B cells shape naive CD8+ T cell programming

Abstract

The presence of B cells is essential for the formation of CD8+ T cell memory after infection and vaccination. In this study, we investigated whether B cells influence the programming of naive CD8+ T cells prior to their involvement in an immune response. RNA sequencing indicated that B cells are necessary for sustaining the FOXO1-controlled transcriptional program, which is critical for homeostasis of these T cells. Without an appropriate B cell repertoire, mouse naive CD8+ T cells exhibit a terminal, effector-skewed phenotype, which significantly impacts their response to vaccination. A similar effector-skewed phenotype with reduced FOXO1 expression was observed in naive CD8+ T cells from human patients undergoing B cell-depleting therapies. Furthermore, we show that patients without B cells have a defect in generating long-lived CD8+ T cell memory following COVID vaccination. In summary, we demonstrate that B cells promote the quiescence of naive CD8+ T cells, poising them to become memory cells upon vaccination.

Keywords: Adaptive immunity; Autoimmunity; Immunology; Immunotherapy; Rheumatology; Vaccines.

Figure 1. B cells promote memory-fated CD8⁺ T cell responses to vaccination and infection.

(A–J) WT or MD4 mice were either vaccinated with a combined-adjuvant subunit vaccine or infected with vaccinia virus (VV) or LCMV. (A) Experimental schematic. (B) Representative tetramer staining for pre-gated on live, CD19⁻CD8⁺ lymphocytes. (C) Histograms showing CD127 expression by tetramer⁺ cells. Naive (CD44^lo) CD8⁺ T cells from WT mice provide a high CD127 reference for the LCMV histograms. (D–F) Seven days after vaccination, spleens were assessed for number (D) and percentage (E) of SIINFEKL tetramer⁺ cells, and CD127 geometric mean fluorescence intensity (gMFI) for tetramer⁺CD127^hi cells (F). (G and H) Seven days after VV infection, spleens were assessed for number (G) and percentage (H) of B8R tetramer⁺ cells. (I and J) Eight days after LCMV-Armstrong infection, spleens were assessed for number (I) and percentage (J) of GP33 tetramer⁺ cells. (K–T) Sublethally irradiated WT or MD4 mice received 1 million CAR or control T cells. After 30 days, mice were vaccinated with the combined-adjuvant subunit vaccine. (K) Experimental schematic. (L) PBMCs were assessed for B cells (CD19⁺) and CAR T cells (TCRβ⁺CD45.1⁺hEGFR⁺). (M) B cell frequencies (left) and CAR T cell frequencies (right). (N–T) Seven days after vaccination, spleens were assessed for relative abundance of splenic tetramer⁺ CD127^hi cells and CD127^lo cells (N and O); tetramer⁺ cells were assessed for CD127 gMFI (P), FOXO1 gMFI (Q), percentage positive for TCF1 (R), percentage positive for granzyme B (S), and IRF4 gMFI (T). Data shown are means ± SEM, n = 4–5 mice per group, representative of 2 experiments. Significance was defined by 2-way ANOVA (D, E, and N–T) or 1-way ANOVA (F) with Holm-Šidák multiple-comparison test; *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 2. B cells shape naive CD8⁺ T cell programming, promoting FOXO1-mediated homeostasis in mice and humans.

(A–D) Bulk RNA sequencing on naive CD8⁺ T cells (sorted on live CD8⁺CD19⁻B220⁻CD44^lo) from 5 WT and 5 MD4 mice. (A) The top 10 most significant gene sets identified by GSEA using the Broad Institute’s Molecular Signatures Database (MSigDB) immunologic signature gene sets (C7) ordered by adjusted P value. (B) ChEA analysis of the 1,027 differentially expressed genes by DESeq2 ranked by combined score. (C) Heatmap of the 242 genes identified by ChEA as being associated with FOXO1 transcriptional activity. (D) Volcano plot where genes known to be differentially expressed in Foxo1-deficient cells (31, 33) are highlighted in black (higher in WT) or maroon (higher in MD4). (E) Flow cytometry staining of CD8⁺ T cells from WT, μMT^−/−, and MD4 mice. Representative histograms are shown for CD44^lo naive cells (top) and for CD44^hi memory and virtual memory cells (middle). A graph plotting the individual gMFI from 5 mice per group for CD44^lo naive cells is shown at bottom. Data shown are means ± SEM, representative of more than 2 experiments. (F) Representative FOXO1 staining in naive CD8⁺ T cells from healthy control versus CAR T cell–treated patients (top) and patients with MS on anti–α₄ integrin versus anti-CD20 therapy (bottom). (G) Summarized results showing the gMFI values (1 × 10³, 10³, 10³, 10², 10³, and 10², respectively) for FOXO1, CD127, IRF4, GZMA, CD49d, and NKG2A protein staining for all 4 groups; n = 19, 8, 20, and 19, respectively; **P < 0.01, ***P < 0.001.

Figure 3. Naive CD8⁺ T cells from B cell–restricted hosts exhibit normal proliferative capacity, but defective survival.

(A) Purified CD8⁺ T cells were plated with or without 5 ng/mL human IL-7. Ratio of cells surviving to number of cells plated is displayed over time. Data are combined from 3 experiments. Analysis by 2-way ANOVA yielded a significant interaction effect, indicating that the slopes of the two lines differ significantly. (B–D) Purified CD8⁺ T cells from WT and MD4 mice were dye-labeled, mixed in equal numbers, and cotransferred into WT or MD4 mice. (B) Experimental schematic. (C) Ratio of transferred cells in blood 2 hours after transfer, and in spleens 7 days later. (D) CD127 and CD122 gMFI was analyzed on transferred cells and endogenous naive CD8⁺ T cells on day 7. (E) Purified CD8⁺ T cells from WT or MD4 mice were stimulated with plate-bound anti-CD3. Representative proliferation dye dilution (top) and proliferation index (bottom). (F–I) Purified OT1 T cells from WT or MD4 mice were dye-labeled, mixed in equal numbers, and transferred into sublethally irradiated recipients. Spleens were analyzed 10 days later. (F) Experimental schematic. (G) Representative proliferation dye dilution as percentage of maximum (left) or as total counts (right). (H and I) Proliferation (H) and expansion (I) indices. (J–M) Purified OT1 T cells from WT and MD4 mice were mixed in equal numbers and transferred to recipient mice, which then received subunit vaccination. (J) Experimental schematic. (K) Ratio of transferred OT1 T cells at day 3. (L) Representative EdU incorporation plots. (M) Quantification of EdU incorporation. Data shown are means ± SEM, representative of ≥2 experiments. Significance was defined by 2-way ANOVA with Holm-Šidák multiple-comparison test; *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 4. The B cell environment in which a naive CD8⁺ T cell develops has significant consequences for its response to vaccination.

Four hundred purified OT1 T cells from WT and MD4 mice were mixed 1:1 and transferred i.v. into WT or MD4 recipients. Mice were then immediately vaccinated with the combined-adjuvant subunit vaccine [OVA, poly(I:C), and anti-CD40]. Seven days later, spleens were analyzed by flow cytometry. (A) Experimental schematic. (B) Representative tetramer staining on CD19⁻CD8⁺ lymphocytes (left), staining for CD45.1 and CD45.2 (middle) to identify tetramer⁺ cell origin, and staining for CD127 (right). (C) Percentage of tetramer⁺ cells positive for CD127. (D) The gMFI of CD127 staining on tetramer⁺CD127^hi cells. (E) Total number of splenic tetramer⁺ cells. Data shown are means ± SEM, n = 5 mice per group, representative of 2 experiments. Significance was defined by 2-way ANOVA with Holm-Šidák multiple-comparison test; *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5. B cells limit effector CD8⁺ T cell expansion following mRNA LNP vaccination, preserving memory pool.

(A–E) Mice were given either a primary mRNA LNP vaccination only (left) or a primary vaccine followed by a booster 30 days later (right). (A) Experimental schematic. (B) Representative dual tetramer staining for CD44^hi CD8⁺ T cells (left) and CD127 × TCF1 plots for tetramer⁺ cells after primary-only (middle) or primary-plus-boost (right) vaccination. (C) Spleens were assessed for the percentage of tetramer⁺CD127^hi cells and tetramer⁺CD127^lo cells out of total CD8⁺ T cells. (D) Percentage of TCF1⁺CD127^hi cells within tetramer⁺ cells. (E) Percentage of tetramer⁺ cells positive for granzyme B. Data shown are means ± SEM, n = 5 mice per group, representative of 2 experiments. Significance was defined by t tests; **P < 0.01, ***P < 0.001. (F and G) Healthy control (HC) subjects or patients with MS receiving anti-CD20 antibody therapy were assessed for antigen-specific T cells 10–12 days after an mRNA LNP COVID-19 vaccine boost. (F) Experimental schematic. (G) Antigen-specific cells were quantified as the percentage of non-naive CD8⁺ T cells positive for the activation-induced markers (AIMs) 4-1BB and IFN-γ and divided into T_CM (CD45RA⁻CD27⁺CCR7⁺), T_EM (CD45RA⁻CD27⁻CCR7⁻ plus CD45RA⁻CD27⁺CCR7⁻), and T_EMRA (CD45RA⁺CD27⁻CCR7⁻) subsets (boxes, 25th–75th percentile; horizontal lines, median). Significance was defined by Mann Whitney tests; *P ≤ 0.05. (H–M) HC subjects or XLA patients were vaccinated, and blood was taken 6 months after the first vaccine dose for single-cell sequencing. (H) Experimental schematic. (I) UMAP visualization of AIM⁺ (4-1BB⁺CD69⁺) CD8⁺ T cells highlighting the cells from 6 months after vaccination. (J) Contour plots of CITE-Seq protein data for CD45RA and CCR7 from HC subjects (top) and XLA patients (bottom). (K) Percentages of T_CM, T_EM, and T_EMRA were analyzed using 2-tailed Mann-Whitney tests; *P ≤ 0.05. (L) Violin plots of genes associated with cytotoxicity and memory. (M) Module scores for cytotoxicity and self-renewal gene signatures as box plots (vertical lines, minimum/maximum; boxes, 25th–75th percentile; horizontal lines, median) were analyzed using 2-tailed Mann-Whitney tests; **P < 0.01, ***P < 0.001.

Figure 6. FOXO1-haploinsufficient CD8⁺ T cells closely resemble those deprived of B cell help.

(A–E) Naive CD8⁺ T cells from unmanipulated Foxo1^fl/WT E8I-Cre⁺ mice (Foxo1^+/−) were compared with naive CD8⁺ T cells from unmanipulated Cre^– (WT) control mice. (A) The total number of splenic naive (CD44^lo) CD8⁺ T cells was calculated for WT and Foxo1^+/− mice. These cells were then assessed for the gMFI of FOXO1 (B), CD127 (C), IRF4 (D), and granzyme B (E). (F–M) OT1 T cells from congenically distinct Foxo1^+/− and control mice were transferred into WT recipients, which were immediately immunized with the combined subunit vaccine. Seven days later, spleens were analyzed by flow cytometry. (F) Experimental schematic. (G) Representative staining for CD45.1 and CD45.2 (left) to identify OT1 cell genotype and CD127 staining on these cells (right). These data were used to calculate the number (H) and percentage (I) of responding OT1 T cells from each genotype. These cells were then assessed for the gMFI of FOXO1 (J), CD127 (K), IRF4 (L), and granzyme B (M). Data shown are means ± SEM, n = 4–5 mice per group, representative of 2 experiments. Significance was defined by t tests (A–E) or paired t tests (H–M); *P < 0.05, **P < 0.01, ***P < 0.001.