mTORC1 and mTORC2 Kinase Signaling and Glucose Metabolism Drive Follicular Helper T Cell Differentiation

Abstract

Follicular helper T (Tfh) cells are crucial for germinal center (GC) formation and humoral adaptive immunity. Mechanisms underlying Tfh cell differentiation in peripheral and mucosal lymphoid organs are incompletely understood. We report here that mTOR kinase complexes 1 and 2 (mTORC1 and mTORC2) are essential for Tfh cell differentiation and GC reaction under steady state and after antigen immunization and viral infection. Loss of mTORC1 and mTORC2 in T cells exerted distinct effects on Tfh cell signature gene expression, whereas increased mTOR activity promoted Tfh responses. Deficiency of mTORC2 impaired CD4(+) T cell accumulation and immunoglobulin A production and aberrantly induced the transcription factor Foxo1. Mechanistically, the costimulatory molecule ICOS activated mTORC1 and mTORC2 to drive glycolysis and lipogenesis, and glucose transporter 1-mediated glucose metabolism promoted Tfh cell responses. Altogether, mTOR acts as a central node in Tfh cells by linking immune signals to anabolic metabolism and transcriptional activity.

Figure 1. T Cell-Specific Loss of PTEN, Raptor, or Rictor Affects GC Formation, Tfh Cell Differentiation and IgA Production in PPs

(A and B) Flow cytometry of GC B cells (GL-7⁺Fas⁺) among B220⁺ cells (A) and Tfh cells (PD-1^hiCXCR5⁺) among B220^–CD4⁺TCRβ⁺ cells (B) in PPs from WT and Cd4^crePten^fl/fl mice. Right shows the frequency of GC B cells or Tfh cells. (C) Frequencies of GC B cells and Tfh cells in PPs from mice treated with vehicle or rapamycin. (D) Frequency of CD4⁺ T cells in PPs from WT, Cd4^creRptor^fl/fl, and Cd4^creRictor^fl/fl mice. (E and F) Flow cytometry of GC B cells (E) and Tfh cells (F) in PPs from WT, Cd4^creRptor^fl/fl and Cd4^creRictor^fl/fl mice. Right shows the frequency and number of GC B cells or Tfh cells. (G) Flow cytometry of IgA⁺ B cells in PPs from WT, Cd4^creRptor^fl/fl, and Cd4^ceRictor^fl/fl mice. Right shows the frequency and number of IgA⁺ B cells. (H) IgA concentration in fecal extracts. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (one-way ANOVA, D–H; Mann-Whitney test, A–C). Data are pooled from two (C and H) or representative of three (A, B, D–G) independent experiments. Error bars represent SEM. See also Figure S1.

Figure 2. Deletion of Raptor and Rictor via OX40-Cre Impairs Tfh Cell Differentiation and Immune Homeostasis in PPs

(A) RNA analysis of Rptor and Rictor in freshly isolated or activated T cells from WT, OX40^creRptor^fl/fl, or OX40^creRictor^fl/fl mice. (B) Naive T cells from WT, OX40^creRptor^fl/fl, and OX40^creRictor^fl/fl mice were stimulated with anti-CD3 alone or anti-CD3 plus anti-CD28 for different days, followed by [³H]-thymidine incorporation analysis. (C) Immunohistochemistry of GCs in PPs of WT, OX40^creRptor^fl/fl, and OX40^creRictor^fl/fl mice. Scale bars represent 200 μm. (D and E) Flow cytometry of GC B cells (D) and Tfh cells (E) in PPs of WT, OX40^creRptor^fl/fl, and OX40^creRictor^fl/fl mice. Right shows the frequency and number of GC B cells or Tfh cells. (F) Flow cytometry of Tfh cells among CD45.2⁺B220^–CD4⁺TCRβ⁺ cells in PPs from mixed BM chimeras constructed by mixing BM cells from WT, OX40^creRptor^fl/fl, or OX40^creRictor^fl/fl mice and CD45.1⁺ mice and injecting into Tcrb^–/–Tcrd^–/– recipient mice. Right shows the frequency of Tfh cells. (G) Frequencies of CD4⁺ T cells in PPs, mesenteric lymph nodes (mLN) and spleen from WT, OX40^creRptor^fl/fl and OX40^creRictor^fl/fl mice. (H) Quantification of anti-CD3 mean fluorescence intensity (MFI) within GCs in PPs based on the immunohistochemistry images in (C). (I) Flow cytometry of IgA⁺ B cells in PPs from WT, OX40^creRptor^fl/fl, and OX40^creRictor^fl/fl mice. Right shows the frequency and number of IgA⁺ B cells. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001 (one-way ANOVA). Data are representative of two (A and B) or at least three (C–I) independent experiments. Error bars represent SEM. See also Figure S2.

Figure 3. mTORC2 Promotes Tfh Cell Differentiation in PPs by Suppressing Foxo1 Activity

(A) Flow cytometry of CD62L, CD127, CCR7, and CD69 expression on CD4⁺ T cells in PPs from WT and OX40^creRictor^fl/fl mice. (B) MFI of CD62L, CD127, CCR7, and CD69 on memory-phenotype (CD44^hiCD62L^lo) T cells from blood, spleen, peripheral lymph nodes (pLN), mesenteric lymph nodes (mLN), and PPs. (C) RNA analysis of S1pr1 and Klf2 expression in CD44^hiCD62L^lo T cells. (D) Confocal microscopy imaging of Foxo1 nuclear exclusion. Activated T cells were rested and re-stimulated with anti-ICOS antibody and then fixed and stained with anti-F-Actin, anti-Foxo1, and DAPI. Right shows calculated percentages of cells with Foxo1 excluded from the nucleus. (E and F) Flow cytometry of Tfh cells (E) and GC B cells (F) in PPs of WT, Cd4^creRictor^fl/fl, and Cd4^creRictor^fl/flFoxo1^fl/+ mice. Right shows the frequency and number of Tfh cells or GC B cells. (G) Comparison of gene-expression changes between WT and Cd4^creRptor^fl/fl versus WT and Cd4^creRictor^fl/fl memory phenotype (MP) T cells. Genes with altered expression (with > 0.5 log₂ fold change) in either Cd4^creRptor^fl/fl or Cd4^creRictor^fl/fl mice were clustered into four groups. R1, concordant expression between Cd4^creRptor^fl/fl and Cd4^creRictor^fl/fl cells; R2, selectively altered in Cd4^creRictor^fl/fl cells; R3, discordant expression between Cd4^creRptor^fl/fl and Cd4^creRictor^fl/fl cells; and R4, selectively altered in Cd4^creRptor^fl/fl cells. (H) GSEA of the Tfh cell gene signature in Cd4^creRptor^fl/fl (left) and Cd4^creRictor^fl/fl (right) MP T cells relative to the expression in WT MP T cells. The heatmaps on the left of GSEA plots show the top hit (leading-edge) genes. NS, not significant; **p < 0.01, ***p < 0.001, ****p < 0.0001 (one-way ANOVA). Data are representative of at least three (A, E, F), 2 (B–D) and one (G and H; n = 5 mice for WT mice, three mice for Cd4^creRptor^fl/fl mice, and four for Cd4^creRictor^fl/fl mice) independent experiments. Error bars represent SEM. See also Figure S3.

Figure 4. mTORC1 and mTORC2 Promote Tfh Cell and GC Responses after LCMV Infection

(A and B) Flow cytometry of GC B cells in the spleen from WT and OX40^creRptor^fl/fl mice (A) or OX40^creRictor^fl/fl mice (B) at 8 days after challenge with LCMV Armstrong strain. Right shows the frequency and number of GC B cells. (C and D) Flow cytometry of IgD^–CD138⁺ plasma B cells in the spleen from WT and OX40^creRptor^fl/fl mice (C) or OX40^creRictor^fl/fl mice (D) at 8 days after LCMV infection. Right shows the frequency and number of plasma B cells. (E and F) Flow cytometry of Tfh cells in the spleen from WT and OX40^creRptor^fl/fl mice (E) or OX40^creRictor^fl/fl mice (F) at 8 days after LCMV infection. Results are gated on B220^–CD4⁺TCRβ⁺ cells. Right shows the frequency and number of Tfh cells. (G and H) The percentages of Tfh cells among CD45.2⁺B220^–CD4⁺TCRβ⁺ donor-derived cells (left) or CD45.1⁺B220^–CD4⁺TCRβ⁺ cells (right) in the spleen from mixed BM chimeras at 8 days after LCMV infection; chimeras were reconstituted with BM cells from CD45.1⁺ mice and WT, OX40^creRptor^fl/fl mice (G) or OX40^creRictor^fl/fl mice (H). *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (Mann-Whitney test for frequencies, unpaired Student’s t test for numbers). Results were pooled from 2 (A, C, and E) or represent 2 (B, D, F, G, and H) independent experiments. Error bars represent SEM. See also Figure S4.

Figure 5. OX40^creRptor^fl/fl and OX40^creRictor^fl/fl Mice Fail to Mount Efficient GC Reaction after Immunization via Cell-Intrinsic and Antigen-Specific Mechanisms

(A and B) Flow cytometry of GC B cells in spleen from WT and OX40^creRptor^fl/fl mice (A) or OX40^creRictor^fl/fl mice (B) at 7 days after intraperitoneal immunization of mice with NP-OVA plus LPS in alum. Right shows the frequency and number of GC B cells. **p < 0.01, ***p < 0.001, ****p < 0.0001 (Mann-Whitney test). (C) Immunohistochemistry of GCs in the mesenteric lymph nodes of WT, OX40^creRptor^fl/fl and OX40^creRictor^fl/fl mice at 7 days after immunization. Scale bars show 100 μm. Right shows calculated GC size. (D) Measurements of anti-NP immunoglobulins in serum from immunized mice, presented as absorbance at 450 nm (A₄₅₀) in ELISA. (E and F) Flow cytometry of Tfh cells in spleen from WT and OX40^creRptor^fl/fl mice (E) or OX40^creRictor^fl/fl mice (F) at 7 days after intraperitoneal immunization of mice with NP-OVA plus LPS in alum. Right shows the frequency and number of Tfh cells. (G and H) Flow cytometry of Tfh cells (PD-1^hiCXCR5⁺ in CD45.2⁺B220^–CD4⁺TCRβ⁺ donor-derived cells) in spleen from mixed BM chimeras reconstituted with BM cells from CD45.1⁺ mice and WT or OX40^creRptor^fl/fl (G) or OX40^creRictor^fl/fl (H) mice at 7 days after intraperitoneal immunization of mice with NP-OVA plus LPS in alum. Right shows the frequency of Tfh cells. (I) Naive CD4⁺ T cells from WT OT-II, OX40^creRptor^fl/fl OT-II or OX40^creRictor^fl/fl OT-II mice were transferred into CD45.1⁺ mice. The recipient mice were immunized with NP-OVA plus LPS in alum through footpad, and Bcl6^hiCXCR5⁺ Tfh cells among CD45.2⁺B220^–CD4⁺TCRβ⁺ donor cells in popliteal lymph nodes were analyzed at 5 days after immunization. Right shows the number of Tfh cells. NS, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 (one-way ANOVA). Data are pooled from three (A–F) or representative of two (G–I) independent experiments. Error bars represent SEM. See also Figure S5.

Figure 6. Raptor and Rictor Are Required for ICOS-Mediated Signaling Events and Anabolic Metabolism

(A–D) Mice were immunized by NP-OVA plus LPS in alum, and splenic Tfh cells were analyzed for the expression of Bcl6 and IRF4 (A), ICOS (B, left), Klf2, Il7r and Ccr7 (C), and CD62L, CD127, and CD69 (D). (B, right) ICOS expression on Tfh cells from PPs under steady state. (E) Naive CD4⁺ T cells from WT, OX40^creRptor^fl/fl or OX40^creRictor^fl/fl mice were activated by anti-CD3 and anti-CD28 for 3 days, rested for 3 hr, and re-stimulated with anti-ICOS. Phosphorylation of S6 and Akt at Ser 473 was examined by flow cytometry at indicated time points. (F–H) Metabolic assays in activated T cells that were re-stimulated with indicated stimuli for 24 hr, with the final 4 hr labeled with ¹⁴C-acetate to measure de novo lipogenesis (F and G) or with [3-³H]-glucose to measure glycolysis (H). (I) Glucose uptake as measured by 2-NBDG labeling in Tfh cells in spleen from WT, OX40^creRptor^fl/fl, and OX40^creRictor^fl/fl mice at 7 days after NP-OVA immunization. (J) Frequencies of GC and Tfh cells in spleen from WT and OX40^creMyc^fl/fl mice at 7 days after NP-OVA immunization. (K) Frequencies of GC and Tfh cells in PPs from WT and OX40^creMyc^fl/fl mice under steady state. *p < 0.05, **p < 0.01, ***p < 0.001 (Mann-Whitey test). Data are representative of two (A, C–J) or at least three (B and K) independent experiments. Error bars represent SEM. See also Figure S6.

Figure 7. Glucose Metabolism Promotes Tfh Cell Differentiation

(A and B) Purified CD4⁺ T cells were polarized in vitro to generate activated T cells (Tact), Tfh-like, Th1, and Th17 cells for high-resolution metabolomics analyses. The top 250 differentially observed metabolites are shown by heatmap (A) and principle component analysis (B). (C) T cells were polarized in vitro to generate Tfh-like cells, in the presence of vehicle, 2-DG (250 μM), or rotenone (5 nM), for 3.5 days, followed by analysis of IL-21 expression. The fold change in the number of IL-21⁺ CD4⁺ cells in comparison to vehicle treated cells is shown. (D and E) T cells were activated or polarized in vitro to generate Tfh-like cells for 3.5 days, followed by analysis of Glut1 expression by flow cytometry (D; right, the MFI of Glut1) or immunoblot (E). (F) Flow cytometry of Glut1 expression on Tfh (B220^–CD4⁺PD-1^hiCXCR5⁺) and non-Tfh cells (B220^–CD4⁺PD-1^loCXCR5^–) from PPs. Right shows the MFI of Glut1. (G) T cells from WT or Slc2a1-Tg mice were polarized in vitro to generate Tfh-like cells. After 3 days, the cells were treated with vehicle (DMSO) or rapamycin (50 nM) for 20 hr, followed by analysis of IL-21 expression. Graph represents the fold change in the number of IL-21⁺ CD4⁺ cells as a result of rapamycin treatment compared with that of vehicle treatment. (H–J) Flow cytometry of GC B cells (H), IgA⁺ B cells (I), and Tfh cells (J) in PPs from WT and Slc2a1-Tg mice. Right shows the frequency and number of indicated cell population. (K) Flow cytometry of Tfh cells in the draining lymph nodes of WT and Slc2a1-Tg mice at day 10 after immunization with KLH. *p < 0.05, **p < 0.01 (Student’s t test). Data are representative of three (C and D) or pooled from at least three (A, B, G–K) or five (E and F) independent experiments. Error bars represent SEM. See also Figure S7.