IL-2 and IL-15 each mediate de novo induction of FOXP3 expression in human tumor antigen-specific CD8 T cells

Abstract

Although FOXP3 is primarily expressed by regulatory CD4 T cells (Treg) in vivo, polyclonal activation of human CD8 T cells can result in the expression of FOXP3 in a fraction of CD8 T cells. However, the cellular lineage and mechanism of FOXP3 induction in CD8 T cells remain unclear. Here, we demonstrate that interleukin-2 (IL-2) induces FOXP3 expression in OKT3-stimulated or antigen-stimulated CD8 T cells, indicating that FOXP3 expression is neither limited to a unique subset of CD8 T cells nor dependent on the mode of T-cell receptor stimulation. In the absence of IL-2, antigen stimulation resulted in T-cell activation and acquisition of effector function without induction of FOXP3, indicating that acquisition of effector function is independent of induction of FOXP3 expression in CD8 T cells. Interestingly, IL-15, but not IL-7 or IL-21, also led to de novo induction of FOXP3 in antigen-specific CD8 T cells, suggesting that signaling by IL-2/IL-15Rbeta chain is pivotal for induction of FOXP3 in human CD8 T cells. These findings indicate that induction of FOXP3 is intrinsic to CD8 T cells that are activated in the presence of IL-2 or IL-15, and in vitro-induced expression of FOXP3 cannot be simply interpreted as an indicator of Treg activity or activation marker.

Figures

FIGURE 1

IL-2 induces FOXP3 expression in both CD4 and CD8 subsets in activated PBMC cultures. A, PBMCs from a healthy adult and a vaccinated melanoma patient were stained for CD8, CD3, and FOXP3. B, PBMC from a healthy donor was activated with soluble OKT3 in the absence (OKT3) or presence of exogenous IL-2 (OKT3/IL-2300 IU), or neutralizing anti-IL-2 (OKT3/anti-IL-2). C, IL-2 was added to PBMC cultures (healthy donors) in the absence of TCR stimulation. All cultures were analyzed after 4 days. The dot plots were gated on CD3+ T cells and the number in upper left quadrant represents %FOXP3+ in CD4 T cells and the number in upper right quadrant represents %FOXP3+ in CD8 T cells. No Stim represents freshly thawed cells without activation. A, Representative of FOXP3 staining in 6 different healthy individuals, and 12 melanoma patients. B and C, Representatives of 3 independent experiments using 3 different healthy donors.

FIGURE 2

IL-2 induces FOXP3 expression in CD4 and CD8 T cells in a dose-dependent manner. PBMCs from healthy donors were stimulated with OKT3 and different doses of IL-2 for 4 days. Activated cells were stained for expression of CD3, CD8, and FOXP3. A, The dot plots were gated on CD3+ T cells and the number in upper left quadrant represents %FOXP3+ in CD4 T cells and the number in upper right quadrant represents %FOXP3+ in CD8 T cells. Results are representative of 3 separate experiments using 3 healthy donors. The percent FOXP3+ cells for CD8 T cells (B) and CD4 T cells (C) from 3 independent experiments from 3 healthy donors were enumerated. For CD4 T cells, CD3+CD8− T cells were considered as CD4 T cells. Each symbol represents one individual.

FIGURE 3

Induction of FOXP3 expression in purified CD8 T cells by IL-2. Purified CD8 T cells were isolated from PBMCs (healthy donors) and were stimulated with immobilized OKT3 antibody for 6 days in the presence or absence of IL-2. Cultured cells were costained for CD8, CD3, and FOXP3, with either (A) CD25, or (B) CTLA-4, CD27, and Ki67 antibodies. The dot plots were gated on CD3+CD8+ T cells. This result is representative of three separate healthy donors. No Stim represents freshly thawed cells without activation.

FIGURE 4

IL-2 induces FOXP3 expression in antigen-stimulated memory CD8 T cells. PBMCs from immunized melanoma patients were stimulated with Ag in the presence or absence of exogenous IL-2 (300 IU/mL) for 6 days. Freshly thawed cells (No Stim) were used as control. The dot plots were gated on tetramer+ CD8+ T cells. This result is representative of 5 independent experiments using 4 different vaccinated patients.

FIGURE 5

FOXP3+ memory CD8 T cells produce effector cytokine. In vitro-generated, Ag-stimulated memory CD8 T cells (as described in Fig. 4) were restimulated with either native g209 peptide (Ag) or irrelevant peptide (not shown) pulsed on T2 cells or PMA and ionomycin (PMA/ION) for 6 to 8 hours. The dot plots were gated on tetramer+ CD8 T cells. This result is representative of the 4 independent experiments.

FIGURE 6

IL-15 induces FOXP3 expression in antigen-specific memory CD8 T cells. PBMCs from immunized melanoma patients (n = 2) were stimulated with Ag alone or with IL-2 (300 IU/mL), IL-15 (100 U/mL), IL-7 (10 ng/mL), IL-21 (10 ng/mL), or IL-4 (100 U/mL) as indicated for 6 days. Freshly thawed cells (No Stim) without activation were used as control. The dot plots were gated on tetramer+ CD8+ T cells. The quadrants were set on the basis of isotype control antibodies for each culture. This result is representative of the 3 independent experiments.

FIGURE 7

Removal of IL-2 abrogates FOXP3 expression in IL-2–induced FOXP3+ CD8 T cells. CD8 T cells were activated with Ag and IL-2 (300 IU/mL) for 6 days as described in Methods. Cultured cells were washed in complete medium 3 times and recultured with (+IL-2) or without (Media) exogenous IL-2 (6000 IU/mL) for an additional 48 or 72 hours before staining for FOXP3 expression. The dot plots were gated on tetramer+ CD8 T cells. This result is representative of the 2 independent experiments. The quadrants were set on the basis of isotype control antibodies for each culture.