henhao
Sign in
ImmunologybioRxiv
Preprint

Tfh dysfunction is associated with poor responses to HBV vaccination

Han M. CheniD, Elizabeth Severa, Xiaomin YaoiD, Tijaana WilliamsiD, Riddhimaa Sinha, Clarice MonteiroiD, Michael TueniD, Thomas Barnett DubenskyiD, Sergei B. KoraloviD, Carla R. NowosadiD, Derek A. OldridgeiD, Laura A. VellaiD, Anoma NelloreiD, Ramin S. HeratiiD

Abstract

The primary goal of vaccination is to induce durable protection through generating antigen-specific humoral immunity and cellular immune memory. Yet the same vaccine may elicit protection in some and not in others. To characterize the molecular and cellular events that underlie differential vaccine responsiveness, we leveraged the known, pronounced variability in vaccine antigen-specific antibody concentration and durability observed after Hepatitis B virus (HBV) immunization. We longitudinally profiled 101 healthy adults undergoing either de novo (n=59) or booster (n=42) HBV immunizations. We then stratified participants into High ([≥]100 mIU/mL) and Low (<100 mIU/mL) Responder groups based on end-of-study titers. From this cohort, we obtained core needle biopsies (CNBx) of vaccine-draining lymph nodes from 10 participants (High Responder n=6, Low Responder n=4), 10-21 days post-final vaccination and performed single-cell RNA sequencing with paired antigen receptor sequencing. Lymph node BCR repertoire analysis revealed convergent, semi-public clonotypes across unrelated participants with elevated somatic hypermutation. Transcriptomic profiling of GC-B cells uncovered functional divergence between response groups, with High Responders showing robust activation across both dark zone (DZ) and light zone (LZ) compartments and enrichment of MYC and mTORC1 signaling. This divergence in B cell responses corresponded to notable differences in the T cell compartment. High Responders harbored a key Tfh subset marked by increased expression of CXCL13, ICOS, and GNG4, supportive of their GC localization and capacity to provide B cell help. Low Responders demonstrated a divergent program in which their GNG4+Tfh displayed aberrant inflammatory transcriptional activity despite expressing coordinated costimulatory genes necessary to provide help to LZ GC-B cells. This inflammatory signaling was associated with reduced specificity of BCL6 regulon activity despite maintained expression of canonical Tfh genes. Our data suggest that variable HBV vaccine responsiveness was associated with transcriptional aberrations in the GNG4+Tfh subset, offering a molecular framework for understanding poor vaccine responses.

PDFDOI
Full-text digest ↓

02 /Specification & Classification

Click any tag to browse matching papers
Citations
0
OpenAlex Topics

In-depth summary

✦ Multi-agent verified full-text synthesis

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (30 Sept 2026) · CC BY 4.0

HBV vaccine-draining lymph node biopsies tie low antibody titers to reduced, interferon-skewed GNG4+ germinal center T follicular helper cells; public B cell clonotypes did not track response.

5-10% of adults fail to reach seroprotective hepatitis B virus (HBV) antibody titers, and blood-based correlates such as circulating T follicular helper (Tfh) cells miss the lymph node germinal center (GC) dynamics behind this variability. By biopsying vaccine-draining lymph nodes and pairing single-cell transcriptomes with B and T cell receptor repertoires, the study ties low responses to molecularly defined GC Tfh dysfunction rather than failed B cell recruitment.

Key results

  1. ShownAntibody gains differed by vaccine regimen. Participants split by final titer into High and Low Responders (≥100 vs <100 mIU/mL) had similar baseline titers but diverged over the course of vaccination, and demographic factors did not explain the split.

    Geometric mean HBsAb fold rise over baseline: 2.8 (HBVimmune), 16 (HBVnaive-Alum), 6.9 (HBVnaive-CpG); High vs Low Responder trajectories diverged (LME interaction P=3.1e-4 in HBVnaive-Alum, P=2.5e-9 in HBVimmune); age, sex, body mass index and smoking were not associated with responder status (all P>0.05, multivariate logistic regression). · Heterogeneous humoral responses following…; Figure legends

  2. ShownVaccination produced convergent and public GC B cell clonotypes across participants, some matching published HBV vaccine BCRs. Within-group public clonotype sharing did not differ from cross-group sharing, which argues against a failure to recruit relevant B cells.

    22,037 B cells from 10 biopsied participants; 129 convergent clusters, 93 (72%) in at least two participants and 47 (36%) in three or more; public clonotypes in 8 of 10 individuals (1726 B cells); within- vs across-group sharing High P=0.1, Low P=0.12 (Wilcoxon signed-rank); 5 clonotypes in 3 participants matched published HBV vaccine sequences. · Germinal center B cell activity and…

  3. SuggestedIn High Responders, light zone (LZ) GC B cells were enriched and GC B cells showed proteostatic and proliferative programs (heat shock, MYC, mTORC1, NFκB). In Low Responders, LZ B cells looked quiescent and kept B cell identity regulons active.

    miloR found LZ GC-B cells enriched in High Responders across 85% (302/355) of iterations showing detectable differences; High Responders upregulated HSP90AA1, HSPH1, HSPD1 and DNAJ co-chaperones; Low Responder LZ cells expressed SELL, SYK, FCRL1, JCHAIN with PAX5, SPI1, SPIB, IKZF3 regulon activity. · Germinal center B cell activity and…; Figure legends

  4. SuggestedTfh cells formed six transcriptional states. The GNG4 subset (CXCL13, POU2AF1, TOX2, BCL6, CXCR5) mapped to the GC core of a reference tonsil, was the only subset with predicted CXCL13-CXCR5 signaling to LZ B cells, and had the strongest association with High Responders.

    2,995 Tfh cells formed six subsets; label transfer onto a Visium HD tonsil showed the GNG4 signature made up nearly all Tfh signal in GC core Cluster 8; by Cohen's d, GNG4+ Tfh had the strongest High Responder association (odds over 10.1 times higher; TNFR2 subset 5.2); miloR also ranked GNG4+ Tfh first. · T follicular helper subsets are spatially…; GNG4+ Tfh cell activity is a hallmark of humoral…; Visium HD analysis

  5. SuggestedIn Low Responders, GNG4+ Tfh cells expressed more IL21, CD40LG and POU2AF1 and had stronger predicted helper-ligand contacts. Their regulon and pathway profiles, however, leaned toward interferon, AP-1 and inflammatory programs rather than BCL6/MAF-anchored helper networks.

    LIANA ranked CD40LG-CD40, IL21-IL21R and CD28-CD80 higher in Low Responders; decoupleR showed BCL6, SMAD2, CREM, STAT4, MAF activity in High versus STAT1, RELA, NFκB1, IRF9, CIITA in Low Responders; GSEA gave TNFa signaling via NFκB NES = 1.796 (q = 2.3e-05) and Interferon Alpha Response NES = -1.84 (q = 2e-4). · GNG4+ Tfh cell activity is a hallmark of humoral…; Figure legends

Methods, data and reproducibility

Partially randomized HBV vaccination trial (CpG or alum series, or alum booster) with longitudinal anti-HBsAg ELISA. Draining axillary node core needle biopsies taken 10-21 days after the final dose underwent immunofluorescence and CITE-seq with paired BCR/TCR, then regulon, pathway, ligand-receptor and spatial label-transfer analyses.

  • Prospective partially randomized HBV vaccination cohort: HBV-naive adults (CpG two-dose or alum three-dose series) and previously vaccinated adults (single alum booster); 531 blood samples · human · n = 101 participants (13 CpG, 46 alum, 42 booster) · NCT04674462
  • Ultrasound-guided core needle biopsies of ipsilateral axillary lymph nodes 10-21 days after final dose, profiled by CITE-seq with BCR/TCR and immunofluorescence · human · n = 10 participants (3 alum, 1 CpG, 6 booster) · GEO GSE349037; scripts Zenodo 10.5281/zenodo.21440760
  • Lymph node B cells with paired BCR repertoires from the biopsies · human · n = 22,037 B cells
  • Lymph node Tfh cells subclustered from CD4+ T cells, with TCR repertoires · human · n = 2,995 Tfh cells
  • Deidentified tonsils (NYU CBRD) used as germinal center size comparators for immunofluorescence · human
  • Public 10x Genomics Visium HD fresh-frozen tonsil section (21-year-old male) used as spatial reference for Tfh label transfer · human
  • IEDB HBV-specific TCR sequences from a chronic HBV infection study, used to flag antigen-specific CD4+ clonotypes · human · IEDB Reference 1035279
Methods & tools
CITE-seq with V(D)J libraries (10x Genomics Chromium X, GEM-X), Cell Ranger v8.0.1 (GRCh38), Seurat (SCTransform, rpca integration), Harmony v.1.2.4, Azimuth v0.4.6, Souporcell and scDblFinder v1.20.2, IgBlast v1.22.0, Change-O v1.3.4, Dowser v2.3, SHazaM v1.3.1, scRepertoire v1.11.0, miloR v.2.6 (consensus over 1000 subsampled seeds), Slingshot v2.14.0, decoupleR v2.12.0 (CollecTRI) and SCENIC v1.3.1, LIANA v0.1.14, SCPA v. 1.6.2, fgsea v. 1.32.4, UCell v2.10.1 (MSigDB Hallmark), lme4/lmerTest linear mixed effects models
Reagents & environment
Cell Ranger v8.0.1 (GRCh38), Harmony v.1.2.4, Azimuth v0.4.6, scDblFinder v1.20.2, IgBlast v1.22.0, miloR v.2.6, decoupleR v2.12.0, SCENIC v1.3.1, LIANA v0.1.14, Space Ranger v3.1.1; R v4.4 and Prism 9.0
Reproducibility gaps
Scripts on Zenodo (doi 10.5281/zenodo.21440760) and data in GEO GSE349037 are given only as identifiers, and GEO is cited with a reviewer token; other resources are via the lead contact; no container or hardware details reported.
Limitations (6)
  • Just 10 of 72 screened people had enlarged draining nodes suitable for biopsy, so selection bias in this subset, or systematic differences in those lacking accessible nodes, cannot be fully ruled out. · stated by the authors
  • A single biopsy time point (10-21 days post-vaccination) cannot tell whether Tfh changes reflect late GC onset, early GC contraction, or a persistent dysfunctional state; serial fine-needle aspiration is suggested. · stated by the authors
  • Too few CpG-adjuvanted participants were biopsied (n=1) to test whether adjuvant choice reshapes the GC and Tfh programs described. · stated by the authors
  • The link between GNG4+ Tfh cells and antibody titers is associative; functional and interventional studies are needed to establish causality. · stated by the authors
  • With 10 biopsied participants, many differential expression, regulon and per-cell Wilcoxon comparisons treat cells as units, risking pseudoreplication; Tfh subset proportions themselves did not differ substantially between responder groups. · noted in this summary
  • GNG4+ Tfh GC positioning is inferred by label transfer onto one external tonsil Visium HD section rather than spatial profiling of the vaccinated lymph nodes, so localization within the biopsies is untested. · noted in this summary

Who should read:Read this if you study human germinal center or Tfh biology, vaccine hyporesponsiveness, or want a model for pairing lymph node biopsies with single-cell BCR/TCR and regulon analyses.

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISElias '20Herrero-Fer '18Sala '25Ahodantin '26Schattgen '23Singhania '21
68% semantic matchImmunology
G. Elias, P. Meysman, E. Bartholomeus+15 more·bioRxiv·Aug 22, 2020