EFB Dengue Literature Review

CD21

10 min read

CD21

Overview

CD21 (Complement Receptor 2, CR2) is a complement receptor expressed on mature naive and memory B cells that binds C3d-opsonised antigen. It forms a co-receptor complex with CD19 and CD81 that lowers the activation threshold of the BCR. CD21 expression is downregulated upon B cell activation; loss of CD21 (CD21^lo or CD21⁻ phenotype) marks activated or anergic B cells and is used in flow cytometry panels to distinguish resting from recently activated B cell populations.

Key Points from Literature

  • CD21⁻ phenotype marks acN cells: In SLE flares, activated naive (acN) B cells (CD19^hi, IgD⁺, CD27⁻, MTG⁺, CD24⁻) are CD21⁻, consistent with downregulation of the complement co-receptor upon sustained B cell activation. In healthy subjects, the CD19⁺IgD⁺CD27⁻ compartment is dominated by CD21⁺ resting naive B cells (see Tipton2015 - ASC Diversity and Origin in SLE, multi-color flow cytometry).

  • Mechanistic basis of CD21 downregulation: B cell activation leads to shedding of complement receptor type II (CR2/CD21) from the cell surface. This is consistent with the CD21⁻ phenotype of acN cells reflecting sustained BCR engagement rather than a developmental stage (see Tipton2015 - ASC Diversity and Origin in SLE, citing Masilamani et al. 2003).

  • CD21^lo B cells previously characterised in SLE: A CD21^lo activated B cell population was previously described in peripheral blood of SLE patients (Wehr et al. 2004, cited in Tipton2015 - ASC Diversity and Origin in SLE). That prior work ascribed this population to CD27⁺ B cells; Tipton2015 clarifies that the relevant activated population is within the CD27⁻ compartment, detectable by MTG and CD24 staining.

  • CD21lo alone is an unreliable marker of AtB/ABC: Sanz (2025) argues that CD21lo cells include several distinct populations — activated naive, DN2, DN3 (pre-plasmablasts), activated memory, and anergic cells — with different functions. Low CD21 in isolation should not be used to identify B cells with pre-defined function (see Sanz2025 - Human Atypical B Cells Overview, review).

  • CD21 and CD11c are not reciprocal: CD21lo cells can be CD11c⁻ (as in DN3 B Cell), and CD11c⁺ cells can be CD21⁺ (as in resting ABC memory). CD21 vs. CD11c is therefore an informative two-dimensional gating strategy, not a single-axis readout (see Sanz2025 - Human Atypical B Cells Overview, review, Figure 2B).

  • Generalised CD21 downregulation in SLE: SLE B cells show a generalised lower level of CD21 expression across all subsets compared with healthy donors, with increased populations of CD21-negative cells. This means CD21lo gating thresholds validated in healthy donors may be inappropriate for SLE samples (see Sanz2025 - Human Atypical B Cells Overview, review, Figure 2C).

  • CD21lo cells have increased TLR7 sensitivity: This could contribute to their recruitment into the ABC compartment and SLE expansion (see Sanz2025 - Human Atypical B Cells Overview, review citing Zhu et al. 2024).

  • CD21lo transitional B cells expanded in mild COVID-19: OUT-C patients showed expansion of CD21lo transitional B cells (>10% of CD19⁺ B cells, with CD21lo cells making up >50% of the gain). CD21lo Tr cells expressed higher IgM, CD24, CD10 and lower CD21, CXCR5 than CD21hi Tr cells, but also showed unexpectedly higher CD138 expression — a marker usually restricted to ASCs (see Woodruff2020 - EF B Cell Responses in COVID-19, spectral FCM).

  • CD21 vs. CD11c gating for DN subdivision in COVID-19 panel: The Woodruff2020 24-marker panel uses CD21 vs. CD11c within the DN gate (IgD⁻CD27⁻CD38⁻CD24⁻) to resolve DN1 (CD21⁺CD11c⁻), DN2 (CD21⁻CD11c⁺), and DN3 (CD21⁻CD11c⁻). Three CD21⁻ populations (aN, DN2, DN3) were all expanded in ICU patients and correlated with ASC expansion — CD21 loss as a shared feature of EF pathway activation (see Woodruff2020 - EF B Cell Responses in COVID-19, Table 1).

  • CD21⁻ marks EF B cells in acute dengue: CD21⁻CD11c⁺ B cells within the IgD⁻CD27⁻ (DN) gate are significantly expanded during acute dengue — the first demonstration of CD21 loss on EF-phenotype B cells in dengue, consistent with the DN2 phenotype defined in SLE and COVID-19 (see Ansari2025 - Peripheral T Helper Subset Drives B Cell Response in Dengue, multi-color FCM, n=170 acute dengue).

  • CD21⁻CD27⁻ gating captures only 44.7% of transcriptomic atBCs: CITE-seq data from Sutton2021 demonstrate that the conventional CD21⁻CD27⁻ flow cytometry gate misses the majority of transcriptomically-defined alternative lineage B cells. CD21 loss is progressive along the alternative lineage pseudotime trajectory — early alternative lineage cells (MBC1) retain CD21, while late-stage atBC1 cells lose it — meaning CD21⁻ gating selectively captures the most activated end of the spectrum. This quantifies the limitation flagged qualitatively by Sanz2025 (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, n=4, CITE-seq).

  • CD21⁺ defines naive B cells in dengue gating: In the GarciaBates2013 panel, naive B cells are defined as CD27⁻CD21⁺ within CD19⁺CD10⁻ mature B cells. Naive B cells contracted to ~30% of B cells in secondary DFC (from ~50% in health), reflecting displacement by the CD21⁻ plasmablast wave. The atypical memory gate (CD27⁻CD21⁻) was also identified, corresponding to DN B cells (see GarciaBates2013 - Plasmablast Response and Dengue Severity, LSRII FCM, n=84 dengue).

  • CD21 loss is part of the review-level definition of human ABCs. Human atypical B cells “typically lack classical MBC markers such as CD27 and CD21” and often express T-bet with CD11c, CXCR3, FCRL4, and FCRL5; the murine counterpart is defined as CD11c⁺CD21^lo “age-associated” B cells (see Glaros2025 - Multilayered Identity of B Cell Memory, review, no original data). ⚠ Note this is a CD27/CD21-negative definition, not an IgD/CD27 one — it does not map one-to-one onto this wiki’s IgD⁻CD27⁻ DN gate, and papers using it will enumerate a partly different population.

  • Repeated antigen exposure progressively drives the CD21⁻ phenotype. Human T-bet⁺CD27⁺CD21⁻ B cells poised for plasma-cell differentiation have been described in antibody-mediated rejection of kidney transplants (see Glaros2025 - Multilayered Identity of B Cell Memory, review, citing Louis 2021 JCI Insight) — a CD27⁺ CD21⁻ population, reinforcing that CD21 loss and CD27 loss are separable axes that do not always co-occur.

  • CD21⁻ was part of the founding murine Hao definition of ABC, alongside loss of CD23, CD95, and CD43 on B220⁺CD19⁺ splenocytes (see Cancro2020 - Age-Associated B Cells, review — no original data; mouse).

  • A CD21⁻-anchored gate over-calls the ABC population by roughly 50%. Within the murine CD21⁻CD23⁻ splenic pool, only ~2/3 of cells are T-bet⁺, and only about half of those are CD11c⁺ — so CD21 loss alone captures at least three overlapping populations, not one (see Cancro2020 - Age-Associated B Cells, review — no original data; mouse, splenic).

  • CD21 loss also marks HIV “tissue-like memory” B cells and VAT “inflammatory B cells.” HIV tissue-like memory cells (Moir et al.) lack both CD21 and CD27; Frasca/Blomberg’s visceral-adipose-tissue T-bet⁺CD21⁻ “inflammatory B cells” / “late memory B cells” increase with age and correlate with ABC numbers (see Cancro2020 - Age-Associated B Cells, review — no original data; human, observational).

  • CD21 was argued into the seven-marker core specifically because its loss reports activation in every parental population — and that same generality is its weakness. The review “strongly recommend[s] the inclusion of CD21 in the core marker set as its expression enables the recognition of activated cells within all parental B cell populations”. It then lists the confounds: CD21 downregulation marks activated memory in normal vaccination responses, HIV, malaria and checkpoint-inhibitor-expanded memory — it is not disease-specific; the CD21ˡᵒ feature is shared by activated naive, DN2, atypical/tissue-based memory and T-bet⁺ B cells alike; early transitional (T1) cells are also CD21ˡᵒ and must be excluded using CD38, CD24 and/or CD10; and the RA/CVID studies calling CD27⁻CD21ˡᵒ cells anergic gated on CD21 without IgD, so they included CD27⁻ switched cells. The review’s own summary is blunt: “the limited use of either CD21, T-bet or CD11c expression is inadequate to identify ABC or other distinct human B cell populations” (see Sanz2019 - Consistent Classification of Human B Cell Populations, review — no original data). Directly relevant to DN2 Gating Strategy.

  • [2026-08-29] CD21 is the second axis of the Emory DN split — and the paper leaves one of its own quadrants unnamed. CD21 × CD11c resolves DN1 (CD21⁺CD11c⁻), DN2 (CD21⁻CD11c⁺) and DN3 (CD21⁻CD11c⁻); activated naive cells are CD21⁻CD11c⁺ within the naive gate. The CD21⁺CD11c⁺ quadrant is defined but not labelled, so this paper does not supply a DN4 and does not arbitrate the standing DN4 marker conflict (see Jenks2021 - B Cell Subset Composition in Cutaneous Lupus, n=207 + 46 HCD, cross-sectional). ⚠ Preanalytical caveat relevant to comparing CD21⁻ frequencies across studies: these are cryopreserved and thawed PBMC, not fresh or fixed whole blood.

  • CD21 loss is part of the same induced module and occurs from a uniform naive starting population. Naive B cells driven with the full cocktail lose CD21 as they acquire T-bet^hi^CD11c⁺FcRL5⁺CXCR5^neg^ (see Zumaquero2019 - IFN-gamma Programs T-bet-hi B Cells for ASC Differentiation, human, n=20 HD + n=40 SLE + in vitro reconstruction). As with CXCR5, the informative point is the starting material: the CD21⁻ phenotype used across the wiki to gate DN2 and aNAV is inducible by a defined cytokine set from ordinary naive cells, so a CD21⁻ gate reports recent cytokine exposure at least as much as it reports a stable cell identity.

Contradictions & Debates

  • ★ Is CD21^lo a lineage marker or an activation state? The wiki treats CD21^lo as part of the stable definition of the atypical/DN2 compartment. The consensus Perspective cautions that “low expression of CD21 and CXCR5 and increased CD11c could indicate recent B cell activation rather than a permanent state of expression/lack of expression” (see Eisenbarth2025 - A Roadmap for Defining Extrafollicular B Cell Responses, consensus Perspective, 12 authors, no primary data). Both readings are supported in the wiki: CD21^lo is durable enough to define reproducible populations across SLE, COVID-19, HIV, and malaria cohorts, yet CD21 is also acutely modulated by activation. The practical consequence is specific to acute-infection cohorts like dengue, where every subject is sampled days into a strong inflammatory response: a CD21^lo expansion at a single acute timepoint has two candidate explanations, and only serial sampling across fever days separates them — a design acute dengue permits and the chronic-disease cohorts that defined these gates did not use.

Activated Naive B Cell, CD19, CD24, CD11c, CXCR5, DN2 B Cell, DN3 B Cell, Memory B Cell, Extrafollicular Response, Conventional Flow Cytometry, Peripheral Helper T Cell, Tissue-Resident Memory B Cell, Early Memory B Cell, GC-Independent Response

Sources

Graph View

Backlinks