EFB Dengue Literature Review

CD27

16 min read

CD27

Overview

CD27 is a member of the TNF receptor superfamily expressed on the surface of B cells and T cells. On B cells, surface CD27 expression has long been used as a canonical marker of memory B cell identity — CD27⁺ B cells consistently carry somatic hypermutation, and CD27 expression is acquired during germinal centre reactions through cognate B–T cell interactions mediated by CD40–CD154. CD27 interacts with its ligand CD70 on activated helper T cells, facilitating memory B cell differentiation into plasma cells.

The discovery of IgD⁻CD27⁻ (double-negative) memory B cells — which carry somatic hypermutation and memory functional properties despite lacking CD27 — has qualified its status as a “universal” memory marker.

Key Points from Literature

  • CD27 expression was the dominant criterion for identifying human memory B cells prior to the characterisation of DN B cells (see Wei2007 - DN Memory B Cells in SLE, citing Klein et al. 1998).

  • CD27 is acquired by B cells following a GC reaction initiated by CD40–CD154-mediated B–T cell interactions; its absence in DN B cells is used as evidence for their GC-independent (extrafollicular) origin (see Wei2007 - DN Memory B Cells in SLE). ⚠ This inference is not licensed as exclusive — see the consensus bullet at the end of this section.

  • CD27–CD70 interactions regulate B cell activation by T cells and enhance plasma cell differentiation; impaired CD27 signalling in DN cells may limit their ability to receive sustained T cell help (see Wei2007 - DN Memory B Cells in SLE, citing Kobata et al. 1995, Jacquot et al. 1997).

  • CD27 expression is upregulated by the majority of proliferating DN cells after CpG DNA stimulation in vitro — suggesting CD27 negativity is not lineage-fixed and can be acquired upon TLR9-driven activation (see Wei2007 - DN Memory B Cells in SLE, in vitro proliferation assay).

  • In SLE, both CD27⁺ memory cells and DN cells tend toward the CD38^dull (early Bm5) phenotype, at the expense of the CD38⁻ (Bm5) fraction seen in healthy donors (see Wei2007 - DN Memory B Cells in SLE).

  • CD27^hi marks the ASC gate: Plasmablasts in peripheral blood are identified as CD19⁺IgD⁻CD27^hiCD38^hi. The CD27^hi criterion (substantially above conventional memory B cell CD27 levels) is a standard component of the Tipton2015/Sanz lab ASC gate, distinguishing ASCs from IgD⁻CD27^lo DN memory cells and from IgD⁺CD27⁺ unswitched memory cells (see Tipton2015 - ASC Diversity and Origin in SLE).

  • DN1/DN2 subdivision resolves the CD27⁻ memory puzzle: The DN1/DN2 subdivision within IgD⁻CD27⁻ cells reveals two distinct reasons for CD27 absence. DN1 cells (CXCR5⁺, CD21⁺, transcriptionally SWM-like with only 22 DEGs) likely represent early switched memory precursors that have not yet acquired CD27 via CD40–CD154 interactions — they are GC-associated. DN2 cells (CXCR5⁻, CD21⁻, CD11c⁺, T-bet⁺) are extrafollicular effector cells whose CD27 absence reflects genuine GC bypass via TLR7-dependent differentiation. CD40L stimulation inhibits DN2 generation but not DN1, consistent with DN1 acquiring CD27 through the conventional CD40 pathway (see Jenks2018 - DN2 B Cells and EF Pathway in SLE, RNA-seq + in vitro differentiation).

  • CD27 can be downregulated upon stimulation: CD27 expression is not lineage-fixed — it can be downregulated by CD70 on activated T cells, TLR ligands, and cytokines. Some CD27⁻ B cells may therefore represent cells that originally expressed CD27 and lost it upon stimulation, rather than cells that never entered GCs (see Sanz2025 - Human Atypical B Cells Overview, review citing Kang et al. 2024).

  • CD27 as sole memory marker is obsolete: Sanz (2025) argues that the prevailing model equating CD27⁺ with memory is now untenable. A substantial memory fraction normally resides in CD27⁻ populations, either because CD27 was never acquired or because it was downregulated. The term “atypical memory” — defined by CD27 absence — is therefore misleading (see Sanz2025 - Human Atypical B Cells Overview, invited review).

  • CD27⁻ status does not cleanly separate alternative from classical lineage: CITE-seq data show that CD27 protein expression overlaps substantially between the transcriptomically-defined alternative and classical lineages. Many alternative lineage cells (especially the quiescent MBC1 cluster) are not CD27⁻ by protein, while some classical lineage cells are. This reinforces that CD27 absence is a consequence of activation state rather than a reliable lineage marker (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, n=4, CITE-seq).

  • CD27 is the marker that separates the ABC superset from the DN compartment. Some ABCs highly express CD27 (Rubtsov 2011), whereas DN cells are CD27⁻ by definition (Jenks 2018; Colonna-Romano 2009). Because the ABC population comprises CD27⁺ + IgD⁺ + predominantly IgD⁻CD27⁻ cells, only its IgD⁻CD27⁻ fraction maps onto DN — the CD27⁺ ABC members are excluded from any DN classification. CD27 absence is thus necessary but not sufficient for ABC identity (see Lamprinou2026 - ABCs and DN B Cells, opinion, citing Rubtsov 2011 / Tangye 2023 / Jenks 2018).

  • ★ Some ABCs express classical memory markers including CD27 — CD27⁻ is not a requirement of the atypical phenotype. While human ABCs “typically lack” CD27 and CD21, the review notes that some ABCs do express classical MBC markers, carry somatically hypermutated BCRs, and respond to antigen restimulation (see Glaros2025 - Multilayered Identity of B Cell Memory, review, no original data). A concrete instance: T-bet⁺CD27⁺CD21⁻ B cells poised for PC differentiation in antibody-mediated kidney-transplant rejection (citing Louis 2021 JCI Insight). This corroborates the existing wiki position that CD27 is obsolete as a sole memory marker, and it means a CD27⁻-gated ABC count is a lower bound.

  • ⚠ Do not confuse “CD27⁻ DN” with the CD80/PD-L2 “DN MBC” of the murine memory literature. That literature’s DN means CD80⁻PD-L2⁻ and is a proxy for GC-independent origin — nothing to do with CD27 or IgD (see Glaros2025 - Multilayered Identity of B Cell Memory, review). See the false-friend note on Atypical B Cell and Early Memory B Cell.

  • CD27 marks the human splenic marginal-zone memory compartment. Unlike rodents — where marginal-zone B cells are thought to be largely naive innate-like cells — a large fraction of human MZ B cells are considered memory cells, judged by CD27 expression and by the fact that most carry somatically hypermutated BCRs (see Glaros2025 - Multilayered Identity of B Cell Memory, review). See Tissue-Resident Memory B Cell.

  • ★ A human exception to the murine two-signal rule — some CD27⁻ blood B cells may already be ABC-poised in vivo. In human peripheral blood, the same TLR-then-cytokine relationships largely held, except that some activated human CD27⁻ B cells expressed T-bet directly induced by IFN-γ, without any concomitant TLR ligand. Cancro’s preferred explanation: some cells within the human CD27⁻ blood pool have already received ABC-poising signals in vivo, before the experimental stimulation — implying part of the “unstimulated” CD27⁻ pool is pre-primed rather than naive with respect to the ABC/atypical programme (see Cancro2020 - Age-Associated B Cells, review — no original data; human, in vitro).

  • HIV “tissue-like memory” B cells lack both CD21 and CD27. These cells (Moir et al.) are hyporesponsive to BCR cross-linking yet proliferate robustly to TLR9 — paralleling murine ABC behaviour — with T-bet expression demonstrated in subsequent work (see Cancro2020 - Age-Associated B Cells, review — no original data; human, observational).

  • CD27 is a TNF receptor superfamily member whose ligand is CD70 on activated helper T cells — the interaction facilitates memory B cell differentiation into plasma cells. This is the molecular reason CD27 acquisition is read as evidence of cognate T cell help (see Beckers2023 - Origins and Functions of DN B Cells, review, citing Agematsu 2000).

  • CD27 loss via immunosenescence is the leading proposed mechanism for the sM→DN route. Chronic antigen stimulation — in normal aging, chronic viral infection, and some autoimmune diseases — is proposed to downregulate CD27 on switched memory cells, generating the population previously described as “exhausted memory B cells.” No transcriptional or epigenetic mechanism for the downregulation is offered; “immunosenescence” is a label for the correlation, not an explanation (see Beckers2023 - Origins and Functions of DN B Cells, review; see Switched Memory B Cell).

  • CD27 loss is not the majority route into the DN compartment. If sM→DN conversion were the dominant origin, DN and sM repertoires should overlap heavily. AIRR sequencing finds only 0.2–2.2% clonal overlap, and the majority of DN and sM cells are clonally distinct with different IgV(D)J family and gene usage. CD27 downregulation is therefore a real but minority contributor — which is what makes the CD27⁻ gate broadly meaningful despite CD27’s instability (see Beckers2023 - Origins and Functions of DN B Cells, review, own AIRR data).

  • IgD loss and CD27 loss are set by different axes — do not read one from the other. It is IgD, not CD27, that is downregulated as a consequence of Ig isotype switching; on that basis the review reads the substantially IgG⁺ composition of the DN compartment as evidence that DN cells underwent isotype switching similar to switched memory cells. Within an IgD⁻CD27⁻ gate, therefore, the IgD⁻ half reports switching while the CD27⁻ half reports something else entirely — GC bypass, or post-hoc CD27 loss (see Beckers2023 - Origins and Functions of DN B Cells, review, own data — Fraussen 2019 — + citing Huang 2020 for the IgD/switching relationship).

  • [2026-08-27] CD27 as a tissue stain — DN identity assigned in situ in archival FFPE. Multiplex immunofluorescence (Opal, clone ab131254) resolved IgD⁻CD27⁻ cells cell-by-cell in COVID-19 lung and lymph node and in IgG4-RD salivary gland, with IgD and CD27 carried in a single combined channel in the four-colour subset panels (CD19 / IgD-CD27 / CXCR5 / CD11c / DAPI) — a channel-economy trick that works precisely because DN identity requires only that both be negative. In IgG4-RD tissue, IgD⁻CD27⁻ DN cells outnumbered IgD⁻CD27⁺ cells in every individual examined (n=7) (see Allard-Chamard2023 - DN3 B Cells Infiltrate Inflamed Tissues, multi-colour IF). See Multi-color Immunofluorescence.

  • [2026-08-27] Splitting the CD27⁺ compartment by CXCR5 reveals a disease shift that total CD27⁺ frequency hides. In IgG4-RD blood (n=38), total switched memory (IgD⁻CD27⁺) was unchanged, but within it the CXCR5⁺ “classic/resting” fraction fell and the CXCR5⁻ fraction rose (both p<0.01). A CD27-only readout would have reported no change (see Allard-Chamard2023 - DN3 B Cells Infiltrate Inflamed Tissues, n=38).

  • ★ [2026-08-28] The field’s consensus position: CD27 absence is not an exclusive marker of an EF response. The twelve-author Perspective states flatly that “CD11c and T-BET expression or lack of CD27 are not exclusive markers of EF B cell responses”, and extends its activation-vs-lineage caution — that low CD21/CXCR5 and high CD11c “could indicate recent B cell activation rather than a permanent state of expression/lack of expression” — explicitly to CD71 as well. Its worked human counter-example: post-influenza-vaccination CD11c⁺T-bet⁺FcRL5⁺ cells at ≤1 week are CD27⁺CD21^lo with high SHM — recent GC emigrants, not EF products; others are CD27⁺ reactivated memory. This consolidates the Sanz2025, Sutton2021 and Glaros2025 bullets above into a field-level position, and it qualifies the second bullet of this section: CD27⁻ is a phenotype consistent with GC-independent origin, not evidence of it. Note the direction of the constraint — what is withdrawn is the location claim and the exclusivity of the marker, not GC-independence itself (see Eisenbarth2025 - A Roadmap for Defining Extrafollicular B Cell Responses, consensus Perspective, twelve authors — no original data).

  • CD27 is core marker (4), and the review’s central argument for the IgD/CD27 scheme is that the Bm1–Bm5 alternative cannot see the DN compartment at all. Applied to blood, Bm1–Bm5 “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”. Within the ASC compartment the review also splits CD27 quantitatively: mature ASC express high CD27, while circulating pre-plasmablasts are CD27ˡᵒ (IgD⁻CD38⁺/⁺⁺CD24⁻, BLIMP-1⁺ Pax5⁺) — a CD27ʰⁱ-anchored ASC gate excludes them (see Sanz2019 - Consistent Classification of Human B Cell Populations, review — no original data). See Bm Classification, Plasmablast.

  • [2026-08-29] The IgD × CD27 quadrant scheme, applied at cohort scale with CD27 intensity carrying a second job. CD19⁺CD3⁻ cells were split into IgD⁺CD27⁻ (naive+transitional), IgD⁺CD27⁺ (unswitched memory), IgD⁻CD27⁺ (switched memory + plasmablasts) and IgD⁻CD27⁻ (DN); plasmablasts were then resolved within the IgD⁻CD27⁺ gate as CD27⁺⁺CD38⁺⁺, i.e. by CD27 intensity rather than positivity. Switched memory was derived by subtraction (see Jenks2021 - B Cell Subset Composition in Cutaneous Lupus, n=207 + 46 HCD, cross-sectional). Practical note: this makes CD27 staining resolution — not just its cutoff — load-bearing for two populations at once.

  • Definitional use in this study’s gating, with the ASC exception stated. The double-negative compartment is B_DN_ = IgD⁻CD27⁻, subdivided by CD11c and CXCR5 into DN2 and DN1; antibody-secreting cells are gated separately as CD38^hi^CD27⁺ (see Zumaquero2019 - IFN-gamma Programs T-bet-hi B Cells for ASC Differentiation, human, n=20 HD + n=40 SLE + in vitro reconstruction). CD27 therefore moves in opposite directions along this pathway — absent on the DN2 pre-ASC, then re-acquired on the ASC it differentiates into — so a CD27⁻ gate captures the precursor and excludes its own product.

  • CD27⁺⁺CD38⁺⁺ as the readout of a de novo differentiation assay — a new assay context for an existing gate. Kwissa2014 took the human plasmablast gate over unchanged from Wrammert2012 - Plasmablast Responses in Acute Dengue (“the CD27⁺⁺CD38⁺⁺ plasmablast B cells were defined as described before”) and applied it as the output measure of a 6-day coculture of resting allogeneic CD19⁺ B cells with DENV-infected monocytes: CD27⁺⁺CD38⁺⁺ cells reached 22.8% of total B cells with DENV-infected monocytes, versus 5.21% with monocyte-derived DCs and 3.83% with B cells alone (see Kwissa2014 - Monocytes Drive Plasmablast Differentiation in Dengue, in vitro, 4 independent experiments with 4 healthy blood donors). The additive point for this page is the assay context — CD27 intensity used to score plasmablast generation from a resting starting population, rather than to classify an ex vivo compartment. The paper reports no new evidence on CD27 biology and made no CD27 measurement on patient B cells beyond the borrowed gate.

  • ★ The CD27 × IgD quadrant analysis that named the DN population in SLE — and the wiki’s earliest source for it. Peripheral B cells were resolved into four quadrants by CD27 and IgD: naive (IgD⁺CD27⁻), memory (IgD⁻CD27⁺), double-negative (IgD⁻CD27⁻) and double-positive/preswitch (IgD⁺CD27⁺). Against healthy controls, SLE patients showed naive lymphopenia (35 ± 17% vs 68 ± 6%, P=0.0008) and DN expansion (14.4 ± 7.9% vs 3.9 ± 1.9%, P=0.01), with DN frequency correlating with VH4.34 autoantibody titre (R²=0.8, P<0.05); both normalised after effective rituximab depletion and reconstitution (see Anolik2004 - Rituximab and B Cell Abnormalities in SLE, n=15-17 SLE + 7 healthy controls, phase I/II open-label trial, 4-colour flow). CD27 also marks the opposite end of the pathway here — plasmablasts were identified as CD27^high^ before CD38 gating was adopted, and the B cells surviving effective depletion were predominantly CD27⁺IgD⁻ switched memory. ⚠ Note the vintage: this is 2004, before CD11c/T-bet, so the DN gate is undivided — DN1 and DN2 are not separated.

Contradictions & Debates

  • CD27 is not absent from all non-memory B cells; it can be transiently upregulated on activated naive B cells and on pre-plasmablasts, adding ambiguity to the memory gate in activated samples.

  • The use of CD27 alone to enumerate memory B cells in dengue cohort studies will miss DN/atypical B cell expansions — a systematic limitation to bear in mind when interpreting older literature.

  • CD27 modulation means that the same cell can appear CD27⁺ or CD27⁻ depending on its activation state and signalling environment — a fundamental challenge for phenotype-based classification (see Sanz2025 - Human Atypical B Cells Overview).

  • CD27^high as original dengue plasmablast criterion: Wrammert2012 used CD27^high combined with CD38^high (within CD19⁺CD3⁻CD20⁻/low) as the core plasmablast gate — the first application of this standard ASC gate to acute dengue. The use of CD27^high (not just CD27⁺) distinguishes plasmablasts from conventional CD27⁺ memory B cells, consistent with the Tipton2015/Sanz lab gate (see Wrammert2012 - Plasmablast Responses in Acute Dengue, 5-color conventional FCM).

  • CD27⁺CD21⁻ gate for plasmablast/activated memory in dengue: GarciaBates2013 used CD27⁺CD21⁻ as the first-tier gate for non-naive memory/effector B cells, then subdivided by CD20/CD38 into plasmablasts (CD20⁻CD38⁺) and activated memory (CD20⁺CD38⁻/lo). This confirms that CD27 upregulation is a consistent feature of dengue plasmablasts across cohorts (see GarciaBates2013 - Plasmablast Response and Dengue Severity, LSRII FCM, n=84 dengue).

  • CD27 as PB/MBC discriminator in dengue clonal analysis: Appanna2016 used CD27^hi (within CD19⁺CD20⁻CD38^hi) for PB sorting and CD27⁺ (within CD19⁺CD20⁺) for MBC sorting. The CD27⁺ MBC gate, which by definition excludes CD27⁻ DN memory B cells, may undercount the total DENV-specific memory compartment. This is acknowledged as a limitation — Singh2026 shows that CD27⁻CD21⁻ (atypical/DN) MBCs are a significant component of DENV-specific memory (see Appanna2016 - Plasmablasts as Subset of Memory B Cell Pool, FACSAria, n=12 dengue).

Double-Negative B Cell, DN2 B Cell, Switched Memory B Cell, Age-Associated B Cell, Memory B Cell, IgD, CD38, Germinal Center, Extrafollicular Response, Plasmablast, Early Memory B Cell, Tissue-Resident Memory B Cell, Atypical B Cell

Sources

Graph View

Backlinks