Kaneko2020 - GC Loss and TFH Block in COVID-19

Full citation: Kaneko N, Kuo HH, Boucau J, Farmer JR, Allard-Chamard H, Mahajan VS, et al. Loss of Bcl-6-Expressing T Follicular Helper Cells and Germinal Centers in COVID-19. Cell. 2020;183(1):143-157.e13. doi:10.1016/j.cell.2020.08.025

Raw file: [[raw/Kaneko2020.pdf]]

Summary

This study examined post-mortem thoracic lymph nodes and spleens from acutely ill COVID-19 patients to understand why humoral immunity to SARS-CoV-2 is often short-lived and of limited quality. Using multi-color immunofluorescence with multispectral imaging and quantitative tissue cytometry, the authors found a striking absence of germinal centers in both early (<10 days from symptom onset) and late (15–36 days) disease, with marked reduction in Bcl-6⁺ GC B cells but preservation of AID⁺ B cells. The key mechanistic finding was a specific block in Bcl-6⁺ GC-type T follicular helper (TFH) cell differentiation, despite relative preservation of CD4⁺CXCR5⁺ pre-GC TFH cells. This was accompanied by TH1 skewing (T-bet⁺ CD4⁺ cell expansion), aberrant TNF-α accumulation in both follicular and extra-follicular zones, and near-absence of T follicular regulatory cells (no Bcl-6⁺/FoxP3⁺ overlap).

Parallel peripheral blood studies in a separate cohort revealed loss of transitional and follicular B cells in severe disease (correlating with CRP), and expansion of disease-related activated B cell populations — including activated naive, IgD⁻CD27⁻ double-negative B cells (DN2 and DN3 subsets), switched memory, and plasmablasts — in severely ill patients. Using dual-fluorophore RBD probes, the authors confirmed that these disease-related B cell populations are SARS-CoV-2 antigen-specific, establishing that the extrafollicular response is antigen-driven. The paper provides the tissue-level histopathological evidence that GC failure in COVID-19 drives humoral immunity through the extrafollicular pathway, producing class-switched but likely non-durable antibody responses.

Study Design

• Type: Post-mortem tissue analysis + cross-sectional peripheral blood cohort
• Sample size: Tissue cohort: 11 COVID-19 patients (5 early, 6 late) + 6 age-matched non-COVID LN controls + 7 control spleens + 10 control tonsils. Blood cohort: 68 patients (39 convalescent, 10 severe CRP-intermediate, 15 severe CRP-high, 4 healthy controls).
• Setting: Brigham and Women’s Hospital / Massachusetts General Hospital, March–May 2020
• Population: Severely ill COVID-19 patients who died (tissue cohort); outpatient/inpatient COVID-19 at various severity stages (blood cohort). No dengue patients.

Key Findings

• Complete absence of germinal centers in thoracic lymph nodes and spleens of COVID-19 patients, both early (<10 days from symptoms) and late (15–36 days). Controls showed germinal centers presumably from ongoing commensal-driven immunity.
• Bcl-6⁺ GC B cells markedly reduced in both LN (p<0.001) and spleen (p<0.01) vs. controls, but AID⁺ B cells preserved and diffusely distributed — indicating T-dependent B cell activation continues outside GCs.
• Specific block in Bcl-6⁺ GC-type TFH differentiation: CD4⁺ICOS⁺ TFH cells diminished; CD4⁺CXCR5⁺ TFH present but reduced; CD4⁺Bcl-6⁺ GC-TFH strikingly depleted (LN: p<0.001; spleen: p<0.01). Changes occurred in both LN and spleen, consistent with circulating factors rather than local tissue effects.
• TH1 expansion: T-bet⁺ CD4⁺ T cells consistently increased in both LN and spleen, early and late. TH2 decreased. TH17 variably increased. FOXP3⁺ T regs accumulated late in disease.
• Aberrant TNF-α: TNF-α⁺ cells markedly increased in COVID-19 LNs vs. controls (early: p<0.05; late: p<0.01), expressed abundantly in both follicular and extra-follicular locations. Controls showed low-level TNF-α limited to follicles.
• Follicular dendritic cells (FDCs) preserved — GC loss not due to FDC absence.
• IgD⁻CD27⁻ double-negative B cells present within and outside follicles in COVID-19 tissue, with T-B conjugates at both sites. IgG⁺ class-switched plasmablasts prominent in follicular and extra-follicular areas.
• Peripheral blood — naive/transitional B cell loss: Total CD19⁺ B cells, naive (IgD⁺CD27⁻), early transitional (T1/T2: IgD⁺CD27⁻CD10⁺CD45RB⁺), and follicular (IgD⁺CD27⁻CD10⁻CD73⁺CXCR5⁺) B cells markedly reduced in severe COVID-19 with high CRP. Loss correlated with CRP level, symptom duration, and hospital length of stay.
• Activated B cell expansion in severe disease: Activated naive (IgD⁺CD27⁻CD21ˡᵒCD11cʰⁱ), DN2 (IgD⁻CD27⁻CXCR5⁻CD11c⁺), DN3 (IgD⁻CD27⁻CXCR5⁻CD11c⁻), switched memory (CD21ˡᵒ), and plasmablasts all expanded in severe CRP-high patients. Sum activated B cells correlated with CRP, symptom duration, and hospitalisation.
• SARS-CoV-2 antigen specificity confirmed: Dual-fluorophore RBD probe staining showed all disease-related B cell populations contained SARS-CoV-2-specific cells. DN2 and DN3 were CXCR5-low; DN1 and DN4 were CXCR5-high. Switched memory cells were the largest RBD⁺ subset (53.6% convalescent, 39.1% severe).

Methods Used

• Multi-color Immunofluorescence (tissue: CD3, CD19, Bcl-6, AID, CD4, ICOS, T-bet, GATA-3, RORγt, FOXP3, CXCR5, IgD, CD27, IgG, TNF-α, CD35, CD8; up to 7-color multispectral)
• Conventional Flow Cytometry (blood: 13-color B cell panel — CD3, CD56, CD19, CD27, IgD, CD38, CD10, CD45RB, CD21, CD73, CD138, CD11c, CXCR5; BD Symphony)
• FACS Sorting (RBD-specific B cell identification with dual-fluorophore probes)

Entities Mentioned

Bcl-6, AID, TNF-alpha, Double-Negative B Cell, DN2 B Cell, DN3 B Cell, Plasmablast, Activated Naive B Cell, CD19, CD27, IgD, IgG, CD38, CXCR5, CD21, CD11c, CD138, CD10, T-bet, ICOS

Concepts Addressed

Germinal Center, Extrafollicular Response, Class Switch Recombination, Somatic Hypermutation

Relevance & Notes

This is the tissue-level histopathological foundation for the EF dominance model in acute viral infection. It provides the mechanistic explanation for why COVID-19 humoral immunity is often non-durable: GCs are absent → high-affinity, long-lived memory and plasma cells are not generated → the humoral response depends on EF-derived plasmablasts that produce class-switched but low-SHM antibodies. The paper was already cited as an external reference in 4 wiki pages (Germinal Center, Extrafollicular Response, Woodruff2020 - EF B Cell Responses in COVID-19, External Citation Audit) prior to ingest.

Relationship to Woodruff2020: This paper provides the tissue evidence that explains the peripheral blood observations of Woodruff2020. Where Woodruff showed EF B cell expansion in COVID-19 blood, Kaneko demonstrates the anatomical basis — there are no GCs for these cells to have come from. The two papers are complementary: Kaneko (tissue) + Woodruff (blood) = complete picture of COVID-19 EF pathway.

Relevance to dengue: The GC loss mechanism (TNF-α-mediated TFH block) may or may not operate in dengue. Ansari2025 shows elevated CXCL13 alongside Tph activation in dengue, tentatively suggesting concurrent GC activity rather than complete GC ablation. However, CXCL13 is not GC-specific (Tph produce it). Whether severe dengue features partial or complete GC suppression — and whether TNF-α plays a role — is unknown. The Kaneko model predicts that any infection producing sufficient cytokine storm (including severe dengue with cytokine dysregulation) could disrupt GC formation.

Limitations: The tissue study is limited to fatal cases; milder disease may permit GC formation. The blood and tissue cohorts are separate (not paired). Small sample sizes throughout. Single imaging methodology for tissue analysis.

Questions Raised

• Does severe dengue produce sufficient TNF-α to disrupt GC TFH differentiation, as occurs in COVID-19? If so, does this explain the low-SHM dengue IgG findings from GodoyLozano2016?
• Are Bcl-6⁺ TFH cells detectable in dengue-draining lymph nodes? This would distinguish dengue (potentially concurrent EF+GC) from COVID-19 (GC ablation).
• The preservation of AID⁺ B cells despite Bcl-6⁺ GC B cell loss suggests AID expression occurs in EF contexts. Does dengue show a similar AID⁺/Bcl-6⁻ B cell profile in acute infection?
• Is the TH1 skewing (T-bet⁺ CD4⁺ expansion) seen in COVID-19 lymphoid tissue analogous to the Tph phenotype in dengue? Both are CXCR5⁻ with TH1 features, but dengue Tph cells provide B cell help (IL-21), while the COVID-19 TH1 cells may primarily block TFH differentiation.
• Can TNF-α blockade (anti-TNF biologics) rescue GC formation in severe viral infections? Murine malaria data (Ryg-Cornejo 2016) suggest yes.