EFB Dengue Literature Review

Follicular Exclusion

7 min read

Follicular Exclusion

Overview

DN2 / atypical B cells do not enter B cell follicles, and this is not incidental — it is the anatomical expression of the extrafollicular commitment. This page covers where these cells are excluded from, where they go instead, and what holds them there.

The story has two halves, and until the 2026-08-18 ingest of Song2022 - Tfh Outside Germinal Centers Drive T-bet CD11c B Cells the wiki carried only the first:

  1. Negative / transcriptional — the cells lose the receptors that would draw them into the follicle (CXCR5, CD21) and repress the TF programme needed for GC entry.
  2. Positive / positional — the cells acquire receptors that draw them elsewhere (CXCR3, S1PR3, CNR2) and adhesion molecules that physically retain them at the splenic marginal zone (LFA-1, VLA-4).

Key Points from Literature

The negative half — losing follicular access

★ The positive half — where they go and what holds them (Song2022)

All findings in this section are murine, from acute LCMV and influenza infection.

  • By histocytometry at day 12 post-infection, ~75% of T-bet⁺CD11c⁺ B cells were in the follicular mantle, ~0% in germinal centres, ~20% in the T zone (see Song2022 - Tfh Outside Germinal Centers Drive T-bet CD11c B Cells)
  • As infection resolved the cells moved to the splenic marginal zone, identified by MadCam-1⁺ sinus-lining cells. Intravenous anti-CD45 labelling — which marks compartments open to circulation — labelled 40.3% of T-bet⁺CD11c⁺ cells at day 12 and 60.5% at day 15, versus 16.8%/23.3% of naive follicular and 2.24%/3.47% of GC B cells
  • Migration receptors acquired during resolution: S1pr3 and Cnr2 (cannabinoid receptor 2) transcripts rose between days 8 and 15, and CXCR3 was higher than on naive follicular or GC cells (its ligands are produced at the marginal zone). See S1PR3 and CXCR3.
  • S1P chemotaxis is S1PR3-dependent in vitro: FTY720 limited migration and the selective S1PR3 inhibitor TY52156 abolished it completely
  • But S1PR3 is dispensable in vivo: mixed bone marrow chimeras showed no competitive advantage of S1pr3⁺/⁺ over S1pr3⁻/⁻ cells at day 15, which the authors attribute to compensation by CNR2 and CXCR3
  • ★ Retention is integrin-dependent. T-bet⁺CD11c⁺ cells showed elevated transcript and surface protein for LFA-1 (αLβ2) and VLA-4 (α4β1), whose ligands ICAM-1 and VCAM-1 are expressed at the marginal zone. In vivo blockade of both alpha subunits for just 3 hours caused significant loss of these cells from spleen with a concomitant increase in blood. See LFA-1 and VLA-4.

Why the positional half matters for interpretation

  • Because retention is active and reversible on a 3-hour timescale, circulating DN2 frequency may partly reflect retention failure or mobilisation rather than pool size. This converges with the caution from Cancro2020 - Age-Associated B Cells that blood and splenic ABC pools are not in equilibrium, and that human blood ABCs fall on HIV antiretroviral therapy.

  • The authors propose the marginal-zone position is functional: it places an antigen-experienced population where blood-borne antigen arrives, enabling rapid response to systemic reinfection (see Song2022 - Tfh Outside Germinal Centers Drive T-bet CD11c B Cells). See Atypical B Cell Effector Output.

  • The same relocation was observed in an Ehrlichia muris model and after influenza infection, per the source’s discussion — so it may be a general property of the T-bet⁺CD11c⁺ compartment rather than an LCMV artefact.

  • [2026-08-27] CXCR5 alone carries the follicular-competence axis in the DN1–DN4 quadrant scheme. With no CD21 and no T-bet in the panel, IgD⁻CD27⁻ cells are partitioned by CXCR5 × CD11c into DN1 (CXCR5⁺CD11c⁻), DN2 (CXCR5⁻CD11c⁺), DN3 (CXCR5⁻CD11c⁻) and DN4 (CXCR5⁺CD11c⁺). DN1 and DN4 are follicle-competent; DN2 and DN3 are not (see Allard-Chamard2023 - DN3 B Cells Infiltrate Inflamed Tissues, n=38, 13-colour FCM).

  • ★ [2026-08-27] CXCR5-negativity does not predict where the cells are actually found. In severe COVID-19 lung, the two CXCR5⁺ subsets accounted for a substantial share of the tissue DN pool (DN1 ~20%, DN4 ~26%), and in IgG4-RD salivary gland the CXCR5⁺ DN1 dominated at ~70% of tissue DN cells. Conversely the CXCR5⁻ DN2 was nearly absent from both tissues. Follicular exclusion by phenotype is not the same as tissue localisation — a CXCR5⁻ cell is excluded from following the CXCL13 gradient into a follicle, which says nothing about whether it enters an inflamed end organ. Any inference from a blood CXCR5⁻ frequency to a tissue-homing claim needs this caveat (see Allard-Chamard2023 - DN3 B Cells Infiltrate Inflamed Tissues, n=6–10 per tissue arm; no significance markers on the subset-level panels).

Contradictions & Debates

In vitro necessity versus in vivo redundancy for S1PR3. TY52156 completely abolished S1P-directed migration in transwell, yet S1pr3⁻/⁻ cells populated the marginal zone as well as wild-type in competitive chimeras. The authors name this as a limitation and propose CNR2/CXCR3 compensation, which they did not test. The wiki should not present S1PR3 as the marginal-zone homing receptor.

Does any of this transfer to humans? All positional data are murine and splenic. Human work in this wiki is almost entirely peripheral blood; Kaneko2020 - GC Loss and TFH Block in COVID-19 provides human tissue but examines GC loss, not marginal-zone retention. Whether human DN2 cells are retained at a marginal-zone equivalent is untested.

CXCR5, CXCR3, CD21, ZEB2, LFA-1, VLA-4, S1PR3, DN2 B Cell, Atypical B Cell, Extrafollicular Response, Germinal Center, Atypical B Cell Effector Output, Extrafollicular T Cell Help, DN3 B Cell

Sources

Graph View

Backlinks