TGF-β Induces Foxp3 + T-Regulatory Cells from CD4 + CD25 − Precursors

Authors


*Corresponding author: Jonathan S. Bromberg, jon.bromberg@msnyuhealth.org

Abstract

CD4 + CD25 + regulatory T cells (Tregs) are potent suppressors, playing important roles in autoimmunity and transplantation tolerance. Understanding the signals necessary for the generation and expansion of Tregs is important for clinical cellular therapy, but only limited progress has been made. Recent reports suggest a role for TGF-β in the generation of Tregs from CD4 + CD25 − precursors, but the mechanism remains unknown. Here, we demonstrate that TGF-β2 triggers Foxp3 expression in CD4 + CD25 − precursors, and these Foxp3 + cells act like conventional Tregs. The generation of Foxp3 + Tregs requires stimulation of the T-cell receptor, the IL-2R and the TGF-β receptor. More importantly, strong costimulation through CD28 prevents Foxp3 expression and suppressive function in an IL-4-dependent manner. Furthermore, TGF-β-driven Tregs inhibit innate inflammatory responses to syngeneic transplanted pancreatic islets and enhance islet transplant survival. Thus, TGF-β is a key regulator of the signaling pathways that initiate and maintain Foxp3 expression and suppressive function in CD4 + CD25 − precursors. TGF-β and signaling through TGF-β receptor, CD28 costimulation and IL-4 may be key components for the manipulation of Treg. The de novo generation of Foxp3 + cells from CD4 + cells has the potential to be used for treatment of autoimmune diseases and induction of transplant tolerance.

Introduction

CD4 + CD25 + regulatory T cells (Tregs) play a critical role in controlling autoimmunity and maintaining transplantation tolerance (1,2). CD4 + CD25 + Tregs are generated in the thymus, and account for only 5–10% of CD4 + cells in both mice and humans (3,4). The administration of sufficient numbers of freshly isolated CD4 + CD25 + Tregs is not therapeutically practical, so the expansion of CD4 + CD25 + Tregs is important for clinical cellular therapy. Several groups including our own have tried to expand CD4 + CD25 + Tregs in vitro, but with only limited success (5–8). In most situations, CD4 + CD25 + Tregs could be expanded 10–30-fold after 7–14 days in culture, but loose their anergic and suppressive properties after long-term culture. Therefore, current techniques are not ideal for the expansion of CD4 + CD25 + Tregs for cellular therapy, probably because little is known about the signals responsible for inducing or maintaining Tregs (9). Thus, a better understanding of signals important for Treg generation will likely result in improved protocols for in vitro expansion of naturally occurring Tregs or the de novo generation of Tregs from naïve precursors.

Transforming growth factor-β (TGF-β) is a multifunctional cytokine regulating T-cell growth and development (10–12). TGF-β inhibits IL-2 production, up-regulates cell-cycle inhibitors, and has potent antiproliferative effects on CD4 + T cells (13,14). TGF-β also has stimulatory effects, inducing the differentiation of Th3 cells and Tr1 cells (15,16). Recent reports suggest a role for TGF-β in the generation and expansion of Tregs (17). TGF-β induces CD4 + T cells to become CD4 + CD25 + Tregs, suppress T-cell cytotoxic activity and inhibit antibody production (18,19). IL-10 and TGF-β induce alloreactive CD4 + CD25 - T cells to acquire regulatory function and prevent GVHD, and the inhibitory function is independent of IL-10 and TGF-β (20). However, the mechanisms and the signal pathways by which TGF-β generates and expands these diverse Treg subsets remain unknown.

The forkhead/winged helix transcription factor Foxp3 is genetically defective in an autoimmune and inflammatory syndrome in humans (IpEX/XLAAD syndrome) and in mice (Scurfy) (21,22). Foxp3 is specifically expressed in naturally occurring CD4 + CD25 + Tregs, distinguishing them from activated CD4 + T cells, which also express CD25 after activation. Retroviral gene transfer of Foxp3 converts naïve T cells to a regulatory T-cell phenotype, similar to that of naturally occurring CD4 + Tregs (23). The lethal autoimmune syndrome observed in Foxp3-mutant (scurfy) and Foxp3-null mice results from CD4 + CD25 + Treg deficiency, and can be rescued by adoptive transfer these cells (24). Furthermore, transgenic mice overexpressing Foxp3 possess more CD4 + CD25 + Tregs, and both CD4 + CD25 - and CD4-CD8+ T cells in these mice also possess suppressive activity (25). Thus, Foxp3 is a key regulatory gene for the development and function of Tregs. It is currently the most specific marker for Tregs and may represent a ‘master switch’ which engages a developmental or regulatory program to generate Tregs. However, the mechanisms that regulate Foxp3 expression are not known, and the forced expression of Foxp3, either via retrovirus or transgenes, is not suitable for understanding the normal biology of Tregs and clinical cellular therapy. Therefore, it is important to define physiologic signals that can induce Foxp3 expression and generate Tregs from precursor CD4 + cells.

While investigating the growth characteristics of Tregs (8), we found that TGF-β is an important growth and survival factor for CD4 + CD25 + Tregs already committed to the Foxp3 program. As recent reports have suggested a role for TGF-β in the generation and expansion of Tregs (17–20), we hypothesized that TGF-β could induce de novo Foxp3 expression in CD4 + CD25 - precursor cells, which have not yet committed to the Foxp3 program. We confirmed this hypothesis by showing that TGF-β induces Foxp3 + CD4 + CD25 + Tregs from CD4 + CD25 − precursors. TCR, IL-2R, and TβR engagement are required to initiate Foxp3 expression and the Treg functional program, while strong costimulation prevents Treg development in an IL-4-dependent fashion. Thus, TGF-β and signaling through the TGF-β-receptor complex, costimulation, and cytokines may be key components to understanding the generation, expansion, and manipulation of Tregs. The generation and regulation of Foxp3 + Tregs has immediate relevance as clinically available techniques.

Methods

Mice

BALB/c (H-2d) male mice (8–10 weeks of age) were purchased from Harlan Sprague-Dawley (Indianapolis, IN). All mice were housed in a specific pathogen-free facility in microisolator cages. Splenic lymphocytes from 5C.C7 TCR Tg × Rag2−/− mice (26) were a kind gift from Dr M. Dustin (NYU). All experiments were performed with age- and sex-matched mice in accordance with IACUC-approved criteria.

Antibodies and reagents

The 145–2C11 hamster antimurine CD3 hybridoma was a gift of Dr J. A. Bluestone (UCSF, CA). The recombinant IL-2, TGF-β1, TGF-β2, BMP-2, BMP-4, BMP-7, IL-4, IL-6, IL-10, IL-12, IFN-γ, and TNF-α were from R & D Systems (Minneapolis, MN). PE, CyChrome and APC-anti-CD4 (RM4-5), FITC-anti-CD8 (53) (6,7), FITC-anti-CD25(7D4), PE-anti-CD62L (l-selectin), PE-anti-CD44 (Pgp-1), PE-anti-CD45RB (16 A), PE-anti-CD69 (H1.2F3), anti-TGF-β2, anti-IL-6, anti-IL-10, anti-IL-10R1, anti-CD28, anti-CTLA4, anti-PD-1, anti-IL-4, anti-IL-12, anti-IFN-γ, Cytofix/Cytoperm intracellular staining kit, and an Annexin V apoptosis detection kit were from BD PharMingen (San Diego, CA). Anti-CD40L was from Bender MedSystems (Vienna, Austria).

Adenovirus vectors

Type 5, E1-depleted, E3-defective adenovirus vectors expressing β-galactosidase under the control of the RSV promoter (AdRSVLacZ) were grown in our vector core. Vector titers were adjusted to 1 × 1010pfu/mL, and had a particle:pfu ratio < 100. All vectors contained < 0.1 unit/mL of endotoxin.

Flow cytometry and cell purification

For cell-surface phenotype analysis, murine splenocytes were stained with CyChrome-anti-CD4, FITC-anti-CD25, PE-anti-CD62L, anti-CD44, anti-CD45RB, or anti-CD69, and analyzed with CELLQuest™ software on a FACSCalibur (Becton Dickinson, San Jose, CA). For total cellular CTLA4 analysis, cells were permeabilized with a Cytofix/Cytoperm intracellular staining kit according to the manufacturer's protocol, and stained with PE-anti-CTLA4. For the detection of apoptosis, cells were stained with the Annexin V apoptosis detection kit according to the manufacturer's protocol. For cell purification, murine splenocytes were stained with CyChrome-anti-CD4 and FITC-anti-CD25, and subjected to a MoFlo cell-sorter (DakoCytomation, Fort Collins, CO). The purity of the sorted CD4 + CD25 − cells was >98%, and the purity of the sorted CD4 + CD25 + cells was >95%.

In some experiments, 5-carboxyfluorescein diacetate succinimidyl ester (CFSE, Molecular Probes, Eugene, OR) was added at a final concentration of 2.5 μM to 107 T cells/mL and incubated at room temperature for 5 min. Cells were washed with 5% FCS in PBS before use.

Primary cell culture

Two × 105 freshly isolated CD4 + CD25 − cells were added to 6-well plates (Costar) in 4-mL volumes in complete RPMI medium (RPMI 1640 supplemented with 10% fetal calf serum, 1 mM sodium pyruvate, 2 mM l-glutamine, 100 IU/mL penicillin, 100 μg/mL streptomycin, 1 × nonessential amino acids, and 2 × 10−5 M 2-mercaptoethanol). Cells were stimulated for 5 days at 37°C and 5% CO2 with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and 5 × 105 irradiated syngeneic T-depleted splenocytes in the presence or absence of TGF-β2 (1 ng/mL). T-depleted splenocytes were prepared by incubating splenocytes with mouse pan T Dynabeads (Dynal Biotech ASA, Oslo, Norway) for 20 min at 4°C and then irradiated at 500 rad. One hundred ng/mL of purified BMP-2, BMP-4, BMP-7, IL-4, IL-12, IFN-γ, TNF-α, or 5 μg/mL of anti-TGF-β2, anti-CD28, anti-CTLA4, anti-CD40L, anti-PD-1 mAb were added as indicated. At day 5, phenotype was analyzed by flow cytometry, apoptosis was detected by Annexin V staining, RNA was extracted for the Foxp3 expression, and the suppressive effects were assessed in secondary cultures after washing and removing the residual cytokines.

To evaluate the time course for TGF-β induction of Foxp3 expression in CD4 + CD25 − cells, 1 ng/mL of TGF-β2 was added to the culture, and Foxp3 expression and suppressive effects were tested at 12, 24, 48, 72 and 96 h. To evaluate the dose-response for TGF-β induction of Foxp3 expression in CD4 + CD25 − cells, different doses of TGF-β1 and TGF-β2 (10−4–101ng/mL) were added to the culture, and Foxp3 expression and suppressive effects were tested on day 5.

Proliferation and suppression assays

To analyze the proliferative capacities of the CD4 + CD25 − cells cultured in the presence or absence of TGF-β2, 2 × 104 cultured CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL) and 5 × 104 irradiated syngeneic T-depleted splenocytes. Cells were cultured for 3 days, pulsed with [3H]TdR for the last 18 h of culture, and collected with a Wallac Harvester (Tomtec, Hamden, CT). The incorporated [3H]TdR was measured with a Wallac Betaplate Counter (PerkinElmer, Boston, MA). In some experiments, exogenous IL-2 (100 U/mL) was added to the culture.

To test the suppressive properties of the CD4 + CD25 – cells cultured in the presence or absence of TGF-β2, the cultured CD4 + CD25 – cells were used as regulatory cells in secondary cultures. 4 × 104 freshly isolated responder CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL) and 1 × 105 irradiated syngeneic T-depleted splenocytes for 3 days, along with 2 × 104 Tregs or control cells (suppressor:responder ratio of 1 : 2). Cells were pulsed with 1 μ Ci [3H]TdR for the last 18 h of culture. Ten μg/mL of control IgG, anti-IL-10, anti-IL-10R, anti-TGF-β, or anti-CTLA4 mAbs were added as indicated.

To assess whether cell contact was necessary for the suppressive function, transwell chambers (Corning Costar, Cambridge, MA) were used. 2 × 105 freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL) and 5 × 105 irradiated syngeneic T-depleted splenocytes in the lower chambers of the plates. 1 × 105 Tregs or control cells was either added directly to the lower chambers or added to the upper chambers with anti-CD3 mAb (1 μg/mL) and 2.5 × 105 irradiated syngeneic T-depleted splenocytes. Cells were cultured for 3 days and pulsed with [3H]TdR for the last 18 h of culture.

In some experiments, freshly isolated CD4 + CD25 – cells were labeled with CFSE, and stimulated with anti-CD3 mAb, IL-2 and irradiated syngeneic T-depleted splenocytes in the presence or absence of TGF-β2. Cell division was monitored by CFSE intensity daily. After 4 days, the cultured CD4 + CD25 – cells were sorted into CFSEhigh and CFSElow subpopulations and used as regulatory cells in secondary cultures.

RT-PCR and real-time RT-PCR for Foxp3 expression

Total RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA), treated with RNase-free DNase I (Invitrogen), and reverse transcribed into cDNA using Sensiscript RT Kit (Qiagen, Valencia, CA) according to the manufacturer's protocol. The relative quantity of Foxp3 mRNA levels was first normalized by semiquantitative PCR. PCR consisted of a 5-min 95°C denaturation step followed by 30 cycles of 45 s at 94°C, 45 s at 56°C and 60 s at 72°C. Foxp3 primers: 5′–CAG CTG CCT ACA GTG CCC CTA– 3′ and 5′–CAT TTG CCA GCA GTG GGT AG–3′; GAPDH primers: 5′–TGA AGG TCG GTG TGA ACG GAT′ and 5′–CAG GGG GGC TAA GCA GTT GGT–3′. Foxp3 mRNA levels were quantified by real-time PCR using QuantiTect SYBR Green PCR Kit (Qiagen) with the LightCycler (Roche, Indianapolis, IN). PCR consisted of a 15-min 95°C denaturation step followed by 45 cycles of 15 s at 94°C, 20 s at 56°C and 15 s at 72°C. Foxp3 primers: 5′–CCC AGG AAA GAC AGC AAC CTT–3′ and 5′–TTC TCA CAA CCA GGC CAC TTG–3′; Cyclophilin A primers: 5′–AGG GTG GTG ACT TTA CAC GC–3′ and 5′–ATC CAG CCA TTC AGT CTT GG–3′. Normalized values for Foxp3 mRNA expression were calculated as the relative quantity of Foxp3 divided by the relative quantity of Cyclophilin A.

Islet isolation, transduction and transplantation

BALB/c mice were anesthetized with ketamine chloride (Fort Dodge Animal Health, Fort Dodge, IA), and injected intraductally of 3 mL of 1.5 mg/mL collagenase P (Roche Diagnosis, NJ). Pancreata were procured and digested at 37°C for 20 min. Islets were purified by discontinuous gradient centrifugation using Ficoll (Sigma, St. Louis, MO) and hand-picked under an inverted microscope. One hundred and fifty freshly isolated islets were transduced with AdRSVLacZ in serum-free CMRL 1066 medium at a multiplicity of infection (MOI) of 105pfu/islet at 37°C for 1 h, washed, mixed with 1.5 × 105 Tregs in complete CMRL 1066 medium with 10% FBS at 37°C for 4 h, washed and then transplanted.

The marginal mass islet transplantation model was used for islet transplantation (27). Briefly, syngeneic BALB/c recipient mice were rendered diabetic by a single intraperitoneal injection of 180 mg/kg of streptozotocin (STZ, Sigma, St. Louis, MO) at least 1 week previously. Diabetes was confirmed by the presence of hyperglycemia (>300 mg/dL blood glucose for two consecutive days). One hundred and fifty freshly isolated or AdRSVLacZ-transduced or AdRSVLacZ-transduced islets plus Tregs were transplanted underneath the subcapsular space of the right kidney. Blood glucose was monitored daily to assess islet graft function. Reversal of diabetes was defined as blood glucose < 200 mg/dL on two consecutive measurements. Grafts were considered to be rejected if blood glucose was >250 mg/dL on two consecutive measurements.

Statistical analysis

Analysis for statistically significant differences was performed with Student's t-test. p < 0.05 was considered a difference, and p < 0.01 was considered a significant difference. Results are expressed as mean ± SEM, and are representative of 3–5 similar experiments. For marginal mass islet transplantation, there were 5–6 mice in each group.

Results

TGF-β2 induces Foxp3 expression in CD4 + CD25 – precursors

Freshly isolated CD4 + CD25 – cells express very little Foxp3, which has already been reported by several other groups (23–25). After 5 days' culture with anti-CD3 mAb plus IL-2 plus TGF-β2, CD4 + CD25 – cells expressed significantly higher levels of Foxp3, whereas CD4 + CD25 – cells cultured with anti-CD3 mAb alone or anti-CD3 mAb plus IL-2 expressed extremely low levels of Foxp3 (Figure 1a). CD4 + CD25 – cells cultured with IL-2 or TGF-β alone, anti-CD3 mAb plus TGF-β, or IL-2 plus TGF-β did not proliferate (not shown). The addition of anti-TGF-β blocked TGF-β-driven Foxp3 expression, proving specificity; but it could not further decrease the low background of Foxp3 expression. For naturally occurring CD4 + CD25 + cells, both cells cultured with anti-CD3 plus IL-2, or anti-CD3 plus IL-2 plus TGF-β2, expressed Foxp3. Real-time RT-PCR showed similar results; CD4 + CD25 – cells cultured with anti-CD3 mAb plus IL-2 plus TGF-β2 expressed high levels of Foxp3 (Figure 1b).

Figure 1.

Figure 1.

TGF-β2 induces Foxp3 expression in CD4 + CD25 – precursors. (A,B) 2 × 104 freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and 5 × 104 irradiated syngeneic T-depleted splenocytes in the presence or absence of TGF-β2 (1 ng/mL). Foxp3 expression was detected on day 5 by RT-PCR (A) and real-time RT-PCR (B). (C,D) For the time course, 1 ng/mL of TGF-β2 was added to the culture and cells collected after 12, 24, 48, 72 or 96 h. Foxp3 expression was assessed by RT-PCR (C) and real-time RT-PCR (D). (E,F) For dose–response, 10−4–101ng/mL of TGF-β2 was added to the culture and Foxp3 expression was assessed after 5 days by RT-PCR (E) and real-time RT-PCR (F).

Figure 1.

Figure 1.

TGF-β2 induces Foxp3 expression in CD4 + CD25 – precursors. (A,B) 2 × 104 freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and 5 × 104 irradiated syngeneic T-depleted splenocytes in the presence or absence of TGF-β2 (1 ng/mL). Foxp3 expression was detected on day 5 by RT-PCR (A) and real-time RT-PCR (B). (C,D) For the time course, 1 ng/mL of TGF-β2 was added to the culture and cells collected after 12, 24, 48, 72 or 96 h. Foxp3 expression was assessed by RT-PCR (C) and real-time RT-PCR (D). (E,F) For dose–response, 10−4–101ng/mL of TGF-β2 was added to the culture and Foxp3 expression was assessed after 5 days by RT-PCR (E) and real-time RT-PCR (F).

A time-course analysis between 12 and 96 h showed no significant Foxp3 expression in CD4 + CD25 – cells during the first 48 h of culture with TGF-β2, but after 72 h there was a large increase in Foxp3 expression (Figure 1c,d). A dose–response analysis demonstrated that both TGF-β1 and TGF-β2 induced Foxp3 expression in CD4 + CD25 – cells in a dose-dependent manner, with as little as 0.1 ng/mL of either ligand inducing Foxp3 expression. One ng/mL TGF-β1 or TGF-β2 induced greater Foxp3 expression, and 10 ng/mL of TGF-β1 or TGF-β2 did not further increase Foxp3 expression levels (Figure 1e,f; TGF-β1 results not shown). Therefore, TGF-β1 and TGF-β2, which have been reported to have similar or disparate biological effects in other systems (28,29), had similar effects in inducing Foxp3 expression; and we chose 1.0 ng/mL as the dose of TGF-β2 for further experiments.

It is possible that there was outgrowth of a contaminating CD4 + CD25 + population in the cultures. However, we believe this is unlikely because the starting CD4 + CD25 – population was greater than 98% pure, virtually all cells were CD25+ at the end of the 5-day cultures (Figure 2a), making complete overgrowth of a minor population unlikely, and the expansion of CD4 + CD25 – cells in these culture conditions was far greater than the expansion of CD4 + CD25 + cells (10–15-fold vs. twofold in 5 days, not shown). Lastly, we purified CD4 + T-cell populations from 5C.C7 TCR Tg × Rag2−/− mice. The TCR in these Tg mice is specific for a pigeon cytochrome C peptide (26) unlikely to be expressed in the thymus. Thus, on a Rag2−/− background, the CD4 + T cells are all TCR Tg monospecific, and CD4 +CD25+ natural Tregs are not selected in the thymus or periphery. Indeed, the results in Table 1 show that there are very few splenic CD4 +CD25+ T cells and that they express extremely low levels of Foxp3. Nonetheless, stimulation of TCR Tg × Rag2−/− CD4 +CD25 T cells with anti-CD3 mAb plus IL-2 plus TGF-β2 results in a substantial increase in Foxp3 expression. These results make it extremely unlikely that there is overgrowth of a contaminating CD4 +CD25+ Treg population in these cultures.

Figure 2.

TGF-β-driven Foxp3 + cells act as naturally occurring CD4 + CD25 + Tregs. (A) The cell-surface phenotype of freshly isolated CD4 + CD25 – cells (dotted line) and 5-day-cultured CD4 + CD25 – cells in the presence (thick line) or absence (thin line) of TGF-β2. (B) 2 × 104 5-day-cultured CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL) and 5 × 104 irradiated syngeneic T-depleted splenocytes, and exogenous IL-2 (100 U/mL) was added as indicated. Wells were pulsed with [3H]-TdR for the last 18 h of the 72-h culture. (C) 4 × 104 freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL) and 1 × 105 irradiated syngeneic T-depleted splenocytes, and 2 × 104 5-day-cultured CD4 + CD25 – cells were added as Tregs. In the transwell plate, the 5-day-cultured CD4 + CD25 – cells were either added directly to the lower chambers or added to the upper chambers with anti-CD3 mAb and irradiated syngeneic T-depleted splenocytes. (D) In secondary cultures to assess suppression, 10 μg/mL of control IgG, anti-IL-10, anti-IL-10R, anti-TGF-β, or anti-CTLA4 mAbs was added as indicated.

Table 1. Foxp3+ cells are generated from highly purified CD4+ CD25 precursors
 Splenocyte source
5C.C7 TCR
Tg × Rag2 –/–

Wild type
% CD4 + splenocytes76.629.3
% CD4 + CD25 + T cells0.225.6
Relative Foxp3 expression to Cyc A 
Fresh CD4 + CD25 +0.3668.59
Fresh CD4 + CD25 –0.154.62
CD4 + CD25 + cx'd with αCD3/IL-2ND45.57
CD4 + CD25 – cx'd with αCD316.118.51
CD4 + CD25 – cx'd with αCD3/IL-25.810.86
CD4 + CD25 – cx'd with αCD3/IL-2/TGF-β241.0725.11

To determine the stability of Foxp3 expression induced by TGF-β, CD4 + CD25 – cells were stimulated with anti-CD3 mAb, IL-2 and T-depleted splenocytes along with TGF-β2 for 5 days, and then washed and restimulated with anti-CD3 mAb plus IL-2 in the absence of TGF-β2 for another 5 days. At the end of secondary culture, cells were harvested and RNA extracted for Foxp3 expression. Real-time RT-PCR showed continued Foxp3 expression in these cells, revealing at least short-term stability of Foxp3 expression in CD4 + CD25 – cells induced by TGF-β (Figure 3c). Furthermore, these cells retained suppressor function Figure 3(d).

Figure 3.

TGF-β2 inhibits cell division and apoptosis. (A) Freshly isolated CD4 + CD25 – cells were labeled with CFSE, stimulated with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and syngeneic T-depleted splenocytes in the presence or absence of TGF-β2 (1 ng/mL) for 4 days. Cell division was monitored by CFSE intensity. (B,C) After 4 days, CD4 + CD25 – cells cultured in the presence or absence of TGF-β2 were sorted into CFSEhigh and CFSElow subpopulations, and Foxp3 expression was detected by real-time RT-PCR (B), and 2 × 104 sorted cells were used as Tregs in secondary cultures (C). Results shown are representative of three similar experiments. (D) Freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and syngeneic T-depleted splenocytes in the presence or absence of TGF-β2 (1 ng/mL). After 4 days, cells were stained for Annexin V. Cells cultured in the presence of TGF-β(thick line), and cells cultured in the absence of TGF-β(thin line).

TGF-β-driven Foxp3 + cells act as conventional CD4 + CD25 + Treg

The cell-surface phenotype of the CD4 + CD25 – cells cultured in the presence or absence of TGF-β2 was analyzed and found to be very different between the two populations (Figure 2a). Before culture, < 2% of the freshly sorted CD4 + CD25 – cells expressed CD25, whereas after culture >95% of the cultured CD4 + CD25 – cells expressed CD25, regardless if they were cultured in the presence or absence of TGF-β2, suggesting that CD25 is a marker for cell activation, that CD25 is not a marker to discriminate Tregs from activated effector T cells, and that CD25 expression does not signify rapid overgrowth of a small contaminating population of CD4 + CD25 + cells. The majority (84%) of the CD4 + CD25 – cells cultured in the presence of TGF-β2 did not express CD62L, whereas the majority (82%) of the CD4 + CD25 – cells cultured in the absence of TGF-β2 expressed CD62L. Furthermore, 70% of the Foxp3 + cells were CD45RBlow, whereas 68% of the Foxp3- cells were CD45RBhigh. Expression of CD25, CD44, CD69, and total cellular CTLA4 were similar in both groups of cells and were all induced after activation.

To determine if the Foxp3 + population behaved like naturally occurring CD4 + CD25 + Treg, CD4 + CD25 – cells cultured in the presence or absence of TGF-β2 were harvested after 5 days and re-stimulated with anti-CD3 mAb and syngeneic T-depleted splenocytes. The CD4 + CD25 – cells cultured in the presence of TGF-β2 (which were Foxp3 +) did not proliferate in response to TCR stimulation, similar to fresh CD4 + CD25 + Treg. CD4 + CD25 – cells cultured in the absence of TGF-β2 (which were Foxp3-) proliferated in response to TCR stimulation, similar to fresh CD4 + CD25 – cells. Proliferation of TGF-β-driven Foxp3 + cells could be restored with exogenous IL-2, showing that these cells were anergic (Figure 2b).

To test the suppressive properties of the CD4 + CD25 – cells cultured in the presence or absence of TGF-β2, they were used as regulatory cells in a secondary culture to determine whether they inhibited the proliferation of freshly isolated responder CD4 + CD25 – cells. At a suppressor:responder ratio of 1 : 2, the CD4 + CD25 – cells cultured in the presence of TGF-β2 significantly inhibited responder-cell proliferation, similar to fresh CD4 + CD25 + Tregs. The CD4 + CD25 – cells cultured in the absence of TGF-β2 did not inhibit T-cell proliferation, similar to fresh CD4 + CD25 – cells (Figure 2c).

It has been shown repeatedly that the suppressive function of CD4 + CD25 + Tregs is cell-contact-dependent and cytokine-independent in vitro (30). We conducted a similar analysis. (Figure 2c) demonstrates that if separated from the CD4 + CD25 – responders in the transwell plate, the CD4 + CD25 – cells cultured in the presence of TGF-β2 would not inhibit T-cell proliferation, similar to fresh CD4 + CD25 + Tregs. The results in Figure 2(d) show that the addition of anti-IL-10, anti-IL-10R, anti-TGF-β or anti-CTLA4 mAbs did not block the suppressive function of the CD4 + CD25 – cells cultured in the presence of TGF-β2, similar to fresh CD4 + CD25 + Tregs.

Therefore, the Foxp3 + cells induced by TGF-β2 from CD4 + CD25 – precursor cells acted like naturally occurring CD4 + CD25 + Tregs in vitro. They were anergic and suppressive via cell-contact-dependent and cytokine-independent mechanisms.

TGF-β2 inhibits cell division and apoptosis

To evaluate the role of TGF-β2 in cell division, freshly isolated CD4 + CD25 – cells were labeled with CFSE, cultured in the presence or absence of TGF-β2, and cell division was monitored by CFSE intensity daily. As reported by others, cells cultured in the presence of TGF-β2 divided less (Figure 4a, Table 2). For cells cultured for 4 days in the absence of TGF-β2, 77.6% cells divided well (CFSElow), 9.5% cells divided only 1–2 cycles (CFSEinter), and 12.9% cells did not divide (CFSEhigh). For cells cultured in the presence of TGF-β2, 61.7% cells divided well (CFSElow), 14.1% cells divided only 1–2 cycles (CFSEinter), and 24.2% cells did not divide (CFSEhigh).

Figure 4.

Anti-CD28 mAb blocks TGF-β-driven Foxp3 expression in CD4 + CD25 – precursors. (A,B) 2 × 104 freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and 5 × 104 irradiated syngeneic T-depleted splenocytes in the presence or absence of TGF-β2 (1 ng/mL). Five μg/mL of purified anti-CD28, anti-CTLA4, anti-CD40L, or anti-PD-1 mAbs were added as indicated. After 5 days of culture, Foxp3 expression was determined by real-time RT-PCR (A), and 2 × 104 cultured cells were used as Tregs in secondary cultures (B). (C,D) 2 × 104 freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and 5 × 104 irradiated syngeneic T-depleted splenocytes, TGF-β2 (1 ng/mL) and/or anti-CD28 mAb (5 μg/mL) were added as indicated. After 5 days, cells in each primary culture condition were divided into four groups, and put into one of four secondary culture conditions. Five days later, Foxp3 expression was determined by real-time RT-PCR (C), and cells were used as Tregs in the suppression assay (D). *p < 0.01.

Table 2. TGF-β inhibits the proliferation of CD4 + CD25 – cells
 % cells in each peak
CD4 + CD25–/
Anti-CD3 + IL-2
CD4 + CD25 –/
Anti-CD3 +IL-2 + TGF-β
M112.924.2
M23.36.1
M36.28.0
M413.714.1
M522.019.3
M626.418.8
M715.59.5
CFSEhigh (M1)12.924.2
CFSEinter
(M2 + M3)
9.514.1
CFSElow
(M4 ∼ M7)
77.661.7

To determine the effect of cell division on Foxp3 expression and Treg function, the cultured CD4 + CD25 – cells were sorted into the CFSEhigh and the CFSElow subpopulations, RNA was isolated for Foxp3 expression and cells were used as regulatory cells in secondary cultures. Real-time RT-PCR and suppressive assay showed consistent results (Figure 4b,c). The CFSEhigh and the CFSElow subpopulations cultured in the presence of TGF-β2 both expressed Foxp3 and suppressed responder cell proliferation, whereas the CFSEhigh and the CFSElow subpopulations cultured in the absence of TGF-β2 neither expressed Foxp3 nor inhibited cell proliferation. Therefore, TGF-β2 inhibited cell division, but T-cell division did not affect Foxp3 expression and Treg function induced by TGF-β2.

TGF-β is a multifunctional protein, and is required for the maintenance of homeostatic balance between cell proliferation and apoptosis (31,32). Several groups have demonstrated that TGF-β has both antiapoptotic and pro-apoptotic effects, depending on the cell type or cellular context (33,34). To determine the role of TGF-β in regulating apoptosis in our culture system, CD4 + CD25 – cells were stimulated with anti-CD3 mAb plus IL-2, cultured in the presence or absence of TGF-β for 4 days, and cells assessed for apoptosis by staining for Annexin V. The results in Figure 4(d) show that TGF-β inhibited apoptosis, with only 2% of CD4 + CD25 – cells cultured with TGF-β being Annexin V +, whereas 21% of CD4 + CD25 – cells cultured without TGF-β were Annexin V + 0. Therefore, TGF-β had antiapoptotic effects in addition to inducing Foxp3 expression.

Anti-CD28 blocks TGF-β-driven Foxp3 expression in CD4 + CD25 – precursors

The data show that signals delivered by the TCR, the IL-2R, and the TβR are required to generate Foxp3 + Tregs. To explore additional signaling pathways that could enhance or prevent the generation of Foxp3 + Treg, CD4 + CD25 – cells were stimulated in the presence or absence of TGF-β2 along with anti-CD28, anti-CTLA4, anti-CD40L, or anti-PD-1 mAbs. After 5 days, RNA was extracted for Foxp3 expression, and cells were used as regulatory cells in secondary cultures. Real-time RT-PCR showed that the addition of stimulatory anti-CD28 mAb completely blocked Foxp3 expression in CD4 + CD25 – cells induced by TGF-β2, but did not affect Foxp3 expression in conventional naturally occurring CD4 + CD25 + Tregs (Figure 3a). The suppressive assay was consistent with the real-time RT-PCR result; the addition of anti-CD28 mAbs completely blocked the acquisition of suppressive function by CD4 + CD25 – cells cultured with TGF-β2, but did not affect the suppressive function of CD4 + CD25 + Tregs (Figure 3b). Addition of anti-CTLA4, anti-CD40L, or anti-PD-1 mAbs did not affect either Foxp3 expression or Treg function in CD4 + CD25 – cells cultured with TGF-β2. Therefore, strong costimulation through CD28 prevented the development of Foxp3 + Tregs from CD4 + CD25 – precursors.

To determine whether anti-CD28 mAb could reverse the suppressive function induced by TGF-β, CD4 + CD25 – cells were stimulated with anti-CD3 mAb, IL-2, TGF-β2 and/or anti-CD28 mAb. After 5 days, cells from each primary culture condition were divided into four groups, placed into secondary cultures for 5 days, and re-stimulated in the presence or absence of TGF-β and/or anti-CD28 mAb. Cells were then harvested and evaluated for Foxp3 expression and suppressor function (Figure 3c,d). Anti-CD28 mAb reversed the suppressive function induced by TGF-β. After culture in the presence of TGF-β for 5 days, CD4 + CD25 – cells became Foxp3 + Tregs, and retained the suppressive function after secondary culture with anti-CD3 mAb plus IL-2, or anti-CD3 mAb plus IL-2 plus TGF-β. The presence of TGF-β in the secondary culture resulted in greater Foxp3 mRNA expression and suppression compared with secondary culture without TGF-β. However, if anti-CD28 mAb was added to secondary culture, CD4 + CD25 – cells lost suppressive function and had decreased Foxp3 expression compared with cultures not receiving costimulation. Moreover, anti-CD28 mAb blocked the subsequent ability of TGF-β to induce Tregs from CD4 + CD25 – precursors. If anti-CD28 mAb was added to primary culture, CD4 + CD25 – cells did not become Tregs after secondary culture with TGF-β. Therefore, the addition of anti-CD28 mAb reversed and blocked the effect of TGF-β to induce Tregs from CD4 + CD25 – precursors.

Anti-CD28 blocks Foxp3 expression in IL-4-dependent manner

As TGF-β inhibits the transcription factors required for expression of IFN-γ and IL-4 (35,36), we wondered whether IL-4 or IFN-γ impaired the signaling pathways that lead to Foxp3 expression. To test this hypothesis, CD4 + CD25 – cells were cultured in the presence or absence of TGF-β2 along with IL-4, IL-12, IFN-γ, or TNF-α. After 5 days, RNA was extracted for Foxp3 expression, and cells were used as regulatory cells in secondary cultures. Real-time RT-PCR showed that addition of IL-4 completely blocked Foxp3 expression in CD4 + CD25 – cells incubated with TGF-β2, while addition of IL-12 or IFN-γ partially blocked Foxp3 expression, and TNF-α did not affect Foxp3 expression (Figure 5a). The suppressive assay showed that the addition of IL-4 completely blocked the acquisition of suppressive function by CD4 + CD25 – cells cultured with TGF-β2, but the addition of IL-12, IFN-γ and TNF-α did not affect suppressive function (Figure 5b). Additional experiments showed that IL-6, IL-10, anti-IL-6, anti-IL-10, and anti-IL-10R1 mAbs did not affect Foxp3 expression or regulatory function (not shown). Therefore, IL-4 provides negative signal(s) for the development of Foxp3 expression in CD4 + CD25 – precursor cells, inhibiting the effects of TGF-β2.

Figure 5.

Anti-CD28 mAb blocks Foxp3 expression in an IL-4-dependent manner. (A,B) 2 × 104 freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and 5 × 104 irradiated syngeneic T-depleted splenocytes in the presence or absence of TGF-β2 (1 ng/mL) along with 100 ng/mL of IL-4, IL-12, IFN-γ, or TNF-α as indicated. Foxp3 expression was detected on day 5 by real-time RT-PCR (A), and 2 × 104 cultured cells were used as Tregs in secondary cultures (B). (C,D) 2 × 104 freshly isolated CD4 + CD25 – cells were stimulated with anti-CD3 mAb (1 μg/mL), IL-2 (100 U/mL) and 5 × 104 irradiated syngeneic T-depleted splenocytes in the presence of TGF-β2 (1 ng/mL) and anti-CD28 (5 μg/mL), anti-IL-4, anti-IL-12, or anti-IFN-γ mAbs (5 μg/mL) were added as indicated. Foxp3 expression was detected on day 5 by real-time RT-PCR (C), and 2 × 104 cultured cells were used as Tregs in secondary cultures (D).

As anti-CD28 mAb blocks TGF-β-driven Foxp3 expression and as strong costimulation enhances cytokine production, we speculated whether anti-CD28 mAb induced cytokines that subsequently blocked acquisition of Foxp3. CD4 + CD25 – cells were cultured in the presence of TGF-β2 and anti-CD28 mAb along with anti-IL-4, anti-IL-12, or anti-IFN-γ mAbs. After 5 days, RNA was extracted for Foxp3 expression, and the cells were used as regulatory cells in secondary cultures. Real-time RT-PCR showed that the addition of anti-CD28 mAb blocked TGF-β-driven Foxp3 expression, and the addition of anti-IL-4 mAb reversed the effect of anti-CD28 mAb (Figure 5c). The suppressive assay was consistent with the real-time RT-PCR result; the addition of anti-CD28 mAb completely blocked the acquisition of suppressive function induced by TGF-β2, while anti-IL-4 mAb reversed this effect (Figure 5d). The addition of anti-IL-12 or anti-IFN-γ mAbs did not affect either Foxp3 expression or Treg function in CD4 + CD25 – cells cultured with TGF-β2 and anti-CD28 mAb. Therefore, anti-CD28 mAb blocks TGF-β-driven Foxp3 expression through an IL-4-dependent pathway.

TGF-β-driven Foxp3 + cells inhibit innate inflammatory responses

Our previous report showed that pancreatic islets produce chemokines and that chemokine production is enhanced by stimuli such as surgical manipulation, ischemia, inflammatory cytokines, or adenoviral gene transfer vectors. Further, these chemokines attract inflammatory infiltrates into transplanted islets, and the infiltrates inhibit the engraftment and function of a marginal islet T-cell mass transplanted to syngeneic diabetic recipients (27). This model is useful for evaluating the regulation of inflammatory and innate immune responses. Several groups have demonstrated that naturally occurring CD4 + CD25 + Tregs can suppress inflammation or innate immune responses (37–40). To evaluate the in vivo function of Foxp3 + cells induced by TGF-β from CD4 + CD25 – precursors, we employed the marginal mass islet transplant model. Transplantation of freshly isolated pancreatic islets reversed diabetes in syngeneic, streptozotocin-induced, diabetic recipients. Adenovirus vector transduction of the islets prevented sustained engraftment of the islets, with return of hyperglycemia within 48 h. Co-transfer of naturally occurring CD4 + CD25 + Tregs along with the islets inhibited innate immunity and permitted islet engraftment and diabetes cure (Figure 6a). Co-transfer of freshly isolated CD4 + CD25 – cells or CD4 + CD25 – cells cultured in the absence of TGF-β did not reverse diabetes. However, cotransfer of Foxp3 + CD4 + CD25 – cells cultured in the presence of TGF-β reversed diabetes (Figure 6b). Therefore, both conventional Tregs and TGF-β-driven Tregs inhibit innate inflammatory responses to transplanted pancreatic islets and enhance islet transplant survival.

Figure 6.

Both naturally occurring and TGF-β-driven Tregs inhibit innate inflammatory responses. Syngeneic BALB/c recipient mice were rendered diabetic by streptozotocin. One hundred and fifty freshly isolated or AdRSVLacZ-transduced islets were transplanted underneath the subcapsular space of the right kidney. Transduces islets were coinjected with Tregs or control T-cell populations. Blood glucose was monitored daily to assess islet graft function. Reversal of diabetes was defined as blood glucose <200 mg/dL on two consecutive measurements. Grafts were considered to be rejected if blood glucose was >250 mg/dL on two consecutive measurements. (A) Naturally occurring CD4 + CD25 + Tregs inhibit innate inflammatory responses. *p < 0.05 vs. control group; **p < 0.01 vs. control group. (B) TGF-β-driven Foxp3 + cells inhibit innate inflammatory responses. *p < 0.05 vs. fresh CD4 + CD25 – group; **p < 0.01 vs. fresh CD4 + CD25 – group.

Discussion

Wahl and her colleagues recently reported that TGF-β induces Foxp3 expression and suppressor function in CD4 + CD25 + naïve T cells (41). Our report confirms and substantially extends those findings. The TCR, IL-2R, and TβR must all be stimulated to induce Foxp3 + Tregs. Failure to engage any one of these receptors prevents the generation of Foxp3 + Tregs. Dose–response and kinetics show that as little as 0.1 ng/mL of TGF-β1 or TGF-β2, and as early as 72 h after culture with TGF-β2, induces strong Foxp3 expression. These Foxp3 + cells act similar to naturally occurring CD4 + CD25 + Tregs: they are anergic; suppress via cell contact-dependent, cytokine-independent mechanisms; and inhibit inflammatory responses in vivo. In our culture system, we used fetal bovine serum instead of serum-free medium. Although the serum contained latent TGF-β, the TGF-β in the serum was not enough to induce Foxp3 expression, and addition of anti-TGF-β to cultures did not decrease further the low background of Foxp3 expression. Only the addition of exogenous TGF-β induced both high-level Foxp3 expression and suppressive function in the CD4 + CD25 – precursor cells. While BMPs regulate Smads and TβR signaling, they do not appear to influence Foxp3 expression and Treg function (data not shown). Conversely, T-cell-specific pathways involving costimulation and cytokines do have a major influence on Foxp3 expression and Treg suppressor function. In our experiments, we observed a low background of Foxp3 expression in CD4 + CD25 – cells. Thus, it is possible that there was outgrowth of a contaminating CD4 + CD25 + population in the cultures. However, we believe this is unlikely because the starting CD4 + CD25 – population was greater than 98% pure, virtually all cells were CD25+ at the end of the 5-day cultures (Figure 2a), making complete overgrowth of a minor population unlikely, and the expansion of CD4 + CD25 – cells in these culture conditions was far greater than the expansion of CD4 + CD25 + cells. Lastly, the experiments with the TCR Tg T cells on a Rag2−/− background (Table 1) definitively showed de novo generation of Foxp3+ cells from a precursor population lacking natural CD4 +CD25+ Tregs.

To date, little is known about the signaling pathways or transcription factors specifically controlling CD4 + CD25 + Treg-cell development. Foxp3 is currently the most specific molecular marker for Tregs. Unlike the cell-surface markers used to identify Tregs such as CD25, CD45RB, CTLA4, and GITR; Foxp3 is not up-regulated upon activation, and discriminates Tregs from activated effector T cells. Several recent reports revealed a strong correlation between Foxp3 expression and Tregs, suggesting that Foxp3 plays a critical role in the development and function of CD4 + CD25 + Tr egs (23–25,42,43). While retroviral transduction of Foxp3 to naïve T cells generates Tregs, the forced expression of Foxp3 can lead to erroneous interpretations about Treg biology, as the molecule is over-expressed in an unregulated fashion. It is critical to define physiologic signals that permit Foxp3 expression. In our study, we showed that TGF-β induced strong Foxp3 expression in CD4 + CD25 – precursor cells, and these Foxp3 + cells were functionally similar to the naturally occurring CD4 + CD25 + Tregs. Thus, it is now possible to analyze the signaling and transcriptional programs that permit Treg development and function. The de novo generation of Foxp3 + Tregs from CD4 + CD25 – precursors has important practical implications. Simple manipulation of an extracellular ligand now enables the generation of a large number of Tregs. This may open up novel treatment strategies for autoimmune disease and transplantation tolerance. An important issue is how stable Foxp3 expression and Treg function remain after withdrawal of TGF-β. In very short-term cultures, Foxp3 expression and suppressor function are retained, however, this must be evaluated over longer time intervals in vivo.

Five isoforms of TGF-βs, termed TGF-β1–5, have been identified to date (31,44). TGF-β1 and TGF-β2 have greater than 97% identity, signal through the same type I (TβR-I) and type II (TβR-II) serine-threonine kinase receptors, and have similar effects on immune cells, including lymphocytes, macrophages, and dendritic cells (45). TGF-β2 has a much lower binding affinity to TβR-II than TGF-β1 (46). Both TGF-β1 and TGF-β2 similarly induced Foxp3 expression in our studies. Considering the much lower binding affinity of TGF-β2 to TβR-II than TGF-β1, TGF-β2 may be relatively more potent in inducing Foxp3 expression, although in vitro presentation of these cytokines may mask affinity differences. It is notable that BMP-2, BMP-4 and BMP-7 did not induce or inhibit Foxp3 expression in the CD4 + CD25 – cells (unpublished data S. Fu and J. S. Bromberg, 2004). It will be important to investigate further the regulation of TβRs and Smads in these cells and elucidate their interaction with TCR, IL-2R, and cytokine receptor-derived signals.

CD28 is necessary for the homeostasis of naturally occurring Tregs in the periphery, and also controls the development of natural Tregs in the thymus (47). However, CD28 is dispensable for the suppressive function of these Tregs in vitro and in vivo. CD4 + CD25+ T cells from CD28-deficient mice exhibit potent suppressor activity in vitro, and prevent intestinal mucosal inflammation induced by CD4 + CD45RBhigh cells, suggesting that the CD4 + CD25 + Treg function in a CD28-independent manner (48,49). CD28 costimulation of natural Tregs attenuates their suppressive activity, and stimulatory anti-CD28 mAb can reverse the anergic phenotype of CD4 + CD25 + Tregs. Removal of anti-CD28 mAb allows Tregs to revert to their original anergic and suppressive state (50,51). In our experiments, the addition of stimulatory anti-CD28 mAb completely blocked TGF-β-driven Foxp3 expression in naïve CD4 + CD25 – precursors, while the addition of anti-CTLA-4, anti-PD-1 or anti-CD40L mAbs did not alter Foxp3 expression or Treg suppressor function. Therefore, there appears to be specificity for CD28-derived costimulatory signals to inhibit Foxp3 expression. Furthermore, IL-4, but not IL-6 or IL-10, also prevented Foxp3 expression, and TGF-β2 did not induce Foxp3 expression in Th1, Th2, or CD8 + cells (data not shown). Thus, strong costimulatory signals, such as CD28, and Th1 or Th2 differentiation, block the generation of Foxp3 + Tregs from CD4 + CD25 – precursors. CD28 costimulation seems to function through induction of Th2 cytokines, as anti-IL-4 mAb completely blocked its effects. Further investigations will be warranted to delineate IL-4/IL-4R/Stat6 signaling pathway involvement in Foxp3 expression, and what alternative pathways Th1 and CD8 + cells use to regulate Foxp3 expression.

Our results may explain the failure of several groups to demonstrate Foxp3 expression in CD4 + CD25 – cells in vitro, either because TGF-β was not added to their culture systems, and/or because costimulation with anti-CD28 mAb or Th1 and Th2 cytokines impaired Foxp3 expression and suppressor function in naïve CD4 + CD25 – T cells (23–25). In contrast to our findings, Chen et al. did find Foxp3 induction in anti-CD28 mAb stimulated cultures but did not seem to find a requirement for IL-2. The reasons for these discrepancies may relate to mouse strains or to the very different culture conditions. Further, we were careful to analyze the same cells for both Foxp3 expression and suppressor function, but that was not reported in their study. The finding that TGF-β induces Tregs and inhibits Th1 and Th2 differentiation, while CD28 costimulation inhibits and extinguishes generation of Tregs from CD4 + CD25 – precursors while stimulating Th1 and Th2 differentiation, suggests a model whereby the kinetics and local concentrations of TGF-β, costimulation, and Th1 and Th2 cytokines may determine the outcome of an immune response. High levels of all these ligands may generate neither Tregs nor effector cells, while a preponderance of one or the other may result primarily in suppression or immunity. Clearly, more studies will have to be performed to determine which naïve CD4 + CD25 – subset(s) can become Tregs, how these cytokine and receptor signals work at a biochemical level, and how different APC subsets may regulate these interactions.

In summary, TGF-β induces de novo Foxp3 expression in CD4 + CD25 – precursor cells, and these Foxp3 + cells act similar to naturally occurring CD4 + CD25 + Tregs. TGF-β is unique in its ability to turn on Foxp3 gene expression, whereas costimulatory and cytokine signals prevent Foxp3 expression. TGF-β and signaling through the TGF-β receptor complex may be key components to understanding the generation, expansion, and manipulation of Tregs. The de novo generation of Foxp3 + Tregs from CD4 + CD25 – cells may have great clinical potential. First, these cells are far more plentiful than CD4 + CD25 + cells and therefore represent a more practical and tractable starting population. Second, ex vivo treatment with a well-known soluble extracellular ligand is a relatively simple process that lacks the potential toxicities associated with gene transfer. Third, the identification of signals that prevent, inhibit or extinguish Foxp3 expression provides readily available tools to modulate the activity of these cells.

Ancillary