Human T-follicular helper and T-follicular regulatory cell maintenance is independent of germinal centers

Abstract

The monoclonal anti-CD20 antibody rituximab (RTX) depletes B cells in the treatment of lymphoma and autoimmune disease, and contributes to alloantibody reduction in transplantation across immunologic barriers. The effects of RTX on T cells are less well described. T-follicular helper (Tfh) cells provide growth and differentiation signals to germinal center (GC) B cells to support antibody production, and suppressive T-follicular regulatory (Tfr) cells regulate this response. In mice, both Tfh and Tfr are absolutely dependent on B cells for their formation and on the GC for their maintenance. In this study, we demonstrate that RTX treatment results in a lack of GC B cells in human lymph nodes without affecting the Tfh or Tfr cell populations. These data demonstrate that human Tfh and Tfr do not require an ongoing GC response for their maintenance. The persistence of Tfh and Tfr following RTX treatment may permit rapid reconstitution of the pathological GC response once the B-cell pool begins to recover. Strategies for maintaining remission after RTX therapy will need to take this persistence of Tfh into account.

Figure 1

RTX effectively depletes naïve circulating B cells but not memory B cells. (A) Flow cytometric contour plots of CD19 and CD20, and numbers of CD20⁺ B cells, (B) CD19⁺ B cells, and (C) on peripheral blood lymphocytes taken from patients who have not (control patients, top) or have been (lower) treated with RTX. (D) Flow cytometric contour plot of naïve, IgD⁺ memory, IgD⁻ memory, plasma cell, and CD27⁻IgD⁻ subsets of CD19⁺ cells and MFI of CD38 on these subsets in a representative untreated patient. (E) Flow cytometric contour plot and MFI of CD38 of the same subsets in a representative RTX-treated patient. Bar graphs of (F) peripheral blood IgD⁺CD27⁻CD19⁺ naïve B cells, (G) IgD⁺CD27⁺CD19⁺ memory B cells, (H) CD38^hiIgD⁻CD27^hiCD19⁺ plasma cells, (I) IgD⁻CD27⁺CD19⁺ memory B cells, and (J) IgD⁻CD27⁻CD19⁺ B cells. For all bar graphs one symbol represents one individual, and the height of the bar represents the mean.

Figure 2

GC B cells are not detectable after RTX treatment. (A) Contour plots of CD19 and CD20 on lymphocytes from iliac lymph node of RTX-treated patients and controls. The number (B) and proportion (C) of CD20⁺ B cells in the lymph nodes of RTX-treated patients and controls. Contour plots (D), the percentage (E), and total number (F) of CD38⁺Bcl-6⁺CD19⁺ GC B cells in the lymph nodes of RTX-treated patients and controls. (G) Mean fluorescence intensity of CD20 on CD38⁻Bcl-6⁻ non-GC B cells and CD38⁺Bcl-6⁺ GC B cells from control patients. (H) Mean fluorescence intensity of CD20 on CD19⁺ lymph node B cells from RTX-treated patients and controls. In bar graphs, one symbol represents one individual, and the height of the bar represents the mean.

Figure 3

Characterization of CD19⁺ B cells in the lymph node suggests the cells resistant to removal are memory cells. Contour plots of cell subsets and MFI of CD38 on CD19⁺ cells from (A) iliac lymph nodes of controls, and (B) RTX-treated patients. Bar graphs of (C) lymph node CD27⁻IgD⁺CD19⁺ naïve B cells, (D) CD27⁺IgD⁺CD19⁺ memory B cells, (E) CD27^hiCD38^hiIgD⁻CD19⁺ plasma cells, (F) CD27⁺IgD⁻CD19⁺ memory B cells, and (G) CD27⁻IgD⁻CD19⁺ B cells. For all bar graphs, one symbol represents one individual, and the height of the bar represents the mean. (H) Detailed analysis of CD19^loCD20^lo cells alone in RTX-treated patients.

Figure 4

Loss of the GC after RTX treatment did not reduce Tfh numbers in the lymph node or peripheral blood. (A) Flow cytometric contour plots, (B) proportion, and (C) total number of CXCR5⁺CD57⁺CD4⁺ Tfh from iliac lymph nodes taken prior to kidney transplantation in RTX-treated patients and controls. (D) Contour plots, and (E) quantitation of CXCR3⁺CCR6⁻Th1-like Tfh (top left quadrant gate, panel D), CXCR3⁻CCR6⁻Th2-like Tfh (lower left quadrant gate, panel D), and CXCR3⁻CCR6⁺Th17-like lymph node Tfh cells (lower right quadrant gate, panel D). (F) Contour plots, and (G) bar graphs of peripheral blood CXCR5⁺CD4⁺ Tfh-like cells from the same patients in (A-E). In graphs, one symbol represents one individual, and the height of the bar represents the mean.

Figure 5

RTX did not alter Treg or Tfr numbers within the lymph node. (A) Flow cytometric contour plots, (B) quantitation of the proportion, and (C) total number of FOXP3⁺CD127⁻CD4⁺ Treg from iliac lymph nodes taken prior to kidney transplantation in RTX-treated patients and controls. (D) Contour plots, (E) proportion, and (F) total number of CXCR5⁺CD57⁺FOXP3⁺CD127⁻CD4⁺ Tfr from iliac lymph nodes. In graphs, one symbol represents one individual, and the height of the bar represents the mean. (G) IgA production in CD19⁺CD27⁺IgD⁻ B cells cultured alone, coculture of B cells with CXCR5⁺CD57⁺CD25⁻CD4⁺ Tfh cells, or coculture of B cells with Tfh and CXCR5⁺CD57⁺CD127⁻CD25⁺CD4⁺ Tfr flow-sorted from iliac lymph nodes. Data in (G) are a representative biological replicate of 3 independent biological replicates. The height of the bars represent the mean of 4 technical replicates, and the error bars show the standard deviation. n.d., not detected above background.