EFB Dengue Literature Review

Thesis Objectives and Grant Pitch

11 min read

Thesis Objectives and Grant Pitch

Scope. This is the strategic layer of the dengue atypical-B-cell project — the central thesis, falsifiable objectives, statistical-honesty framing, confounder pre-emption, and the pilot→grant trajectory. The wet-lab protocol (panel, compensation, FMOs, gating) lives in the sibling page Research Plan - DN B Cell Expansion in Dengue; the gating rationale in DN2 Gating Strategy. This page is the pitch; those are the bench.

Research Question

Can a budget-frozen, surface-only 11-color flow panel generate enough signal — across dengue disease, infection history, and severity — to show that the extrafollicular / DN2-phenotype B-cell compartment is the cellular substrate of dengue’s low-fidelity antibody output, and thereby justify a funded study of atypical B cells in dengue severity that merges neutralizing-antibody and autoantibody (ANA) readouts?

Sources Used

Synthesis

Central thesis

Dengue’s pathogenic antibody problem and its autoimmune signal may be two faces of one low-fidelity antibody property — cross-reactive, polyreactive, near-germline IgG. One face is non-neutralization / ADE (the property that makes severe secondary dengue dangerous); the other is autoreactivity via molecular mimicry (ANA, anti-platelet/-endothelial reactivity). The proposed cellular source of both is the atypical / extrafollicular compartment — DN2-phenotype B cells and the Plasmablast wave they co-vary with — generated through the extrafollicular pathway that produces class-switched but poorly mutated antibody (see GodoyLozano2016 - Lower IgG SHM Rates in Acute Dengue; Ansari2025 - Peripheral T Helper Subset Drives B Cell Response in Dengue).

Honesty flag: the cells→autoantibody link is imported from SLE and unproven in dengue. This pilot does not claim to prove it; it tests whether the cellular substrate exists, tracks infection history and severity, and correlates with autoreactivity — the minimal evidence needed to make the mechanistic study fundable. The cross-wiki synthesis of this arm lives in ../bridge-wiki/.

The pitch in one paragraph

Atypical (DN2-phenotype) B cells expand in acute dengue and, like their counterparts in SLE and severe COVID-19, are predicted to secrete the cross-reactive, near-germline antibodies that drive both non-neutralizing/ADE-relevant responses and autoreactivity. Using an existing, low-cost surface flow panel plus banked serum, this pilot will (1) quantify the EF/DN2-phenotype compartment and the DN2:DN1 ratio in acute dengue, (2) show it correlates with antinuclear antibody (ANA) reactivity and with cross-reactive (non-type-specific) neutralization, and (3) generate the effect-size estimates needed to power a definitive study. A positive pilot converts a plausible SLE-borrowed hypothesis into a dengue-specific, fundable program that sorts these cells, sequences their B-cell receptors, and tests whether they make the antibodies that cause severe disease.

Current sample inventory

DFDHFTotal
Male — n (median age)3 (27)7 (24)10 (25.5)
Female — n (median age)5 (25)4 (35)9 (29)
Total — n (median age)8 (26)11 (29)19 (27)

Cross-sectional; day 5–8 post-fever-onset; fresh blood (flow) + serum (single timepoint). Target ≥10–15 per arm; recruitment ongoing.

⚠ Age and sex are first-order biology here, not nuisance covariates. Atypical/age-associated B cells are age-defined, and the T-bet⁺ ABC program is female-biased via X-linked TLR7 (see Lamprinou2026 - ABCs and DN B Cells, Age-Associated B Cell). The current arms are imbalanced on both — DHF skews male (7M/4F) and older; DF skews female (5F/3M) and younger (and the 4 DHF females are markedly older, median 35). A naive DF-vs-DHF “atypical B cell” contrast therefore partly measures age and sex. At n≈10–15/arm, multivariable adjustment is weak (~1 covariate per 10–15 subjects), so the fix is balanced recruitment now, not statistics later. Recruit the remaining cases to even sex and age across arms; match the healthy/comparator arm the same way.

Fixed constraints

What we can measure

Cellular (flow): total B cells; plasmablasts (CD27⁺⁺CD38⁺⁺); the IgD×CD27 quadrants (naive, unswitched memory, switched memory, DN); transitional (CD24⁺⁺CD38⁺⁺); activated-naive-phenotype (IgD⁺CD27⁻CD21⁻CD11c⁺CD24⁻); DN subgated CD21×CD11c → DN2-phenotype / DN1-like / DN3-like.

Derived metrics (free — ratios multiply endpoints):

Antibody layer (banked serum) — both faces of the thesis now measurable in-pilot:

  • ANA (HEp-2 IIF: titer and pattern) — the autoreactive face.
  • FRNT against all four serotypes — DENV2 = presumed homotypic, others heterotypic → breadth / cross-reactivity index + OAS signature (heterotypic > homotypic neutralization in secondary cases; see Priyamvada2016 - Cross-Reactive Memory Plasmablasts in Secondary Dengue) → the non-neutralizing/ADE face.
  • IgG/IgM → primary vs secondary classification.

Objectives

  1. O1 — primary, novel (the differentiating arm). Across the acute cohort, the EF metric (DN2:DN1, EF composite) correlates with ANA (titer + HEp-2 pattern). This is the cells→autoantibody bridge and the genuinely unbuilt link in the dengue literature (neutralization/ADE/OAS are already well-covered; cellular origin of autoreactivity is not).
  2. O2 — coherence. ANA and cross-reactive (non-type-specific) neutralization co-occur in the same patients — one low-fidelity output showing both faces.
  3. O3 — biological axis. EF and antibody metrics differ by primary vs secondary serostatus (secondary = memory recall / OAS / cross-reactive non-neut antibody — upstream of severity and central to the thesis).
  4. O4 — severity (exploratory / effect-size-generating). EF and antibody metrics across WHO-2009 binary severity, within the d5–8 window, age/sex/day-adjusted. Job: generate the variance and effect sizes that power the grant’s sample-size calculation.
  5. O5 — feasibility. Reproducible DN2-phenotype gating with validated FMOs (the CD11c-PE / PE-Cy7→PE spillover risk; see DN2 Gating Strategy).

Analysis hierarchy & statistical honesty

  • Lead with continuous, full-cohort correlations (O1/O2), not severity group-contrasts. At n≈10–15/arm a DF-vs-DHF contrast is powered only for large effects; a correlation across all ~25–30 patients uses full n and detects moderate-strong effects.
  • Pre-specify 1–2 primary tests (nominate O1) and declare everything else exploratory — multiplicity discipline. A reviewer who sees 30 correlations and one asterisk discounts the lot; one pre-registered prediction confirmed is fundable.
  • State the pilot’s job plainly: not to prove the biology but to deliver feasibility, directional signal, and effect-size/variance estimates for the grant’s power calc. This reframes “small n” from a weakness into the point.

Confounders we pre-empt

ConfounderStatus / mitigation
Age & sexFirst-order for ABCs → balance by recruitment now; record + adjust where possible; report sex-stratified; match comparator arm. (Headline issue.)
Day-of-samplingLargely controlled by design — narrow d5–8 window straddling the PB peak; this is the design strength that distinguishes the study from the day-confounded severity claim in Ansari2025 - Peripheral T Helper Subset Drives B Cell Response in Dengue. Still record exact day as covariate.
SerotypeNo reliable PCR at d5–8 → epidemiological assignment to DENV2; attempt RT-PCR/NS1 on the d5 subset to anchor it. Constrains the OAS interpretation only — not the cellular spine or the ANA bridge.
CD27 sheddingHigh-TNF dengue milieu can shed CD27 → inflate the DN gate; unmeasurable on this panel; acknowledge (sCD27 in serum could quantify it later).
ANA transienceAcute-infection ANA can be transient/polyclonal; single timepoint can’t show resolution → specificity rests on the within-cohort correlation (ANA tracks the EF metric and serostatus), not prevalence. Convalescent-resolution = a grant aim.
”DN2-phenotype” ≠ DN2No CXCR5/T-bet/FCRL5; Ansari2025-comparable; confirmation = grant aim.

Severity scheme decision (flagged — curator’s call)

WHO 1997 (DF / DHF I–IV / DSS)WHO 2009 (D / DWS / SD)
Anchored onPlasma leakage + thrombocytopeniaSevere leak or bleed or organ impairment
Research strengthMechanistically homogeneous severe arm — leak is the immune-driven endpoint → tightest signalCaptures all severe phenotypes; sensitive
WeaknessRigid; misses non-leak severe disease; needs serial HCTSD lumps mechanistically distinct phenotypes (dilutes a leak-specific signal); DWS is soft/subjective
StandingOlder, being retiredCurrent clinical standard; reviewer/clinician expectation

Recommendation: pre-register WHO-2009, collapsed to binary (non-severe vs severe) as primary (translatable, reviewer-expected, better powered than 4-way grading at this n); run the WHO-1997 leak-defined contrast as a declared sensitivity analysis (free, since both classifications exist). Classify on the full clinical course, not status-at-sampling (some d5 “non-severe” patients declare leak by d7). Do not fish across schemes for the best p-value.

See Dengue Severity Classification for the two WHO schemes side by side and the evidence behind this trade-off — Narvaez2011 (n=544) shows WHO-2009 is far more sensitive for intervention-level disease (92.1% vs 39.0%) but dissolves the plasma-leakage entity, and the two schemes agree only at κ=0.25.

Serology plan

  • IgG/IgM: LFA on all samples (cheap first-pass serostatus) → capture-ELISA IgM/IgG ratio on a calibration subset (~25–30 spanning the spectrum + all LFA-borderline/discordant) → compute LFA↔ELISA concordance (κ) against a pre-set threshold. High κ → trust LFA for the bulk; low κ → ELISA broadly. (The statistically sound inversion of “LFA to check if ELISA is worth it” — only ELISA can validate LFA.)
  • FRNT×4: homotypic (DENV2) vs heterotypic titers → cross-reactivity index + OAS check. Laborious (4 serotypes × n × replicates) — if n grows, tier it (FRNT the EF-metric extremes + a serostatus/severity-balanced subset).

Trajectory to the grant

Pilot delivers: the EF/DN2-phenotype substrate is present, tracks serostatus/severity, and co-occurs with ANA and cross-reactive neutralization — plus effect sizes and a validated, low-cost assay set. Grant tests the mechanism: sort DN2-phenotype / EF cells → BCR-sequence → express monoclonals → test directly for ANA reactivity and for neutralization/ADE — i.e., do these cells make the low-fidelity antibodies that drive severe disease? Add what the pilot lacked: convalescent + longitudinal sampling (kinetics, ANA resolution), confirmatory markers (CXCR5, T-bet, FCRL5), and intracellular capacity.

Open Questions

  • Is there a healthy / convalescent comparator arm, and is it age/sex-matched? (Needed for the DN-expansion-vs-baseline claim; the within-cohort correlations do not strictly require it.)
  • Is a convalescent draw feasible on any subset? (Unlocks ANA resolution + kinetics — a major pilot strengthener.)
  • Final n per arm and the primary/secondary split once recruitment closes — determines whether O3 and O4 are simultaneously analyzable.
  • Serotype PCR/NS1 yield on the d5 subset — how much of the DENV2 assumption can be empirically anchored?
  • Should the Research Plan - DN B Cell Expansion in Dengue be advanced to Rev 5 to fold in the antibody assays (ANA, FRNT×4, IgG/IgM), the d5–8 window, realistic n, and age/sex balancing?

Research Plan - DN B Cell Expansion in Dengue, DN2 Gating Strategy, B Cell Panel Variant 1, DN2 B Cell, Atypical B Cell, Age-Associated B Cell, Plasmablast, Activated Naive B Cell, Extrafollicular Response, Peripheral Helper T Cell, Antibody-Dependent Enhancement, Original Antigenic Sin, FRNT, CD11c, CD21, CXCR5, T-bet, TLR7, Ansari2025 - Peripheral T Helper Subset Drives B Cell Response in Dengue, Woodruff2020 - EF B Cell Responses in COVID-19, Jenks2018 - DN2 B Cells and EF Pathway in SLE, Priyamvada2016 - Cross-Reactive Memory Plasmablasts in Secondary Dengue, GodoyLozano2016 - Lower IgG SHM Rates in Acute Dengue

Graph View

Backlinks