MHC Class II Expression Identifies Functionally Distinct Human Regulatory T Cells

Cell culture reagents and Abs

Cells were cultured in RPMI 1640 medium supplemented as described previously (17) with 5% human AB serum (SeraCare Life Sciences) in 96-well U-bottom plates (Costar) or Transwell plates (Nunc). The anti-CD3 (UCHT1 for bead and Hit3a for soluble stimulation) and anti-CD28 (28.2) Abs were purchased from BD Pharmingen. The anti-CD2 Ab (BMA 0111) was purchased from Dade Behring. To generate different stimuli, tosyl-activated beads (Dynal Biotech) were covalently bound with anti-CD3 alone, anti-CD3 and anti-CD2, or anti-CD3 and anti-CD28 at 1 μg/10⁷ beads per the manufacturer’s instructions and were used at 1.5 × 10⁴ beads/well. In certain experiments, anti-IL-10 (JES3–9D7; BD Pharmingen) was added at 5 μg/ml. Staining for intracellular cytokines was performed on cells stimulated with anti-CD3/anti-CD2 for 48 h, during which GolgiStop (BD Pharmingen) and PMA and ionomycin were added for the last 12 h before they were fixed and permeabilized (Cytofix/Cytoperm reagents; BD Pharmingen), and stained separately with allophycocyanin-labeled anti-IL-4, anti-IL-10, or anti-IFN-γ Abs. Nonspecific background levels (control cytokine staining) were determined for each cytokine by incubating the permeabilized cells with each unlabeled cytokine specific mAb for 10′ before adding the same, but fluorochrome-labeled anti-cytokine mAb. Staining for intracellular Foxp3 protein was performed on cells directly after ex vivo isolation with the PCH101 anti-Foxp3 Ab (eBioscience) per the manufacturer’s instructions.

Cell isolation

Whole mononuclear cells were isolated from human blood drawn (from 16 different healthy control individuals) in green-capped, heparinized tubes by Ficoll-Hypaque (Amersham Biosciences) gradient centrifugation. Total CD4 T cells were isolated via the CD4⁺ T cell isolation kit II (Miltenyi Biotec). CD4 T cells were stained for four-color sorting with anti-CD4 (clone RPA-T4) (optional), anti-CD25 (clone M-A251), anti-CD62L (clone Dreg 56), and anti-HLA-DR (clone L243). All Abs were purchased from BD Pharmingen and used after they had been dialyzed to remove azide. The specific DR⁺ and DR⁻ Treg populations were isolated by sorting with a FACS Aria (BD Biosciences) to typically >98% purity in postsort analysis. The sorted cells were gated first on small lymphocytes by forward and side scatter, then on CD4 (optional, due to their sample being preselected to be ∼99% CD4 T cells) and CD62L^high expression, and then on CD25 and HLA-DR expression. In samples not directly stained for CD4, a separate control was analyzed with CD4 to demonstrate the typical 95–99% purity for CD4.

In vitro micrococulture assay

In all assays using cell populations isolated directly ex vivo, the CD4⁺CD25⁻ (Tresp) cells were plated at 2.5 × 10³/well, while the DR⁺ or DR⁻ Treg were plated at 1.25 × 10³/well, and cocultures were at a 2:1 ratio. No irradiated accessory cells were added. A total of 1.5 × 10⁴ beads coated with anti-CD3 alone, anti-CD3/anti-CD2, or anti-CD3/anti-CD28 was added per well. The day on which the triplicate assays were set up was considered day 0. To determine proliferation on days 3, 4, and/or 5, half of the culture supernatant (100 μl) was replaced with 1 μCi of [³H]thymidine (NEN) at 24-h intervals for a 24-h pulse so that the entire final 3 days of the culture would be monitored. All assays exhibited <10% SEM and were repeated a minimum of three times using blood from different donors. In experiments using CFSE and/or fluorescence analysis of protein expression, the Tresp were labeled with 0.25 μM CFSE as described (26), and fluorescence was assessed on a FACSCalibur (BD Biosciences) using CellQuest Pro software (BD Biosciences).

Generating clones

For the single-cell cloning experiments, CD4⁺CD62L^high cells that also gave the profiles of DR⁻ CD25⁻ (Tresp), DR⁺ CD25^−/low, DR⁺ Treg, and DR⁻ Treg were sorted at one cell per well and stimulated with soluble anti-CD3 (Hit3a, 2.5 μg/ml) and anti-CD28 (28.2, 1 μg/ml), 1–5 × 10⁴ irradiated (3000 rad) PBMC, and 100 U/ml IL-2 (rhIL-2, Tecin; NCI), as modifications to previously published procedures (18, 27). Half of the medium was replaced with fresh medium containing 100 U/ml IL-2 starting at day 10 and every 3–4 days thereafter. The Treg clones were generated by stimulation with anti-CD3/anti-CD28/IL-2, because it gave a significantly higher cloning frequency than PHA (data not shown). After 21 days, the cells were restimulated once with PHA-P (REMEL) at 0.5 μg/ml with a mixture of irradiated PBMC (25% allogeneic, 75% autologous) at 1–5 × 10⁴/well. The clones were again given fresh IL-2 medium every 3–4 days and were analyzed after 21 days.

Cytokine analyses by cytometric bead array (CBA)

The supernatants that were removed before addition of [³H]thymidine were analyzed without dilution for the presence of IFN-γ, IL-10, and IL-4 via the Human Th1/Th2 Cytokine CBA kit (BD Pharmingen) per the manufacturer’s instructions. The beads were analyzed on a FACSCalibur using the CBA software purchased from BD Pharmingen. Upon initial comparison with ELISA analysis of the cytokine levels, the CBA data were very similar, highly reproducible, but much more sensitive.

Real-time PCR analysis

RNA, isolated from the indicated cell populations either directly ex vivo or after 3 and 5 days of stimulation by the RNeasy Micro, RNase-free DNase procedure (Qiagen), was converted to cDNA via RT by random hexamers and Multiscribe RT, via the TaqMan Gold RT-PCR kit (Applied Biosystems), diluted 1/10, and used in triplicate 25 μl of TaqMan PCR to amplify human Foxp3, TGFβ, or IL-10 and the β₂m endogenous control on an ABI 7700 machine (Applied Biosystems). The primer and probe combinations for all of these transcripts were purchased from Applied Biosystems. The values are presented as the difference in C_t values normalized to β₂m for each sample as per the following formula: Normalized expression = (power(2,−ΔCt)) × 1000.

Frequency and Foxp3 expression by DR⁺ and DR⁻ cells within the CD4⁺CD25^highCD62L^high population

To address the potential significance of MHC II expression on CD4⁺CD25^high Treg, highly purified DR⁺ and DR⁻ subsets of CD4⁺CD25^highCD62L^high Treg were isolated from negatively selected CD4 peripheral blood T cells and examined for functional activity in vitro. CD4 cells were first gated on high expression of CD62L (L-selectin), as described previously (18, 22) (Fig. 1⇓A). The CD62L⁻ population, representing in vivo-activated CD4 T cells (13% of total CD4 cells), express intermediate levels of CD25. In contrast, the CD62L^high subset, representing ∼82% of CD4 T cells that have presumably not been activated in vivo (23, 28), contains the CD4⁺CD25^high population that exhibits in vitro regulatory activity (15, 18). In this analysis, 2.5 and 32% of the peripheral blood CD4⁺CD62L^high T cells are CD25^high and CD25^low, respectively. The DR⁺ (30%) and DR⁻ (70%) subsets were sorted from the total CD4⁺CD25^highCD62L^high Treg population and are designated as Treg. The DR⁺ Treg selected by this gating represent approximately half of the DR⁺ cells in the CD4⁺CD62L^high population and correspond to ∼0.6% of total CD4 T cells. The remaining DR⁺ CD4⁺CD62L^high are CD25^−/low. In contrast, while there are no CD25^high DR⁺ T cells in the CD62L⁻ population, the DR⁺ CD62L⁻ cells are CD25^low and represent ∼0.7% of total CD4 T cells. Thus, while two-thirds of all the peripheral blood CD4⁺ DR⁺ T cells are CD62L^high, approximately one-half of them reside in the CD25^high regulatory population.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

MHC-DR is expressed on human peripheral blood CD4⁺CD62L^highCD25^high Treg. A, CD4⁺ peripheral blood T cells were stained for expression of CD4, CD25, CD62L, and HLA-DR. Specific T cell subsets were sorted by sequentially restricting the cells to small lymphocytes via forward and side scatter, to the CD4⁺CD62L^high population (gate shown in left histogram), and ultimately to individual populations expressing HLA-DR and/or CD25^high (gates shown in right histogram). The dashed lines denote the gates used to sort the DR⁺ Treg (upper right), DR⁻ Treg (upper left), the DR⁺ Tresp (lower right), and DR⁻ Tresp (lower left). B, The specific sorted populations were tested for their relative expression of Foxp3 mRNA directly ex vivo by real-time PCR using β₂m as the endogenous control. This pattern of relative Foxp3 expression was consistently observed in 14 experiments. Statistical analysis (Wilcoxon matched-pairs test) indicates that the increased expression of Foxp3 in the DR⁺ Treg population as compared with the DR⁻Treg subset is significant (p < 0.0067). C and D, Directly after ex vivo isolation, DR⁻CD25⁻ (Tresp), DR⁺ CD25^high, and DR⁻ CD25^high cells were permeabilized and stained for Foxp3 (PCH101 mAb). The histograms (C) represent the overlay of Foxp3 (solid line) and isotype control (dashed lines), except for the fourth histogram, which represents the Foxp3 staining of the DR⁺ (solid) and the DR⁻ (dashed) Treg for comparison. Dot plots (D) show that, regardless of their level of DR expression, all DR⁺ CD25^high cells express a similar high level of Foxp3. The staining for Foxp3 is representative of three experiments.

We first examined the Treg and DR⁻ Treg subsets for their relative expression of Foxp3, which has been shown to be involved in the regulatory function of natural CD4⁺CD25⁺ Treg (29, 30, 31). Total RNA isolated from total CD4⁺ cells, the CD62L^high-restricted populations of DR⁻CD25⁻ (Tresp), DR⁺ Treg, and DR⁻ Treg, were analyzed for relative levels of Foxp3 mRNA by real-time RT/PCR (Fig. 1⇑B). Compared with β₂m to control for differences in cDNA amount, Foxp3 mRNA was expressed at significantly higher levels in both the DR⁺ and DR⁻ Treg, compared with the Tresp and the other potentially in vivo-activated populations (CD62L⁻). From similar analyses on 14 different donor-derived Treg subsets, the DR⁺ Treg population always expressed a significantly higher level of Foxp3 message than the DR⁻ Treg (p < 0.0067, Wilcoxon matched-pairs test). In total, these data demonstrated that both the DR⁺ and DR⁻ Treg contained >40-fold greater levels of Foxp3 message, compared with ex vivo CD4⁺CD25⁻ T cells.

To assess Foxp3 expression at the protein level, DR⁺ Treg and DR⁻ Treg were permeablilized and stained for Foxp3 directly after their ex vivo isolation (Fig. 1⇑C). The FACS profile histograms in the top row demonstrate that the DR⁺ Treg are highly homogeneous for Foxp3 as all are Foxp3⁺, while 12% of the DR⁻ Treg population do not stain over background. When the levels of Foxp3 protein expressed by the DR⁺ and DR⁻ Treg populations are compared (fourth histogram), not only are the DR⁺ Treg more homogeneous for Foxp3 expression, they also express higher amounts of the protein. Two-parameter dot-plot analysis comparing the expression of HLA-DR and intracellular Foxp3 indicated that, although the DR⁺ Treg exhibit a range of HLA-DR expression, they all contain the same high level of Foxp3 protein (Fig. 1⇑D). Thus, although all DR⁺ Treg express high amounts of Foxp3 protein, there is no relationship between the extent of HLA-DR expression and the extent of Foxp3 expression. In further characterizations, although the DR⁺ Treg are highly homogeneous for Foxp3, ∼40% of the DR⁺ Treg are CD45RA⁺RO⁻, while <10% of the DR⁻ Treg are CD45RA⁺RO⁻ (data not shown). Thus, these data suggest that the DR⁻ Treg (CD4⁺CD25^highCD62L^high) population is different from a newly described DR⁻ CD45RA⁺RO⁻ regulatory population (32).

Distinct regulatory activity of DR⁺ and DR⁻ Treg populations

We then examined whether the DR⁺ and DR⁻ Treg populations were both capable of suppressing the proliferation of CD4⁺CD25⁻ Tresp in the in vitro coculture assay. In these experiments, because the total CD25^highCD62L^high T cells that represent only 2.5% of the total CD4 population were further sorted into separate DR⁺ and DR⁻ populations, the cell separations typically produced ∼0.8 × 10⁵ DR⁺ Treg and ∼2 × 10⁵ DR⁻ Treg from ∼2 × 10⁸ total PBMC. As a result, these cell populations were tested for regulatory function in a highly defined in vitro coculture microassay system that lacked APCs and included beads coated with αCD3 alone (Fig. 2⇓C), αCD3/αCD28 (Fig. 2⇓B) to model interactions with APCs expressing CD80/86, or αCD3/αCD2 (Fig. 2⇓A) to model signals generated through CD2 by interaction with the high levels of CD58 that are expressed on many types of cells, including accessory cells, endothelial cells, and activated T cells (33, 34). Proliferation and cytokine secretion were measured by [³H]thymidine incorporation and by a highly sensitive cytokine bead array. In response to αCD3 stimulation, the 2500 CD4⁺CD25⁻ Tresp proliferated, while the DR⁺ and DR⁻ Treg subsets did not (Fig. 2⇓C). In contrast, cocultures of Tresp stimulated in the presence of half as many DR⁺ or DR⁻ Treg exhibited no proliferation. Because this αCD3 stimulation alone is weak, proliferation could be detected only at day 5 and was accompanied by the secretion of IFN-γ only. These data demonstrate that both DR⁺ and DR⁻ Treg subsets can exhibit in vitro suppressive function under conditions of weak strength of signal associated with αCD3 cross-linking alone.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

MHC-DR expression identifies Treg CD4⁺CD62L^highCD25^high populations that mediate different functional activities in response to different types of costimulation. In this microculture system, 2500 Tresp were cocultured with 1250 DR⁺ or DR⁻ Treg per well (1:1/2) ratio, and stimulated with αCD3/αCD2 (A), αCD3/αCD28 (B), or αCD3 (C) beads. The cultures were examined for proliferation and cytokine production at days 3 and 5. With stimulation by αCD3/αCD2 beads, the DR⁺ Treg mediate early (day 3) suppression of CD4 T responders, while DR⁻ Treg induce IL-10 and IL-4 secretion. D, CFSE-labeled DR⁻ Tresp, DR⁺ Treg, and DR⁻ Treg were stimulated alone or in coculture (1:1/2) with αCD3/αCD2/IL-2 (100U/ml) for 60 h. GolgiStop and PMA/ionomycin were added for the last 12 h of culture, and the cells were stained for intracellular levels of IFN-γ, IL-4, or IL-10. The specific cytokine staining was compared with control staining that resulted from the preincubation of the cells with unlabeled anti-cytokine mAbs, as per Materials and Methods. Representative of two experiments.

To better model human in vivo stimulatory conditions, we examined the effects of costimulation on the anergic and suppressive characteristics of the DR⁺ and DR⁻ Treg populations. Human T-T cell interactions involve engagement of both TCR by MHC and CD2 by CD58 (6, 33, 34), while typical T cell–APC interactions involve engagement of the T cell’s TCR and CD28. To address these more physiologically relevant conditions, assays were performed assessing the regulatory activity of DR⁺ and DR⁻ Treg in response to αCD2 or αCD28 costimulation (Fig. 2⇑, A and B). In contrast with the day 5 measurement of proliferation in cultures stimulated with αCD3 alone, the presence of a costimulatory signal resulted in an earlier Tresp proliferation that started at day 3 and was accompanied by increased cytokine secretion.

The DR⁺ and DR⁻ Treg exhibited a striking difference in their kinetics of suppression subsequent to the engagement of CD2. Compared with the strong proliferation of Tresp, αCD3/αCD2-stimulation did not induce the proliferation of either the DR⁺ or DR⁻ Treg subsets (Fig. 2⇑A). The cocultures established with the DR⁺ Treg under αCD3/αCD2 stimulation exhibited a robust and early suppression (day 3) as indicated by the strong inhibition of proliferation and secretion of all measured cytokines, including IFN-γ, IL-10, and IL-4. In marked contrast, cocultures of DR⁻ Treg and Tresp did not induce a similar early contact-dependent suppression, but rather, induced secretion of IL-4 and IL-10. This was followed by a late contact-dependent suppression (day 5) of both proliferation and IFN-γ secretion but not IL-4 or IL-10 production. Importantly, the late suppression was shown not to be due to culture exhaustion, because control cultures of Tresp alone were still exhibiting increasing proliferation at this point. Furthermore, the difference in kinetics of suppression by the DR⁺ and DR⁻ Treg subsets was not simply a reflection of a delay in DR expression by DR⁻ Treg as <5% of the DR⁻ Treg expressed DR during the course of these 5-day cultures.

To determine the source of the increase in secreted IL-4 and IL-10 in supernatants from DR⁻ Treg cocultures, CFSE-labeled Tresp, DR⁺ Treg, DR⁻ Treg, or the specific Treg/CFSE Tresp cocultures were stimulated with αCD3/αCD2 and exogenous IL-2 and stained for intracellular cytokine production (Fig. 2⇑D). IL-2 was added to augment cytokine production, although it also abrogated coculture suppression as expected. Also, although the DR⁺ Treg do not typically secrete any detectable cytokines, they were similarly analyzed to allow the comparison between the DR⁺ and DR⁻ Treg populations. These data demonstrate that, while both the DR⁻ Treg and Tresp subsets can produce IL-4 and IL-10 under these conditions, a greater proportion of DR⁻ Treg produced IL-4 (25%) and IL-10, while a greater proportion of the Tresp produced IFN-γ (18%). Thus, the DR⁻ Treg population was more prone to produce Th2 cytokines, while the Tresp population was primarily responsible for the production of Th1 cytokines. In additional studies, half of the Th2 (IL-4) cytokine-producing DR⁻ Treg were found to express Foxp3 protein (data not shown), indicating that at least a significant portion of the Th2 cytokines produced in these cultures are derived from DR⁻, Foxp3⁺ natural Treg.

The qualitatively different and stronger signals associated with αCD28 costimulation again altered the behavior of these functionally distinct Treg populations. With αCD28 costimulation, the Tresp exhibited greater levels of proliferation and cytokine production (Fig. 2⇑C), while the DR⁺ and DR⁻ Treg populations still remained unresponsive. In contrast with CD2 costimulation, DR⁺ Treg cocultures established with αCD3/αCD28 stimulation did not induce early suppression. Rather, αCD3/αCD28 stimulated cocultures containing either DR⁺ or DR⁻ Treg subsets exhibited the same delayed suppression of Tresp proliferation. Importantly, while these αCD3/αCD28 cocultures exhibited the same late suppression of proliferation, only the cocultures established with DR⁻ Treg also induced an increase in Th2 cytokine secretion. Because the Tresp-only cultures continued to increase their proliferation during this time, the decrease in proliferation in the 4- and 5-day cocultures was not due to culture exhaustion.

Differential expression of IL-10 and kinetics of superinduction of Foxp3 by stimulated DR⁺ and DR⁻ Treg subsets

Due to the distinct kinetics of inhibition by DR⁺ and DR⁻ Treg, we examined whether these regulatory populations alter their expression of Foxp3 mRNA over time in relation to their temporal changes in suppressive activity. Because αCD3/αCD2 costimulation produces a striking difference in the kinetics of suppression by these two Treg subsets, cultures were set up in which Tresp alone, DR⁺ Treg alone, or DR⁻ Treg alone were stimulated with αCD3/αCD2 in the presence or absence of IL-2. The cells were recollected after 3 and 5 days of stimulation with and without IL-2. Real-time PCR analysis of the mRNA isolated from these subsets indicated that, after 3 days of stimulation in the absence of IL-2, the levels of Foxp3 mRNA did not change appreciably from direct ex vivo levels (Fig. 3⇓A). In contrast, when stimulated in the presence of IL-2, the DR⁺ Treg exhibit an additional 8-fold superinduction of Foxp3 message on day 3 when they also exhibit strong suppression. Yet, after 5 days of stimulation, the level of DR⁺ Treg Foxp3 expression is greatly decreased. In comparison, in response to the same IL-2 supplemented stimulation, the DR⁻ Treg do not exhibit their maximal induction of Foxp3 mRNA until day 5, which correlates with their kinetics of maximal suppressive activity. Low-level Foxp3 mRNA also was induced in similarly stimulated Tresp, although the levels were significantly lower than those found in the Treg populations. Thus, these data indicate that, while Foxp3 mRNA can be expressed in activated, human nonregulatory CD4 T cells, as reported by others (35), it is expressed at >20-fold higher levels by activated Treg populations in an IL-2-dependent fashion, and its induction correlates with early suppression in DR⁺ Treg cocultures and delayed regulation in DR⁻Treg cocultures. The similar analysis of RNA isolated from DR⁺ and DR⁻ Treg that had been reisolated from suppressive αCD3/αCD2 stimulated cocultures (no IL-2) showed a highly similar further induction of Treg Foxp3 message (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

The regulatory cell populations were analyzed for Foxp3 and IL-10 mRNA expression by TaqMan PCR. A, Total mRNA samples, isolated from CD4⁺CD62L^high populations that were DR⁻ CD25⁻ (Tresp), DR⁺ Treg, or DR⁻ Treg, directly after their ex vivo isolation (day 0), or after 3 and 5 days of stimulation with αCD3/αCD2 in the presence or absence of IL-2, were analyzed for expression of Foxp3 mRNA. The similar profile of IL-2-enhanced induction of Foxp3 was observed in two additional experiments with cells from different donors. In the first repeat, the relative Foxp3 expression values comparing day 0 values to 3 days of stimulation demonstrated an increase in Tresp from 2 to 29, in DR⁺ Treg from 84 to 597, and in DR⁻ Treg from 60 to 241. In the second repeat, the relative expression increased in the Tresp subset from 1 to 7, the DR⁺ Treg from 24 to 172, and the DR⁻ Treg from 17 to 47. B, Relative levels of IL-10 as normalized to β₂m also were assessed in total RNA isolated either directly ex vivo or after 3 or 5 days of αCD3/αCD2 (±IL-2) stimulation of Tresp, DR⁺ Treg, or DR⁻ Treg subsets. Representative of three experiments.

The DR⁺ Treg, DR⁻ Treg, and DR⁻ Tres cells also were assessed for their ability to express IL-10, an inhibitory cytokine that has been shown to be involved in the function of CD4⁺CD25⁺ and Tr1 regulatory cells in both mouse and man (27). Although the cells did not express detectable amounts of IL-10 mRNA when analyzed directly ex vivo (day 0), IL-10 message could be detected 3 and 5 days after in vitro activation. The values of IL-10 amplification were again normalized to the levels of β₂m by real-time PCR (Fig. 3⇑B). As indicated, αCD3/αCD2-stimulated DR⁺ Treg did not express IL-10 mRNA regardless of the presence or absence of exogenous IL-2, while both the DR⁻ Treg and Tresp populations produced message for IL-10 within 3 days of stimulation.

IL-10 is a component of the suppressive repertoire of DR⁻ Treg

Studies were then performed to assess the contact dependence and potential involvement of IL-10 in DR⁺ or DR⁻ Treg suppression because the classical CD4⁺CD25⁺ Treg are known to suppress via a cell contact-mediated mechanism, while the ability of the DR⁻ Treg to produce IL-10 suggests that they could suppress via its secretion. Contact dependence was assessed by Transwell assays in which the Tresp were cultured in the lower chamber and the DR⁺ or DR⁻ Treg were stimulated either separately in the upper chamber or together with the Tresp in the lower chamber in the presence (open) or absence (filled) of neutralizing αIL-10 mAb (Fig. 4⇓). Both upper and lower chambers received αCD2/αCD3 stimulation, and proliferation in the lower chambers was assessed at day 5 when both the DR⁺ and DR⁻ Treg cocultures show inhibition. The addition of αIL-10 had no effect on the proliferation of the Tresp alone. The separate stimulation of DR⁺ Treg in the upper chamber also had no affect on the proliferation of Tresp in the lower compartment. However, if the DR⁺ Treg were stimulated in the lower chamber in contact with the Tresp, proliferation was suppressed by >50%. Thus, these data indicated that the DR⁺ Treg require cell contact to exert suppression.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Late functional suppression by DR⁻ Treg involves both contact-dependent and IL-10-mediated suppression, while DR⁺ Treg only exert contact-dependent suppression. Cultures were set up in Transwell plates where Tresp were placed in the bottom chamber and DR⁺ Treg or DR⁻ Treg were either added to the bottom chamber together with the Tresp or to the top chamber alone, in triplicates. Both the upper and lower chambers received αCD3αCD2 beads. The cultures received either neutralizing αIL-10 mAb (5 μg/ml) or rat IgG isotype control (5 μg/ml). Cultures were pulsed with [³H]thymidine on day 4 and are representative of three experiments

In contrast with the suppression by the DR⁺ Treg subset, the Transwell analysis of DR⁻ Treg function indicated that these cells can mediate suppression via the secretion of IL-10 as well as via a cell contact-dependent mechanism. The DR⁻ Treg were able to directly inhibit Tresp proliferation via the secretion of IL-10, because the Tresp proliferation decreased by 40% when the DR⁻ Treg were stimulated in the upper chamber in the absence of αIL-10. However, when IL-10 was neutralized in replicate wells, the ability of the DR⁻ Treg to suppress across the membrane was abrogated as it resulted in full Tresp proliferation. Thus, the DR⁻ Treg subset can directly suppress the proliferation of Tresp via the secretion of IL-10. However, these data also show that DR⁻ Treg can induce a more vigorous suppression via an IL-10-independent, cell contact-dependent mechanism, because a significantly greater suppression occurred in the cocultures where cell contact was allowed that was only marginally reduced by neutralizing IL-10 (Fig. 4⇑). Thus, suppression by DR⁺ Treg absolutely requires cell contact, while DR⁻ Treg can suppress by either the secretion of IL-10 or via a cell contact-dependent mechanism.

T cell clones from DR⁺ and DR⁻ Treg populations exhibit an increased propensity to constitutively express CD25 and HLA-DR

A fundamental question in the biology of human T cells is whether the expression of MHC-DR on directly ex vivo-isolated CD4 T cells is a stable phenotype of a subset of T cells that would maintain DR expression with longer-term culture or whether DR is simply acquired with activation. It has been reported previously that suppressive human CD4⁺CD25⁺ Treg clones, generated by PHA/IL-2 stimulation, maintain high levels of CD25 on their surface 2 wk after in vitro stimulation, while CD25 expression is at baseline levels on non-Treg 2 wk after activation (27). We previously generated clones of CD4⁺CD25^high cells under the same stimulatory conditions and observed similar findings (18). In this study, T cell clones were generated by single cell cloning of CD4⁺CD62L^high T cells sorted on the basis of CD25 and DR expression. Single-cell cloning was undertaken to enable the coordinate assessment of function and changes in DR expression after in vitro culture by individual T reg subsets representing the ex vivo DR⁺ and DR⁻ populations.

DR⁺ Treg, DR⁻ Treg, DR⁺ CD25^−/low, or DR⁻ CD25⁻ (Tresp) cells from the CD4⁺CD62L^high population were sorted into single wells and expanded as described. Three weeks after a single restimulation, regardless of their level of expansion, every clone was tested for its ability to suppress the proliferation of freshly isolated autologous Tresp and was thus classified as suppressive (filled) or nonsuppressive (open) (Fig. 5⇓). The two Treg populations exhibited markedly different cloning capacity as the DR⁺ Treg consistently grew less well, while over two-thirds of wells seeded with DR⁻ Treg produced clones (Table I⇓). When tested for functional activity, the DR⁺ and DR⁻ Treg subsets gave rise to the highest frequencies of suppressive clones. However, in contrast with the assay of bulk ex vivo DR⁻ Treg, the suppressive clones derived from the DR⁻ Treg population no longer induced IL-4 or IL-10 during culture (Fig. 5⇓B). Furthermore, when tested for surface expression of CD25 and DR 24 days after their last activation (Fig. 5⇓D), 40% of the suppressive clones derived from the DR⁻ Treg population now expressed DR (13 of 32), while 44% of those from the DR⁺ Treg population expressed DR. In striking contrast, no clones derived from the DR⁺ CD25^−/low population and only 2% (1 of 50) of the DR⁻ CD25⁻-derived clones expressed DR. Thus, the initial ex vivo expression of high CD25 levels strongly correlated with subsequent clonal expression of DR, while the initial expression of DR on CD25^−/low cells did not predict long-term expression of DR. In addition, although both the DR⁺ and DR⁻ CD25^high populations gave rise to suppressive and nonsuppressive clones that did not express DR, virtually all the DR expressing clones were suppressive (4 of 5 and 13 of 14, respectively). Thus, DR expression is associated with a strong suppressive activity by ex vivo CD4⁺CD25^high cells and by roughly half of the suppressive clones generated from either the DR⁺ or DR⁻ Treg populations. These clone data indicate that the functional distinctions between the DR⁺ Treg and DR⁻ Treg populations are observed only ex vivo, as further differentiated DR⁺ Treg or DR⁻ Treg no longer function in a distinct fashion. Thus, these data lead us to hypothesize that DR may identify a mature and highly active regulatory CD4⁺CD25^high cell and not a separate T cell lineage.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Ex vivo expression of CD25^high, but not DR, on CD4⁺CD62L^high T cells identifies functional Treg that exhibit an increased tendency to become constitutive for DR with long-term culture. T cell clones were generated by expansion of single cells from CD4⁺CD62L^high T cells expressing DR or CD25^high via stimulation with soluble αCD3/αCD28 and IL-2 (see Table I⇓ for cloning efficiency data). Twenty-one days after a single restimulation, each clone was examined for expression of CD25 and HLA-DR (D) and tested for suppressor function in a 3-day, αCD3-stimulated, dose-response coculture assay with 3 × 10³ Tresp/well measuring proliferation (A) and cytokines (B and C). Each clone was classified as suppressive (•) or responder (○) based on coculture proliferation. Long-term DR expression depended on initial ex vivo expression of CD25^high and was associated with suppression by ex vivo CD4⁺CD25^high cells and by long-term T cell clones.

In an attempt to elucidate the suppressive mechanism(s) used by these clones, real time PCR analysis for Foxp3, IL-10 and TGFβ was performed on RNA isolated from a set of representative clones from each group (Fig. 6⇓). Thus, although the DR⁺ and DR⁻ Tresp ex vivo populations express little Foxp3 message as a whole, 18% of the clones derived from these populations expressed high levels of Foxp3 mRNA (Fig. 6⇓A) after their clonal expansion, even though they were not suppressive in the in vitro assays. In contrast, 85% of the clones derived from the DR⁺ and DR⁻ Treg populations were Foxp3^high. These data also indicate that high levels of Foxp3 are maintained in the clones derived from the CD25^high populations regardless of their functional suppressive ability. The amount of TGFβ mRNA (Fig. 6⇓B) produced by each clone did not correlate with either its suppressive ability or its original cloning group. In contrast, the results of the IL-10 expression analysis (Fig. 6⇓C) not only corroborate the lack of IL-10 production in cocultures with Treg clones, but also suggest that at least a subset of the suppressive clones derived from the DR⁻ Tresp population may be Tr1-like in their mechanism of action, as they express IL-10 but lack Foxp3.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

Long-term clones generated from the DR⁺ or DR⁻ CD25^high populations maintain high levels of Foxp3 mRNA regardless of their in vitro suppressive ability. RNA was isolated from those T cell clones that were generated and assayed in Fig. 5⇑, and that had expanded to sufficient numbers. These 6–10 clones per subset contained both suppressive (•) and nonsuppressive (○). Real-time PCR was used to comparatively assess the relative expression of Foxp3 (A), TGFβ (B), and IL-10 (C) in these clones, compared with endogenous β₂m. The levels of Foxp3 mRNA in the clones derived from the CD25^high populations were significantly higher than that expressed by the clones derived from the CD25⁻ populations (p < 0.0001, Mann-Whitney U test), while there was no significant differences between the groups in their expression of TGFβ or IL-10.