T follicular helper 17 (Tfh17) cells are superior for immunological memory maintenance

Abstract

A defining feature of successful vaccination is the ability to induce long-lived antigen-specific memory cells. T follicular helper (Tfh) cells specialize in providing help to B cells in mounting protective humoral immunity in infection and after vaccination. Memory Tfh cells that retain the CXCR5 expression can confer protection through enhancing humoral response upon antigen re-exposure but how they are maintained is poorly understood. CXCR5⁺ memory Tfh cells in human blood are divided into Tfh1, Tfh2, and Tfh17 cells by the expression of chemokine receptors CXCR3 and CCR6 associated with Th1 and Th17, respectively. Here, we developed a new method to induce Tfh1, Tfh2, and Tfh17-like (iTfh1, iTfh2, and iTfh17) mouse cells in vitro. Although all three iTfh subsets efficiently support antibody responses in recipient mice with immediate immunization, iTfh17 cells are superior to iTfh1 and iTfh2 cells in supporting antibody response to a later immunization after extended resting in vivo to mimic memory maintenance. Notably, the counterpart human Tfh17 cells are selectively enriched in CCR7⁺ central memory Tfh cells with survival and proliferative advantages. Furthermore, the analysis of multiple human cohorts that received different vaccines for HBV, influenza virus, tetanus toxin or measles revealed that vaccine-specific Tfh17 cells outcompete Tfh1 or Tfh2 cells for the persistence in memory phase. Therefore, the complementary mouse and human results showing the advantage of Tfh17 cells in maintenance and memory function supports the notion that Tfh17-induced immunization might be preferable in vaccine development to confer long-term protection.

Trial registration: ClinicalTrials.gov NCT05009134.

Keywords: Tfh; antibody; human; immunology; inflammation; memory; mouse; vaccine.

Figure 1.. The in vitro differentiation of induced Tfh1, Tfh2 and Tfh17-like (iTfh1, iTfh2, iTfh17) cells.

(A) Splenocytes from WT mice were analyzed and representative FACS plot for Tfh1, Tfh17, and Tfh2 cells was shown. (B–H) OT-II cells were co-cultured with WT splenocytes as antigen-presenting cells (APCs) in the presence of OVA peptide, indicated cytokines and blocking antibodies for three days before phenotypic analysis. Experiment design (B), representative FACS plots for the expression of Tfh markers CXCR5, PD-1 and BCL6 (C) transcription factors T-bet, RORγt and GATA3 (E), CXCR3 vs CCR6 expression (G) and statistics (D, F, H). (I–L) 5×10⁴ cultured OT-II iTh0, iTfh1, iTfh2, and iTfh17 cells were FACS-purified and separately transferred into CD28KO recipients, followed by OVA-Alum immunization. The spleens were collected on day7 post-immunization for FACS analysis. Experiment design (I), representative FACS plot for Tfh percentage in OT-II cells (J), statistics of Tfh percentage in OT-II cells (K) and statistics of CXCR3/CCR6⁺ percentage in OT-II Tfh cells (L). The p values were calculated by two-way ANOVA for (D) and one-way ANOVA for (F, H, L). The results in (D, F, H) were pooled from three independent experiments. The results in (K, L) were pooled from two independent experiments. Source data for the statistics can be found in Figure 1—source data 1.

Figure 1—figure supplement 1.. iTfh1/2/17 cells show lower BCL6 expression than GC-Tfh cells.

OT-II cells were transferred into congenic WT mice followed by OVA in alum immunization. On day7, the BCL6 expressions for splenic OT-II Non-Tfh (CD44⁺CXCR5^-PD-1^-) and GC-Tfh cells (CD44⁺CXCR5^highPD-1^high) were analysed and compared with iTfh1/2/17 cells. Representative FACS plots (A) and statistics (B) comparing the expressions of BCL6 of indicated cell populations. The numbers in (A) indicate GMFI values. The results were pooled from two independent experiments. The p values were calculated by one-way ANOVA. Source data for the statistics can be found in Figure 1—figure supplement 1—source data 1.

Figure 2.. iTfh17 cells are superior in memory maintenance.

(A–H) 5×10⁴ FACS-purified OT-II iTfh1, iTfh17, or iTfh2 cells were separately transferred to CD28KO recipients. The early immunization group was immunized by OVA or NP-OVA in alum one day after the adoptive cell transfer. The late immunization group was immunized by the same antigens 35 days after the adoptive cell transfer. Spleens or serum were collected on day 7 after the immunization. Experiment design (A), representative FACS plots (B, C, D) and statistics (E, F, G) showing the percentages of Tfh cells in OT-II cells, the percentages of B_GC in total B cells and the percentages of B_ASC in total B cells. For antibody titers, statistic (H) showing OD405 values of anti-NP2 and anti-NP23 total IgG. (I–J) Tfh1/2/17 cells from mouse splenocytes were analyzed for CCR7 and CD44 expression. Representative FACS plots (I). and statistics (J) showing the expressions of CCR7 and CD44. The p values were calculated by one-way ANOVA. The results in (E, F, G, H, I, J) were both pooled from two independent experiments. Source data for the statistics can be found in Figure 2—source data 1.

Figure 3.. Human cTfh_CM and cTfh_EM subsets phenotypically and functionally resemble T_CM and T_EM subsets respectively.

(A–E) Naive, T_CM, T_EM, cTfh_CM , and cTfh_EM cells were FACS-purified from PBMC of two healthy donors and bulk RNA-seq was performed for differentially expressed genes analysis and gene set enrichment analysis (GSEA). (A) Representative FACS plot showing the gating strategy for indicated subsets. (B) MDS plot showing sample distribution. (C) Heatmap of the top 50 variable genes normalized by z-score. (D) Summarized normalized enrichment score (NES) of significantly enriched (p<0.05, FDR <0.25) hallmark gene sets by either cTfh_EM vs cTfh_CM or T_EM vs T_CM. (E) GSEA on selected gene sets were performed on cTfh_EMvs cTfh_CM and T_EM vs T_CM and the number indicates NES. (F–G) FACS-purified T_CM, T_EM, cTfh_CM and cTfh_EM cells were rested in complete RPMI for 3 days. Representative FACS plots (F) and statistics (G) showing the percentages of viable cells. (H–I) FACS-purified T_CM, T_EM, cTfh_CM, and cTfh_EM cells were labelled with CFSE and stimulated by anti-CD3/CD28 for 2.5 days. Representative FACS plots (H) and statistics (I) showing the CFSE fluorescence intensity and the division index. The p values were calculated by Wilcoxon matched-pairs signed-rank test. The results in (G, I) were pooled from ive healthy individuals with each conducted in three technical replicates. Source data for the statistics can be found in Figure 3—source data 1.

Figure 4.. Human cTfh_CM cells are enriched with the cTfh17 subset whereas cTfh_EM cells are enriched with the cTfh1 subset.

(A–C) Human PBMC samples from 33 healthy blood donors were analyzed. Representative FACS plots and statistics showing the percentages of cTfh1, cTfh2, and cTfh17 cells in cTfh_CM (A) or cTfh_EM (B) subsets. cTfh_CM/cTfh_EM ratios for cTfh1/2/17 in each individual were calculated (C). (D–E) FACS-purified cTfh_EM and cTfh_CM from five healthy individuals were analyzed for the expressions of indicated transcription factors by qPCR. The statistics for relative gene expression 2^-ΔΔCt (normalized to cTfh_EM) (D) and cTfh_CM/cTfh_EM ratios (E). (F–G) PBMC from 13 healthy individuals were analyzed for the secretions for indicated cytokines post PMA/ionomycin stimulation. The statistics for the percentages of cytokine⁺ cells (F) and the cTfh_CM/cTfh_EM ratios (G). FC: average fold change. The p values were calculated by Friedman test. Source data for the statistics can be found in Figure 4—source data 1.

Figure 5.. HBV antigen-specific cTfh17 cells are preferentially maintained in memory phase.

Blood samples from HBV vaccinated healthy individuals (N=38) were collected on indicated time points before/after HBV vaccination, and serum was diluted 10 times to analyse the anti-HBVSA antibody titer by ELISA. PBMC were also isolated and cultured with or without 20 µg/mL HBVSA for 18 hr, followed by FACS to analyse the phenotype of HBVSA-specific cTfh cells. Experiment design (A) and representative FACS plot (B) showing the gating strategy to detect HBVSA-specific cTfh cells by AIM assay. Representative FACS plot (C) and statistics (D) showing the percentage of cTfh1/2/17 cells in HBVSA-specific cTfh cells before vaccination. Classification (E) of 38 individuals into four groups was based on their anti-HBVSA antibody titers. Representative FACS plot (F) for an early responder showing the percentage of cTfh1/2/17 cells in HBVSA-specific cTfh. Statistics (G) showing the percentage of cTfh1/2/17 cells in HBVSA-specific cTfh on day 7 and day 28 after the vaccination in all defined groups (N=30, 8 samples with poor signals in AIM assay were excluded). The p values were calculated by Wilcoxon matched-pairs signed-rank test. Source data for the statistics can be found in Figure 5—source data 1.

Figure 5—figure supplement 1.. AIM assays to identify antigen-specific cTfh cells.

(A–C) FACS-purified cTfh1/2/17 cells from five individuals were stimulated by αCD3/CD28 for 18 hr and were analysed by FACS. Statistics (A) showing the cellular sizes before/after αCD3/CD28 stimulation. Representative FACS plots (B) and statistics (C) showing the percentages of cTfh1/2/17 cells before/after stimulation. (D–E) PBMC were cultured with or without HBVSA for 18 hr followed by FACS analysis. Example (D) of one sample excluded from the analysis. Example (E) of one sample included in the analysis. The numbers of total cTfh cells were then normalized to 10⁶ and the absolute counts of (AIM⁺) PD-L1⁺OX40⁺CD25⁺ cTfh cells were calculated, and the absolute counts of antigen-specific cTfh1, cTfh2, cTfh17, and cTfh1/17 were obtained. Any count values less than 0 were changed to 0 because cell count cannot be negative. At last, the normalized percentage was calculated based on the absolute count of antigen-specific cTfh1, cTfh2, cTfh17, and cTfh1/17. The p values were calculated by paired t-tests. Source data for the statistics can be found in Figure 5—figure supplement 1—source data 1.

Figure 5—figure supplement 2.. The cTfh responses in HBV vaccinated individuals.

(A) Blood samples were obtained from 38 individuals who received HBV vaccines on day –7, day 7, and day 28 before/after HBV vaccination. ELISA was performed to determine the titer of anti-HBVSA IgG antibody in serum. Participants were separated into four groups according to their responses to the vaccine. (B–E) The PBMC were isolated and cultured for 18 hr with or without 20 µg/mL HBVSA, followed by FACS analysis. Representative FACS plot of one early responder (B) showing the percentage of cTfh_CM and cTfh_EM in HBVSA-specific cTfh cells on indicated time points. The statistics (C) showing the percentages of HBVSA-specific cTfh_EM cells on indicated time points in four groups of responders. The statistics of early responders (D) showing the percentages of HBVSA-specific cTfh1/2/17 cells on day 7 and day 28 post HBV vaccination. The statistics of early responders (E) showing the percentages of cTfh1/2/17 in total cTfh on day 7 and day 28 post HBV vaccination. The p values were calculated by Friedman test. Source data for the statistics can be found in Figure 5—figure supplement 2—source data 1.

Figure 6.. Influenza virus-specific cTfh cells show cTfh1 signatures in effector phase but cTfh17 signatures in memory phase.

The single-cell RNA-seq dataset (GSE152522, the experiment design A) was analyzed to identify CXCR5-expressing cTfh clusters (B), which contain 12 major clones with a total of 249 cells. Comparison of cTfh1 and cTfh17 signature scores between effector and memory phase cTfh cells based on each cell or clone were shown in (C, D) and (E, F). The signature score of each clone was calculated as the mean value of the signature scores of all the cells in this clone. The p values were calculated by unpaired t-tests for (C, D) and paired t-tests for (E, F). Source data for the statistics can be found in Figure 6—source data 1.

Figure 6—figure supplement 1.. scRNA-seq analysis for influenza-specific cTfh cells.

(A) Top 100 signature genes for cTfh1 and cTfh17 were generated by Limma package based on bulk RNA-seq dataset GSE123812. The heatmap showing the expressions of cTfh1 and cTfh17 signature genes in cTfh1/2/17 samples. (B) cTfh1 and cTfh17 signature scores in were separated by individuals (H01, H02 and H04) and each clone was colour-coded accompanied by the CDR3 amino acid sequences. The statistics comparing the cTfh1 and cTfh17 signature scores between effector and memory phase influenza-specific cTfh cells. One individual (H03) was not included since lacking abundant clones. The p values were calculated by student t-tests. Source data for the statistics can be found in Figure 6—figure supplement 1—source data 1.

Figure 7.. cTfh17 cells are long-lived and accumulate with aging.

(A–C) PBMC samples from 20 healthy individuals were cultured for 18 hr with or without indicated antigens, followed by FACS to detect the phenotype of antigen-specific cTfh cells. Experiment design (A), representative FACS plot (B) and statistics (C) showing the percentage of cTfh1/2/17 cells in antigen-specific cTfh cells against tetanus toxin or measles. (D–F) PBMC samples from individuals of different ages were analysed. Representative FACS plots (D) showing the percentages of cTfh1/2/17 cells in total cTfh cells in individuals of different ages. Correlations tests (E) between the biological age with the percentages of cTfh1/2/17 cells in total cTfh cells. Cord blood samples were excluded from the correlation tests because of insufficient cTfh cell numbers. The p values were calculated by Friedman test for (C) and Pearson correlation for (E). Source data for the statistics can be found in Figure 7—source data 1.

Figure 7—figure supplement 1.. SARS-CoV-2-specific cTfh17 cells have superior persistence.

Peripheral blood was collected from 14 healthy donors and 13 convalescent Covid-19 patients at 10–12 months post the infection, and ELISA or AIM assays were performed. The SARS-CoV-2specific IgG antibody titers were shown in (A), representative FACS plots and statistics for cTfh1/2/17’s percentages in AIM⁺ cTfh in convalescent Covid-19 patients were shown in (B, C). One samples with poor signals in AIM assay were excluded. The p values were calculated by Friedman test. Source data for the statistics can be found in Figure 7—figure supplement 1—source data 1.

Figure 8.. iTfh17 cells are superior in survival and differentiation into GC-Tfh cells after resting.

(A–C) 5×10⁴ FACS-purified OT-II iTfh1, iTfh17, or iTfh2 cells were separately transferred to CD28KO recipients, and the spleens were FACS analysed on day1 and day35. Experimental design (A), representative FACS plots (B) and statistics (C) showing the total and normalized numbers of transferred iTfh1/2/17 cells in the spleens. (D–G) FACS-purified 1×10⁴ B1-8 B cells and 5×10⁴ OT-II iTfh1, iTfh17 or iTfh2 cells were co-transferred to CD28KO recipients. The early immunization group was immunized by NP-OVA in alum 1 day after the adoptive cell transfer. The late immunization group was immunized by the same antigens 35 days after the adoptive cell transfer. Spleens were collected on day 5 after the immunization. Experiment design (D). Statistic showing the numbers of OT-II cells in the spleen (E). Representative FACS plots (F) and statistics (G) showing the percentages of GC Tfh cells in OT-II cells. The p values were calculated by one-way ANOVA. The results in (C, E, G) were both pooled from two independent experiments. Source data for the statistics can be found in Figure 8—source data 1.

Figure 8—figure supplement 1.. iTfh1/2 cells retain polarized Th1/2 cytokine profiles and promote specific isotype class switching.

FACS-purified 1×10⁴ B1-8 B cells and 5×10⁴ OT-II iTfh1, iTfh17 or iTfh2 cells were co-transferred to CD28KO recipients. The early immunization group was immunized by NP-OVA in alum 1 day after the adoptive cell transfer. The late immunization group was immunized by the same antigens 35 days after the adoptive cell transfer. Spleens were collected on day 5 after the immunization followed by FACS analyses, and PD-1^hi CXCR5^hi OT-II GC-Tfh cells were FACS-purified for qPCR analyses. Total CD4⁺ T cells were purified by MACS enrichment from naive mice as control. Statistics (A) showing the expression of indicated genes. Statistics (B) showing the numbers of indicated B1-8 populations in the spleens. Statistic (C) showing the expression of Ifng and Il4 in all samples. Statistics (D) showing the expression of indicated antibody isotypes. The p values were calculated by one-way ANOVA. Results were pooled from two independent experiments. Source data for the statistics can be found in Figure 8—figure supplement 1—source data 1.