patients sampled later (median 8±4 vs 5±2 days)
Tph identityAccepted as TphTh1 signature, not canonical Tph; authors acknowledge
Coculture T cell sourceNot specifiedMemory T cells from seropositive donors (acute cells died)
FRNT serotypeNot specifiedDENV-2 only (strain S-16803)
Activated gate definition”CD38⁺HLA-DR⁺“ICOS⁺Ki67⁺ for the ~75% figure; ~53.4% of CD45RA⁻ CD4⁺
Missing citationsNot assessedTipton2015, Wei2007, Scharer2019, Sanz lab reviews absent

COUNCIL RECOMMENDATION

Treat Ansari2025 as landmark for identifying the Tph→IL-21→memory B cell axis in dengue. The coculture data — now confirmed to include head-to-head Tfh comparison — is solid mechanistic evidence showing functional superiority of the CXCR5⁻PD-1⁺ subset.

Do NOT treat as having:

  • Confirmed DN2 identity of the CD21⁻CD11c⁺ cells (requires T-bet/CXCR5/FCRL5)
  • Resolved the SHM paradox (GodoyLozano2016 vs. Priyamvada2016)
  • Established a causal severity model (day-of-sampling confounder unresolved)
  • Confirmed concurrent EF+GC activity (CXCL13 is non-specific)
  • Validated scRNA-seq subclusters as generalizable biology (n=3)
  • Proven these cells are canonical Tph (Th1 signature, not Rao2017 Tph)

For wiki purposes:

  • The source page accurately captures the paper’s findings but should note the Th1 vs. canonical Tph tension
  • The wiki’s framing of these cells as “DN2-like” (rather than confirmed DN2) is appropriate
  • The state.md Watch Items correctly flag the missing T-bet/CXCR5 confirmation
  • The “concurrent EF+GC” claim in wiki pages should be softened to “suggested by CXCL13 elevation”
  • New Watch Item recommended: The memory-vs-naive precursor question — whether the coculture memory preference reflects genuine in vivo biology or the use of seropositive donor T cells
  • Previous concern partially resolved: The Tfh coculture comparison arm for plasmablast generation exists (Figure 6H) — wiki content should reflect this. However, the IL-21 blocking experiment (Figure 6J) lacks a parallel Tfh arm, so whether Tfh help also depends on IL-21 (vs. a different cytokine) remains untested

Council report finalised 2026-05-14 by Council Head (Claude Opus 4.6) This is a PDF-based review superseding the earlier source-page-based review. Raw deliberation minutes available in: Claude-council/raw thinking minutes-Ansari2025.md

Graph View