Natural and Induced CD4⁺CD25⁺ Cells Educate CD4⁺CD25⁻ Cells to Develop Suppressive Activity: The Role of IL-2, TGF-β, and IL-10

Medium

Aim V serum-free medium supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, and 10 mM HEPES (all from Invitrogen, Carlsbad, CA) was used for generation of CD4⁺ T_reg or control cells. RPMI 1640 medium supplemented as above with 10% heat-inactivated FCS (HyClone Laboratories, South Logan, UT) was used for all other cultures.

Abs and cytokines

The following Abs were used: anti-CD3, anti-CD4, anti-CD8, anti-CD25, anti-CTLA-4, anti-CD122, anti-CD62L, anti-HLA-A2, and respective mouse and rat isotype controls (all from BD PharMingen, San Diego, CA); anti-glucocorticoid-induced TNF family related receptor (GITR) and control IgG (R&D Systems, Minneapolis, MN); anti-CD8, anti-CD11b, and anti-CD74 (all hybridomas from American Type Culture Collection, Manassas, VA); and anti-CD16 (3G8; kindly provided by J. Unkeless, Mount Sinai Medical School, New York, NY). Abs used for intracellular cytokine staining were FITC- or PE-conjugated anti-IL-2, anti-IL-10, anti-IFN-γ, and anti-TNF-α (all from eBioscience, San Diego, CA). Unconjugated anti-IL-2, anti-IL-10, anti-TGF-β, and matched control IgG Abs (all from R&D Systems) were used for the neutralization experiments; anti-CD3 and anti-CD28 beads (the gift from C. June, University of Pennsylvania, Philadelphia, PA) were used for polyclonal activation of T cells. Recombinant human IL-2 and TGF-β1 were purchased from R&D Systems.

Lymphocyte isolation

T cells and T-depleted (stimulator) cells were prepared from heparinized venous blood of healthy adult volunteers, as described previously (9). CD4⁺ or naive CD4⁺ cells were prepared from T cells that were stained with Abs to CD8 or CD8 and anti-UCHL-1 (CD45RO), then isolated by negative selection using immunomagnetic beads (Dynal Biotech, Great Neck, NY). CD4⁺CD25⁺ and CD4⁺CD25⁻ cells were isolated from these populations by cell sorting gating on CD25 bright cells using a FACSDiVa (BD Biosciences, San Jose, CA). In some experiments, the CD25⁺ and CD25⁻ subsets were isolated by CD25 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) with >95% purity.

Generation of CD4⁺ T_reg cell subsets

We have used three sequential cell cultures to determine the role of the few CD25⁺ cells in the naive fraction of CD4⁺ cells in TGF-β-induced T_reg activity, and to document effects of these T_reg cells on the induction of new T_reg cells from CD25⁻ precursors. In primary cultures, purified naive CD4⁺ cells were added to irradiated (30 Gy) allogeneic stimulator cells for 6 days ±TGF-β (0.1–10 ng/ml). In some experiments, rIL-2 (10 U/ml) was also added to the cultures. The CD4⁺CD25⁺ cells were then isolated by cell sorting. Those primed with TGF-β are called T_reg1, and those activated without this cytokine are called T_con1.

To track the response of CD4⁺CD25⁺ cells to alloantigens, we used donors who differed in HLA-A2 expression. Naive CD4⁺CD25⁺ or CD4⁺CD25⁻ cells (2 × 10⁵) from one donor and 2 × 10⁶ CD4⁺CD25⁻ cells from another donor were activated with third party allogeneic stimulator cells. After 6 days of coculture, the CD4⁺ cells were stained with Abs to CD25, CD122, GITR, or isotype controls and analyzed by flow cytometry.

To determine whether T_reg1 could induce other CD4⁺CD25⁻ cells to become suppressor cell, they were added to a fresh allostimulatory culture. To distinguish between the two CD4 populations, the CD4⁺CD25⁻ cells were labeled with CFSE (Molecular Probes, Eugene, OR). The labeled CD4⁺CD25⁻ cells (2 × 10⁶) were activated with equivalent numbers of irradiated allogeneic stimulator cells for 6 days. In some cultures, 2–4 × 10⁵ T_reg1 or T_con1 were added to these cultures. Six days later, three distinct cell populations could be identified by flow cytometry. The undivided cells stained brightly for CFSE (CD4⁺CFSE^high), divided cells had diluted the CFSE (CD4⁺CFSE^mod), and primed CD4⁺CD25⁺ cells were unlabeled (CD4⁺CFSE⁻). The different populations were isolated by cell sorting. We obtained CD4⁺CFSE⁻ cells (T_reg1 or T_con1), CFSE moderately positive cells (T_reg2 or T_con2), or nondivided, CFSE bright cells (T_reg3 or T_con3).

Cytokine assays

Primed CD4⁺ cells were extensively washed and restimulated with either allogenic stimulators or anti-CD3/28 beads (1:20) for 24, 48, 72, 96, and 120 h in serum-free AIM V medium for TGF-β production, and with complete medium for production of other cytokines. Active TGF-β was determined by mink lung epithelial cells transfected with a luciferase gene construct (14). Supernatants were also tested in duplicate using ELISA kits (BioSource International, Camarillo, CA) for IL-10 and IFN-γ, according to the manufacturer’s instructions. For analysis of intracellular cytokine production, CD4⁺ cells were stimulated with 100 ng/ml PMA and 5 μM ionomycin for the last 6 h of culture. Brefeldin A (10 μg/ml) was added for last 5 h of culture. Cells were harvested, then fixed and permeabilized (Fix and Perm; Caltag Laboratories, Burlingame, CA), and stained with cytokine-specific Abs or isotype controls.

Assays for suppressive activity

To determine the suppressive activity of the various T_reg subsets, we measured their ability to inhibit the generation of autologous alloreactive CD8⁺ cells in a CTL cytotoxicity assay, as previously described (8). The ability of these cells to inhibit the CD8⁺ cell activation and proliferation induced by allogeneic stimulator cells was determined by measuring CD25 expression or cycling of CFSE-labeled CD8⁺ cells.

FoxP3 expression by real-time RT-PCR

Total RNA was prepared with TRIzol LS reagent (Invitrogen). First strand cDNA was synthesized using Omniscript TR kit (Qiagen, Valencia, CA) with random hexamer primers (Invitrogen). Real-time PCR was performed with a LightCycler (Roche, Mannheim, Germany), and message levels were quantified using the LightCycler Fast Start DNA Master SYBR Green I Kit (Roche), according to the manufacturer’s instructions. Amplification was conducted for 45 cycles. The recovered PCR product and amplicon were checked by agarose gel electrophoresis for a single band of the expected size. The relative expression of FoxP3 was determined by normalizing expression of each target to GAPDH. Primer sequences were as follows: GAPDH, 5′-CCACATCGCTCAGACACCAT-3′ and 5′-GGCAACAATATCCACTTTACCAGAGT-3′; FoxP3, 5′-GAAACAGCACATTCCCAGAGTTC-3′ and 5′-ATGGCCCAGCGGATGAG-3′.

Costimulatory effects of TGF-β on naive CD4⁺CD25⁺ cells

We have previously reported that TGF-β has costimulatory effects on alloactivated naive CD4⁺ T cells, as indicated by up-regulation of CD25 and CTLA-4 and an expansion of these cells (8). These costimulatory effects were dependent upon a sufficient amount of IL-2 to overcome the inhibitory effects of TGF-β. In this study, we report that a target of these costimulatory effects is the CD4⁺CD25⁺ cell subset, even though they constitute only 1% of naive CD4⁺ cells. Fig. 1⇓ shows that a dose-dependent increase in CD25 expression was completely abolished by depletion of CD25⁺ cells in the starting population. TGF-β also significantly enhanced expression of CD122, the β-chain of the IL-2R. Signaling through CD122 is essential for the suppressive activity of CD4⁺CD25⁺ cells (15, 16). TGF-β-mediated enhancement of CD122 expression was apparent at day 3 and was maximal 1 day later. The few CD25⁺ cells present in the naive fraction resulted in a 4-fold expansion of CD122-bearing cells. Like CD25 expression, depletion of the CD25⁺ cells abolished this enhancement.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Costimulatory effects of IL-2 and TGF-β on alloactivated naive CD4⁺ T cells. The costimulatory effects are dependent upon CD4⁺CD25⁺ T cells. One million CD4⁺CD45RO⁻ cells or naive CD4⁺ cells depleted of CD25⁺ cells from donor A were mixed with equal numbers of irradiated non-T cells from donor B with IL-2 (10 U/ml) ± TGF-β1 in the doses shown. Two separate experiments are shown. Top, The cells were incubated for 6 days, counted, stained for CD25, and examined by flow cytometry (CD25 mean fluorescence intensity, total CD4⁺ cells ± TGF-β (10 ng/ml) 405 vs 578; CD25⁻CD4⁺485 vs 511). Bottom, Samples of the cells were harvested at 3, 4, and 5 days, counted in triplicate, and stained for CD122. Depletion of CD25⁺ cells completely abolished the TGF-β-mediated enhancement of IL-2R α- and β-chains. The experiments shown are representative of at least eight experiments examining CD25 and CD122.

The next step was to learn whether TGF-β increased the expansion and potency of the CD25⁺ subset, or whether the CD25⁺ cells acted on CD25⁻ cells. Pilot experiments on allostimulated purified CD4⁺CD25⁺ cells revealed that expansion of these cells was only minimal, and that the combination of IL-2 and TGF-β did not substantially increase their suppressive activity (results not shown). To distinguish the progeny of CD4⁺CD25⁺ cells and CD4⁺CD25⁻ cells, we used cells from donors that differed in HLA-A2 expression. These studies conclusively demonstrated that CD25⁺ cells costimulated the CD25⁻ cells. Naive CD4⁺CD25⁺ or CD25⁻ cells from one donor were added to CD25⁻ cells from another donor in a 1:10 ratio, and these cells were stimulated by irradiated non-T cells from a third donor. The presence of CD25⁺ cells enabled IL-2 and TGF-β to increase the number of CD25⁺ cells by 2-fold, and GITR⁺ cells by 5-fold (Fig. 2⇓A). This expansion, however, was derived predominantly from CD25⁻ precursors. Depletion of the CD25⁺ T_reg cells resulted in a more vigorous MLR response, but this effect was not modified by the addition of IL-2 or TGF-β.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

TGF-β-stimulated CD4⁺CD25⁺ cells expand CD4⁺CD25⁻ cells and augment their suppressive activity. A, IL-2 and TGF-β alloactivated CD4⁺CD25⁺ cells induce CD4⁺CD25⁻ cells to expand and express markers characteristic of T_reg cells. Naive CD4⁺CD25⁺ cells or control CD4⁺CD25⁻ were prepared from donor A, who was HLA-A2 negative, and CD4⁺CD25⁻ cells were prepared from donor B, who was HLA-A2⁺. One million CD4⁺CD25⁻ cells from donor B were mixed with CD4⁺CD25⁺ or CD4⁺CD25⁻ cells from donor A in a 1:10 ratio, and cultured with one million irradiated non-T cells from donor C. Some wells contained medium only, TGF-β1 (10 ng/ml), IL-2 (10 U/ml), or both TGF-β and IL-2. The cells were harvested after 6 days, counted, and triple stained for CD4, HLA-A2, and CD25 or GITR. Gating on CD4⁺ cells, the CD25⁺ and GITR⁺ cells from each donor could be counted and the solute numbers determined. This experiment was repeated twice with similar results. B, Naive CD4⁺ cells from one donor were separated into CD25⁺ and CD25⁻ cells. One million total naive CD4⁺ cells, CD25-depleted cells, or CD25⁺ and CD25⁻ cells mixed together in a 1:10 ratio were cultured with irradiated stimulator cells ± TGF-β1 (10 ng/ml) without additional IL-2 in this experiment. After 6 days, the cells were harvested, each well was counted, and each subset was mixed with CFSE-labeled autologous T cells in a 1:10 ratio. The cells were restimulated for 6 days, counted, and stained for CD8, and the number of CD8⁺ cells that had divided was determined by flow cytometry. The panels show the numbers of each subset in cultures with and without TGF-β. Left, Changes in CD4⁺ cells in cultures with and without TGF-β. The horizontal line denotes the number at baseline. Right, Suppressive effects of naive CD4⁺ cells primed with TGF-β on dividing, CFSE-labeled CD8⁺ cells. TGF-β- and control-primed CD4⁺ cells were mixed with CFSE-labeled T cells in a 1:10 ratio. Bottom, Histogram of gated CD8⁺ cells showing the inhibitory effects of CD4⁺ cells primed with TGF-β. One of six similar experiments is shown.

The experiment shown in Fig. 2⇑B shows that naive CD4⁺CD25⁺ cells not only enable TGF-β to increase the cell number of CD4⁺CD25⁻ cells, but probably also augment their suppressive activity. In a primary MLR, 10% CD4⁺CD25⁺ cells resulted in a 60% increase in cell numbers in the presence of TGF-β. Although TGF-β could induce CD25-depleted CD4⁺ cells to suppress the proliferative response of CD8⁺ cells to alloantigens, this activity was markedly enhanced if the starting population also contained CD25⁺ cells. Suppression of both the percentage and total numbers of dividing CD8⁺ cells was increased.

The finding that CD4⁺CD25⁺ cells have direct effects on CD25⁻ cells is consistent with the results of Jonuleit et al. and Dieckmann et al. (12, 13). These workers reported that preactivated human CD4⁺CD25⁺ cells induce CD4⁺CD25⁻ cells to become cytokine-producing suppressor cells by a contact-dependent, cytokine-independent mechanism. In this study, the naive CD4⁺CD25⁺ cells were not preactivated, but the costimulatory effects of TGF-β enabled these cells to act on CD25⁻ cells. The cytokine profile of these cells will be discussed below.

CD4⁺CD25⁺ T_reg cells generated ex vivo induce CD4⁺CD25⁻ cells to become T_reg cells

Naive CD4⁺ cells induced to become CD25⁺ T_reg cells with TGF-β also have the capacity to induce other CD4⁺CD25⁻ cells to become T_reg cells. CD4⁺CD25⁺ T_reg cells were prepared, as described previously, by stimulating naive CD4⁺ cells with alloantigen and TGF-β1 for 5–6 days, then sorting the CD25⁺ cells. These TGF-β-conditioned cells, called T_reg1, have cytokine-independent suppressor activity. Control CD4⁺CD25⁺ cells (T_con1) were prepared similarly, except that TGF-β was left out of the cultures. T_reg1 or T_con1 CD25⁺ cells were added to fresh autologous CFSE-labeled CD4⁺CD25⁻ cells in a 1:5 ratio and stimulated with the same alloantigens for another 4–6 days. As expected, T_reg1 significantly decreased the numbers of CD4⁺ cells proliferating in response to alloantigens (Fig. 3⇓A). The remaining cells showed decreased activation and more cell death in comparison with controls (Fig. 3⇓B). Nonetheless, in cultures with T_reg1 cells, there was an identifiable subset of CFSE-labeled CD4⁺CD25⁻ cells that had divided (Fig. 3⇓C).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

T_reg1 inhibit most allostimulated CD4⁺CD25⁻ cells, but allow a subset to proliferate. A, Cytokine-independent inhibitory effects of T_reg1. A total of 4 × 10⁵ CD4⁺ cells, T_reg 1 cells, or T_con 1 cells and 2 × 10⁶ CFSE-labeled CD4⁺CD25⁻ cells from donor A was stimulated with 2 × 10⁶ irradiated non-T cells from donor B for 6 days. Some wells contained anti-TGF-β, anti-IL-10, or control IgG. The mean of 10 separate experiments ± SEM shows that T_reg1 decreased the numbers of proliferating CD4⁺ cells in comparison with T_con1. B, Flow cytometry scatter profile of a representative experiment showing the number of viable cells. T_reg1 cells resulted in markedly fewer viable cells. C, The cells were stained for CD4 and analyzed by flow cytometry. Wells that contained CFSE-labeled CD4⁺CD25⁻ responder cells without primed CD4⁺ cells are shown along with wells containing T_con1 cells or T_reg1 cells. CFSE-labeled CD4⁺ cells that had divided are called T_con2 cells or T_reg2 cells, and those that had not divided are called T_con3 cells or T_reg3 cells. After pilot studies revealed that T_con3 and T_reg3 did not have inhibitory effects, they were not studied further.

From these secondary cultures, the different CD4⁺ cell subsets were obtained by cell sorting and tested for suppressive activity on autologous CD8⁺ cells. When these alloprimed cells were mixed with fresh autologous T cells in a 1:10 ratio, both T_reg1 cells and CFSE-labeled CD4⁺CD25⁻ cells that had undergone cell division (T_reg2) had an equivalent ability to suppress alloactivation of CD8⁺ cells and the development of CTL activity (Fig. 4⇓A). As expected, neither T_con1 nor CFSE-labeled T_con2 cells had acquired any suppressive activity.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

T_reg1 have the ability to induce new CD4⁺CD25⁺ cells to become T_reg2 through a mechanism that both is contact dependent and requires TGF-β and IL-10. A, The generation of T_reg2 cells is contact dependent. In secondary cultures, T_reg1 or T_con1 (>90% CD25⁺) and CFSE-labeled CD4⁺CD25⁻ cells from donor A were mixed in a 1:10 ratio and cultured with irradiated allostimulator cells from donor B for 5 days. In three of eight experiments, T_reg1 cells were cultured with stimulator cells in Transwell chambers, where they were separated from responder cells by a semipermeable membrane. At the end of the culture, CFSE-negative T_reg1, CFSE moderate T_reg2, CFSE-negative T_con1, and CFSE moderate T_con2 were isolated by cell sorting and mixed with CFSE-labeled autologous T cells from donor A in a 1:10 ratio. These cells were restimulated with irradiated non-T cells from donor B for 5–6 days in tertiary cultures. Suppressive activity was assessed by determining the absolute numbers of activated CD8⁺ (CD25⁺) cells, or by measuring cytotoxic lymphocyte activity against Con A blasts from donor B in a standard 4-h chromium release assay. Left panel, Shows the mean ± SEM of eight experiments; right panel, shows the CTL activity of a representative experiment. B, The generation of T_reg2 cells is also cytokine dependent. The effect of anti-TGF-β and anti-IL-10 Abs on the generation of T_reg2 in secondary cultures was assessed. Anti-cytokine Abs were added in secondary cultures. The suppressive effect of T_reg2 cells was assessed in tertiary cultures. The experiment shown indicates inhibitory effects on allostimulated, proliferating CD8⁺ cells. The numbers of CFSE-labeled CD8⁺ cells that had divided are indicated. Each of the anti-cytokine Abs abolished the generation of new suppressor cells, although the suppressive effects of T_reg1 were cytokine independent. The result shown is representative of four independent experiments.

The ability of T_reg1 cells to induce CD4⁺CD25⁻ cells to develop suppressive activity was contact dependent. Separation of these two CD4⁺ T cell subsets from each other by a semipermeable membrane completely abolished their inducer effect (Fig. 4⇑A).

FoxP3 expression by CD4⁺ regulatory cells induced by IL-2 and TGF-β and by secondarily educated suppressor cells

The forkhead/winged helix transcription factor FoxP3 is specifically expressed by T_reg cells and programs their development and function (17, 18, 19). Using FoxP3 levels in fresh CD4⁺CD25⁺ and CD4⁺CD25⁻ cells as reference controls, we assessed levels of this transcription factor in T_reg1 after allostimulation of naive CD4⁺ cells with IL-2 and TGF-β, and in T_reg2 educated by T_reg1 from CD4⁺CD25⁻ precursors. Fig. 5⇓ shows dramatic up-regulation of FoxP3 mRNA in both of these CD4⁺ regulatory subsets. FoxP3 levels in T_reg2 were even greater than reference control CD4⁺CD25⁺ cells. The TGF-β-induced up-regulation of FoxP3 in human CD4⁺ cells is in agreement with the recent findings of Chen et al. (20), who documented that this cytokine increased FoxP3 gene expression in activated mouse CD4⁺CD25⁻ cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Both the natural-like CD4⁺ regulatory cells induced by IL-2 and TGF-β and the secondarily educated CD4⁺ cytokine-secreting suppressor cells express FoxP3. Naive CD4⁺ cells were prepared from donors A and B, and each was stimulated with allogeneic irradiated non-T cells from the respective donors. T_reg1 or T_reg2 and T_con1 or T_con2 cells from each donor were prepared, as described above. FoxP3 expression was determined in cDNA samples by a real-time RT-PCR method using GAPDH as an endogenous reference gene.

The importance of TGF-β and IL-10 in the generation and function of new suppressor cells

Both TGF-β and IL-10 contributed to the generation of CD4⁺ T_reg2 cells. In secondary cultures, either anti-TGF-β or anti-IL-10 blocked the ability of T_reg1 cells to induce T_reg2 (Fig. 4⇑B). In sharp contrast, the ability of T_reg1 to suppress the proliferative response of CD8⁺ cells to alloantigens remained intact even in the presence of these neutralizing anti-cytokine Abs.

Unlike T_reg1 cells, the suppressive activity of T_reg2 cells was cytokine dependent. We assessed the suppressive activity of these cells on CD8⁺ cell activation, proliferation, and development of allo-CTL activity. In these studies, the anti-cytokine Abs were added to the tertiary cultures. Fig. 6⇓A is representative of 10 experiments in which T_reg2, in a ratio of 1:10 to responder cells, inhibited CD8⁺ cell alloactivation by >75%. By contrast, T_con2 enhanced the number of activated CD8⁺ cells by >50%. Remarkably, anti-TGF-β and anti-IL-10, used singly and in combination, not only abolished this inhibitory effect, but also enhanced the number of activated CD8⁺ cells to levels higher than T_con1. We had previously observed this phenomenon with superantigen-induced CD4⁺ T_reg cells. In that study of T cell-dependent Ab production, neutralization of TGF-β not only abolished suppression, but also significantly enhanced IgG production (9). Taken together, our studies suggest that the T_reg we induce with TGF-β are markedly activated and can become helper cells rather than suppressor cells if they lose the capacity to produce these cytokines, or they are neutralized. In support of this hypothesis, lymphocyte production of TGF-β is impaired in systemic lupus erythematosus, and CD8⁺ T cells from these patients provide B cell help instead of inhibiting Ab production (21, 22).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

T_reg2 cells have a cytokine-dependent mechanism of action. A, Suppressive effects on alloactivated CD8⁺ cells. The experimental design was described in the legend for Fig. 2⇑. Sorted T_reg2 cells from secondary cultures were mixed with CFSE-labeled autologous T cells in a 1:10 ratio and restimulated with alloantigen for 5 days. From cell counts and the percentage of CD8⁺CD25⁺ cells by FACS, the absolute numbers of activated CD8⁺ are shown. The baseline (labeled as Nil) is determined by CD8⁺CD25⁺ cells from the MLR without added primed T cells. The effect of anti-TGF-β and anti-IL-10 added singly and together is shown. The result shown is representative of 10 independent experiments. B, Suppressive effects on proliferating CD8⁺ cells. The experimental design is the same as that described in Fig. 3⇑A. The difference is that the anti-cytokine Abs were added to tertiary cultures. C, Suppressive effects on the generation of CTL activity. The effects of T_reg2 or T_con2 cells on the ability of T cells to develop allo-CTL activity are shown in a standard 4-h chromium release assay at various E:T cell ratios.

The suppressive effects of T_reg2 cells on the proliferative response of CD8⁺ cells to alloantigens are also cytokine dependent. The inhibition was reversed by neutralization of TGF-β and IL-10. In this experiment and in most of the six experiments conducted, the inhibitory effects of TGF-β were greater than IL-10. In two experiments, anti-TGF-β, but not anti-IL-10, reversed the inhibition. In some experiments, anti-TGF-β only partially inhibited suppression (Fig. 6⇑B). Finally, Fig. 6⇑C shows that the development of allo-CTL activity by CD8⁺ cells is also blocked by T_reg2, in a cytokine-dependent manner. In this experiment, reversal of the suppression by anti-cytokine Abs was partial, but in others it was complete.

We considered the possibility that T_reg1 cells could serve as a source for the TGF-β and IL-10 needed for the generation of T_reg2. T_reg1 cells, accordingly, were prepared and restimulated with either T-depleted PBMC cells or beads coated with anti-CD3 and anti-CD28 without TGF-β. Time course studies revealed down-regulation of IL-2 production by day 3, poor IFN-γ production at all times, and an increase in IL-10-producing cells beginning at day 3 (Fig. 7⇓A). The flow cytometry profile at day 3 is shown in Fig. 7⇓B. Not shown is an absence of IL-4 and no difference between TNF-α production between T_reg1 and T_con1 cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

Cytokine profile of T_reg1 cells. A, Naive CD4⁺ cells from donor A were stimulated with irradiated non-T cells for donor B with IL-2 (10 U/ml) ± TGF-β (10 ng/ml) for 6 days. The primed CD4⁺ cells were restimulated with allogeneic cells, IL-2, but without TGF-β. At various times (days 1, 3, and 5), the capacity of these cells to produce IL-2, IFN-γ, and IL-10 was determined by stimulating the cells with PMA and ionomycin for 6 h and intracellular staining of permeabilized cells, as described in Materials and Methods. In comparison with control alloprimed (Figure legend continues) naive CD4⁺ cells, those primed with TGF-β had a decreased capacity to produce IL-2 by day 3, an inability to produce IFN-γ, but an enhanced ability to produce IL-10. IL-4 was also measured, but was undetectable. B, FACS analysis of IFN-γ and IL-10 production at day 3. The example shown is representative of 10 separate experiments. C, TGF-β and IL-10 production is restricted to CD4⁺CD25⁺ cells. After primary cultures, CD4⁺ cells were sorted into CD25⁺ T_reg1 cells and T_con1 cells and their respective CD25⁻ subsets. Each was restimulated with anti-CD3- and anti-CD28-coated beads for 5 days, and the cell supernatants were tested for TGF-β and IL-10 production.

Separation of T_reg1 and T_con1 cells into CD25⁺ and CD25⁻ subsets by cell sorting revealed that by day 5 T_reg1 CD25⁺ cells produced IL-10 and high levels of TGF-β in the biologically active form (Fig. 7⇑C). Remarkably, this was the only T cell subset that produced these cytokines. These studies, therefore, provide evidence that T_reg1 can produce the cytokines needed for the induction of new CD4⁺ suppressor cells.

This study may help to resolve apparently conflicting observations regarding the role of cytokines in the suppressive activities of CD4⁺CD25⁺ cells isolated from lymphoid tissues. Although some workers have reported that suppression is independent of TGF-β and IL-10, others have claimed that suppression can be reversed by anti-cytokine-neutralizing Abs, especially at high concentrations (23). It has become evident that CD4⁺CD25⁺ T_reg cells constitute heterogeneous populations. Our studies support the view that natural CD4⁺ T_reg cells have a cytokine-independent mechanism of action (1, 2, 3), but that CD4⁺CD25⁺ T_reg cells induced in the periphery may have either cytokine-dependent effects or cytokine-independent effects.

An important new finding is that the target of the costimulatory effects of TGF-β are CD4⁺CD25⁺ cells. Whereas previously we had reported that IL-2 and TGF-β enhance expression of CD25, CTLA-4, and CD40L (8, 24), in this study we demonstrate that expression of CD122 and GITR is also enhanced. Depletion of the few CD25⁺ cells found in the naive fraction of CD4⁺ cells completely abolished these costimulatory effects.

Remarkably, the effects of IL-2 and TGF-β on CD4⁺CD25⁺ cells did not significantly expand these cells in vitro, but stimulated the expansion of CD4⁺CD25⁻ cells and induced them to become suppressor cells. We and others have reported that CD4⁺CD25⁻ cells activated in the presence of TGF-β become T_reg cells (9, 25). In this study, naive, natural CD4⁺CD25⁺ cells treated with IL-2 and TGF-β greatly amplified the effect of these cytokines on CD4⁺CD25⁻ cells, thus increasing the numbers of T_reg generated. Although their suppressive effects were also cytokine independent in vitro, upon restimulation they produced both TGF-β and IL-10, and these cytokines had an important role in the generation of new T_reg cells.

The next new finding is that CD4⁺CD25⁺ T_reg1 cells could induce other CD4⁺CD25⁻ cells to become suppressor cells. Unlike the generation of T_reg1 cells that required exogenous cytokines for their development, the TGF-β and IL-10 produced by T_reg1 were sufficient for the alloactivated CD4⁺CD25⁻ cells to become T_reg2 cells. Although the generation of T_reg2 cells required cell contact, the presence of TGF-β and IL-10 was also necessary for this phenomenon. Physical interaction between the CD4⁺CD25⁺ inducer and CD4⁺CD25⁻ cells was also required for their differentiation to suppressor cells (12, 13). In our recent study, in which we induced CD4⁺CD25⁻ cells to become TGF-β-producing suppressor cells, their suppressive activity was also contact dependent, but nonetheless abolished by anti-TGF-β (9). Others have also reported that surface-bound TGF-β may mediate the suppressive effects of CD4⁺CD25⁺ cells (23).

In this study, we added TGF-β to CD4⁺ cells to initiate T_reg differentiation, but in vivo both TGF-β and IL-10 are produced by cells of the innate immune system. It is likely that in the gastrointestinal tract, the respiratory tract, and the female reproductive system, immature APCs produce these cytokines that drive Ag-activated T cells in these organs to become cytokine-producing suppressor cells analogous to those we generated ex vivo. A model suggesting that the effects of IL-2 and TGF-β on both natural CD4⁺CD25⁺ cells and CD4⁺CD25⁻ cells trigger a continuous loop that sustains the renewal of CD4⁺CD25⁺ T_reg cells is shown in Fig. 8⇓. IL-2 and TGF-β costimulate Ag-activated natural T_reg cells, and these cells, together with these two cytokines, induce naive CD4⁺CD25⁻ cells to become suppressor cells that can produce IL-10 and the active form of TGF-β. The stimulated CD4⁺CD25⁻ cells, therefore, produce the cytokines that have previously been shown to be involved in the generation of T_reg cells (4, 5, 6, 7, 8, 9). The TGF-β- and IL-10-producing CD4⁺CD25⁺ T_reg cells are then able to induce other Ag-activated CD4⁺ cells to become T_reg cells. The sustained renewal of T_reg cells that express FoxP3 would insure the maintenance of nonresponsiveness to self Ags and exogenous Ags presented by mucosal APCs.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

A model for the effects of IL-2, TGF-β, and IL-10 in a circuit for the sustained renewal of CD4⁺ T_reg cells. IL-2 and TGF-β induce Ag-activated CD4⁺CD25⁻ cells to become CD4⁺CD25⁺ T_reg1. These cytokines also costimulate natural CD4⁺CD25⁺ cells that do not proliferate, but greatly expand the developing CD4⁺CD25⁺ T_reg1 that have a contact-dependent, cytokine-independent mechanism of action. Restimulated T_reg1 also produce both TGF-β and IL-10 and induce other Ag-activated CD4⁺CD25⁻ cells to become T_reg2 by a contact-dependent mechanism that also requires both of these cytokines. Ag-activated T_reg2 also has the ability to produce TGF-β and IL-10 and can induce other CD4⁺CD25⁻ cells to become T_reg3, which can also induce T_reg4, etc. This circuit depends upon repeated exposure of Ag, and IL-2 produced by Ag-stimulated T cells. It probably also depends upon specialized APCs that remain to be characterized.

We have recently reported that a single transfer of T_reg cells generated ex vivo with TGF-β and IL-2 will double the survival of mice that had developed a lupus-like syndrome (26). We suggest that the long-term effects of adoptively transferred natural-like CD4⁺CD25⁺ regulatory cells induced ex vivo are due to their ability to generate new cytokine-producing CD4⁺ T_reg cells in vivo. Our studies raise the possibility, therefore, that the adoptive transfer of CD4⁺ T_reg cells generated ex vivo with IL-2 and TGF-β as a treatment for autoimmune diseases may have sustained long-term beneficial effects in vivo.