EFB Dengue Literature Review

DN2 B Cell

29 min read

DN2 B Cell

Overview

DN2 B cells are a subset of IgD⁻CD27⁻ (double-negative) B cells defined by the phenotype CXCR5⁻, CD21⁻, CD11c⁺, CD19^hi. They were formally defined by Jenks et al. (2018) as the dominant expanded DN population in active SLE, distinct from DN1 cells (CXCR5⁺, CD21⁺, CD19 intermediate) which transcriptionally resemble switched memory B cells.

DN2 cells are pre-plasmablasts: they express a T-bet/ZEB2 transcriptional programme, have high IRF4 and BLIMP-1, lack BACH2 and FOXO1, and differentiate into autoantibody-secreting plasmablasts in response to TLR7 + IL-21 + IFN-γ without requiring BCR stimulation or extensive cell division. They share phenotypic markers and a near-identical transcriptome with activated naive (aNAV) B cells, forming the intermediate step in the extrafollicular differentiation pathway: rNAV → aNAV → DN2 → plasmablast.

Key Points from Literature

  • Full phenotype: IgD⁻, CD27⁻, CXCR5⁻, CD21⁻, CD11c⁺, CD19^hi, CD24⁻, CD38⁻, CD62L^lo, MTG⁺, FCRL4⁻, FCRL5⁺. Higher expression of CD32b, CD22, CD69, HLA-DR, CD86 relative to NAV and SWM. Surface IgG 50% lower than DN1 or SWM (see Jenks2018 - DN2 B Cells and EF Pathway in SLE, two SLE cohorts N=90 + HCD N=21).

  • IgG3 enrichment: DN2 cells have a higher frequency of IgG3⁺ cells than SWM or DN1 in both HCD and SLE. IgG3 is the human equivalent of murine IgG2a (the dominant subclass in T-bet⁺ ABC-driven autoimmunity) and is the dominant IgG subclass deposited in lupus kidneys (see Jenks2018 - DN2 B Cells and EF Pathway in SLE).

  • Frequency in SLE: DN2 cells are a minor fraction of CD19⁺ and DN B cells in HCD (mean + 2 SD = 4.01% of CD19⁺), but represent the majority of DN cells and may become the dominant non-plasmablast CD19⁺ population in active SLE. Absolute DN2 counts are greatly elevated in SLE patients with high DN2 frequency (see Jenks2018 - DN2 B Cells and EF Pathway in SLE).

  • Transcriptional identity (RNA-seq): Over 1,000 DEGs separate DN2 from NAV and SWM. DN2-high: TBX21, ZEB2, IRF4, ITGAX, FCRL5, FCGR2B, PRDM1, SLAMF7, SPI1. DN2-low: BACH2, FOXP1, FOXO1, BCOR, TRAF5, TNFAIP3, BCL2, ZEB1, ETS1, TCF7, CXCR5, CR2. DN2 and aNAV cells have highly similar transcriptomes. DN1 and SWM differ by only 22 DEGs (see Jenks2018 - DN2 B Cells and EF Pathway in SLE, RNA-seq of sorted populations).

  • Poised for PC differentiation: High IRF4/low IRF8 ratio; low ETS1; BLIMP-1 (PRDM1) protein elevated above all B cells except PC; PRDM1 locus open by ATAC-seq; enrichment for IRF4 target genes expressed in PC by GSEA. SLAMF7 (PC marker) upregulated in DN2 and aNAV but no other B cells (see Jenks2018 - DN2 B Cells and EF Pathway in SLE).

  • TLR7 hyper-responsiveness: R848 (TLR7 agonist) induces strong pERK and pMAPKp38 phosphorylation in DN2 and aNAV but not in SWM, DN1, or NAV. CD40L stimulation increases CD25 in NAV but not DN2. Mechanistic basis: low TRAF5 (the main negative regulator of TLR signalling in B cells) (see Jenks2018 - DN2 B Cells and EF Pathway in SLE, phospho-flow cytometry n=5–10).

  • In vitro differentiation from rNAV: rNAV + TLR7 + IFN-γ + IL-21 generates aNAV (day 3), DN2 (day 3–5), and PC (day 5–7). Equal efficiency from SLE and HCD rNAV cells. IL-4 substitution blocks generation. CD40L inhibits aNAV/DN2 but not DN1 generation. aNAV cells directly differentiate into DN2 cells in 3-day cultures (see Jenks2018 - DN2 B Cells and EF Pathway in SLE, in vitro n=5).

  • DN2 → PC without BCR stimulation or cell division: DN2 cells stimulated with TLR7 + IL-21 + IFN-γ (signal 3 only) generate PC robustly, in the presence or absence of BCR stimulation, without extensive cell expansion. IgG output per cell equivalent to SWM cultures. TLR7 is required — removing R848 causes >95% cell death by day 7 (see Jenks2018 - DN2 B Cells and EF Pathway in SLE).

  • Autoantibody production: DN2 cell cultures produce anti-Sm, anti-RNP, and anti-Ro autoantibodies at titers comparable to SWM (see Jenks2018 - DN2 B Cells and EF Pathway in SLE, LIPS assay).

  • Clonal connectivity with aNAV and PC: BCR sequencing demonstrates clonal sharing between aNAV, DN2, and PC populations — in vivo evidence of the aNAV → DN2 → PC developmental pathway. IgG mutation rate in DN2 is similar to PC but lower than SWM, arguing against derivation from memory cells (see Jenks2018 - DN2 B Cells and EF Pathway in SLE).

  • Clinical associations: DN2 expansion is most prominent in African-American patients, patients with active nephritis, high SLEDAI, anti-Sm/RNP/RNA autoantibodies. No age dependence (present in a 5-year-old; Jenks2018 reported patients as young as 6). DN2 cells can contribute up to 70% of all blood CD19⁺ B cells in active SLE. Modestly correlated with type I IFN activity. Specifically linked to anti-Sm and anti-RNP titers by LIPS assay (see Jenks2018 - DN2 B Cells and EF Pathway in SLE; see also Sanz2025 - Human Atypical B Cells Overview, review).

  • DN2/DN1 ratio as an informative index: SLE patients typically display an elevated DN2/DN1 ratio even when total DN frequency is not greatly increased, making this calculation an informative index of disturbed B cell homeostasis (see Sanz2025 - Human Atypical B Cells Overview, review).

  • DN2 in RA synovium: DN2 cells in the rheumatoid synovium represent the main precursor of ASC (Wing et al. 2023, cited in Sanz2025 - Human Atypical B Cells Overview, review).

  • Cross-disease naïve-derived DN2: Naïve-derived DN2 cells prominently contribute to the early response to primary SARS-CoV-2 infection, with significant participation in generating neutralizing antibodies but also substantial autoreactivity through dual-reactive ASC with low or no SHM. In healthy subjects, these autoreactive responses subside within months — self-limited EF autoreactivity (see Sanz2025 - Human Atypical B Cells Overview, review citing Woodruff et al. 2020, 2022).

  • ABC as antigen-presenting cells: Across mouse and human studies, ABC/DN2 cells function as powerful APCs. This APC function may be important for TFH induction and sustainment of secondary GC responses. Excessive ABC APC activity has been proposed to cause abnormal TFH regulation, defective antigen-specific GC responses, and induction of autoreactivity (see Sanz2025 - Human Atypical B Cells Overview, review).

  • Distinction from HIV exhausted memory cells: SLE DN2 cells lack FCRL4 (which defines exhausted memory cells in HIV and tissue-resident memory in tonsil). DN2 cells express FCRL5 but retain intact proximal BCR signalling (BLNK phosphorylation after anti-IgG stimulation), unlike the functionally exhausted FCRL4⁺ cells in HIV (see Jenks2018 - DN2 B Cells and EF Pathway in SLE).

  • First infection context: DN2 expansion in severe COVID-19 mirrors active SLE: Critically ill COVID-19 patients (CoV-A cluster) showed DN2 expansion to ~80% of the DN compartment — matching active SLE. DN2 frequency of total CD19⁺ B cells was significantly higher in ICU-C than in OUT-C or HD (P ≤ 0.001). T-bet and CD11c expression (by intracellular staining) were highest in aN and DN2 populations. The EF cluster (aN + DN2 + DN3 + ASC expansion) defined the critically ill patient profile by hierarchical clustering (see Woodruff2020 - EF B Cell Responses in COVID-19, 24-marker spectral FCM, n=10 ICU-C, n=7 OUT-C, n=17 HD).

  • DN2:DN1 ratio in COVID-19 indistinguishable from SLE: log₂(DN2:DN1) in CoV-A was not significantly different from active SLE (both P ≤ 0.0001 vs. HD and CoV-B). This ratio is the most robust single metric of EF pathway activation across disease contexts (see Woodruff2020 - EF B Cell Responses in COVID-19).

  • DN2 frequency correlates with disease severity biomarkers: DN2 frequency within the DN compartment correlated with log(CRP) (r² = 0.39, P = 0.022). CRP in turn correlated with IL-6 (r² = 0.84) and IP-10/CXCL10 (r² = 0.58). This positions DN2 expansion as a direct cellular correlate of the inflammatory cascade associated with COVID-19 morbidity and mortality (see Woodruff2020 - EF B Cell Responses in COVID-19).

  • Chemokine receptor switch on DN2 cells: EF populations (aN, DN2) in CoV-A patients showed decreased CXCR5 and increased CXCR3 relative to follicular populations (rN, DN1) — consistent with homing to IFN-γ-inflamed tissue rather than B cell follicles (see Woodruff2020 - EF B Cell Responses in COVID-19).

  • EF activation not attributable to demographic baseline: African-American HD (n=24) showed slightly higher baseline DN2:DN1 ratios than the primary HD cohort, but remained significantly different from ICU-C — confirming that the DN2 expansion in severe COVID-19 is disease-driven (see Woodruff2020 - EF B Cell Responses in COVID-19).

  • EPIGENETIC CHARACTERISATION — DN2 closest to ASC by DNA methylation: Phylogenetic analysis of differentially methylated loci (DMLs) placed DN2 cells closest to ASCs in both SLE and HC, confirming their position as the most terminally differentiated non-ASC B cell subset. Progressive global hypomethylation from rN → T3 → aN → SM → DN2 → ASC establishes a linear epigenetic differentiation trajectory (see Scharer2019 - Epigenetic Programming in SLE B Cells, RRBS of 5 sorted subsets, n=9 SLE + n=12 HC).

  • DN2 chromatin driven by T-BET + AP-1 + EGR motifs: Motif enrichment of ATAC-seq DARs between DN2 and switched memory identified T-BET, ISGF3, AP-1 (JUN/FOSB/FOSL1/FOSL2), and EGR as the top DN2-enriched motifs. Switched memory was enriched for EBF, NF-κB, and OCT2 — confirming epigenetically distinct differentiation endpoints (see Scharer2019 - Epigenetic Programming in SLE B Cells).

  • T-BET chromatin programme shared by HC and SLE DN2 cells: The T-BET motif enrichment in DN2 accessible chromatin was present regardless of disease status, confirming that T-BET-driven epigenetic programming is a normal feature of DN2 differentiation. T-BET bound to its own TBX21 locus (autoregulatory loop) in both conditions (see Scharer2019 - Epigenetic Programming in SLE B Cells, ATAC-seq + ENCODE ChIP-seq).

  • AP-1/EGR amplification is SLE-specific: While T-BET motifs were shared, AP-1 and EGR accessibility was enhanced in SLE DN2 and aN cells relative to healthy counterparts. This identifies AP-1/EGR as the disease-specific epigenetic layer superimposed on the normal T-BET DN2 programme (see Scharer2019 - Epigenetic Programming in SLE B Cells).

  • ATF3 is a key SLE DN2-specific regulator: ATF3 was maximally expressed in SLE DN2 cells, validated at mRNA and protein levels. 98 ATF3 target genes were disease DEGs (87% upregulated). ATF3 motif accessibility was highest in SLE DN2 cells. ATF3 heterodimerises with Jun family members (JUN, JUNB, JUND — all upregulated in SLE DN2), shifting the equilibrium from repression to activation (see Scharer2019 - Epigenetic Programming in SLE B Cells).

  • DN2 cells lack G2/M checkpoint and apoptosis enrichment: While all other SLE B cell subsets showed enrichment for G2/M checkpoint and apoptosis pathways (GSEA), DN2 cells uniquely showed negative enrichment — potentially explaining their expansion in SLE by resistance to cell cycle arrest and apoptosis (see Scharer2019 - Epigenetic Programming in SLE B Cells).

  • PD-1 highest on DN2 cells: PDCD1 promoter and cis-regulatory elements were highly accessible in DN2 vs. switched memory. PD-1 protein was ~60% positive on DN2 cells vs. ~10–20% on other subsets (flow cytometry, n=4 SLE) (see Scharer2019 - Epigenetic Programming in SLE B Cells).

  • FIRST DENGUE EVIDENCE — CD21⁻CD11c⁺ B cells in acute dengue: Within the IgD⁻CD27⁻ (DN) gate, CD21⁻CD11c⁺ B cells — phenotypically consistent with DN2 — are significantly expanded in acute dengue vs. HD and convalescence. These cells emerge in the context of massive Peripheral Helper T Cell (CXCR5⁻PD-1⁺) activation providing IL-21, the same cytokine required for the aNAV→DN2→plasmablast pathway in SLE. Formal confirmation requires T-bet and CXCR5 staining within the DN gate, but the CD21/CD11c phenotype and cytokine milieu are consistent with DN2 identity (see Ansari2025 - Peripheral T Helper Subset Drives B Cell Response in Dengue, multi-color FCM, n=170 acute dengue adults).

  • SARS-CoV-2 antigen specificity of DN2 confirmed by dual-fluorophore RBD probes: In the peripheral blood of COVID-19 patients, dual-fluorophore RBD probe staining demonstrated that DN2 cells (alongside DN3, aN, SWM, and PB populations) contain SARS-CoV-2-specific cells. DN2 and DN3 were CXCR5-low; DN1 and DN4 were CXCR5-high. This independent confirmation (separate cohort from Woodruff2020) establishes that the DN2 expansion in COVID-19 is antigen-driven, not bystander. The parallel tissue study showed IgD⁻CD27⁻ DN B cells present at both follicular and extra-follicular sites with T-B conjugates (see Kaneko2020 - GC Loss and TFH Block in COVID-19, blood: n=68, 13-color FCM + RBD probes; tissue: n=11, multi-color immunofluorescence).

  • ALTERNATIVE LINEAGE FRAMEWORK — DN2/atBC cells are NOT pre-plasmablasts in healthy/infection contexts: scRNA-seq of >12,000 B cells from malaria-exposed and non-exposed donors placed atBC populations (which overlap with DN2 by phenotype: T-bet⁺, CD11c⁺, FCRL5⁺) on an “alternative lineage” pseudotime branch, separate from classical memory. No atBC cluster upregulated PC maintenance genes (XBP1, IRF4, PRDM1), and PCs were detached from the pseudotime manifold with no intermediate population. Sutton reconciles this with Jenks2018 as context-dependent: in SLE, chronic TLR7 can drive atBCs to PCs, but in healthy/vaccination/infection contexts they are an alternative memory lineage rather than obligate pre-plasmablasts (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, n=4, 10x Chromium).

  • MBC1 cluster = quiescent alternative lineage memory (“memory DN2”): The MBC1 cluster at the base of the alternative lineage pseudotime branch represents a quiescent memory state within the T-bet⁺/CD11c⁺ lineage. This provides transcriptomic evidence for the “memory DN2” population predicted by Sanz2025 and Faliti2024 (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, n=4, 10x Chromium).

  • CD21⁻CD27⁻ gating captures only ~45% of transcriptomic atBCs: CITE-seq demonstrated that the conventional DN gate misses the majority of transcriptomically-defined alternative lineage cells. CD11c protein is a superior single marker. This implies prior estimates of DN2 frequency by flow cytometry are underestimates (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, n=4, CITE-seq).

  • Alternative lineage primed by vaccination: PfSPZ vaccination (n=15) and influenza vaccination (n=9) both activate alternative lineage B cells, with repeated boosting shifting cells toward increasingly atypical surface phenotype (CD21⁻CD27⁻). This establishes that the DN2/atBC phenotype is a normal vaccination response, not restricted to pathological contexts (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection).

  • DN2 is the human counterpart of the murine ABC — and the only DN subset that is ABC-like. In current usage “ABC” denotes murine cells and human DN2 is regarded as the closest counterpart; among DN subsets, DN2 is the only subset that highly expresses T-bet and efficiently differentiates into plasma cells, so the ABC↔DN correspondence runs specifically through DN2 (and only through the IgD⁻CD27⁻ fraction of the heterogeneous ABC superset) (see Lamprinou2026 - ABCs and DN B Cells, opinion, citing Chung 2023 / Ricker 2021 / Satterthwaite 2021 / Jenks 2018).

  • Caveat — ABC ≈ DN2 phenotypically but not transcriptomically. Comparative transcriptomics show ABCs are distinct from other CD11c⁺ B cells including DN2, with elevated cytokine/chemokine expression not seen in DN2 — so the equivalence is a phenotypic approximation, not a molecular identity (see Lamprinou2026 - ABCs and DN B Cells, opinion, citing Maul 2021; see Age-Associated B Cell Contradictions).

  • ★ The soluble-antigen hyporesponsiveness caveat extends to plasma-cell differentiation specifically — with a mechanism. The wiki already carried, via Sanz2025, that atypical B cells may respond to membrane-associated antigens and immune complexes while appearing hyporesponsive to soluble antigen (see Contradictions below). Glaros2025 - Multilayered Identity of B Cell Memory supplies the mechanism and extends it to the PC-differentiation literature: soluble anti-Ig fails to drive PC differentiation from ABCs, whereas membrane-bound anti-Ig succeeds, because inhibitory receptors — FcγRIIB, and per the review FCRL5 — must be physically excluded from the B cell immune synapse for BCR signalling and PC differentiation to proceed, which only membrane-associated antigen achieves (review, citing Ambegaonkar 2020 Sci Adv; primary abstract independently verified at ingest 2026-08-16). The foundational “ABCs differentiate poorly into PCs” results (Portugal 2015 eLife; Sullivan 2015 PLoS Pathog, both malaria-associated human atypical MBCs) used soluble stimulation.

  • Positive evidence that ABC/DN2-like cells are primed for PC fate. ABCs express PC-associated genes and show increased PC-differentiation propensity in T cell coculture (see Glaros2025 - Multilayered Identity of B Cell Memory, review, citing Song 2022 mouse; Wang 2018 Nat Commun SLE CD11c^hi T-bet⁺ B cells + IL-21; Louis 2021 JCI Insight T-bet⁺CD27⁺CD21⁻ B cells in kidney-transplant rejection). Note the two human primaries are CD27⁺ or SLE-context — neither is a clean IgD⁻CD27⁻ DN2 preparation.

  • But not obligate pre-plasmablasts: ABC-phenotype cells are plastic. Adoptive transfer within an autoimmune mouse strain shows they can also differentiate into GCBCs and self-renew (see Glaros2025 - Multilayered Identity of B Cell Memory, review, citing Nickerson 2023 J Exp Med). Combined with the above, the defensible position is that DN2/ABC cells are capable and often primed for PC differentiation without being committed to it — which is closer to Sutton2021’s alternative-lineage reading than to a strict pre-plasmablast model.

  • In acute viral infection, most antigen-specific ABC-phenotype cells are GC-independent in origin. GC-specific genetic fate mapping in mice assigns the majority to a GC-independent pathway (see Glaros2025 - Multilayered Identity of B Cell Memory, review, citing Song 2022, mouse, one acute viral model — the review flags generalization as unresolved). This is the closest thing the wiki holds to direct origin evidence for the DN2-phenotype expansion seen in acute dengue (Ansari2025 - Peripheral T Helper Subset Drives B Cell Response in Dengue), though it is murine and the dengue cells were never origin-mapped.

  • ★ The murine two-signal model (F1) as mechanistic backbone for DN2-like generation. In the mouse, TLR7 or TLR9 signalling is necessary to poise a B cell for the ABC/DN2-like fate but is not sufficient — poising must be followed by IFN-γ or IL-21, with both requisites cell-intrinsic (coculture-proven). BCR ligation, alone or together with CD40 costimulation, does not enable the fate (though it synergises with TLR signalling for proliferation) — directly consistent with this wiki’s existing finding that DN2 differentiates without BCR stimulation (Jenks2018 - DN2 B Cells and EF Pathway in SLE). T-bet expression begins within 12 h, before the first division (see Cancro2020 - Age-Associated B Cells, review — no original data; mouse).

  • ★ TLR9-tolerance-rescue as an alternative origin story for DN2-like cells (§8). B cells that receive a TLR9 ligand via BCR-mediated internalisation normally undergo cell-cycle arrest and programmed death after an initial proliferative burst — a peripheral tolerance checkpoint (Sindhava et al. 2017, J Clin Invest). This death is circumvented by survival cytokines or CD40 costimulation, and in the presence of IFN-γ or IL-21 the rescued cells assume the ABC phenotype — so ABC/DN2-like generation can arise not only from productive activation but from cells that would otherwise have been deleted. Cancro’s interpretation: the ABC–autoimmunity association may partly reflect failure of, or rescue from, this checkpoint (see Cancro2020 - Age-Associated B Cells, review — no original data; mouse).

  • DN2 has the largest BCR signalling capacity of the three DN subsets — it is not a hyporesponsive cell. DN1, DN2 and DN3 cells from HD, mild and severe COVID-19 patients, and post-SARS-CoV-2-vaccination donors all maintained BCR signalling after IgG stimulation; among them DN2 showed the highest expression of activation markers (CD69, CD86) and the largest BCR signalling capacity. TLR7 stimulation further increased CD25, HLA-DR and CD86 on DN2 specifically (see Beckers2023 - Origins and Functions of DN B Cells, review, citing Castleman 2022 / Jenks 2018; see Phospho-Flow Cytometry).

  • DN2 in SLE out-expresses switched memory on antigen-presentation machinery. SLE DN2 cells expressed higher HLA-DR, CD69 and CD86 than SM B cells — the strongest phenotypic case in the DN literature for an antigen-presenting function, though no ingested source demonstrates DN2-driven T cell activation directly (see Beckers2023 - Origins and Functions of DN B Cells, review, citing Jenks 2018; see Atypical B Cell Effector Output).

  • The ASC-differentiation programme, stated as transcription-factor logic: DN2’s transcriptomic profile includes IRF4 (essential for ASC differentiation) and lacks Ets-1 and BACH2 — both of which normally prevent plasma cell differentiation. The DN2→PB route is therefore a de-repression as much as an induction (see Beckers2023 - Origins and Functions of DN B Cells, review, citing Jenks 2018).

  • DN2 lacks CD62L as well as CXCR5. Both molecules are required for migration into lymphoid follicles; their joint absence is the migration-level argument for DN2’s extrafollicular confinement (see Beckers2023 - Origins and Functions of DN B Cells, review; see Follicular Exclusion).

  • DN2 across the disease spectrum beyond SLE and COVID-19. Review-level tabulation places DN2 or DN2-like (CD11c⁺ / CD21⁻) expansions in: RA (CD11c⁺ DN2 by mass cytometry, associated with anti-citrullinated protein autoantibodies), JIA (synovial fluid, especially in ANA⁺ patients), axSpA (a minority of CD27⁻CD38^low CD21^low cells are T-bet⁺CD11c⁺), pSS, malaria (CD21⁻ atypical memory, “suggesting overlap with DN2”), HIV (CD21⁻T-bet⁺ expansion), acute sepsis, CVID, and obesity (elevated DN2 in young obese individuals, linked to autoantibody production). Frequencies are not comparable across these studies — see Conventional Flow Cytometry (see Beckers2023 - Origins and Functions of DN B Cells, review, Table 1).

  • The “EF” in “EF pathway DN2 cell” is a claim about origin, not location — and only the origin claim is supported. A twelve-author consensus Perspective, co-authored by Ignacio Sanz (in whose laboratory the DN2 framework was defined), states that GC-independent derivation of DN2 is suggested by severe-COVID tissue studies and lupus-nephritis kidney biopsies, but that “direct visualization of EF foci with DN2 cells in the splenic bridging channel or LN medullary cords has not been done.” DN2 cells may instead arise at non-classical sites — the T-B border, interfollicular zone, or subcapsular sinus. The paper’s conclusion is explicit: “the EF designation of this human DN2 cell refers to its presumed GC-independent origin rather than its location.” Nothing in the TLR7 + IFN-γ + IL-21 differentiation programme, the ZEB2/T-bet transcriptional signature, or the clonal connectivity to aNAV is affected — those are origin evidence. What is withdrawn is the licence to infer an anatomical site from a blood phenotype (see Eisenbarth2025 - A Roadmap for Defining Extrafollicular B Cell Responses, consensus Perspective, 12 authors, no primary data). See GC-Independent Response.

  • Annotation layer — the process-based label proposed for this cell is “primary switched non-GCB” (or “primary switched GCB” where evidence warrants), to be used when location is unknown. The wiki retains “DN2” as its working name for searchability against the existing literature; curator decision, 2026-08-27 (see Eisenbarth2025 - A Roadmap for Defining Extrafollicular B Cell Responses).

  • ★ [2026-08-27] DN2 is scarce in inflamed tissue — a hard constraint on the DN2-as-tissue-effector story. In severe COVID-19 thoracic lymph nodes DN2 was ~7 cells/mm² against DN3’s ~400 (n=6), and in IgG4-RD salivary gland ~6 cells/mm² and ~3% of the tissue DN pool (n=10). The authors state that DN2 cells “are not abundant in COVID-19 lymph nodes and are relatively sparse in both IgG4-RD and COVID-19 end organs.” DN2 was significantly increased in the blood of both diseases. The subset carrying the wiki’s extrafollicular case is a blood finding; it is DN1 and DN3 that are found in the tissue. Candidate explanations — rapid transit, retention in other compartments, or loss of CD11c detectability in FFPE — are untested (see Allard-Chamard2023 - DN3 B Cells Infiltrate Inflamed Tissues, n=6 autopsy + n=10 SMG, multi-colour IF).

  • [2026-08-27] DN2 gated without CD21 or T-bet still reproduces its SLE transcriptome. In IgG4-related disease, DN2 defined only as IgD⁻CD27⁻CXCR5⁻CD11c⁺ (13-colour panel, no CD21, no T-bet) yielded a sorted transcriptome that broadly matched the DN2 transcriptome described in SLE by Scharer2019 - Epigenetic Programming in SLE B Cells. This is modest independent support that the CXCR5⁻CD11c⁺ gate captures the same cell as the fuller Jenks/Emory definition — in a third disease (see Allard-Chamard2023 - DN3 B Cells Infiltrate Inflamed Tissues, n=4 bulk RNA-seq).

  • [2026-08-27] DN2 carries a cytotoxic gene module and the highest SLAMF7 of any DN subset. Sorted DN2 expressed GZMA, GZMH, GZMB, GNLY, NKG7, KLRB1, KLRD1, KLRF1 and FCGR3A — genes characteristic of cytotoxic CD8⁺/CD4⁺ T cells and NK cells — shared with DN3. SLAMF7 was confirmed at RNA and protein level, with DN2 the highest of the four subsets (MFI 2123); the authors note SLAMF7 on DN2 was also reported in SLE transcriptomics. Functional significance is unaddressed: no cytotoxicity assay was performed (see Allard-Chamard2023 - DN3 B Cells Infiltrate Inflamed Tissues, n=4).

  • [2026-08-27] In IgG4-RD blood, DN2 tracks plasmablasts less closely than DN3 does. Spearman against plasmablast frequency (n=38): DN2 r = 0.4496, p = 0.0067, versus DN3 r = 0.6566, p < 0.0001. Logistic regression discriminating IgG4-RD from healthy controls: DN2 OR 5.381 (95% CI 1.58–18.32), p = 0.0071 (see Allard-Chamard2023 - DN3 B Cells Infiltrate Inflamed Tissues, n=38).

  • The canonical Sanz-lab phenotype table — and the function it assigns is a hypothesis, by the review’s own standard. Table 1 defines DN2 as IgD⁻CD27⁻CD38⁻CD24⁻CD21⁻, T-bet⁺CD11c⁺FcRL5⁺SLAMF7⁺CXCR5⁻, IgM/IgG/IgA⁺, with function given as “extrafollicular ASC precursors”. Two cautions the review itself supplies: that functional annotation is exactly the kind of phenotype-to-function attribution its own introduction warns against; and DN2 shares its entire activation signature with activated naive cells, differing by IgD alone (see Sanz2019 - Consistent Classification of Human B Cell Populations, review — no original data). See DN2 Gating Strategy.

  • [2026-08-29] The DN2:DN1 log2 ratio is used again as a reported outcome, by the lab that defined it. Figure 1E of a 207-patient lupus-spectrum cohort reports DN2 relative to DN1 as an explicit log2-transformed ratio panel, not an incidental derivation — the third wiki source to report the metric, after Jenks2018 - DN2 B Cells and EF Pathway in SLE (SLE) and Woodruff2020 - EF B Cell Responses in COVID-19 (acute COVID-19, ratio indistinguishable from active SLE); all three are Sanz-lab. DN2 is gated CD21⁻CD11c⁺ within IgD⁻CD27⁻ cells, with no T-bet and no CXCR5 in the panel. DN2 and DN3 together account for the DN expansion in all three lupus groups, with reversal of the normal DN1 predominance (see Jenks2021 - B Cell Subset Composition in Cutaneous Lupus, n=207 + 46 HCD, cross-sectional). This is precedent for the ratio as a stratifying outcome in a large clinically heterogeneous cohort — the situation acute dengue presents; see Thesis Objectives and Grant Pitch.

  • [2026-08-29] DN2 expansion is present but smaller in primary cutaneous lupus than in SLE. Effector expansion (activated naive, DN2, DN3, plasmablasts) was shared by primary CCLE and both SLE groups at lower magnitude in CCLE, alongside the same unswitched-memory contraction. The effector phenotype is therefore not SLE-specific; it grades with systemic involvement (see Jenks2021 - B Cell Subset Composition in Cutaneous Lupus, n=69 primary CCLE vs n=85 SLE without skin disease).

  • A proposed mechanism for DN2 plasmablast efficiency: brake-release rather than PC-programme induction. DN2 cells are efficient plasmablast precursors on stimulation (Jenks2018 - DN2 B Cells and EF Pathway in SLE) yet show no PC-programme upregulation at rest in the alternative-lineage framework (Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection). Stone’s mouse model reconciles these without requiring either to be wrong: T-bet’s contribution to ASC fate is permissive — repressing the IFN-γ-induced NF-κB/TLR/IRF inflammatory programme that blocks terminal differentiation — rather than instructive, since T-bet regulates neither early Prdm1 nor Irf4 nor Xbp1 (see Stone2019 - T-bet Promotes ASC Differentiation by Limiting IFN-gamma Inflammation, mouse, Be1/Be2 co-culture + ATAC-seq). ⚠ Untested in human DN2 cells; see BLIMP-1 Contradictions for the open form of this question.

  • ⚠ A tension with the wiki’s TLR7-driven DN2 model that is worth stating plainly. TLR7 signalling is central to DN2 generation in SLE and is hyper-responsive in DN2 cells. But in Stone’s mouse system, sustained TLR7/9 ligation actively suppresses ASC formation — adding R848 or CpG to wild-type Be1 cultures from day 2 significantly reduced ASC and IgG-secreting cell numbers without affecting proliferation (see Stone2019 - T-bet Promotes ASC Differentiation by Limiting IFN-gamma Inflammation, mouse, in vitro). If both hold, TLR7 signalling would have to be timed: driving DN2 generation early, then withdrawing or being overridden for those cells to secrete. The wiki has no source that resolves the timing. See TLR7, Toll-like Receptor Signaling in B Cells.

  • ★ DN2 cells are pre-programmed primary effectors, not resting precursors — the strongest functional evidence in the wiki. Sorted SLE patient subsets were stimulated for only 2.5 days with R848 + IFN-γ + IL-21 + IL-2 and no anti-Ig: conventional switched memory and DN1 memory formed IgG ASCs efficiently, naive B cells failed entirely, and T-bet^hi^ DN2 gave ≥50-fold more ASCs than naive cells and only 2–3-fold fewer than memory (see Zumaquero2019 - IFN-gamma Programs T-bet-hi B Cells for ASC Differentiation, 3 independent experiments, B cells sorted from 3 SLE donors, one-way ANOVA with Tukey). Two things follow. DN2 differentiation is BCR-independent once IL-21 arrives — no anti-Ig was present. And DN2 cells sit functionally closer to memory than to naive despite lacking CD27, which is the clearest available answer to whether the DN2 phenotype is an activation state or a differentiated effector state.

  • ★ The murine counterpart of DN2 requires Tfh help — which is not the human DN2 model, and the two should not be collapsed. Song2022 shows that T-bet⁺CD11c⁺ B cells in acute viral infection need canonical CXCR5⁺PD-1^hi^ Tfh acting outside the germinal centre: sorted Tfh transferred into infection-matched Tcrb⁻/⁻ mice induced them, sorted Th1 did not, and Icos⁻/⁻, Sh2d1a⁻/⁻ and CD4^Cre^Bcl6^fl/fl^ models all reduced generation (see Song2022 - Tfh Outside Germinal Centers Drive T-bet CD11c B Cells, mouse, LCMV-Armstrong + influenza PR8, n=3–5 mice/group). In human SLE the DN2 pathway is driven by TLR7 with IL-21 and IFN-γ (Jenks2018 - DN2 B Cells and EF Pathway in SLE), with CXCR5⁻ Peripheral Helper T Cells implicated as the helper in dengue (Ansari2025 - Peripheral T Helper Subset Drives B Cell Response in Dengue). ⚠ Treat T-bet⁺CD11c⁺ as the murine analogue of the DN2/ABC phenotype, not as DN2 itself.

Contradictions & Debates

  • DN2 cells were defined in SLE, where dysregulated TLR7 signalling is a disease-intrinsic feature. Whether the same DN2 phenotype and TLR7 hyper-responsiveness characterise EF B cells in acute viral infections (dengue, SARS-CoV-2, malaria) is not yet established. In infections, the TLR7 ligand is exogenous viral ssRNA rather than endogenous self-RNA; the signalling outcome may differ.
  • The relationship between DN2 cells and the “atypical memory B cells” described in malaria and HIV is phenotypically overlapping (both CD21⁻, CD27⁻, T-bet⁺) but functionally distinct: DN2 cells have intact BCR signalling and robust PC differentiation capacity, whereas malaria atypical memory cells show impaired BCR signalling and poor effector function. However, the exhaustion phenotype in malaria has been challenged — AtB cells may respond strongly to membrane-associated antigens and immune complexes even when hyporesponsive to soluble antigens (see Sanz2025 - Human Atypical B Cells Overview, review citing Holla et al. 2019).
  • Memory vs. effector DN2 cells: Post-SARS-CoV-2 vaccination, antigen-specific DN2 cells persist >1 year, accounting for >50% of all spike/RBD⁺ cells. This establishes the existence of durable CD27⁻ memory with DN2 phenotype. The full properties of memory DN2 (including their extended phenotype, T-bet expression, separation from effector DN2, and derivation from EF vs. GC pathways) remain to be elucidated (see Sanz2025 - Human Atypical B Cells Overview, review citing Faliti et al. 2024).
  • Whether the autoreactivity attributed to DN2 cells in SLE is a general property of the phenotype or specific to the autoimmune context is unresolved. Sanz (2025) argues that autoreactivity is context-dependent and should not be assumed from phenotype alone (see Sanz2025 - Human Atypical B Cells Overview). [2026-08-16] A mechanistic constraint now bears on this. The verified Ambegaonkar 2020 abstract states atypical MBCs “robustly respond to antigens that associate with cell surfaces, such as antigens in immune complexes, but are unable to respond to fully soluble antigens, such as self-antigens” (via Glaros2025 - Multilayered Identity of B Cell Memory). If so, a DN2 cell bearing an autoreactive BCR would be poorly activated by soluble self-antigen — meaning DN2-derived autoantibody production would have to run through membrane-associated or immune-complexed self-antigen. This is directly relevant to the bridge-wiki/ cells→autoantibody thesis, which currently assumes the link without specifying antigen form. Secondary dengue is immune-complex-rich, so the route exists — but it must be argued, not assumed.
  • ★ Does the “no PC genes in atBCs” finding survive the stimulation-mode reframing? Partly, and the two are less opposed than they look. Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection reports no PC-maintenance gene upregulation (XBP1, IRF4, PRDM1) in atBC clusters — a steady-state transcriptomic observation. Glaros2025 - Multilayered Identity of B Cell Memory reports that ABCs can differentiate into PCs when the BCR is engaged by membrane-associated antigen, and cites primaries finding PC-associated gene expression and PC propensity in T cell coculture — stimulated-state functional observations. A cell need not express PC genes at rest to be capable of the fate on the right signal, so the two coexist; but they do settle differently against the strong claim. Net: the “obligate pre-plasmablast” model is unsupported by both; the “incapable of PC fate” reading is now clearly wrong; the defensible middle is conditional capacity, with the condition being antigen presentation form and T cell help. Note also that the wiki already held the BCR-signalling half of this via Sanz2025 (Holla 2019) — Glaros’s contribution is connecting it to the PC-differentiation debate and supplying the immune-synapse mechanism.

Double-Negative B Cell, Age-Associated B Cell, Atypical B Cell, Activated Naive B Cell, Plasmablast, T-bet, CD11c, CXCR5, FCRL5, IRF4, BLIMP-1, BACH2, TRAF5, TLR7, ZEB2, ATF3, EGR, PD-1, Extrafollicular Response, Germinal Center, Early Memory B Cell, Mechanistic Case for DN and DN2 Cells in Dengue, GC-Independent Response, DN3 B Cell, SLAMF7

Sources

Graph View

Backlinks