Journal Logo

Original Articles: Immunobiology and Genomics

Requirement of CD28 Signaling in Homeostasis/Survival of TGF-β Converted CD4+CD25+ Tregs from Thymic CD4+CD25 Single Positive T Cells

Liu, Yongzhong; Amarnath, Shoba; Chen, WanJun

Author Information
doi: 10.1097/01.tp.0000232330.46688.37

Abstract

Transforming growth factor (TGF)-β is a critical factor in the regulation of T cell-mediated immune responses and immune tolerance (1–4). In addition to its potent and diverse effects on T cells, TGF-β, in the context of T-cell receptor (TCR) stimulation, is able to convert peripheral CD4+CD25 naïve T cells to CD4+CD25+ T regulatory cells (Tregs) by induction of transcription factor Foxp3 (5–10). Foxp3 is not only a specific marker for CD25+CD4+ Tregs, but also required for their development (9, 11–14). Although TGF-β induction of Foxp3 in CD4+CD25 naïve T cells has been validated, the underlying molecular events are largely unknown. Moreover, it is also not known whether TGF-β induces Foxp3 expression in thymic CD4+CD25 single positive (SP) T cells and therefore converts them to CD4+CD25+ Tregs.

CD28 is the best characterized costimulatory molecule that is not only required for naïve T cell proliferation and differentiation through interleukin (IL)-2 production, but is also implied in the development and homeostasis of natural CD4+CD25+ Tregs (15, 16). CD28−/− mice exhibit profound reduction of CD4+CD25+ Tregs in the peripheral lymphoid tissues, which is responsible at least in part for the autoimmune-like inflammation in the CD28−/− nonobese diabetic (NOD) mice (15–17). Because TGF-β induced CD4+CD25+ Tregs share many phenotypical and functional features with the natural CD4+CD25+ Tregs (5), it is of importance to study whether CD28 signaling is involved in TGF-β induction of Foxp3 and consequently the conversion of CD4+CD25 naïve T cells to CD4+CD25+ Tregs.

In this report, we show that TGF-β and TCR co- stimulation induces Foxp3 expression in thymic CD4+CD25 SP T cells and converts them to CD4+CD25+ Foxp3+ T regulatory cells. We further provide evidence that CD28 signaling is primarily required for the survival/homeostasis of TGF-β converted thymic CD4+CD25+ Tregs, although engagement of CD28 contributes to Foxp3 induction by TGF-β. Importantly, TGF-β induces alloantigen specific CD4+CD25+ Tregs from thymic CD25SP cells, which also requires CD28 signaling to maintain their survival.

MATERIALS AND METHODS

Mice

WT C57BL/6, Rag1−/− (C57BL/6), BALB/c, CD28−/− and B6.PL-Thy1a/CyJ mice were obtained from The Jackson Laboratory and maintained at the National Institute of Dental and Craniofacial Research.

Antibodies and Reagents

The following reagents were from PharMingen (San Diego, CA): FITC- anti-CD45RB, purified anti-CD3 (145-2C11, NA/LETM), PE-anti-murine CTLA-4, hamster IgG isotypic control, FITC- or biotinylated-anti-murine CD25 (Clone 7D4), FITC-rat IgM, PE-, APC- or purified anti-CD25 (Clone PC61, NA/LETM), PE-anti-murine Thy1.1, FITC-anti-murineThy1.2 and anti-FcRII/III (CD16/32). FITC or PE-anti-murine CD4, anti-CD8α, and the respective isotypic control mAbs and streptavidin-FITC, PE or Tricolor were purchased from Caltag (San Francisco, CA). Recombinant human TGF-β1 and murine IL-2 and anti-TGF-β1, 2,3 mAb were purchased from R&D Systems (Minneapolis, MN). 7-amino-actinomycin D (7-AAD) was purchased from Calbiochem. PE- or FITC- anti-mouse/rat Foxp3 Staining Set (clone FJK-16s) was purchased from eBioscience (San Diego, CA).

Cell Preparation

Thymi were gently minced in complete DMEM containing 10% FBS (Biowhittaker, Walkersville, MD) and CD4+ T cells purified using a mouse CD4+ T Cell Column System (R&D Systems). Whole spleen cells (irradiated 3000 rad) from C57BL/6 mice were used as antigen presenting cells (APCs) as indicated. CD4+CD25 and CD4+CD25+ T cells were isolated as reported before (5) or with mouse CD4+CD25+ T cell isolation kit (Miltenyi Biotec, Auburn, CA) following manufacturer's recommendations. The purity of CD4+CD25 and CD4+CD25+ T cells were usually >98–99% and 80–90%, respectively.

Cell Culture and Induction of Foxp3

For induction of Foxp3, purified thymic CD4+CD25 T cells from C57BL/6 or CD28−/− mice were stimulated with plate-bound anti-CD3 (5 μg/ml) and soluble anti-CD28 (2 μg/ml) in the absence and presence of TGF-β1 (2 ng/ml) or plus IL-2 (2 ng/ml) for 24, 48 and 72 hr. In some experiments, anti-CD3 (10 μg/ml) and anti-CD28 (5 μg/ml) or anti-CD2 (5 μg/ml) were coated onto plate to stimulate T cells. Cells were harvested and RNA was isolated for RT-polymerase chain reaction (PCR) analysis. For induction of T regulatory cells by TCR and TGF-β stimulation, purified CD4+CD25 thymic SP cells from C57BL/6 or CD28−/− mice were cultured with syngeneic splenic APCs and anti-CD3 mAb (0.5 μg/ml)) or with plate-coated anti-CD3 (5 μg/ml) and soluble anti-CD28 (2 μg/ml) without APCs in the presence or absence of TGF-β1 (2 ng/ml) in complete DMEM at 37°C and 5% CO2 for a week. In some cultures, rIL-2 (2 ng/ml) was added as indicated. CD4+ live T cells were then isolated for further analysis.

Co-Culture of CD4+ T Cells

Freshly isolated C57BL/6 thymic or splenic CD4+CD25 responder T cells were cultured in 96-well plates with anti-CD3, splenic APCs, and TGF-β-induced Tregs or control-cultured cells or natural CD4+CD25+Tregs for proliferation assay. In some experiments, co-culture assay was performed with anti-CD3 (coated) and anti-CD28 (soluble) in the absence of APCs to test the suppressive activity of TGF-β-induced Tregs or nTregs. Cells were cultured at 37oC in 5% CO2 for 72 hr and pulsed with 1 μCi 3H thymidine for the last 6–16 hr. Radioactivity incorporated was counted using a faltbed β counter (Wallac). For carboxy-fluorescein diacetate succinimidyl ester (CFSE) labeling assay, CD4+ CD25 T cells (107/ml) were incubated with 2 μM CFSE in PBS at 37oC for 15 min. Cells were then washed three times and resuspended in complete DMEM (without Phenol red) for cell culture.

Flow Cytometry Analysis

T cells were resuspended in PBS containing 1% BSA (Irvine, Santa Ana, CA) and 0.1% sodium azide (Sigma). For the staining of surface antigens, cells were incubated with FITC, PE or Tricolor conjugated mAbs or their negative control antibodies as indicated for 30 min on ice. Intracellular staining of CTLA-4 was performed as reported (5). Intracellular Foxp3 staining was performed with PE anti-mouse/rat Foxp3 staining set following manufacturer's recommendation.

Adoptive Transfer Experiments

TGF-β converted CD4+CD25+ T cells from C57BL/6 and CD28−/− cultures were injected i.p. into nonirradiated syngeneic B6.PL-Thy1a/Cy (Thy1.1+) or Rag1−/− mice. The recipients were bled at days one, two, three, and five post cell transfer. Red blood cells were lysed and the white blood cells were stained with anti-CD4-PE and FITC-anti-Thy1.2 antibodies. The mice were sacrificed on day five and cells from spleens and mesenteric lymph nodes were isolated and stained ex vivo with PE-anti-CD4 and FITC-Thy1.2 mAbs to determine the number of injected TGF-β converted Tregs (CD4+Thy1.2+) in vivo.

Semiquantitative and Real-Time PCR

RT-PCR for Foxp3 and IL-2 expression was performed as described (5, 18). The primer sequences were as follows and synthesized by Invitrogen: GAPDH: 5′-CCATGGAGAAGGCTGGG-3′ and 5′-CAAAGTTGTCATGGATGACC-3′; Foxp3: 5′-CAG CTG CCT ACA GTG CCC CTA G-3′ and 5′-CAT TTG CCA GCA GTG GGT AG-3′; IL-2: 5′-ATGTACAGCATGCAGCTCGCATC-3′ and 5′-GGCTTGTTGAGATGATGCTTTGACA-3′. Normalized values for Foxp3 mRNA expression in each sample were calculated as the relative quantity of Foxp3 divided by the relative quantity of GAPDH. Foxp3 mRNA was also quantified by real-time PCR using the ABI 7500 Real-time PCR system (PE Applied Biosystems, Foster City, CA). Experiments were performed using primers, the fluorescent TaqMan probe specific to Foxp3 and an internal control HPRT gene. Foxp3 primers: 5′-CCC AGG AAA GAC AGCAAC CTT-3′ and 5′-TTC TCA CAA CCA GGC CAC TTG-3′; Foxp3 probe: 5′-FAM-ATCCTA CCC ACT GCT GGC AAA TGG AGT C-3′; HPRT was analyzed according to the protocol of Taqman gene expression assay kit (PE Applied Biosystems). PCR reaction system contained 0.5 μM primers and 0.2 μM TaqMan probe. PCR amplification was preceded by a 10 min denaturation step at 95°C and the amplification step consisted of 40 cycles of 15s at 95°C and 60s at 60°C. The amount of Foxp3 was normalized to the endogenous internal control HPRT using the 2(-Delta Delta C[T]) method (2−ΔΔct), and presented as a relative expression ratio.

Statistical Analysis

Student's t tests were used for the significance of data comparison.

RESULTS

TGF-β Induces Foxp3 in CD4+CD25 Single Positive Thymocytes

We first examined whether TGF-β was able to induce Foxp3 expression in CD4+CD25 SP thymocytes. Stimulation of highly purified C57BL/6 (B6) thymic CD4+CD25 SP cells with anti-CD3 and anti-CD28 in the presence of TGF-β1 induced Foxp3 mRNA expression (Fig. 1, A and B). This TGF-β induction of Foxp3 mRNA was detected within 24 hr (Fig. 1A) and continued to increase dramatically at 48–72 hr (Fig. 1B and data not shown). As in peripheral CD25 T cell conversion (5), neither TCR nor TGF-β stimulation alone could result in Foxp3 expression in CD4+CD25 SP thymocytes. Stimulation of thymic CD4+CD25+ SP natural Tregs with TCR and TGF-β in the presence of high dose IL-2 failed to further enhance their Foxp3 expression (data not shown), consistent with that in peripheral Treg (5). To further explore this TGF-β induction of Foxp3 in thymic SP cells, intracellular Foxp3 protein was stained by anti-Foxp3 mAb and analyzed by flow cytometry. Highly purified CD4+CD25 SP thymocytes (>98–99%) contained 1–3% of Foxp3+ cells (Fig. 1C), whereas majority of natural CD4+CD25+ Tregs expressed Foxp3 (9, 14, 19) (unpublished data). As in natural Tregs, Foxp3 protein of CD4+CD25 thymic SP was also confined to the nucleus. Notably, when thymic CD4+CD25 SP cells were cultured with TCR and TGF-β for only 16–24 hr, 8–10% of the stimulated cells were already Foxp3+ CD25+, whereas only 3–4% of Foxp3+ CD25+ cells was seen in the control cultures (without TGF-β). As expected, there was no detectable T cell proliferation by 24 hr and the total CD25+ T cells were compatible between TGF-β treated and control cultures (40–50%). Strikingly, when cells were continued in culture for 72 hr and stained for Foxp3, 40–50% of TGF-β treated cells became Foxp3+ CD25+, whereas only 0.3–2% of control cultured cells were Foxp3+CD25+(Fig. 1C, 2). This induction of Foxp3 was clearly dependent on the concentration of TGF-β in the cultures (Fig. 2). As low as 80–400 pg/ml of TGF-β1 already resulted in significant CD4+CD25+Foxp3+ Tregs (7–10%)(Fig. 2). When an aliquot of these cells were visualized under the immunofluorescence microscopy, it was found that Foxp3 protein (red) is indeed localized in the nucleus in TGF-β treated cells (Fig. 1D right). In contrast, Foxp3+ cells were hardly seen in the control cultures (Fig. 1D, left). When total Foxp3+ was calculated according to the percentage of Foxp3+ cells and the total number of the viable cells in each well of the plate, it was found that TGF-β treated cultures contained 2.22x105 of 25+Foxp3+ T cells that was about 15-fold higher than that in control cultured cells (Foxp3+: 0.14x105) and more than 55-fold increase compared to that in the freshly isolated CD4+CD25 SP thymocytes (Foxp3+; 0.04x105) (Fig. 1C,D). Thus, TGF-β and TCR costimulation induces Foxp3 expression in thymic CD4+CD25 SP cells. The data strongly argues for the TGF-β de novo induction of the Treg specific gene in the naïve CD4+CD25 SP thymocytes and converts CD25 Foxp3 into CD25+Foxp3+ T cells (9, 20).

FIGURE 1.
FIGURE 1.:
TGF-β induces Foxp3 expression in thymic CD4+CD25– SP T cells. (A) CD4+CD25– cells sorted from C57BL/6 (B6) thymocytes (1–2×106/well in 24-well plate) were cultured with plate-bound anti-CD3 (5 μg/ml) and soluble anti-CD28 (2 μg/ml) in the absence or presence of TGF-β1 for 16–24 hr. RNA was isolated and cDNA synthesized for assessing the expression of Foxp3 by RT-PCR. GAPDH was used as internal control. (B) Real-time PCR of Foxp3 expression of stimulated CD4+CD25− SP cells (48 hr). Values are expressed as normalized ratio of Foxp3 to HPRT (host gene). The experiments were repeated more than three times with similar results. Med: Medium; T: TGF-β1; αCD3: anti-CD3 mAb; αCD28: anti-CD28 mAb. (C) Analysis of Foxp3 protein expression at the single-cell level by FACS. Freshly isolated CD4+CD25− SP thymocytes (2×105/well) were cultured with plate-coated anti-CD3 and soluble anti-CD28 in the absence or presence of TGF-β for 72 hr. The total number of viable cells per well was then counted and shown in parenthesis. Freshly isolated CD4+CD25− SP thymocytes or cultured cells were stained with FITC-anti-CD25 (surface) and PE-anti-Foxp3 (intracellular) and analyzed on FACScalibur. The data are shown as profile of CD25 vs. Foxp3. The quadrant gates were set according to the negative isotype control antibodies in the respective cells. (D) An aliquot of cultured cells in (C) was visualized under immunofluoresence microscopy. Green: CD25; Red, Foxp3.
FIGURE 2.
FIGURE 2.:
(A) Dose dependence of TGF-β1 on Foxp3 induction in CD4+CD25− SP thymocytes. CD25− SP thymocytes were cultured as in Figure 1C in the presence of indicated concentrations of TGF-β1 for 72 hr. Cells were then stained with FITC-anti-CD4, APC-anti-CD25 and PE-anti-Foxp3 (intracellular) and analyzed on FACScalibur. Viable CD4+ T cells were gated and the profile of CD25 vs. Foxp3 is displayed. (B) APCs are dispensable in TGF-β induction of CD4+CD25+Tregs from CD4+CD25− T cells. Thymic CD4+CD25− T cells (1×106) were cultured with immobilized anti-CD3 (5 μg/ml) and soluble anti-CD28 (2 μg/ml) antibodies and TGF-β (2 ng/ml) for 6 days. After washed extensively, the harvested CD4+CD25+ T cells (5×104) were then added into the co-cultures containing CD4+CD25− responder T cells (5×104, prelabeled with CFSE) stimulated with either anti-CD3 and APCs (αCD3+APCs) or plate-coated anti-CD3 and soluble anti-CD28 antibodies (αCD3+αCD28) for 72 hrs. The dilution of CFSE fluorescence intensity was measured to detect the proliferation of CD4+CD25− responder T cells.

TGF-β Converts CD4+CD25 SP Thymocytes to CD4+CD25+hi Suppressor T Cells

To study whether TGF-β-induced Foxp3+ SP thymocytes possessed immunosuppressive ability, thymic CD4+CD25 SP cells were cultured with anti-CD3 and APCs in the presence or absence of TGF-β for 7 days and viable CD4+ T cells were isolated and analyzed. Although CD4+ T cells harvested from both TGF-β treated and control cultures were CD25+ (>90–95%), the pattern of CD25 expression in the two groups was different. While most of CD4+ T cells in control cultures were CD25+low, majority of CD4+ T cells in TGF-β treated cultures were CD25+hi (Fig. 3A). Inclusion of exogenous IL-2 in the primary cultures did not change the expression pattern of CD25 in TGF-β treated cells (Fig. 3A), although it increased total number of recovered CD4+ T cells (Fig. 6A). CD4+CD25+ T cells derived from the TGF-β treated thymic cultures were smaller in size and expressed CD4+low, CD45RB−/low, intracellular CTLA-4+ and GITR+ (Fig. 3A and data not shown) consistent with the TGF-β converted splenic CD4+CD25+ Tregs (5) as well as the natural CD4+CD25+ Tregs (20–24). Strikingly, by real-time PCR analysis, TGF-β converted thymic CD4+CD25+ T cells expressed much higher levels of Foxp3 (30- to 60-fold) than that in CD4+CD25+ T cells obtained from the control cultures at the end of one week (Fig. 3B). Significantly, when co-cultured with freshly isolated thymic or splenic CD4+CD25 SP responder T cells, TGF-β-converted CD4+CD25+ T cells markedly inhibited the proliferation of the responder T cells, whereas the CD4+CD25+ T cells from control cultures had only marginal effect on responder T cells (Fig. 3C). Interestingly, increase in the cell number of TGF-β induced Tregs (CD25 cells:Tregs, 1:2, 1:3) could not further enhance their suppressive ability (Fig. 3D). As expected, neither TGF-β converted Tregs nor natural CD4+CD25+ Tregs (nTregs) could significantly inhibit T cell proliferation (CD4+ CD25 T cells alone: 116666±5773 cpm; plus TGF-β-Tregs [1:1]: 77972±1100 cpm; plus nTregs[1:1]: 72264±15880 cpm) in co-cultures with immobilized anti-CD3 and soluble anti-CD28 without APCs.

FIGURE 3.
FIGURE 3.:
TGF-β converted CD4+CD25+Foxp3+ T cells are suppressor cells. Thymic CD4+CD25– SP cells were cultured with soluble anti-CD3 (0.5 μg/ml) or plus IL-2 (2 ng/ml) and APCs with or without TGF-β for one week and analyzed for CD4 vs. CD25 expression (A) and assessed for Foxp3 expression by real-time PCR (B). (C). Freshly isolated thymic CD4+CD25− SP T cells (5×104) prelabeled with CFSE (2 μM) were stimulated with anti-CD3 and APCs in the absence (αCD3) or presence of the viable CD4+CD25+ T cells (5×104) derived from TGF-β treated cultures (αCD3+TGFβ 25+ cell) or control cultures (αCD3+Con 25+ cell) for 72 hr and analyzed on FACScalibur. The data shown are the histograms of CFSE positive cells. The experiments were repeated twice with similar results. (D) Increase in the number of TGF-β converted CD25+ Tregs in cultures failed to further enhance the suppressive effect on CD4+CD25− responder T cells. Freshly isolated CD4+CD25− (5×104)T cells were cultured with anti-CD3 (0.5 μg/ml) and APCs in the presence of indicated cell numbers of TGF-β converted CD25+ Tregs or of natural CD4+CD25+treg (nTreg, 5×104) for 72h. The cells were pulsed with 1μCi 3H thymidine for the last 16h. The values are expressed as Mean±SD of triplicate wells for 3H incorporation (CPM).
FIGURE 6.
FIGURE 6.:
CD28 deficiency results in reduction of TGF-β converted CD4+CD25+ thymic Tregs in vitro and in vivo. (A) In vitro, CD4+CD25− SP thymocytes (5×105/well, 12-well plate) were cultured with TGF-β and APCs and in the presence and absence of IL-2 for one week. Total number of viable CD4+ T cells are counted and shown. White bar: wildtype cells; Black bar: CD28−/− cells. (B) CD4+CD25+ T cells (1.5×106) isolated from CD28−/− or wildtype cultures were injected i.p. into B6.PL-Thy1a/Cy mice (Thy1.1+, n=2 per group). Three days later, peripheral blood leukocytes were stained with FITC-Thy1.2 and PE-CD4 antibodies and analyzed by FACScalibur. CD4+ T cells were gated and the profile of CD4+ vs. Thy1.2+ cells is displayed. The values are shown as representative mouse in each group. (C) The kinetics of TGF-β converted WT or CD28−/− CD4+CD25+Tregs (Thy1.2+) in B6.PL-Thy1a/Cy mice (Thy1.1+). The blood was collected at days one, two, and five post injection. The cells were then stained with anti-CD4 and anti-Thy1.2 and analyzed by FACScalibur. Each group contained three mice and the values are shown as Mean±SD of the mice each group at indicated time points. (D and E) On day five, the recipients were sacrificed and cells from spleens and mesenteric lymph nodes (MLN) were stained with anti-CD4 and anti-Thy1.2 antibodies. Each symbol represents one mouse except the symbol (open circle) indicating the pooled cells of two mice. Data are shown as the percentages of CD4+Thy1.2+ T cells within CD4+ T cells. (F and G) CD4+CD25+ Tregs (1×105) generated from TGF-β treated WT or CD28−/− CD25− thymocytes were injected (i.p.) into Rag1−/− (C57BL/6) mice. The injected cells were then monitored by staining with anti-CD4 antibody in blood PBMC (F, Mean±SD, n=2–3 per group, days 1, 2, and 5) and in spleens (G, each symbol represents one mouse, day 5). The mononuclear cells were gated and the percentages of CD4+ T cells are displayed.

We then examined whether APCs were required for induction of CD4+CD25+ Tregs by TGF-β. CD4+CD25 thymocytes were stimulated with immobilized anti-CD3 and soluble anti-CD28 antibodies and TGF-β in the absence of APCs for 6 days. CD4+CD25+ cells were then isolated and tested for their suppressive ability on CD4+CD25 responder T cells in a co-culture assay. Interestingly, TGF-β induced CD25+ Tregs from the APC-free cultures (anti-CD3 and anti-CD28) inhibited CD4+CD25 T cell proliferation (Fig. 2B), which was as potent as natural Tregs (nTreg) (Fig. 2B) and TGF-β converted Tregs from the APC-containing cultures (anti-CD3 and APCs) (Fig. 3C, D). The suppression was observed in the co-cultures in which CD4+CD25 responder T cells were stimulated with anti-CD3 and APCs but not by coated anti-CD3 and soluble anti-CD28 antibodies without APCs. Thus, thymic CD4+CD25 SP T cells can be converted into CD4+CD25+ suppressor T cells by TGF-β through induction of Foxp3. APCs are dispensable for the induction of CD25+ Tregs by TGF-β, but important for the suppressive activity of TGF-β converted Tregs in vitro.

CD28, but not CD2, Contributes to TGF-β Induction of Foxp3 in CD4+CD25 T Cells

Because CD28 signaling has been indicated in homeostasis of natural CD4+CD25+ Treg (15–17), we then sought to determine whether CD28 engagement had an impact on TGF-β induced Foxp3 expression in CD4+CD25 T cells. Thymic CD4+CD25 SP cells were stimulated with immobilized anti-CD3 alone or plus anti-CD28 mAb in the presence of active TGF-β1 and Foxp3 mRNA was then determined by RT-PCR and real-time PCR. Costimulation of CD28 by anti-CD28 mAb enhanced Foxp3 expression induced by TGF-β in thymic CD4+CD25 SP T cells (Fig. 4, A and C). The TGF-β up-regulation was specific for Foxp3, because IL-2 mRNA in the same TGF-β treated cultures was significantly suppressed (data not shown). In contrast to CD28 costimulation, co-engagement of CD2, a molecule required for thymic development of T cells in in vitro models (25), failed to enhance Foxp3 expression in TGF-β treated cultures (Fig. 4A). The same result was observed whether anti-CD3 (Fig. 4A) or anti-TCR antibodies (BD, Clone 3&23, not shown) were used to stimulate TCR in the cultures. As expected, TCR and CD2 stimulation in the absence of TGF-β showed no Foxp3 expression (Fig. 4A). Thus, CD28 engagement enhances the effect of TGF-β on Foxp3 expression. CD2 costimulation, however, plays no role in TGF-β induction of Foxp3 in CD4+CD25 SP thymocytes.

FIGURE 4.
FIGURE 4.:
CD28 signaling contributes to Foxp3 induction by TGF-β in CD4+CD25− SP thymocytes. (A) Co-engagement of CD28, but not CD2 up-regulates TGF-β induction of Foxp3 expression. CD4+CD25− SP thymocytes (B6) were stimulated with plate-bound anti-CD3 (10 μg/ml) and anti-CD28 (5 μg/ml) or anti-CD2 (5 μg/ml) antibodies in the presence and absence of TGF-β1 for 24 hr and Foxp3 expression was detected by RT-PCR. GAPDH was used as host gene. (B and C) CD28−/− CD4+CD25− SP thymocytes express transit lower levels of Foxp3 induction by TGF-β than that in the wildtype cells. CD4+CD25− SP thymocytes from CD28−/− and wildtype (B6) mice were stimulated with plate-bound anti-CD3 plus soluble anti-CD28 or IL-2 with or without TGF-β for 48 hr and Foxp3 mRNA was determined by real-time PCR. Values expressed as normalized ratio of Foxp3 to HPRT. The experiments were repeated more than three times with similar results.

To further explore the role of CD28 signaling in TGF-β induction of Foxp3, CD28−/− CD4+CD25 SP thymocytes were cultured with TCR and TGF-β. Stimulation CD4+CD25 SP thymocytes from CD28−/− mice showed lower levels of Foxp3 mRNA than that in the wildtype thymic CD4+CD25 SP cells during the first 24–48 hr of cultures (Fig. 4, B and C). Because CD28−/− T cells produced lower levels of IL-2, exogenous rIL-2 was added into the cultures. Inclusion of IL-2 however failed to restore the defect in TGF-β induction of Foxp3 in CD28−/− T cells completely (Fig. 4B,C). As in wildtype mice, TGF-β inhibited IL-2 mRNA in the CD28−/− CD4+CD25 SP thymocytes, indicating intact TGF-β signaling in the CD28−/− T cells. Similar results were also observed in CD28−/− splenic CD4+CD25 T cells (data not shown). Interestingly, when the cultures were continued for 60–72 hr, Foxp3 expression induced by TGF-β in the CD28−/− T cells was restored to the levels of TGF-β-treated WT cells. Consequently, there was similar amount of Foxp3 expression in TGF-β treated cells between CD28−/− and WT mice at the end of cultures (day 7) as determined by real-time PCR (Fig. 5B). The data demonstrates that CD28 signaling contributes to, particularly in the early phase of stimulation, but is not essential for TGF-β induced Foxp3 expression in CD4+CD25 SP thymocytes, indicating that other molecules and/or compensatory pathway(s) may also be involved.

FIGURE 5.
FIGURE 5.:
TGF-β and TCR co-stimulation converts CD28−/− CD4+CD25− SP thymocytes to CD4+CD25+hi suppressor T cells. CD28−/− CD4+CD25− SP thymocytes were cultured with TGF-β and soluble anti-CD3 and APCs in the absence of IL-2 (αCD3+TGF-β) and presence of IL-2 (αCD3+IL-2+TGF-β) for one week. (A) The profile of CD4 vs. CD25 is shown as dot plots. Majority of CD4+CD25+ T cells expresses high levels of CD25. (B) Real-time Foxp3 expression of TGF-β converted CD4+CD25+ T cells between CD28−/− and wildtype mice. The value of Foxp3 in the wildtype CD3+TGF-β treated cultures (Fig. 1D) was set to one and relative ratios of Foxp3 in the indicated treatments were shown. (C) TGF-β converted CD28−/− CD4+CD25+ T cells are suppressor cells. Freshly isolated CD4+CD25− SP thymocytes (5×104) prelabeled with CFSE were stimulated with anti-CD3 and APC (αCD3) and co-cultured with TGF-β converted CD4+CD25+ T cells in CD28−/− mice in complete DMEM (without Phenol red) for 72 hr. The cells in triplicate wells were pooled and analyzed on FACScalibur. Data are shown as histograms of CFSE dilution (FL-1) in responder cells. αCD3+koTGFβ cell: CD28−/− CD4+CD25+ cells derived from the primary cultures with anti-CD3+APCs+TGF-β; αCD3+koTGFβ/IL-2 cell: CD28−/− CD4+CD25+ from the primary cultures with anti-CD3+APCs +TGF-β +IL-2.

TGF-β Converted CD28−/− Thymic CD4+CD25+ Cells Are Suppressor T cells

We then examined whether TGF-β was able to convert CD28−/− thymic CD4+CD25 T cells into CD4+CD25+ Treg in vitro. CD4+CD25 SP thymocytes from CD28−/− and wildtype mice were cultured with TCR and TGF-β for one week and viable CD4+ T cells were then isolated and analyzed. As seen in wildtype cultures, TCR and TGF-β co-stimulation converted CD28−/− thymic CD4+CD25 SP T cells to CD4+CD25+ T cells that expressed high levels of CD25 (Fig. 5A). In contrast, majority of the CD4+ T cells from control cultures (without TGF-β) showed CD25+low (data not shown). Importantly, TGF-β converted CD4+CD25+hi T cells expressed similar levels of Foxp3 compared to that in wildtype mice determined by real-time PCR (Fig. 5B). Significantly, when co-cultured with freshly isolated wildtype thymic CD4+CD25 responder T cells that had been labeled with CFSE, TGF-β converted CD28−/− CD4+CD25+ T cells inhibited TCR mediated responder T cell proliferation (Fig. 5C), which was as efficient as TGF-β induced WT CD4+CD25+ Tregs (Fig. 3C). Addition of exogenous IL-2 in the primary cultures did not change the levels of CD25, Foxp3 expression and suppressive function of CD28−/− CD4+CD25+ Tregs (Fig. 5). Thus, CD28 −/− mice preserve the potential of TGF-β conversion of thymic CD4+CD25 SP cells to CD4+ CD25+ Tregs.

CD28 Is Required for the Homeostasis/Survival of TGF-β Induced CD4+CD25+ Treg In Vitro and In Vivo

Although TGF-β converted CD28−/− CD4+CD25+ Tregs were similar in phenotype and function as compared with wildtype natural and TGF-β induced Tregs, the number obtained was significantly lower. This prompted us to examine whether CD28 deficiency had any impact on homeostasis/survival of TGF-β induced Tregs. Thymic CD4+CD25 SP cells from both CD28−/− and wildtype mice were cultured with anti-CD3 antibody in the presence or absence of TGF-β for one week and viable CD4+ CD25+ T cells were counted. Strikingly, the total number of TGF-β converted CD4+CD25+ Tregs in CD28−/− cultures was profoundly decreased compared to that in the wildtype cultures (Fig. 6A), although the percentage of Tregs were similar between the two groups (Fig. 3A, 5A). The reduction was not specific to TGF-β induced Tregs, because the total number of viable CD4+ T cells in control cultures of CD28−/− mice was also markedly decreased (data not shown). Inclusion of exogenous IL-2 in the primary cultures increased the over all number of CD4+ T cells in both CD28−/− and wildtype mice, but the number of Tregs in CD28−/− cultures still remained significantly lower than that in wildtype cultures (Fig. 6A). Thus, CD28 signaling is required for the maintenance and/or survival of TGF-β induced CD4+CD25+ Tregs in vitro, although it may be dispensable for TGF-β induction of Foxp3.

We then sought to identify whether CD28 deficiency had an impact on homeostasis/survival of TGF-β induced CD4+CD25+ Tregs in vivo. TGF-β converted thymic CD4+CD25+ Tregs from CD28−/− and wildtype mice (Thy1.2+) were adoptively transferred i.p. into syngeneic B6.PL-Thy1a/CyJ (Thy1.1+) without further challenge. The injected Tregs were monitored by FACS staining with PE-CD4 and FITC-Thy1.2 antibodies. On days one, two, three and five after cell transfer, peripheral blood was collected from the recipients and the CD4+ Thy1.2+ donor cells were analyzed. A reduction of CD4+Thy1.2+ CD28−/− Tregs was observed compared to the injected wildtype Tregs (Fig. 6B–E). When the recipients were sacrificed on day 5 after injection, the recovered CD4+Thy1.2+ CD28−/− T cells were lower in spleen (Fig. 6D) and in mesenteric lymph nodes (Fig. 6E). When TGF-β converted CD4+CD25+ Tregs from CD28−/− or WT thymic CD25cells were injected into the syngeneic Rag1−/− (C57BL/6) mice, similar results were obtained (Fig. 6F.G). The data indicates that CD28−/− Tregs also have low survival capacity than those from wildtype mice in vivo.

Finally, we examined whether TGF-β could induce alloantigen-specific CD25+ Foxp3+ Tregs from CD4+CD25 SP thymocytes and if CD28 deficiency had any effect on such alloantigen-specific Treg generation. Highly purified CD4+CD25 SP thymocytes from CD28−/− and wildtype C57BL/6 mice were cultured with irradiated spleen cells from BALB/c mice in the presence of TGF-β and IL-2 for one week. Cells were then harvested, counted and stained with anti-CD4, CD25 and Foxp3 antibodies to identify Tregs. TGF-β efficiently converted CD25 naïve T cells into CD25+Foxp3+ Tregs upon alloantigen stimulation in the WT cultures (about 80% of CD25+ T cells) (Fig. 7A), although the percentage of alloantigen-specific CD25+Foxp3+ T cells was lower than that in the cultures with anti-CD3 stimulation, CD28 deficiency markedly reduced TGF-β converted CD25+Foxp3+ Tregs, although the ratio between Foxp3+ and Foxp3 T cells within CD25+ population (about 60%) was similar to that in WT mice (Fig. 7A). Consequently, the total number of recovered CD25+Foxp3+ Tregs was significantly lower in CD28−/− cultures than that in the WT controls (Fig. 7B). As expected, the CD25+Foxp3 non-Tregs also decreased in CD28−/− cultures (Fig. 7B). Thus, TGF-β induces alloantigen-specific Foxp3+ Tregs that also require CD28 for survival.

FIGURE 7.
FIGURE 7.:
CD28 deficiency causes decrease in TGF-β converted alloantigen-specific CD4+CD25+Foxp3+ Tregs. CD4+CD25− SP thymocytes (5×105) from C57BL/6 or CD28−/− mice were stimulated with irradiated spleen cells (2×106) from BALB/C mice in the presence of TGF-β1 (2 ng/ml) and IL-2 for seven days. Cultured cells were then counted and stained with FITC-anti-CD4, APC-anti-CD25 and PE-anti-Foxp3 antibodies and analyzed by flow cytometry. (A) Viable CD4+ T cells were gated and the profile of CD25 vs. Foxp3 is displayed. (B) The values are shown as the total number of the cells recovered from the indicated cultures.

DISCUSSION

We have shown here that TGF-β and TCR co- stimulation converts thymic CD4+CD25 SP T cells into CD4+CD25+hiFoxp3+ T regulatory cells. The TGF-β converted Tregs are phenotypically and functionally indistinguishable from the TGF-β induced peripheral CD4+CD25+ Tregs (5–10) as well as the natural CD4+CD25+ Tregs (20–24). Several novel conclusions can be drawn from the current study. First, TGF-β can induce the expression of Foxp3 in CD4+CD25 SP thymocytes along with TCR co-stimulation. The TGF-β mediated Foxp3 increase is indeed due to de novo gene induction rather than selective proliferation of a small number of Foxp3+ subset within the CD4+CD25 T cells. This is evident from the fact that Foxp3 expression is clearly seen within 24 hr culture in the absence of obvious T cell proliferation. In addition, immunofluorescent staining clearly demonstrates that TGF-β treatment has already increased Foxp3+ cells in the culture to three- to five-fold within 24 hr. Furthermore, since almost half of TGF-β treated cells express Foxp3+CD25+ and this increase represents about 55-fold increase in total Foxp3+ cells by 72 hr compared to freshly isolated SP thymocytes, it is impossible that TGF-β induces the original minor population of Foxp3+CD25 to proliferate 55-fold within three days. Therefore, CD25+Foxp3+ T cells in TGF-β treated cultures are, at least in great part, converted from Foxp3 SP thymocytes. It is important to point out that as low as 80–400 pg/ml of TGF-β in the culture was already able to induce about 10% CD25+Foxp3+ Tregs, which might have implications in the physiological situation in vivo. Significantly, TGF-β converted thymic Tregs exhibit similar suppressive capacity as the natural Tregs (nTregs) on normal CD4+ T cell proliferation. This suppression seems to require the presence of APCs, since neither TGF-β converted Tregs nor natural CD4+CD25+ Tregs could significantly inhibit T cell proliferation in co-cultures with immobilized anti-CD3 and soluble anti-CD28. Moreover, TGF-β induces Foxp3+CD25+ T cells from CD4+CD25 thymocytes stimulated with anti-CD3 and anti-CD28 antibodies in the absence of APCs. The TGF-β induced Tregs derived from APC-free cultures are equally potent in suppressing normal CD4+ T cell proliferation as those from APC-containing cultures, suggesting that APCs are dispensable in TGF-β induction of Foxp3+ Tregs. Since TGF-β converted CD4+CD25+ Treg share basic features of “natural” Tregs, our data provide additional insight in understanding the development/generation of natural Tregs in the thymus (26–28).

Second, TGF-β converted CD4+CD25+ Tregs express very high levels of CD25 on the surface that are significantly distinct from the low levels of CD25 on CD4+ T cells in the control cultures (without TGF-β). The expression of high levels of CD25 on CD4+ T cells is associated with TGF-β stimulation, but is unlikely to be related to CD28 expression on the T cells or exogenous IL-2 presence in the cultures. This finding has identified an undeniable link between TGF-β signaling and high levels of CD25 surface expression on CD4+ T cells that represents one unique feature of natural murine and human CD4+CD25+ Treg (22, 23, 29). Addition of exogenous IL-2 to the cultures results in an insignificant effect on CD25 expression of TCR stimulated CD4+ T cells, nor did it influence the profile of high levels of CD25 on TGF-β converted CD4+CD25+ Treg, however it indeed increased the total number of viable cells. Although mechanistically elusive, it is likely that TGF-β signaling maintains CD25hi expression by keeping rather than up-regulating its expression. In this regard, the levels of CD25 expression on CD4+ T cells in TGF-β treated cultures are similar and even slightly lower compared to that in the control cultures during initial one, two, and three days of stimulation (5) (Fig. 1 and data not shown). Alternatively, TGF-β may enhance the survival of CD25+hi T cells by preventing them from activation induced cell death after TCR signal induced high expression of CD25, because TGF-β is indispensable in protecting T cell from apoptosis (30, 31).

Third, CD28 signaling contributes to TGF-β induced Foxp3 expression, which is evident by anti-CD28 antibody upregulation of TGF-β mediated Foxp3 in wildtype thymic CD4+CD25 SP cells and by a transient defect in TGF-β induction of Foxp3 in CD28−/− thymic SP cells. The CD28 involvement in TGF-β induction of Foxp3 is however time-dependent and dispensable, because the lower levels of Foxp3 in CD28−/− T cells are only seen in the early phase of cultures (24–48 hr). Furthermore, the early transient defect in Foxp3 induction had no significant impact on TGF-β conversion of CD4+CD25+ Tregs, since TGF-β converted CD4+CD25+ Tregs are phenotypically and functionally indistinguishable between CD28−/− and wild-type mice. The data indicates a limited role for CD28 signaling in the induction and/or conversion phase of Tregs, at least in vitro. Fourth, CD2 signaling is not involved in TGF-β induction of Foxp3 in thymic SP T cells, unlike its requirement, at least in vitro, for the maturation of CD4+CD8+ DP thymocytes toward CD4+or CD8+ SP thymocytes (25). Most importantly, the profound reduction in the number of TGF-β induced CD4+CD25+ Tregs in CD28−/− mice strongly argue for the indispensable role of CD28 in maintaining the homeostasis /survival of T regulatory cells. Although the detailed mechanisms remain to be elucidated, CD28 signaling may function through both increasing proliferation and/or preventing apoptosis (17) of TGF-β induced CD4+CD25+ Tregs, which may not necessarily rely solely on IL-2 production, because exogenous IL-2 fails to restore the number of CD28−/− Tregs to normal levels. CD28 signaling is shown to prevent T cell death by up-regulating anti-apoptotic BCL-XL expression (32), which may also occur in the process of TGF-β mediated conversion of CD4+CD25 T cells to CD4+CD25+ Tregs. Since TGF-β converted CD4+CD25+ Tregs resemble the basic phenotypical and functional features of natural CD4+CD25+ Tregs, our study will help define the molecular events involved in the development of natural Tregs and offer an explanation for the lack of natural CD4+CD25+ Tregs in CD28 as well as B7–1/B7–2 deficient mice.

Finally, TGF-β conversion of naïve CD4+CD25 T cells to CD25+Foxp3+ Tregs upon alloantigen stimulation opens a new avenue to generate alloantigen-specific Tregs from normal cells in vitro, which has implications in the prevention and inhibition of allograft rejection and graft against host reaction. The finding that CD28 is also essential for the maintenance of TGF-β converted alloantigen-specific Foxp3+ Tregs not only confirms our data in pan-TCR stimulation, but also helps to understand the molecular mechanisms for CD28 in immune tolerance in transplantation (33). The fact that natural CD4+CD25+Foxp3+ Tregs are difficult to isolate and expand in vitro makes our approach to induce alloantigen-specific T regs from normal naïve CD4+ T cells particularly attractive for future clinical applications in transplantation.

ACKNOWLEDGMENTS

The authors thank Drs. Sharon M. Wahl, Songtao Shi, and Sylvain Perruche from the NIDCR for critical reading of the manuscript. This work was carried out in the Intramural Division at the National Institute of Dental and Craniofacial Research, National Institutes of Health.

REFERENCES

1. Chen W, Wahl SM. TGF-β: receptors, signaling pathways and autoimmunity. Curr Dir Autoimmun 2002; 5: 62.
2. Gorelik L, Flavell RA. Transforming growth factor-β in T-cell biology. Nature Rev Immunol 2002; 2: 46.
3. Strober W, Kesall B, Fuss I, et al. Reciprocal IFN-g and TGF-b response regulate the occurrence of mucosal inflammation. Immunol Today 1997; 18: 61.
4. Weiner HL. Oral tolerance, an active immunologic process mediated by multiple mechanisms. J Clin Invest 2000; 106: 935.
5. Chen W, Jin W, Hardegen N, et al. Conversion of peripheral CD4+CD25- naive T cells to CD4+CD25+ regulatory T cells by TGF-β induction of transcription factor Foxp3. J Exp Med 2003; 198: 1875.
6. Cobbold SP, Castejon R, Adams E, et al. Induction of foxP3+ regulatory T cells in the periphery of T cell receptor transgenic mice tolerized to transplants. J Immunol 2004; 172: 6003.
7. Fantini MC, Becker C, Monteleone G, et al. Cutting edge: TGF-β induces a regulatory phenotype in CD4+CD25− T cells through Foxp3 induction and down-regulation of Smad7. J Immunol 2004; 172: 5149.
8. Fu S, Zhang N, Yopp AC, et al. TGF-β induces Foxp3 + T-regulatory cells from CD4 + CD25− precursors. Am J Transplant 2004; 4: 1614.
9. Wan YY, Flavell RA. Identifying Foxp3-expressing suppressor T cells with a bicistronic reporter. Proc Natl Acad Sci U S A 2005; 102: 5126.
10. Zheng SG, Wang JH, Gray JD, et al. Natural and induced CD4+CD25+ cells educate CD4+CD25− cells to develop suppressive activity: the role of IL-2, TGF-β, and IL-10. J Immunol 2004; 172: 5213.
11. Fontenot JD, Gavin MA, Rudensky AY. Foxp3 programs the development and function of CD4(+)CD25(+) regulatory T cells. Nat Immunol 2003; 4: 330.
12. Hori S, Takahashi T, Sakaguchi S. Control of autoimmunity by naturally arising regulatory CD4+ T cells. Adv Immunol 2003; 81: 331.
13. Khattri R, Cox T, Yasayko SA, Ramsdell F. An essential role for Scurfin in CD4(+)CD25(+) T regulatory cells. Nat Immunol 2003; 4: 337.
14. Fontenot JD, Rasmussen JP, Williams LM, et al. Regulatory T cell lineage specification by the forkhead transcription factor foxp3. Immunity 2005; 22: 329.
15. Salomon B, Lenschow DJ, Rhee L, et al. B7/CD28 costimulation is essential for the homeostasis of the CD4+CD25+ immunoregulatory T cells that control autoimmune diabetes. Immunity 2000; 12: 431.
16. Tai X, Cowan M, Feigenbaum L, Singer A. CD28 costimulation of developing thymocytes induces Foxp3 expression and regulatory T cell differentiation independently of interleukin 2. Nat Immunol 2005; 6: 152.
17. Tang Q, Henriksen KJ, Boden EK, et al. Cutting edge: CD28 controls peripheral homeostasis of CD4+CD25+ regulatory T cells. J Immunol 2003; 171: 3348.
18. Christ M, McCartney-Francis NL, Kulkarni AB, et al. Immune Dysregulation in TGF-b1 deficient mice. J Immunol 1994; 153: 1936.
19. Sakaguchi S. Naturally arising Foxp3-expressing CD25+CD4+ regulatory T cells in immunological tolerance to self and non-self. Nat Immunol 2005; 6: 345.
20. Chen W, Wahl SM. TGF-β: the missing link in CD4(+)CD25(+) regulatory T cell-mediated immunosuppression. Cytokine Growth Factor Rev 2003; 14: 85.
21. Fontenot JD, Rudensky AY. A well adapted regulatory contrivance: regulatory T cell development and the forkhead family transcription factor Foxp3. Nat Immunol 2005; 6: 331.
22. Sakaguchi S. Regulatory T cells: key controllers of immunologic self-tolerance. Cell 2000; 101: 455.
23. Shevach EM. CD4+ CD25+ suppressor T cells: more questions than answers. Nat Rev Immunol 2002; 2: 389.
24. Thompson C, Powrie F. Regulatory T cells. Curr Opin Pharmacol 2004; 4: 408.
25. Yu Q, Erman B, Bhandoola A, et al. In vitro evidence that cytokine receptor signals are required for differentiation of double positive thymocytes into functionally mature CD8+ T cells. J Exp Med 2003; 197: 475.
26. Apostolou I, Sarukhan A, Klein L, von Boehmer H. Origin of regulatory T cells with known specificity for antigen. Nat Immunol 2002; 3: 756.
27. Jordan MS, Boesteanu A, Reed AJ, et al. Thymic selection of CD4+CD25+ regulatory T cells induced by an agonist self-peptide. Nat Immunol 2001; 2: 301.
28. van Santen HM, Benoist C, Mathis D. Number of T reg cells that differentiate does not increase upon encounter of agonist ligand on thymic epithelial cells. J Exp Med 2004; 200: 1221.
29. Viglietta V, Baecher-Allan C, Weiner HL, Hafler DA. Loss of functional suppression by CD4+CD25+ regulatory T cells in patients with multiple sclerosis. J Exp Med 2004; 199: 971.
30. Chen W, Jin W, Tian H, et al. Requirement for Transforming Growth Factor β1 in Controlling T Cell Apoptosis. J Exp Med 2001; 194: 439.
31. Genestier L, Kasibhatla S, Brunner T, Green DR. Transforming growth factor β1 inhibits Fas ligand expression and subsequentactivation-induced cell death in T cells via downregulation of c-Myc. J Exp Med 1999; 189: 231.
32. Boise LH, Minn AJ, Noel PJ, et al. CD28 costimulation can promote T cell survival by enhancing the expression of Bcl-XL. Immunity 1995; 3: 87.
33. Clarkson MR, Sayegh MH. T-cell costimulatory pathways in allograft rejection and tolerance. Transplantation 2005 80:555–563.
Keywords:

Alloantigen specific Treg; Foxp3; Interleukin-2; CD2; Natural CD4+CD25+ Treg; Apoptosis

© 2006 Lippincott Williams & Wilkins, Inc.