EFB Dengue Literature Review

Memory B Cell

27 min read

Memory B Cell

Overview

Memory B cells are long-lived antigen-experienced B cells that mediate accelerated, higher-affinity secondary immune responses upon antigen re-encounter. Their hallmarks are either genetic (somatic hypermutation of rearranged immunoglobulin genes) or functional (fast recall responses of higher affinity than primary responses). Surface phenotypic markers — most prominently CD27 — have been used to identify them by flow cytometry, but the discovery of CD27-negative memory B cells (DN B cells) has complicated this picture.

Human memory B cells are heterogeneous. Classical subsets include isotype-switched CD27⁺ cells (IgG or IgA surface expression, GC-derived), unswitched CD27⁺ cells (IgM⁺IgD⁺ or IgM-only, partly marginal zone-related), and the more recently defined IgD⁻CD27⁻ double-negative memory B cells with proposed extrafollicular origin.

Key Points from Literature

  • CD27 was historically considered a universal memory B cell marker; CD27⁺ cells consistently carry significant somatic hypermutation (see Wei2007 - DN Memory B Cells in SLE, citing Klein et al. 1998).

  • Two diversification pathways: Isotype-switched CD27⁺ memory cells are thought to originate primarily from GC reactions; IgM/IgD CD27⁺ cells may develop through GC-independent, CD40-independent pathways (cited as Weller et al. 2001 in Wei2007 - DN Memory B Cells in SLE).

  • DN (IgD⁻CD27⁻) memory B cells represent a third subset, present at ~5% of PBL B cells in healthy subjects and expanded in SLE. They carry somatic hypermutation, fail to extrude R123, and respond to CpG without BCR crosslinking — all functional memory hallmarks (see Wei2007 - DN Memory B Cells in SLE, n=29 healthy + n=36 SLE cross-sectional).

  • IgM⁺ memory B cells within the CD27⁺ unswitched subset predominantly co-express IgD; by contrast, IgM⁺ cells within the DN subset lack IgD — a phenotypic distinction used to separate the two (see Wei2007 - DN Memory B Cells in SLE).

  • Proliferating DN cells upregulate CD27 after CpG stimulation, suggesting that CD27 expression is not lineage-fixed but can be acquired upon activation — complicating its use as a stable memory marker (see Wei2007 - DN Memory B Cells in SLE, in vitro assay).

  • The Bm1–Bm5 classification (IgD/CD38 axes) provides an alternative to CD27-based gating; Bm5 (IgD⁻CD38⁻/dull) cells substantially overlap with classical memory cells but also include a CD27⁻ fraction corresponding to DN cells (see Wei2007 - DN Memory B Cells in SLE).

  • Autoreactive memory B cells are expanded in SLE and normalise after treatment: VH4.34-encoded autoreactive B cells within the switched memory compartment (CD27⁺IgD⁻) were 16.2 ± 11.9% of memory cells in SLE vs. 1.3 ± 0.3% in healthy controls (P=0.03, n=6 SLE). After effective rituximab depletion and immune reconstitution, levels returned to 1.92 ± 0.7% (n=9 depletors), consistent with restored GC tolerance checkpoints (see Anolik2004 - Rituximab and B Cell Abnormalities in SLE).

  • Residual memory B cells are the dominant surviving population after B cell depletion: After 95–99% B cell depletion with rituximab, residual CD19⁺ B cells are predominantly switched memory phenotype (CD20⁺, CD38^low, CD27⁺, IgD⁻), with absolute counts dramatically reduced (0.21 cells/µl vs. 76.8 cells/µl at baseline). Their persistence despite near-total depletion may reflect resistance of memory cells in peripheral lymphoid tissue (see Anolik2004 - Rituximab and B Cell Abnormalities in SLE).

  • SM and DN2 share methylation state but diverge epigenetically: DNA methylation phylogenetics show SM and DN2 as the two B cell subsets closest to ASCs, yet their chromatin accessibility is driven by different TF programmes: SM by NF-κB/EBF/OCT2 (GC-associated), DN2 by T-BET/AP-1/EGR (EF-associated). This confirms that SM and DN2 are genuinely distinct differentiation endpoints — SM is GC-derived memory, DN2 is an EF effector — not different maturation stages of the same lineage (see Scharer2019 - Epigenetic Programming in SLE B Cells, RRBS + ATAC-seq, n=9 SLE + 12 HC).

  • DN1/DN2 resolution resolves the DN-memory relationship: The DN1/DN2 subdivision (Jenks2018) resolves the longstanding question of whether DN cells are memory cells. DN1 cells (CXCR5⁺, CD21⁺) share a near-identical transcriptome with SWM (only 22 DEGs by RNA-seq), express TCF7 (central memory TF) and CXCR5 (follicle-homing receptor), and are proposed as early SWM precursors that have not yet acquired CD27. DN2 cells (CXCR5⁻, CD21⁻, CD11c⁺, T-bet⁺) are transcriptionally and functionally distinct from memory cells — they are EF effectors poised for immediate plasmablast differentiation, with high IRF4/BLIMP-1, absent BACH2/FOXO1, and open PRDM1 chromatin by ATAC-seq. DN2 cells are not quiescent memory cells but activated pre-plasmablasts (see Jenks2018 - DN2 B Cells and EF Pathway in SLE, RNA-seq + ATAC-seq + in vitro differentiation).

  • CD40L promotes memory, TLR7 promotes EF effectors: In vitro, CD40L stimulation promotes DN1 generation (memory-like) but inhibits aNAV/DN2 generation (EF effectors). This antagonistic regulation between CD40 (GC/memory) and TLR7 (EF/effector) pathways determines whether activated naive B cells enter the memory or effector branch (see Jenks2018 - DN2 B Cells and EF Pathway in SLE).

  • CD27 is not a reliable universal memory marker: Sanz (2025) argues that CD27 as the sole memory marker is now untenable. CD27 can be downregulated upon stimulation with CD70 (on activated T cells), TLR ligands, and cytokines — meaning CD27⁻ status does not necessarily indicate a cell never expressed CD27. Furthermore, the definitive identification of memory B cells would require demonstration of long-term survival and participation in recall responses at an antigen-specific level — a standard most studies fail to meet (see Sanz2025 - Human Atypical B Cells Overview, invited review).

  • Durable CD27⁻ memory with DN2 phenotype exists: Post-SARS-CoV-2 mRNA vaccination, antigen-specific DN2 cells persist >1 year in the absence of intercurrent infections, accounting for >50% of all spike/RBD⁺ cells. This establishes that DN2 is not exclusively an effector phenotype — durable memory cells with DN2 markers exist. Whether memory DN2 cells are generated via EF or GC pathways, and whether they share the full transcriptional/epigenetic profile of effector DN2, remains unresolved (see Sanz2025 - Human Atypical B Cells Overview, review citing Faliti et al. 2024).

  • T-bet⁺/FcRL5⁺ memory ABC vs. canonical memory: Within the CD27⁺ memory compartment, T-bet⁺/FcRL5⁺ memory ABC cells are poised for ASC differentiation (transcriptional, epigenetic, and metabolic programmes), while T-bet⁻/FcRL5⁻ canonical memory cells have stem-like central memory properties. Memory ABC correlate with long-lived antibody responses, suggesting they contribute to antibody response durability (see Sanz2025 - Human Atypical B Cells Overview, review citing Nellore et al. 2023).

  • “Atypical” memory B cells are not atypical: Sanz (2025) argues the AtB/ABC label should be abandoned. Cells classified as “atypical memory” are in fact a normal component of immune responses. The memory vs. effector identity of any CD11c⁺/T-bet⁺ B cell depends on the context: in primary responses, the majority are naïve-derived effectors; in recall settings, they may represent genuine memory (see Sanz2025 - Human Atypical B Cells Overview).

  • IgM-only memory as GC-independent first layer: The elevated frequency of IgM sequences in the IgD⁻CD27⁺ memory compartment in SLE (20.9–68.1%) compared with vaccinated controls (1.5–37.5%) is attributed to the active generation of IgM-only memory from newly recruited naive B cells via EF pathways — representing the first GC-independent memory layer before class switch recombination occurs (see Tipton2015 - ASC Diversity and Origin in SLE, citing Dogan et al. 2009).

  • acN cells as activated-naive precursors distinct from memory: The activated naive (acN) population (IgD⁺CD27⁻, CD19^hi, MTG⁺, CD24⁻) described in Tipton2015 represents a recently activated state within the naive compartment — not yet a memory B cell — but functionally it is the most important short-term precursor of circulating ASCs during SLE flares. The temporal relationship: naive → acN → ASC → (potentially) IgM-only memory → isotype-switched memory. acN cells may be the gateway through which naive B cells enter both EF and GC differentiation programs (see Tipton2015 - ASC Diversity and Origin in SLE).

  • Unswitched memory contraction in severe COVID-19: usM (IgM⁺IgD⁺) cells were significantly reduced in CoV-A patients — matching the pattern consistently observed in SLE and other autoimmune diseases. This contraction accompanies EF pathway activation and may reflect either consumption of the usM pool as EF precursors or suppression of the GC pathway that generates usM cells (see Woodruff2020 - EF B Cell Responses in COVID-19).

  • DN1/sM overlay on UMAP: UMAP projections of the composite COVID-19 dataset confirmed DN1 and switched memory (sM) cells occupying overlapping space, consistent with the Jenks2018 finding that DN1 and sM differ by only 22 DEGs. This overlay was clearly separable from the distinct DN2 cluster, reinforcing the dual-lineage model of DN1 (GC-derived/memory) vs. DN2 (EF-derived/effector) (see Woodruff2020 - EF B Cell Responses in COVID-19, Fig. 2c–d).

  • Memory B cells, not naive cells, are the dominant Tph responders in dengue: In T-B coculture experiments, Peripheral Helper T Cell (CXCR5⁻PD-1⁺) cells drive class-switched memory B cell → plasmablast differentiation via IL-21. Naive B cells respond poorly to Tph-provided help. This contrasts sharply with SLE, where naive (acN) cells are the dominant EF ASC precursors during flares (Tipton2015). The difference likely reflects the endemic setting — most patients have pre-existing cross-reactive memory from prior DENV exposure, making memory recall the primary substrate for EF activation in dengue. Caveat: the coculture system used seropositive donor memory T cells, not acute-phase T cells (which died in culture), so the memory B cell preference may partly reflect the experimental system rather than in vivo biology (see Ansari2025 - Peripheral T Helper Subset Drives B Cell Response in Dengue, T-B coculture + anti-IL-21 blocking, n=170 cohort).

  • FIRST DENGUE DATA — Qualitative reprogramming of DENV-specific MBC compartment: In the first comprehensive phenotyping of DENV-specific MBC subsets, total DENV-specific B cell frequency does not differ between 1° and 2° dengue (median 54 vs 66 per 10,000 B cells; n.s.), but specific MBC subsets — IgG+, IgM+, atypical (CD27⁻CD21⁻), and class-switched IgD⁻ MBCs — are significantly higher in 2° immunity (p<0.001 for IgG+ and IgD⁻). These subsets are durable to 18 months and accumulate with exposures, representing a bona fide DENV-specific memory compartment. Critically, 2° immunity is driven by subset-level qualitative changes, not total frequency — measuring only total DENV-specific B cells masks this reprogramming (see Singh2026 - DENV-Specific Memory B Cell Subsets, n=58 samples, 18 pediatric patients, longitudinal to 18M).

  • DENV-specific IgM+ MBCs as underappreciated memory compartment: DENV-specific IgD⁻IgM+ MBCs are significantly higher in 2° than 1° dengue at the acute timepoint (p<0.05) and at 18 months (p<0.05), and are the only significantly different MBC subset during acute infection — suggesting IgM+ MBCs are recalled from memory during secondary DENV infection. This is distinct from naïve-like IgD+/IgM+ cells and connects to the broader literature on virus-specific IgM+ memory in SARS-CoV-2, YFV, and rotavirus (see Singh2026 - DENV-Specific Memory B Cell Subsets, n=4/group longitudinal).

  • Naïve-like DENV-specific IgD+/IgM+ B cells persist: A population of DENV-specific B cells with naïve phenotype (CD20+/IgD+/IgM+) persists to 18 months, constituting ~14% (1°) and ~8% (2°) of all DENV-specific B cells. Whether these represent true antigen-experienced cells with SHM, germline-encoded polyreactive binders, or anergic cells is unresolved (see Singh2026 - DENV-Specific Memory B Cell Subsets).

  • Delayed MBC peaks in 2° dengue suggest ongoing maturation: Peak frequency of several DENV-specific MBC subsets occurs >3 months post-infection in 2° but not 1° responses (p<0.05 for 3 subsets), and an uptick at 12–18M was observed in 3/4 secondary cases. This is consistent with prolonged GC reactions or tissue-peripheral redistribution dynamics in 2° immunity (see Singh2026 - DENV-Specific Memory B Cell Subsets, n=4/group — small sample).

  • Different initiating compartments by infection history: Temporal correlation analysis suggests IgD+ naive B cells and IgG+ MBCs are the most influential early responders in 1° dengue immunity, whereas IgM+ MBCs and atypical MBCs are the early responders in 2° immunity — fundamentally different MBC subsets initiate the response depending on prior exposure (see Singh2026 - DENV-Specific Memory B Cell Subsets, correlation network model).

  • Anamnestic IgG-dominant plasmablast response implies memory recall in secondary dengue: Wrammert2012 showed that the acute dengue plasmablast burst is almost entirely IgG-secreting (IgM near-absent) in a cohort of 42/46 secondary infections. The 4 primary responders had detectable IgM, confirming that IgG dominance reflects class-switched memory B cell recall. The magnitude (47% of B cells) and kinetics (day 6–7 peak, barely detectable at day 2–3) are consistent with rapid anamnestic activation of pre-existing cross-reactive memory, later mechanistically attributed to the Tph→IL-21→memory B cell→plasmablast axis by Ansari2025 (see Wrammert2012 - Plasmablast Responses in Acute Dengue, n=46 cohort, ELISpot).

  • Convergent CDR3s in acute dengue derive from affinity-matured memory B cells — BCR-level confirmation of memory recall: VH sequencing of unsorted PBMCs from 60 dengue patients identified specific CDR3 amino acid sequences shared across individuals during acute infection but absent post-recovery. These convergent CDR3s carried 4.4–6.9% V gene mutation (well above naive threshold), were encoded by multiple V gene families (convergent evolution), and were significantly more prevalent in secondary than primary dengue (p≤0.05). This provides repertoire-level evidence — orthogonal to the flow cytometry and coculture data — that pre-existing affinity-matured memory B cells are preferentially expanded during acute dengue, particularly in secondary infection. The convergence across individuals further suggests that dengue antigens impose strong structural constraints on the antibody response, selecting shared CDR3 solutions from the memory repertoire (see Parameswaran2013 - Convergent Antibody Signatures in Dengue, 454 pyrosequencing, n=60 cohort).

  • Higher B cell clonality in secondary dengue — quantitative memory recall signature: P(collision), a replicate-based clonality metric, was significantly higher in secondary vs. primary acute dengue (p=0.0409) but did not differ between primary and secondary convalescent or post-convalescent samples. This temporal profile — clonality elevated only during the acute phase of secondary infection — is the expected signature of memory B cell recall: pre-existing cross-reactive clones expand massively upon re-encounter, producing a more oligoclonal repertoire than the polyclonal primary response (see Parameswaran2013 - Convergent Antibody Signatures in Dengue).

  • Plasmablasts represent a small, non-representative subset of the memory B cell repertoire in dengue: Longitudinal single-cell and 454 BCR sequencing in 12 DENV-2 patients showed very few CDR3 sequences shared between acute-phase plasmablasts (CD19⁺CD20⁻CD27^hiCD38^hi) and convalescent DENV-binding MBCs (CD19⁺CD20⁺CD27⁺). The rare shared clones were exclusively IgM. Specificity diverged sharply: PB-derived mAbs were 85% E-specific (none prM), while MBC-derived mAbs were 56% complex epitope-specific and 24% prM-specific. This establishes that the acute PB wave samples only a narrow slice of the broader MBC repertoire — measuring plasmablasts alone during acute dengue does not capture the full memory response (see Appanna2016 - Plasmablasts as Subset of Memory B Cell Pool, n=12 cohort, mAb cloning + sequencing).

  • Comparable VH mutation between clonally distinct PBs and MBCs — both populations are affinity-matured: Despite being clonally unrelated, PBs and DENV-binding MBCs showed similar VH nucleotide mutation frequencies and N-addition counts. This argues against a model where MBCs undergo substantially more GC maturation than PBs and suggests both derive from comparably matured precursors that diverged in antigen specificity (E-specific for PBs, prM/complex epitope for MBCs). Combined with the Ansari2025 Tph→memory B cell model, this is consistent with E-specific IgG⁺ memory cells being selectively recruited into the PB compartment (see Appanna2016 - Plasmablasts as Subset of Memory B Cell Pool, IMGT mutation analysis).

  • Paradox: lower SHM in secondary than primary dengue — challenges pure memory recall model: GodoyLozano2016 found SHM significantly lower in secondary than primary acute dengue (p<0.001). If the plasmablast wave were driven entirely by GC-matured memory recall (as suggested by the Ansari2025 Tph→memory B cell model and the anamnestic IgG-dominant kinetics from Wrammert2012), one would expect higher SHM in secondary infections — more rounds of GC maturation from prior exposures. The inverse pattern suggests either: (a) original antigenic sin activates cross-reactive memory clones with low SHM that bypass further GC maturation; (b) a GC-independent pathway producing new germline-coded responses dominates in secondary infection; or (c) both pathways contribute, but the GC-independent pathway is proportionally stronger in secondary infection. This complicates the dichotomy between memory-recall and naive-derived plasmablasts (see GodoyLozano2016 - Lower IgG SHM Rates in Acute Dengue, n=19 acute dengue, 454 pyrosequencing).

  • MBC1 cluster — transcriptomic evidence for quiescent alternative lineage memory: scRNA-seq identified a quiescent MBC1 cluster at the base of the alternative lineage pseudotime branch (alongside atBC1, atBC2, atBC3). MBC1 cells express memory markers without activation markers and are T-bet⁺/FCRL5⁺ — providing transcriptomic evidence that durable memory cells exist within the alternative (T-bet⁺/CD11c⁺) lineage. This confirms the Sanz2025/Faliti2024 prediction and establishes MBC1 as a candidate “memory DN2” population (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, n=4, 10x Chromium scRNA-seq).

  • Alternative lineage memory primed by vaccination: PfSPZ vaccination primes Pf-specific alternative lineage memory cells that are recalled upon boosting, with progressive acquisition of CD21⁻CD27⁻ surface phenotype upon repeated antigen exposure. Influenza vaccination similarly activates the alternative lineage. This demonstrates that alternative lineage memory is generated during normal vaccine responses, not only in pathological contexts (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, n=15 PfSPZ + n=9 influenza vaccine cohorts).

  • Isotype-fate segregation predicts selective MBC recruitment: IgM⁺ memory B cells tend to re-initiate germinal centers for affinity maturation, while IgG⁺ memory B cells preferentially yield plasmablasts and plasma cells (Seifert et al. 2015, cited in Bhattacharya2016 - Memory B Cell Subset Selection in Secondary Dengue, commentary). This provides a mechanistic basis for the Appanna2016 clonal disconnect: E-specific IgG⁺ memory cells may be intrinsically predisposed to PB differentiation during secondary dengue, while prM/complex epitope-specific IgM⁺ memory cells maintain memory identity or re-enter GCs. The framework also explains why the dengue PB wave is overwhelmingly IgG (Wrammert2012) and why Tph cells preferentially drive memory (not naive) B cells to PB fate (Ansari2025) — the responders are pre-existing IgG⁺ memory cells.

  • PB analysis alone is an insufficient correlate of long-term DENV immunity: Because only a subset of DENV-specific memory B cells is recruited into the circulating PB pool, plasmablast magnitude and specificity do not capture the full memory response. The unmeasured compartments — MBC diversity, bone marrow-resident long-lived plasma cells, and potentially tissue-retained PBs in secondary lymphoid organs — may be equally or more important for protection or pathogenesis (see Bhattacharya2016 - Memory B Cell Subset Selection in Secondary Dengue, commentary on Appanna2016).

  • STRONGEST FUNCTIONAL EVIDENCE OF MEMORY ORIGIN — OAS in secondary dengue plasmablasts: Priyamvada2016 provides the most direct functional proof that secondary dengue plasmablasts are memory-derived. In 2/4 patients with secondary DENV2 infection, plasmablast-derived mAbs preferentially neutralised DENV1 (the presumed prior serotype) over the infecting DENV2 — Original Antigenic Sin. DENV1-specific mAbs were more potent (FRNT₅₀ 0.16 µg/ml) than DENV2-specific (1.2 µg/ml), and showed binding preference for DENV1 virions by capture ELISA. This serotype bias can only arise from recall of memory B cells generated during prior DENV1 exposure, not from de novo activation by the current DENV2 infection. Combined with high SHM (mean 18.1 VH mutations, CDR R:S >2.9) and 23% clonal relatedness, this constitutes the strongest composite evidence for memory B cell origin of secondary dengue plasmablasts — converging flow cytometry (Wrammert2012), coculture (Ansari2025), repertoire (Parameswaran2013), and now functional mAb data (see Priyamvada2016 - Cross-Reactive Memory Plasmablasts in Secondary Dengue, n=4 secondary DHF, 53 mAbs).

The layered-identity framework (Glaros2025)

  • ★ Memory identity is set by four independent axes, not one. A synthesis review reframes the MBC compartment as a set of overlapping “layers”: developmental pathway (GC-derived vs GC-independent), duration and strength of antigenic stimulation, anatomical and microanatomical localization, and timing of generation in ontogeny. Combinations of these — not any single marker — determine what an MBC does on reactivation (see Glaros2025 - Multilayered Identity of B Cell Memory, review, no original data).

  • ★ Activated B cells are tripotent, and GC-independent memory is the LARGER compartment. Activated precursors at the T–B border can become early plasma cells, early MBCs (eMBCs), or GC B cells; limiting-dilution transfer shows one naive B cell can seed all three. GC-specific genetic fate mapping (S1pr2/Gcsam-driven Cre) then showed, unexpectedly, that eMBCs outnumber GC-derived MBCs across multiple immunization scenarios (see Glaros2025 - Multilayered Identity of B Cell Memory, review, mouse fate mapping — ⚠ the review attributes this jointly to Glaros 2021 Immunity and to the authors’ own unpublished results — the numerical-dominance claim is not independently published, and Glaros 2021 is a self-citation. Treat as the authors’ position, not as an established finding; do not let it carry weight alongside the published fate-mapping origin result (Song 2022).). Humans with CD40L deficiency — unable to form functional GCs — still generate detectable MBC populations. See Early Memory B Cell.

  • ★ The surrogate markers the field used to infer GC origin do not work. Class-switch status fails because CSR occurs primarily before GC entry, with switched cells then gaining a competitive advantage inside GCs; SHM load fails because the ranges overlap in both directions; and recent transcriptomics finds eMBCs and gcMBCs “closely related at the transcriptional level, with only relatively subtle differences” (see Glaros2025 - Multilayered Identity of B Cell Memory, review). Wiki consequence: any claim in this wiki that infers GC vs EF origin from isotype or mutation load alone is weaker than it reads.

  • The CD80/PD-L2 subsetting axis — and its “DN” naming trap. MBCs are commonly split into DP (CD80⁺PD-L2⁺), SP (CD80⁻PD-L2⁺), and DN (CD80⁻PD-L2⁻). Fate mapping shows DP is enriched for gcMBCs and DN consists almost exclusively of eMBCs — though a considerable proportion of DP cells also arise non-GC. ⚠ This “DN” means CD80⁻PD-L2⁻ and is unrelated to this wiki’s IgD⁻CD27⁻ Double-Negative B Cell. See the false-friend note on Atypical B Cell.

  • Reactivation fate splits by subset — but naive B cells supply most of the secondary GC. IgG1⁺ and DP MBCs (gcMBC-enriched) are biased toward PC differentiation; IgM⁺ and DN MBCs (eMBC-enriched) more readily reenter secondary GCs. However, GC-specific fate mapping showed fate-mapped gcMBCs contributed only minimally to secondary GCs while the non-fate-mapped pool dominated — and a follow-up study found those secondary GC cells derived primarily from naive B cells. Net: gcMBCs have the greater intrinsic capacity for secondary GC entry, but naive B cells fill most of the actual pool (see Glaros2025 - Multilayered Identity of B Cell Memory, review).

  • ★ Epigenetic “recording” of cumulative stimulation biases MBCs toward PC fate. Repeated or prolonged antigen/T-help stimulation progressively increases chromatin accessibility at PC-associated loci including Prdm1, raising baseline BLIMP-1 expression and biasing MBCs toward PC differentiation over GC reentry — routed through epigenetic regulation of BACH2 and BLIMP1. The proposed model is that MBC fate on reactivation is programmed not by a binary GC vs non-GC origin but by the cumulative history of stimulation recorded epigenetically (see Glaros2025 - Multilayered Identity of B Cell Memory, review; the authors flag it as “intriguing… remaining to be tested”). This bears directly on the secondary-dengue SHM paradox above — see Original Antigenic Sin and the source page’s Questions Raised for the (wiki-generated, untested) application.

  • gcMBCs are enriched for LOW-affinity clones, and that is the point. The GC-derived memory compartment is seeded across a broad affinity spectrum including very low-affinity cells, giving it greater clonal diversity than the GC-derived PC compartment. The review argues this diversity is what enables cross-variant recall — protective antibodies against flavivirus and influenza variants arise primarily from rapid MBC→PC differentiation rather than from long-lived plasma cells laid down in primary infection (see Glaros2025 - Multilayered Identity of B Cell Memory, review). Relevant to dengue serotype cross-reactivity.

  • Anatomy is a memory layer in its own right. MBCs are non-uniformly distributed — biased to the lymph-node outer follicular area near the subcapsular sinus, resident in lung/liver/gut, distinct in bone marrow, and (in humans) heavily represented among splenic marginal-zone B cells. Draining-LN MBCs reenter secondary GCs more efficiently than non-draining-LN MBCs, so the anatomical site of a booster can influence response quality (see Glaros2025 - Multilayered Identity of B Cell Memory, review). See Tissue-Resident Memory B Cell.

  • Ontogeny is a memory layer too. Clones generated in the first two weeks of life — with a distinct, low-TdT, germline-proximal repertoire — contribute substantially to adult antigen-experienced compartments across organs, most strongly at mucosal sites, and are enriched among splenic MZ B cells (see Glaros2025 - Multilayered Identity of B Cell Memory, review). A large fraction of the adult MBC pool may therefore predate any adult exposure.

  • ABCs are framed as the first Bmem pool defined by a transcriptional programme rather than surface markers, and are BAFF-independent like other Bmem. ABCs are treated as a durable, antigen-experienced memory B cell subset — notable as arguably the first Bmem population defined primarily by a characteristic transcriptional regulatory programme (the T-bet/IFN-γ/IL-21 axis) rather than by surface phenotype alone. ABCs express BAFFR/TACI but are largely BAFF-independent (BAFF-blocking antibody eliminates FO and MZ pools but leaves ABC numbers unchanged) — a property they share with conventional memory B cells (see Cancro2020 - Age-Associated B Cells, review — no original data; mouse).

  • T-bet⁺ memory B cells are multipotential, and interchange with the T-bet⁻ Bmem pool is an open question. In murine Ehrlichia muris infection, protective T-bet⁺ memory B cells give rise to multiple effector B cell lineages on serial adoptive transfer — i.e., are multipotential rather than lineage-restricted. Whether there is appreciable interchange between the T-bet⁺ ABC pool and the T-bet⁻ conventional Bmem pool is explicitly listed as unresolved (see Cancro2020 - Age-Associated B Cells, review — no original data; mouse). This bears directly on this page’s existing “memory DN2” discussion (Sanz2025/Sutton2021 MBC1 cluster) — whether that durable T-bet⁺ population interconverts with canonical Bmem is untested in any species.

  • The nomenclature case: memory cannot be read off CD27 alone, and Bm1–Bm5 makes it worse. The review’s “first critical distinction separates IgD⁺CD27⁺ and IgD⁻CD27⁺ cells”, and it holds that CD27⁺ B cells are the main source of recall responses and are endowed with enhanced ASC differentiation and a distinct transcriptional programme. Applied to blood, however, Bm1–Bm5 “coalesces different types of memory B cells” and “does not distinguish between conventional CD27⁺ memory cells and IgD/CD27 double negative cells, which in turn contains a heterogeneous population of cells including atypical/tissue-based/exhausted memory cells and activated extrafollicular PB precursors” (see Sanz2019 - Consistent Classification of Human B Cell Populations, review — no original data). See Bm Classification, Double-Negative B Cell.

  • [2026-08-29] Unswitched-memory contraction is a cross-disease autoimmune signature, not an SLE marker. All three lupus groups — including skin-limited disease — showed loss of IgD⁺CD27⁺ unswitched memory, and the authors note the same contraction in Sjögren’s syndrome, rheumatoid arthritis, vasculitis and inflammatory bowel disease, where it can be restored by TNF inhibition. In Sjögren’s it is an early feature that correlates with serological autoimmunity and disease progression. Candidate explanation offered without direct evidence: loss of a marginal-zone-equivalent population with apoptotic-clearance, IL-10-mediated regulatory and autoreactivity-diluting functions (see Jenks2021 - B Cell Subset Composition in Cutaneous Lupus, n=207 + 46 HCD, cross-sectional; cross-disease comparisons are relayed from other groups’ work, not measured here).

  • [2026-08-29] The memory family clusters as a unit — and DN1 belongs to it. Hierarchical clustering resolved three subset families: early (resting naive, early transitional), memory (unswitched memory, switched memory and DN1) and effector (activated naive, DN2, DN3, plasmablasts). DN1 grouping with the memory family is consistent with its description as transcriptionally linked to resting switched memory (see Jenks2021 - B Cell Subset Composition in Cutaneous Lupus).

  • ★ Memory maintenance and memory differentiation have separable transcriptional requirements. Tamoxifen-induced deletion of Tbx21 from established day-90 influenza memory B cells did not change memory B cell numbers, frequencies or isotype distribution over a 10-day period — but after heterologous X31 challenge, NP⁺ ASC numbers fell ~10-fold versus challenged controls, and CXCR3 expression was lost from NP⁺ memory B cells (see Stone2019 - T-bet Promotes ASC Differentiation by Limiting IFN-gamma Inflammation, mouse, inducible Tbx21^fl/fl^hCD20-TAM-cre, n=3–6/group). A memory B cell can therefore persist normally while having lost the capacity to execute a recall response. Enumerating a memory population says nothing about its recall competence, which is a direct caution for cross-sectional cohort designs that measure subset frequencies at a single timepoint. See T-bet, CXCR3.

  • A memory population that fails the memory signature test — and can still re-enter a germinal centre. Despite a memory-like surface phenotype (CD38⁺CD138⁻IgD^lo^GL-7⁻, maintained to at least day 35), gene-set enrichment at day 12 did not match conventional memory B cell signatures in existing molecular databases, which the authors attribute to the subset’s low frequency in the memory pool (~5% of BrdU⁺ memory at day 30). On sorted transfer and re-challenge, all transferred subsets — naive, memory and T-bet⁺CD11c⁺ — gave rise to GC B cells at day 10 (see Song2022 - Tfh Outside Germinal Centers Drive T-bet CD11c B Cells, mouse, LCMV-Armstrong, n=3–5 mice/group). A GC-independent origin therefore does not fix the cell outside the GC on re-exposure, which bears on whether the EF/GC distinction is a property of the response or a fixed lineage boundary.

Contradictions & Debates

  • CD27-based memory B cell identification misses the DN subset, which can represent the majority of the memory compartment in SLE and may be relevant in acute infections. Studies that rely solely on CD27 to enumerate memory B cells will undercount this population.
  • The DN1/DN2 subdivision partially resolves the ontological question: DN1 cells are likely GC-derived memory precursors (SWM-like), while DN2 cells are EF effectors and not memory cells per se. Whether DN1 cells complete full GC reactions or represent early GC emigrants remains unresolved.
  • New complication — memory DN2 cells: The discovery of durable antigen-specific DN2 cells post-vaccination blurs the effector/memory distinction. DN2 may be a phenotypic state that can be occupied by both effector cells (in acute responses) and memory cells (long-term), with different underlying transcriptional and epigenetic programmes. Studies using DN2 phenotype alone cannot distinguish these two states (see Sanz2025 - Human Atypical B Cells Overview).
  • ★ Is “GC-independent memory” the same thing as “the extrafollicular response”? No, and the wiki should keep them apart. Glaros2025 - Multilayered Identity of B Cell Memory’s eMBC is a quiescent, low-SHM, largely unswitched memory cell formed by default when antigen runs out — the review is explicit that it “seems to reflect not an active differentiation process but rather a passive transition to quiescence.” The EF arm this wiki tracks in SLE/COVID/dengue is a proliferative, antibody-secreting effector programme (aNAV → DN2 → plasmablast) with an active transcriptional identity. Both are GC-independent and both descend from the same tripotent precursor, but they are different outputs of that branch point — eMBC is the memory limb, DN2→PB the effector limb. Equating “GC-independent” with “extrafollicular effector” would be a category error, and it would make the wiki’s EF claims look better supported than they are.
  • How much weight can murine fate-mapping bear for a human dengue cohort? The two claims this wiki would most like to import — GC-independent origin of ABCs, and eMBC numerical dominance — both rest on mouse GC-specific fate mapping, a technique with no human equivalent. The human corroboration (CD40L deficiency, BCL6-deficiency) establishes that GC-independent memory exists in humans, not that it dominates. The review does not overclaim here; the wiki should not either.

Double-Negative B Cell, DN2 B Cell, Switched Memory B Cell, Activated Naive B Cell, CD27, IgD, Somatic Hypermutation, Extrafollicular Response, Germinal Center, Class Switch Recombination, Plasmablast, Original Antigenic Sin, Antibody-Dependent Enhancement, Early Memory B Cell, Tissue-Resident Memory B Cell, Atypical B Cell, BACH2

Sources

Graph View

Backlinks