EFB Dengue Literature Review

Somatic Hypermutation

16 min read

Somatic Hypermutation

Overview

Somatic hypermutation (SHM) is the process by which activation-induced cytidine deaminase (AID) introduces point mutations at a rate ~10⁶-fold above background into the variable regions of rearranged immunoglobulin heavy and light chain genes. SHM is the primary genetic hallmark of antigen-experienced memory B cells and occurs predominantly within germinal centres, where it drives affinity maturation through iterative cycles of mutation and selection. However, SHM has also been demonstrated in extrafollicular settings (murine studies), which is mechanistically important for the origin of DN B cells.

Key Points from Literature

  • SHM is detected in IgD⁻CD27⁻ (DN) B cells, confirming antigen experience despite the absence of CD27 (see Wei2007 - DN Memory B Cells in SLE, VH3 family sequencing of sorted IgG⁺ cells).

  • Mutation rates: DN IgG⁺ cells carry ~3.2% nucleotide mutation rate in healthy donors and ~2.6% in SLE patients; CD27⁺ IgG⁺ memory cells carry ~5.4% (healthy) and ~5.1% (SLE). DN cells are consistently less mutated than their CD27⁺ counterparts (see Wei2007 - DN Memory B Cells in SLE, n≈28 per group — note: small sample, VH3 family only).

  • The lower SHM rate in DN vs CD27⁺ cells is the principal evidence cited for their GC-independent or abortive-GC origin; GC-derived cells undergo extended rounds of mutation-selection cycling, accumulating more mutations (see Wei2007 - DN Memory B Cells in SLE).

  • Replacement:silent (R:S) mutation distribution is consistent with antigen-driven selection in DN cells — CDR1 R:S ratios are elevated — suggesting these cells experienced some degree of antigen selection, even at a lower overall mutation rate (see Wei2007 - DN Memory B Cells in SLE).

  • Murine precedent for extrafollicular SHM: William2002 - Extrafollicular Somatic Hypermutation in Autoimmune Mice showed SHM can occur outside GCs at GC-comparable rates — cited as key support for EF origin of DN cells (cited in Wei2007 - DN Memory B Cells in SLE).

  • Quantitative SHM benchmarks for GC vs. EF-derived circulating ASCs in human peripheral blood: Deep NGS of sorted ASC populations provides the most precise available numbers. In SLE acute flares, average VH mutation rate in circulating plasmablasts is 4.98%; ~30–33% of ASC sequences have <3% VH mutation. In contrast, vaccination-elicited ASCs (which are predominantly memory-recall, GC-derived) average 7.33% VH mutation, with only ~10–12% sequences below 3%. These differences are statistically significant (p<0.05) and are consistent across CD138⁻ and CD138⁺ ASC subsets (see Tipton2015 - ASC Diversity and Origin in SLE, n=5 SLE, n=8 vaccinated controls, Illumina MiSeq NGS).

  • Activated naive (acN) cells carry intermediate SHM: acN cells show 2.37% average VH mutation vs. 0.95% for resting naive B cells — consistent with AID activity in extrafollicular precursors before ASC differentiation (see Tipton2015 - ASC Diversity and Origin in SLE).

  • Germline SHM rate does not constrain autoreactivity: ASC clone 652-F6 had zero VH and VL mutations yet was strongly autoreactive to ANA, dsDNA, chromatin, and ribosomal P. This shows that germline-encoded BCR sequences themselves — without any SHM — can confer full lupus-specific autoreactivity. The association between SHM accumulation and autoantigen reactivity does not hold universally (see Tipton2015 - ASC Diversity and Origin in SLE).

  • Epigenetic data confirms EF vs. GC origin distinction for DN2 vs. SM: DNA methylation phylogenetics place both DN2 and SM closest to ASCs (progressive hypomethylation at the PRDM1 locus and globally), yet their chromatin accessibility profiles diverge: SM is enriched for NF-κB/EBF motifs (GC-associated TFs), DN2 for T-BET/AP-1/EGR motifs. This epigenetic bifurcation independently supports the BCR sequencing-based conclusion that DN2 cells accumulate lower SHM because they bypass GC reactions, not because they are selected against in GCs (see Scharer2019 - Epigenetic Programming in SLE B Cells, RRBS + ATAC-seq, n=9 SLE + 12 HC).

  • DN2 mutation rate comparable to PC but lower than SWM: BCR sequencing of DN2 cells shows IgG mutation rates similar to co-circulating plasmablasts (PCs) but significantly lower than SWM cells. This argues against DN2 derivation from GC-experienced memory cells and is consistent with the EF pathway model where DN2 cells accumulate limited SHM during EF activation before differentiating into PCs (see Jenks2018 - DN2 B Cells and EF Pathway in SLE, BCR sequencing of sorted populations).

  • LANDMARK: First direct demonstration that SHM occurs outside germinal centres at GC-comparable rates. In MRL/lpr lupus-prone mice, RF B cells at the splenic T zone–red pulp border (not in GCs) accumulated somatic mutations at ~0.3 mutations per Vκ gene per generation — a rate comparable to GC estimates by the same mathematical modelling method. Genealogical trees from microdissected clusters showed shared trunk mutations with unique branch mutations, the definitive signature of ongoing in situ diversification. The weighted average across 8 mice was 4.3 mutations per unique sequence (3.2 trunk + 0.7 unique). Replacement:silent mutation distribution was consistent with antigen-driven selection. Id⁻ GCs in the same spleens yielded no mutated Vκ8 sequences, and some mice had active EF mutation without any GCs at all. This is the foundational murine evidence for GC-independent SHM, cited as the key precedent by Wei2007, Sanz2025, and throughout the EF literature (see William2002 - Extrafollicular Somatic Hypermutation in Autoimmune Mice, in vivo murine model, 8 mice, 305 sequences, 125 unique).

  • SHM and isotype switch are not restricted to GCs: Sanz (2025) emphasises that both SHM and CSR occur in naïve-derived EF effector B cell responses independently of GC reactions. The assignment of “memory” status based on SHM or isotype switch is therefore faulty — it ignores the EF pathway (see Sanz2025 - Human Atypical B Cells Overview, review, citing MacLennan et al. 2003, William et al. 2002, Roco et al. 2019).

  • DN2 cells with intermediate SHM relative to SWM — cross-disease evidence: SLE DN2 display intermediate levels of SHM relative to CD27⁺ switched memory cells across studies. Strikingly lower SHM in circulating plasmablasts relative to pre-existing memory cells is consistent with a large contribution from newly activated naive B cells (see Sanz2025 - Human Atypical B Cells Overview, review).

  • T-bet⁺ B cells in partial RAG deficiency show extrafollicular maturation with autoreactivity: In patients with partial RAG deficiency, T-bet⁺ B cells with substantial autoreactivity undergo strong extrafollicular maturation — consistent with GC-independent SHM and selection (see Sanz2025 - Human Atypical B Cells Overview, review citing Csomos et al. 2022).

  • >50% germline VH clonotypes in COVID-19 ASCs (n=1 patient — not generalizable): Single-cell V(D)J of ASCs from a single critically ill COVID-19 patient showed >50% of clonotypes had exclusively germline (unmutated) VH genes, especially in the IgG1 and IgA1 compartments. This germline fraction exceeds even the ~30–33% low-SHM (<3% VH mutation) benchmark from SLE (Tipton2015). The co-existence of germline clonotypes with some highly mutated, branching lineages suggests concurrent EF and pre-existing memory contributions to the same ASC pool. However, this derives from a single patient’s repertoire — generalizability to the broader CoV-A cluster or to other EF-dominant infections is unestablished (see Woodruff2020 - EF B Cell Responses in COVID-19, 10x Chromium scV(D)J, 2,017 clonotypes, n=1 patient).

  • Germline VH4-34 clones with intact autoreactive FR1 patch: 85% of VH4-34-expressing ASC clonotypes retained the germline FR1 hydrophobic patch — the autoreactivity-mediating motif that is normally eliminated by SHM-driven “clonal redemption” in healthy GC reactions. Expanding VH4-34 lineages had SHM but retained the FR1 patch, indicating that SHM was not acting as a tolerance mechanism in these clones (see Woodruff2020 - EF B Cell Responses in COVID-19).

  • Limited further affinity maturation in flavivirus immunity: Singh2026 cites Wong et al. (2020) showing that affinity-restricted memory B cells dominate recall responses to heterologous flaviviruses, with limited further SHM accumulation. This suggests diversification through pre-existing cross-reactive clones rather than iterative GC refinement — compatible with the low SHM rates found in many flavivirus neutralizing mAbs (Godoy-Lozano et al. 2016) and consistent with an EF or early-GC origin for dengue-specific MBCs (see Singh2026 - DENV-Specific Memory B Cell Subsets, discussion citing Wong et al. 2020, Immunity).

  • Wrammert2012 does not address SHM directly — a critical gap: The 5-color flow panel and ELISpot design in this foundational dengue PB study provide magnitude and specificity data but no repertoire information. The SHM status of dengue acute-phase plasmablasts remains unknown. Whether the massive PB wave is dominated by germline (EF-derived) or somatically mutated (GC/memory-derived) clonotypes is the key unresolved question that would connect dengue to the Tipton2015 <3% VH mutation EF benchmark and the Woodruff2020 >50% germline COVID-19 ASC signature (see Wrammert2012 - Plasmablast Responses in Acute Dengue, n=46 cohort — no BCR sequencing performed).

  • FIRST DENGUE BCR DATA — Convergent CDR3-bearing B cells carry 4.4–6.9% V gene mutation: In the first BCR repertoire analysis from dengue patients, VH sequences associated with convergent CDR3s (shared across multiple individuals in acute dengue) had median somatic mutation rates of 4.4–6.9% (7–11 nucleotide changes), with very few sequences below the naive threshold (<1.9% mutation, ≤3 changes). This was not attributable to V gene segment bias — B cells using the same V genes but encoding other CDR3s had significantly more naive-range sequences (p≤0.0001). The 4.4–6.9% range is intermediate: above the EF benchmark (<3% per Tipton2015) but below full GC maturation (~7.3% per Tipton2015 vaccination data). This is consistent with memory B cell origin of the convergent clones rather than de novo naive recruitment, but does not definitively distinguish EF-matured from GC-matured memory (see Parameswaran2013 - Convergent Antibody Signatures in Dengue, 454 pyrosequencing of gDNA from unsorted PBMCs, n=60 dengue patients — caveat: not sorted plasmablasts, so mutation rates reflect all B cells carrying these CDR3s).

  • Higher V gene mutation in secondary than primary dengue (convergent clones): The convergent CDR3s were significantly more prevalent in secondary than primary acute dengue (p≤0.05), consistent with recall of affinity-matured memory B cell populations. Combined with the intermediate mutation levels, this supports a model where repeated dengue exposure selects and expands pre-existing memory clones rather than generating new responses from naive cells — aligning with the Ansari2025 finding that Tph preferentially drive memory (not naive) B cell differentiation (see Parameswaran2013 - Convergent Antibody Signatures in Dengue).

  • Comparable SHM between clonally distinct dengue PBs and MBCs — similar affinity maturation despite separate lineages: Appanna2016 demonstrated that PB-derived and MBC-derived antibodies from the same individuals had comparable VH nucleotide mutation frequencies and N-addition counts (not significantly different by ANOVA), despite being clonally unrelated. This argues that the PB precursors are not less GC-experienced than the MBCs that emerge at convalescence — both have undergone similar levels of SHM, consistent with derivation from comparably matured memory B cell pools with different specificities rather than from different stages of affinity maturation (see Appanna2016 - Plasmablasts as Subset of Memory B Cell Pool, IMGT analysis, n=12 dengue, Sanger + 454).

  • FIRST DIRECT SHM MEASUREMENT IN DENGUE IgG PLASMABLASTS — paradoxically low SHM in acute phase: High-throughput VH cDNA sequencing (454-Roche, IgG-specific 5ʹ RACE-PCR) of 19 acute dengue patients showed globally lower pM-VH% during acute infection vs. 6-month post-convalescence (p<0.001). Monte Carlo simulation confirmed that acute-phase reads derive predominantly from ASCs (probability of sampling memory B cell lineage = 0.015 when plasmablasts are 56% of IgG⁺ cells). SHM was significantly lower in DWS+ than DWS− (p<0.001), and paradoxically lower in secondary than primary infections (p<0.001). This contrasts sharply with influenza vaccination, where SHM increases at day 7. The data establish that the dengue acute-phase IgG plasmablast repertoire is dominated by poorly mutated clones — quantitatively below the 4.4–6.9% range reported for unsorted PBMCs in Parameswaran2013 — and are compatible with a substantial contribution from GC-independent (extrafollicular) B cell differentiation (see GodoyLozano2016 - Lower IgG SHM Rates in Acute Dengue, n=19 acute + n=11 paired post-convalescent + n=10 controls, 454 pyrosequencing, 385,206 lineages).

  • IGHV segment-specific SHM reduction: IGHV1-2, IGHV1-18, and IGHV1-69 all showed significantly lower SHM during acute dengue vs. post-convalescence (p<0.001 for each). IGHV1-2 SHM was particularly reduced in DWS+ vs. DWS− (p<0.001) and in secondary vs. primary infection. These segments were simultaneously overrepresented in the acute repertoire — expanded and hypomutated — consistent with germline-encoded “innate-like” antiviral recognition rather than affinity-matured recall (see GodoyLozano2016 - Lower IgG SHM Rates in Acute Dengue).

  • Convergent CDRH3s carry lower SHM than non-convergent clones: De novo identified convergent CDRH3s (shared in ≥3 individuals during acute dengue, 8.3-fold enrichment vs. post-convalescence) had significantly lower SHM than non-convergent clonotypes (p<0.001), with a predominant 10-residue CDRH3 signature (ARQxGNWFDx). Two specific CDRH3s were shared by 52% of individuals (10/19). This extends the Parameswaran2013 convergent CDR3 finding by showing these shared clones are specifically hypomutated — consistent with germline-coded BCR-driven selection rather than GC affinity maturation (see GodoyLozano2016 - Lower IgG SHM Rates in Acute Dengue).

  • All non-naive alternative lineage clusters carry significant SHM — post-GC origin evidence: BCR analysis from Smart-seq2 scRNA-seq showed that all memory and atBC clusters (atBC1, atBC2, atBC3, MBC1) carry somatic hypermutation, consistent with post-GC origin rather than purely extrafollicular derivation. Combined with the absence of PC maintenance genes, this suggests these cells exited GCs and adopted the T-bet⁺ programme without committing to terminal differentiation. The SHM in alternative lineage cells is consistent with Wei2007’s original observation of lower SHM in DN vs CD27⁺ cells, but reinterprets it: the SHM is GC-acquired, and the T-bet⁺ programme is adopted after GC exit (see Sutton2021 - Alternative Lineage B Cells in Vaccination and Infection, Smart-seq2, n=11, 163 cells — limited throughput).

  • HIGH SHM IN SORTED SECONDARY DENGUE PLASMABLASTS — contrasts with GodoyLozano2016 bulk data: Single-cell VH sequencing of FACS-sorted plasmablasts from 4 secondary DHF patients (all DENV2) showed per-patient average VH mutations of 14.5–21.7 (overall mean 18.1, range 5–39 per sequence). This is significantly higher than IgG⁺ GC B cells (p<0.005) and comparable to influenza recall responses. CDR R:S ratios >2.9 in all patients confirmed antigenic selection. These SHM levels (~6.5% nucleotide mutation) are above the Tipton2015 vaccination benchmark (~7.3%) and far above the GodoyLozano2016 acute dengue bulk IgG data. The discrepancy is methodological: Priyamvada2016 sorted plasmablasts from exclusively secondary DHF (all memory-derived), while GodoyLozano2016 captured bulk IgG cDNA from a mix of primary and secondary infections. The Priyamvada data demonstrate that the memory-recall component of secondary dengue plasmablasts is genuinely highly mutated — the low global SHM from GodoyLozano2016 must therefore reflect additional low-SHM clones (possibly de novo EF-derived) that are diluted out by the cell-sorting approach (see Priyamvada2016 - Cross-Reactive Memory Plasmablasts in Secondary Dengue, n=4 secondary DHF, single-cell VH sequencing of sorted PBs, 53 mAbs).

  • ABCs carry SHM throughout — including the unswitched IgD⁺ fraction. The age-associated B cell population expresses a diverse Ig repertoire marked by somatic hypermutation and antigen-driven activation, and even its IgD⁺ members carry SHM and are clonally related to IgD⁻ cells (i.e., are antigen-experienced, not naive). This is one basis for the proposed GC-experienced origin of at least a subset of ABCs — though homeostatic and EF routes are not excluded — and it cautions that SHM presence cannot by itself assign GC vs EF origin within the atypical cluster (see Lamprinou2026 - ABCs and DN B Cells, opinion, citing Cancro 2020 / Maul 2021; Age-Associated B Cell).

Contradictions & Debates

  • The lower SHM in DN cells is consistent with EF origin but not conclusive — it could also reflect GC selection against autoreactive high-SHM clones, releasing cells with fewer mutations into the periphery. The Jenks2018 DN1/DN2 resolution helps: DN2 cells specifically (not DN1) show the low-SHM pattern, and their TLR7-dependent generation is independently confirmed by in vitro differentiation, making the EF origin model substantially more parsimonious for DN2 cells.
  • Three-way SHM tension in dengue: high in sorted secondary PBs (Priyamvada2016) vs. globally low in bulk IgG (GodoyLozano2016) vs. PB-comparable-to-MBC (Appanna2016). Priyamvada2016 reports mean 18.1 VH mutations (~6.5% nucleotide mutation) in sorted plasmablasts from 4 secondary DHF patients — significantly higher than GC IgG⁺ B cells and comparable to influenza recall. GodoyLozano2016 reports globally lower SHM during acute dengue vs. post-convalescence from unsorted IgG cDNA. Appanna2016 reports comparable mutation between sorted PBs and DENV-specific MBCs. Reconciliation: (1) Priyamvada2016 captures the memory-derived fraction of PBs in secondary infection (high SHM from prior GC maturation); (2) GodoyLozano2016 captures the full IgG pool including both memory-derived and de novo EF-derived PBs — the latter with low SHM pulling the average down; (3) Appanna2016 compares DENV-specific subsets that are similarly memory-derived regardless of compartment. The three findings are compatible with a dual-pathway model (working hypothesis — neither study was designed to test it): memory recall produces high-SHM PBs while concurrent de novo EF differentiation produces low-SHM PBs, and their ratio depends on infection history and the methodology used to capture them.
  • Previously framed as two-way: GodoyLozano2016 reports globally lower SHM during acute dengue vs. post-convalescence (p<0.001) from unsorted IgG cDNA. Appanna2016 reports comparable VH mutation frequencies between FACS-sorted plasmablasts and DENV-specific memory B cells (no significant difference by ANOVA). These findings are in apparent tension. Likely reconciliation: GodoyLozano2016 captures the full IgG⁺ B cell pool (dominated by plasmablasts in acute phase, pulling the average down), while Appanna2016 compares two sorted populations that may represent similarly matured cells selected for DENV binding. The Appanna2016 result does not contradict low global SHM — it shows that the subset of PBs that are DENV-specific have comparable mutation to DENV-specific MBCs, consistent with both being memory-derived. The GodoyLozano2016 result captures additional non-DENV-specific, germline-coded PBs that dilute the global average.

Double-Negative B Cell, DN2 B Cell, Age-Associated B Cell, Activated Naive B Cell, Memory B Cell, Germinal Center, Extrafollicular Response, BCR Sequencing, Class Switch Recombination, Immunohistochemistry

Sources

Graph View

Backlinks