(A) Comparison of the percentage of CD4 cells expressing FoxP3 in PBMC (n= 14), benign tonsils (n=12) and FL nodal biopsies (n=11) (P<0.0001). (B) Tumor-infiltrating regulatory T cells exhibit suppressive function. Sorted CD4⁺CD25⁻ cells (Tresp) from a FL specimen were labeled with CFSE. Subsequently, these cells were co-cultured with sorted autologous CD4⁺CD25⁺ cells (Treg) in a ratio of 1: 0; 1: 0.25; 1:0.5; and 1:1. Proliferation of Tresp under the stimulation of pooled allogeneic antigen presenting cells (in 1:1 ratio) was measured by the percentage of CFSE^dim cells, as indicated. The CFSE-labeled Tresp had a distinct scatter pattern due to allogeneic stimulation. As such, the histogram represented gated CFSE-positive cells. Results are representative of 3 experiments.

(A) Expression of chemokine receptors, CCR4 and CCR5, on Treg from FL and tonsil specimens. The MFI of each staining was indicated. The open histograms represent follicular lymphoma samples, the shaded histograms represent tonsil specimens and the blackened histograms represent the isotype control. Results are representative of 3 experiments. (B) Comparing the production of chemokines, CCL17 (open bar) and CCL22 (hatched bar), by purified B cells from either tonsil or FL specimens. B cells were cultured without stimulation for 4 days (left panel) or with stimulation of soluable CD40L at 700 ng/ml for 4 days (right panel).

(A) B cells from FL, but not B cells from PBMC, are sufficient to induce FoxP3 expression in autologous conventional T cells without artificial T cell receptor (TCR) stimulation. Sorted CD4⁺CD25⁻ T cells from PBMC (left panel) or from a FL specimen (right panel) were cultured in medium alone or with purified autologous B cells, as indicated, for 96–120 hours. The cells were harvested and the percentage of CD4⁺FoxP3⁺ cells was assayed. Results are representative of 5 experiments. (B) Induction of Treg requires cell-cell contact. The same set-up as (A), except that B and T cells in each well were separated by a 3μm transwell membrane. Results are representative of 3 experiments. (C) Treg conversion is attributed to unique properties of malignant B cells, rather than that of tumor-infiltrating T cells. Tumor-infiltrating CD4⁺CD25⁻ T cells isolated from FL were co-cultured with B cells from PBMC (left panel) for 96–120 hours. Conversely, CD4⁺CD25⁻ T cells isolated from PBMC were co-cultured with B cells from FL (right panel) for 96–120 hours. The cells were then harvested and the percentage of FoxP3 expressing CD4 cells was assayed. Results are representative of 3 experiments. All data collected for this set of experiments were summarized in . At the end of 96–120-hour co-culture, cells were stained with CD4-FITC and FoxP3-APC before subjected to FACS analysis. Dead cells were excluded from analysis by their small forward scatter. Cells were first gated on CD4, and subsequently gated on FoxP3. The CD4+ and FoxP3+ gates were drawn based on their isotype controls.

Sorted CD4⁺CD25⁻ cells from a FL or PBMC specimen were co-cultured with purified respective autologous B cells for 96–120 hours. Then, the cell mixture was stimulated with PMA and ionomycin for 3 hours, after which intracellular staining of FoxP3, IL-2 and IFN-γ were performed. The frequency of FoxP3⁺ cells was shown as a percentage with B cells and CD4⁺ cells as the denominator. Cytokine production by FoxP3-negative CD4 cells depicted in the right lower quadrant of each plot served as an internal positive control. Cells were first gated on FoxP3 and subsequently on IL-2 or IFN-γ. Results are representative of 2 experiments.