Circulating T follicular regulatory and helper cells have memory-like properties

Abstract

Follicular Tregs (Tfr cells) inhibit antibody production, whereas follicular Th cells (Tfh cells) stimulate it. Tfr cells are found in blood; however, relatively little is known about the developmental signals for these cells or their functions. Here we demonstrated that circulating Tfr and Tfh cells share properties of memory cells and are distinct from effector Tfr and Tfh cells found within lymph nodes (LNs). Circulating memory-like Tfh cells were potently reactivated by DCs, homed to germinal centers, and produced more cytokines than did effector LN Tfh cells. Circulating memory-like Tfr cells persisted for long periods of time in vivo and homed to germinal centers after reactivation. Effector LN Tfr cells suppressed Tfh cell activation and production of cytokines, including IL-21, and inhibited class switch recombination and B cell activation. The suppressive function of this population was not dependent on specific antigen. Similar to LN effector Tfr cells, circulating Tfr cells also suppressed B and Tfh cells, but with a much lower capacity. Our data indicate that circulating memory-like Tfr cells are less suppressive than LN Tfr cells and circulating memory-like Tfh cells are more potent than LN effector Tfh cells; therefore, these circulating populations can provide rapid and robust systemic B cell help during secondary antigen exposure.

Figure 9. Circulating Tfr cells have lower suppressive capacity.

(A–C) Suppression assays. Tfr cells (CD4⁺ICOS⁺CXCR5⁺GITR⁺CD19^–; see Supplemental Figure 3) were cultured with dLN Tfh cells and B cells. (A) GL7 and IgG1 expression on B cells. Left: Plots pregated on CD19⁺IA⁺; number indicates percent of cells in gate. Right: Quantification. (B) Quantification of GL7⁺ B cells. (C) Ki67 staining in Tfr cells. Histograms are pregated on CD4⁺CD19^–FOXP3⁺; number indicates percent of cells in gate. (D and E) Blood Tfh and Tfr cells elicit more potent B cell responses. 2 × 10⁴ blood or dLN CD4⁺ICOS⁺CXCR5⁺ cells (Tfh and Tfr) from actin-CFP Foxp3-GFP mice were transferred to Cd28^–/– mice that were immunized. dLNs were harvested, and B cells were stained for (D) GL7 and FAS (pregated on CD19⁺) or (E) CD138 and CD19 (pregated on live cells). Left: Representative plots; number indicates percent of cells in gate. Right: Quantification of GC B cells. (F) Similar experiments as in D, except serum was collected on d10 after transfer. (G) Mice were infected with influenza PR8 virus, and blood or dLN CD4⁺ICOS⁺CXCR5⁺ cells were transferred to Cd28^–/– mice and infected with PR8 influenza. Serum IgG was measured 12 days later. Data are mean ± SEM of pooled data from 3 (A, B, and D) or 2 (G) independent experiments or are representative of 3 independent experiments (C and F). *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Student’s t test (G) or 1-way ANOVA with Tukey post-test (A and D).

Figure 8. Tfr cells can suppress GC B cells and do not require specific antigen for suppressive capacity.

(A–D) Tfr cells suppress GC B cells. 3 × 10⁴ Tfh cells and 1.5 × 10⁴ Tfr cells or 1.5 × 10⁴ non-Tfr Tregs (see Supplemental Figure 3) were cultured with 5 × 10⁴ GC B cells (sorted as CD19⁺GL7⁺CD4^– from the dLN of WT mice immunized 12 days previously with NP-OVA) in the presence of NP-OVA for 6 days. (A) Quantification of Ki67⁺ Tfh cells. (B) Quantification of IgG1⁺GL7⁺ B cells. (C) Quantification of GL7⁺ B cells. (D) Quantification of B7-1 expression on B cells. (E and F) In vitro antigen-specific suppression assays. 3 × 10⁴ Tfh and 5 × 10⁴ total B cells from the dLN of NP-OVA immunized mice were cultured with 1.5 × 10⁴ dLN Tfr cells (sorted as CD4⁺ICOS⁺CXCR5⁺GITR⁺CD19^–) from either NP-OVA– or NP-HEL–immunized mice in the presence of NP-OVA for 6 days. (E) B cell GL7 and IgG1 expression. Left: Representative plots pregated on CD19⁺IA⁺; number indicates percent of cells in gate. Right: Quantification. (F) Quantification of Ki67⁺ Tfh cells. Data indicate mean ± SEM of 3–4 replicate wells and are representative of 3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, 1-way ANOVA with Tukey post-test.

Figure 7. Tfr cells suppress Tfh and B cells.

(A–E) dLN Tfr suppression assays. 3 × 10⁴ dLN Tfh cells (CD4⁺ICOS⁺CXCR5⁺GITR^–CD19^–; see Supplemental Figure 3) were plated with 1.5 × 10⁴ Tfr cells (CD4⁺ICOS⁺CXCR5⁺GITR⁺CD19^–) and either 5 × 10⁴ B cells or 5 × 10⁴ DCs (isolated from the dLN of WT immunized mice) for 5–6 days with anti-CD3 and anti-IgM. (A) Tfh cells were gated as CD4⁺CD19^–FOXP3^– and stained for CXCR5. (B and C) Tfh cells (cultured and gated as in A) were intracellularly stained for (B) BCL6 or (C) Ki67. (D) Intracellular cytokine staining in Tfh cells from suppression assays with B cells. (E) Quantification of IL-4 and IL-21 in culture supernatants. (F–J) 3 × 10⁴ Tfh cells (CD4⁺ICOS⁺CXCR5⁺FOXP3^–CD19^–) and 5 × 10⁴ B cells were plated with 1.5 × 10⁴ Tfr cells (CD4⁺ICOS⁺CXCR5⁺FOXP3⁺CD19^–) or 1.5 × 10⁴ conventional dLN Tregs (CD4⁺ICOS^–CXCR5^–FOXP3⁺) for 6 days. (F) Ki67 in Tfh cells from cultures. Left: Plots pregated on CD4⁺FOXP3^–CD19^–; number indicates percent of cells in gate. Right: Quantification. (G) B cells were quantified for GL7 expression. (H) IgG1 expression in B cells. (I) Dose response of suppression of IgG1 expression in B cells cultured as indicated. (J) B7-1 expression on B cells. Data are mean ± SEM of 3–4 replicate wells and are representative of 3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Student’s t test (A–E) or 1-way ANOVA with Tukey post-test (F–J).

Figure 6. Circulating Tfh cells require DCs for enhanced cytokine production.

(A) In vitro activation assay. dLN or blood Tfh cells (sorted as CD4⁺ICOS⁺CXCR5⁺GITR^–CD19^–; see Supplemental Figure 3) were plated with B cells or CD11c⁺MHCII⁺ DCs (sorted from dLN of WT immunized mice) plus anti-IgM and anti-CD3 for 6 days. Left: Plots pregated on CD4⁺CD19^–; number indicates percent of cells in gate. Right: Quantification of BCL6⁺Ki67⁺ cells. (B and C) Intracellular cytokine staining of Tfh cells from cocultures with B cells or DCs as in A, except samples were stimulated with PMA/ionomycin for 4 hours in the presence of Golgistop prior to staining. (B) Representative plots; numbers indicate percentage of cells in the each respective quadrant. (C) Quantification. (D) Quantification of cytokines measured in culture supernatants of cultures as in A. ND, not detected. Graphs represent mean ± SEM of pooled data from 3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Student’s t test (C and D) or 1-way ANOVA with Tukey post-test (A).

Figure 5. Circulating Tfr and Tfh cells are memory-like cells that persist in vivo.

(A and B) Parabiosis experiments. (A) Actin-CFP Foxp3-GFP mice were immunized with NP-OVA and surgically joined via parabiosis to WT mice. 6 days after separation, the WT mate was immunized. (B) CFP⁺ population in basal or CD4⁺CXCR5⁺ populations in the dLN, nondraining cervical LN, spleen, blood, and skin. (C) Blood Tfr cells dominated the CXCR5⁺FOXP3⁺ population. Experiments were performed as in A, except actin-CFP Foxp3-GFP mice were joined with WT Foxp3-GFP mice. The indicated populations were analyzed for CFP chimerism and compared with the respective basal populations (CXCR5^–FOXP3⁺ or CXCR5^–FOXP3^–) in the nondraining LN. Comparison of Tfr cell incidence between immunized mice and unimmunized controls is also shown. (D–I) Memory transfer experiments. Circulating Tfh and Tfr cells persisted for 30 days in vivo. (D) 2 × 10⁴ blood CFP⁺ICOS⁺CXCR5⁺ (Tfr and Tfh) cells (see Supplemental Figure 3) from blood of immunized mice were transferred to recipients that were immunized 30 days later with NP-OVA. (E) Representative plots; number indicates percent of cells in gate. (F) CFP⁺ cells (percentage of CD4⁺ T cells) in dLN and in skin at the immunization site. (G) Tfr cells (percentage of CFP⁺ cells) in dLN. (H) CXCR5 expression on CFP⁺ cells in dLN. End, endogenous cells. (I) Transferred blood CD4⁺ Treg populations (percentage of total FOXP3⁺ cells). Data are pooled from 3 (A and B) or 5 (C) replicate surgeries, or pooled from 4 (D–F) or 2 (I) independent transfers. *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Student’s t test.

Figure 4. Circulating Tfr cells migrate to LNs and interact with GC B cells.

(A–C) Actin-CFP Foxp3-GFP mice were immunized with NP-OVA, and 2 × 10⁴ sorted blood CXCR5⁺ cells (containing Tfh and Tfr cells; see Supplemental Figure 3 for sorting) were transferred to Cd28^–/– mice that were immunized with NP-OVA. (A) Transferred cells (CFP⁺) were identified 7 days later according to gates (iLN, inguinal dLN; aLN, axillary LN; cLN, cervical LN; mLN, mesenteric LN; Skin, skin at immunization site). Plots are pregated on CD4⁺CD19^–; number indicates percent of cells in gate. (B) Dual-color overlay plots showing ICOS and CXCR5 expression on CFP⁺ transferred cells (red) or endogenous Cd28^–/– CD4⁺ T cells (blue). Numbers indicate percent of cells in gate for each cell type. (C) Quantification of FOXP3⁺ cell percentage. (D) 1 × 10⁴ blood Tfr cells from NP-OVA immunized CD45.2 mice were transferred to CD45.1 mice that were immunized. 7 days later, dLNs were stained for CD45.2 (blood Tfr), IgG1, and either IgD (left) or FOXP3 (right). Original magnification, ×200. (E) 2 × 10⁴ blood Tfh cells from NP-OVA immunized CD45.2 mice were transferred to CD45.1 mice that were then immunized. 7 days later, dLNs were stained for CD45.2 (blood Tfh), IgD, and IgG1. Data are from transfers of Tfr or Tfh cells sorted from 20 pooled mice into a single recipient and are representative of ≥2 (A–C) or ≥3 (D and E) individual experiments. Scale bars: 20 μm; 5 μm (E, inset).

Figure 3. Circulating Tfr cells have a distinct phenotype when exiting the LN.

(A and B) Tfr cells (A) or Tfh cells (B) in the dLN and per ml blood after FTY720 treatment for 5 days (d2, d4, and d6 after immunization, followed by analysis on d7). (C) Tfr and Tfh cells (per ml blood) after FTY720 treatment during the last 3 or 6 hours of a 7-day immunization. (D) Representative plots of total CD4⁺ICOS⁺CXCR5⁺ cells from the dLN, efferent lymph (Lymph), and blood of immunized mice. Plots are pregated on CD4⁺CD19^–; number indicates percent of cells in gate. (E and F) Percentage of Tfr (E) or Tfh (F) cells in the dLN, efferent lymph, and blood of WT mice immunized (imm) or not (un) 7 days previously. (G–I) ICOS (G), Ki67 (H), and CXCR5 (I) expression on dLN, lymph, and blood Tfr and Tfh cells from experiments as in D. Data are mean ± SEM with 5 mice per group and representative of ≥3 independent experiments (A–C) or with mice from 5 replicate surgeries (D–I). *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Student’s t test (A, B, E, and F) or 1-way ANOVA with Tukey post-test (C, G, H, and I).

Figure 2. Circulating Tfr cells require DCs for development.

(A–C) B cells are not required for circulating Tfr and Tfh cell differentiation. WT or μMT mice were immunized s.c. with NP-OVA; 7 days later, CD4⁺ICOS⁺CXCR5⁺ cells were analyzed. (A) Representative plots (pregated on CD4⁺CD19^–; number indicates percent in gate) and quantification. (B) Tfr (CD4⁺ICOS⁺CXCR5⁺FOXP3⁺CD19^–) and (C) Tfh (CD4⁺ICOS⁺CXCR5⁺FOXP3^–CD19^–) cells after immunization. (D and E) WT or CD11c-DTR BM chimeras were immunized s.c. with NP-OVA, and DT was administered on d0, d2, d4, and d6. (D) dLN analysis of CD11c⁺MHCII⁺ DCs; number indicates percent of cells in gate. (E) Quantification of Ki67⁺ Tfr cells (percentage of total FOXP3⁺ cells) and Ki67⁺ Tfh cells (percentage of total CD4⁺ T cells) in dLN and blood. (F–H) DC transfer experiments. WT BMDCs were pulsed with NP-OVA and transferred s.c. to WT mice (DC Trans); 7 days later, dLNs and blood were analyzed. WT mice that received no transfer were included as controls. (F) Representative plots (pregated on CD4⁺CD19^–; number indicates percent of cells in gate). (G and H) Quantification of dLN and blood Tfr (G) and Tfh (H) cells. Data are mean ± SEM with 5 mice per group and representative of ≥3 independent experiments. **P < 0.01, ***P < 0.001, unpaired Student’s t test.

Figure 1. Phenotype of circulating Tfh and Tfr cells.

(A–J) WT mice were immunized s.c. with NP-OVA, and dLN and blood were analyzed for Tfh (A–E) and Tfr (F–J) cells on day 7. Unimm., unimmunized control. (A and F) Plots pregated on CD4⁺FOXP3^–CD19^– (A) or CD4⁺FOXP3⁺CD19^– (F); number indicates percent in gate. (B and G) Quantification of cell subpopulations from plots. (C–E and H–J) Quantification of CXCR5 (C and H), ICOS (D and I), and Ki67 (E and J) expression. Total CD4⁺CD19^– (Total CD4; C–E) and total CD4⁺FOXP3⁺CD19^– cells (Total FOXP3⁺; H–J), are included as controls. (K) Tfh and Tfr cells in dLN and blood on d10 after PR8 influenza (PR8 Flu) infection. (L) Tfh and Tfr cells in spleen and blood on d7 after LCMV CL13 infection. Uninf., uninfected control. Data are mean ± SEM with 5 mice per group and representative of ≥3 independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001, unpaired Student’s t test.