Figure 1.
Depletion of splenic B cells in Sle1-hCD19Tg mice leads to a significant reduction of Tfh cells.
(A) Spleens were harvested from 5 to 12-month old Sle1-hCD19Tg mice and age-matched C57BL/6 (B6) control mice. Numbers of GC B cells (B220+CD19+PNA+FAShighIgDlow) and Tfh cells (CD4+CXCR5+PD1high) were enumerated by flow cytometry analysis and plotted in the graph. Each dot represents a single mouse of indicated genotype. (B) Immunofluorescence images of frozen spleen sections from a 10-month old Sle1-hCD19Tg mice. Cryosection are stained with PNA (for GC B cells, blue), IgD (naïve B cells, magenta), CD4 (green), and PD1 (red). Tfh cells are CD4 and PD1 double positive and therefore appear yellow. Majority of Tfh cells colocalize with PNA+ GC B cells. (C–E) Sle1-hCD19Tg mice were treated with a single dose of an anti-CD19 depleting MAb (MEDI-551) or control MAb (10 mg mg/kg) and spleen cells were collected for FACS analysis seven days after MAb administration. (C) Bar graphs show the number of B cells (B220+CD19+), GC B cells (PNA+FAShighIgDlow) and Tfh cells (CXCR5+PD1high or CXCR5+Bcl-6+). Bars represent the mean value for each group and error bars are standard error of the mean. (D) Representative FACS contour plots showing CXCR5 versus PD1 staining gated on CD4+CD44high cells from Sle1-hCD19Tg mice with indicated treatment. Gates have been drawn around the populations representing Tfh cells. (E) Bar graph showing the number of non-Tfh cells within CD4+CD44high effector and memory T cell gate (CD44highCXCR5−PD1low/−). *** P<0.001 **, P<0.01 *, P<0.05. Data in (C–E) are pooled from two independent experiments.
Figure 2.
Reduction of GC B cells decreases the number of Tfh cells in NZB/W F1 mice.
(A–F) Female NZB/W F1 mice at 5–6 months of age were treated with anti-CD20 (1 mg mg/mouse), anti-CD40L-TM (1 mg mg/mouse), combination of anti-CD20 and CD40L-TM (1 mg each mg each/mouse) or isotype control antibodies (1 mg each mg each/mouse) intravenously on days 0 and 2. Mice were sacrificed and splenocytes were analyzed by FACS at day 7 post first antibody dosing. Bar graph shows numbers per spleen for (A) B cells (CD19+B220+), (B) GC B cells (PNA+FAShighIgDlow), (C) Tfh cells (CD4+CD44highCXCR5+PD1high) and (D) Tfh cells (CD4+CD44highCXCR5+Bcl-6+). (E) Bar graph shows the mean fluorescence intensity of Bcl-6 for Tfh (CD4+CD44highCXCR5+Bcl-6+) cells. ***P<0.001 **, P<0.01 *, P<0.05. Bars represent the mean value for each group and error bars are standard error of the mean. Data are pooled from three independent experiments. (F) Representative FACS contour plots showing PD1 versus CXCR5 staining gated on CD4+CD44high cells from mice with the indicated treatment 7 days after MAb dosing. Oval gates show the Tfh (CXCR5+PD1high) subsets.
Figure 3.
Depletion of B cells in SRBC immunized mice with anti-CD20 plus -CD40L-TM treatment leads to accelerated loss of mature Tfh cells.
(A) Diagrammatic representation of experimental protocol. BALB/c mice were immunized with SRBC and treated with either anti-CD20 (0.25 mg mg/mouse) + anti-CD40L-TM (0.40 mg mg/mouse) or control antibodies at day 9 post treatment. Mice were sacrificed on days 5 and 8 post antibody treatment (14 and 17 days post immunization). (B) Relative expression of IL-21 and Bcl-6 in FACS sorted CD4+CD44highCD62Llow cells from mice treated with either anti-CD20+anti-CD40L-TM or control antibodies. Data are normalized with expression of housekeeping gene 18S. (C) Graphs show number of total B cells (B220+CD19+), GC B cells (PNA+FAShighIgDlow), Tfh cells (CXCR5+PD1high and CXCR5+Bcl-6+ cells) gated on CD4+CD44high T cells. ***P<0.001 **, P<0.01 *, P<0.05. Data are pooled from two independent experiments. Bars represent the mean value for each group and error bars are standard error of the mean. (D) FACS contour plots show CXCR5 versus PD1 and CXCR5 versus Bcl-6 after gated on CD4+CD44high T cells from mice with the indicated treatment. Oval gates show the Tfh subsets.
Figure 4.
Elimination of GC B cells in established GC reduces the number of mature Tfh cells in SRBC immunized BALB/c mice.
(A) A schematic view of SRBC immunization and anti-CD40L-TM treatment protocol: a cohort of BALB/c naïve mice were immunized with SRBC at day 0 and were treated on days 9 and 11 with indicated amounts of either anti-CD40L-TM or isotype control. (B) FACS contour plots show the gate and percentage of GC (PNA+FAShigh) cells after gated on live B cells, from mice treated with MAbs or naïve mice. Bar graph shows number of splenic GC B cells in mice treated with increasing doses of anti-CD40L-TM or isotype control. (C–D) Mice are same as in (B). FACS contour plots show the gate and percentage of Tfh cells (CXCR5+PD1high) in (C) and (CXCR5+Bcl-6+) in (D) after gating on CD4+CD44high cells and bar graphs show numbers of Tfh cells per spleen. ***P<0.001 **, P<0.01, ns indicates not significant. Bars represent the mean value for each group and error bars are standard error of the mean. Data are pooled from three or more independent experiments.
Figure 5.
LtβR-Ig treatment in SRBC immunized mice disrupts FDC but does not lower the number of Tfh cells.
(A) Diagrammatic representation of experiment protocol. BALB/c mice were immunized with SRBC on day 0 and treated with 0.25 mg of recombinant Lt mg of recombinant LtβR-Ig or PBS on days 9 and 11. Spleens were harvested on day 17 post immunization and subjected to histology and FACS analysis. (B) Cryosections of spleens stained with PNA (green) to detect GC B cells, IgD (blue) to detect follicular B cells and CD35 (red) to detect FDC. Images were captured and analyzed by microscopy. Bar scale represents 500 µm. (C) Sections were stained with PNA (green), PD1 (red) to detect Tfh, IgD (blue) and CD35 (magenta). Bar scale represents 50 µm. (D) The bar graphs show numbers of GC B cells (PNA+FAShigh) and Tfh (CXCR5+PD1high) cells from either LtβR-Ig or PBS treated mice. Data are pooled from at least eight mice analyzed in two experiments and each dot represents an individual mouse. Bars represent the mean value for each group and error bars are standard error of the mean.
Figure 6.
Blocking ICOS but not CD28 signaling results in decreased Tfh in the maintenance phase of the SRBC immunization model.
(A) BALB/c mice were immunized with SRBC on day 0 and treated with 0.5 mg of recombinant CTLA4 mg of recombinant CTLA4-Ig, anti-ICOS-TM MAb or isotype control MAbs on days 9, 11 and 13. Spleens were harvested on day 14 and 16 post immunization and subjected to flow cytometry analysis. (B) Graphs show the numbers of GC B cells (PNA+FAShighIgDlow) (A) and Tfh cells (CXCR5+Bcl-6+) over time in mice treated with MAbs or PBS. Cell numbers on day 9 were pooled form untreated BALB/c mice receiving the same immunization. Cell numbers on day 14 and 16 from CTLA-4 Ig or anti-ICOS-TM MAb treated mice were compared to mice dosed with PBS. Data are pooled from two independent experiments with n = 4∼8 per group. Error bars are standard error of the mean. Numbers of GC B cells and Tfh cells in anti-ICOS-TM treated mice on day 14 and 16 were significantly lower than those in PBS treated mice (***, p<0.001).
Figure 7.
Maintenance of Tfh cells in NZB/W F1 mice requires ICOS signaling.
Female NZB/WF1 mice at 6 months of age were treated with CTLA4-Ig (1 mg mg/mouse) anti-ICOS-TM (1 mg mg/mouse), or PBS at days 0, 2 and 4. Mice were sacrificed and splenocytes were analyzed by FACS at day 7 post first antibody dosing. (A–D) Graphs show number of total B cells (B220+CD19+) in (A), GC B cells (PNA+FAShighIgDlow) in (B), Total CD44high cells in (C) and Tfh cells (CXCR5+PD1high) in (D). **, P<0.01 *, P<0.05. Data are representative of three independent experiments. Bars represent the mean value for each group and error bars are standard error of the mean.
Figure 8.
Model of Tfh differentiation and maintenance.
(A) Our data suggest that in both immunization and autoimmune mouse models, GC B cells are actively providing signals to maintain optimal Tfh numbers during the course of the GC reaction. In the immunization model, ICOSL:ICOS but not B7.2:CD28 signaling, is required for Tfh maintenance. Interestingly, in the spontaneous autoimmune model both signals are required to maintain optimal numbers of mature Tfh. (B) Previous studies have shown that Tfh differentiation is a multistage process: Dendritic cells initiate the differentiation of naïve T cells into pre-Tfh. Upon upregulation of CXCR5, these pre-Tfh are able to enter the follicle and interact with cognate B cells. This interaction with B cells drives the full polarization of Tfh into mature Tfh which is required for the germinal center reaction. Our data in an immunization model demonstrate that once fully mature, Tfh continue to require signals from GC B cells to sustain their maintenance. We further show that the same B cell dependence continues to play a role in spontaneous models of autoimmunity.