Selective Survival of Naturally Occurring Human CD4⁺CD25⁺Foxp3⁺ Regulatory T Cells Cultured with Rapamycin¹

A growing body of evidence suggests that regulatory T (Treg)³ cells are capable of inhibiting most types of immune responses. They play key roles not only in the maintenance of immunological self-tolerance, but also in limiting responses to foreign Ags. Thus, Treg cells have been implicated in autoimmune diseases, organ transplantation, and infectious diseases (1, 2, 3). Importantly, Treg cells can also dampen antitumor T cell immunity and have been proposed as a potential mechanism of immune escape (4). Previously, increased proportions of Treg cells among the total CD4⁺ T cell populations have been described in the circulation and tumor tissues of patients with different types of cancer (5, 6, 7). In ovarian cancer, accumulations of Treg cells in the tumor predicted a significant reduction in survival (8). To date, at least two types of CD4⁺ Treg cells have been partly characterized in humans: 1) Treg type 1 and Th3 cells, which are induced in the periphery upon encountering cognate Ags and secrete IL-10 and TGF-β, respectively (9, 10) and 2) naturally occurring CD4⁺CD25^highFoxp3⁺ T cells (naturally occurring CD4⁺CD25⁺ Treg (nTreg) cells) which arise directly in the thymus and have the ability to suppress responses of both CD4⁺CD25⁻ and CD8⁺CD25⁻ T cells (11, 12, 13, 14). nTreg cells express the transcription factor, forkhead box p3 (Foxp3), which has been associated with their development and suppressive function (15, 16). In addition, they express glucocorticoid-induced TNFR family related gene (GITR) and intracellular CTLA-4. More recent studies indicate that CD4⁺CD25^highFoxp3⁺ T cells are a heterogeneous population within CD4⁺ T cells, endowed with different regulatory functions, including regulation of Ag-specific immune responses (17, 18). As our knowledge of Treg-mediated immunosuppression has increased, the potential for therapeutic alterations in immunity through in vivo and in vitro manipulation of Treg cell numbers and/or functions in disease holds great promise.

Currently, isolation and expansion of human nTreg cell subsets into functionally active, disease-specific T cells is, however, difficult due to the following: 1) the paucity of nTreg cells in the peripheral blood; 2) the lack of specific identity markers for nTreg cells; and 3) the absence of protocols for rapid and extensive expansion of pure nTreg cells with suppressive activity from peripheral blood. In humans, CD4⁺CD25⁺ T cells include suppressor CD4⁺CD25^high T cells as well as CD4⁺CD25^low T cells which are nonsuppressive, activated CD4⁺ T cells. Furthermore, CTLA-4 and GITR are activation Ags and because levels of their expression vary depending on cell activation, they are not useful for discriminating nTreg from effector T cell populations. Similarly, Foxp3 expression—though more specific for Treg—may also be up-regulated on effector cells following activation (19), and due to its intracellular localization, Foxp3 cannot be used for separation of living Treg cells. As a consequence, most current studies using CD4⁺CD25⁺ populations for analysis of the phenotype and functions of nTreg cells are not interpretable due to the presence of contaminating effector CD4⁺ T cells, which reduce or even completely camouflage suppressive functions of nTreg cells present in T cell cultures.

It has been shown that the immunosuppressive drug rapamycin (sirolimus) is a powerful pharmacological agent which blocks intracellular signaling in response to T cell growth factors, e.g., IL-2, required for the progression of activated T cells from G₁ into S phase (20). Though rapamycin has been reported to favor CD4⁺CD25⁺ T cell-dependent immunoregulation in vitro and in vivo (21, 22, 23, 24, 25), the exact mechanism of its action on nTreg cells remains largely unknown. In particular, a direct effect of rapamycin on survival and proliferation of human nTreg cells has not been demonstrated.

To gain insights into the phenotypic and functional characteristics of expanded, phenotypically homogeneous, and highly suppressive populations of human CD25^high nTreg cells, we have used rapamycin to discriminate between CD4⁺CD25^highFoxp3⁺ T cells with a constitutive expression of CD25 from CD4⁺CD25⁺ effector T cells and CD4⁺CD25⁻ resting T cells. In this study, we provide new evidence that in the presence of IL-2, rapamycin alone or rapamycin combined with stimulation signals delivered via TCR and CD28 is capable of fostering the selective survival and expansion of CD4⁺CD25^highFoxp3⁺ nTreg cells, which are endowed with potent suppressive activities and thus may serve as pure populations of human nTreg cells for future investigations and therapy.

All Abs were purchased from BD Pharmingen except anti-GITR (R&D Systems) and Foxp3 (eBioscience). Rapamycin, LPS, PHA, PMA, ionomycin, and saponin were obtained from Sigma-Aldrich. IL-2 and IL-7 cytokines were obtained from R&D Systems. Monensin was obtained from BD Pharmingen. T cells were cultured in RPMI 1640 (Invitrogen Life Technologies) supplemented with human AB serum (8% v/v), l-glutamine (200 mM), sodium pyruvate (1%), nonessential amino acids (1%), penicillin/streptomycin (1%) (all obtained from Invitrogen Life Technologies), 2-ME (1%) (Sigma-Aldrich), and fluoroquinolon (1%) (ciproxin infusion; Bayer).

Peptides were synthesized by standard solid-phase chemistry on a multiple peptide synthesizer (Biosynthesis). The following synthetic peptides were used: tetanus toxoid_947–967 (FNNFTVSFWLRVPK) (TT_947–967), MAGE-A3_247–258 (TQHFV QENYLEY) (MAGE-A3_247–258), Melan-A/MART1_26–35 (ELAGIGILTV) (Melan-A_26–35), HIV-POL_476–484 (ILKEPVHGV) (HIV_476–484). Synthesis of PE-labeled HLA-A2/Melan-A_26–35 multimeric complexes was performed as previously described (26).

PBMC were isolated from buffy coat preparations derived from the whole blood of healthy volunteers (Suisse Blood Center, Zurich, Switzerland and Central Blood Bank, Pittsburgh, PA) by density sedimentation on Ficoll-Hypaque gradients. Cells recovered from the gradient interface were washed twice in RPMI 1640 medium, counted, and immediately used for MACS and staining. Briefly, CD4⁺ T cells were negatively selected from the total PBMC using the CD4 isolation kit (Miltenyi Biotec), yielding a population of CD4⁺ cells with purity of 92–98%. Positive selection on anti-CD25 magnetic microbeads was then used to separate the negative fraction containing CD4⁺CD25⁻ T cells from the CD4⁺CD25⁺ T cell fraction, using the CD4⁺CD25⁺ T Regulatory Cell Isolation kit from Miltenyi Biotec. Cells were then applied to a second magnetic column, washed, and eluted again. This procedure led to the complete positive selection of CD4⁺CD25⁺ T cells (purity ≥96%), and negative depletion of CD4⁺CD25⁻ T cells, as measured by flow cytometry. Both cell subsets were used for in vitro expansion.

Initially, purified CD4⁺CD25⁺ or CD4⁺CD25⁻ T cells were cultured in Terasaki plates (10 cells/well) with syngeneic irradiated (3000 rad) PBMC (20,000 cells/well), PHA (1 μg/ml), and 50 IU IL-2/ml for 7–8 days to increase their numbers. Next, aliquots of cells (150,000 cells/well) were transferred to wells of a 96-well plate and cultured with anti-CD3/anti-CD28 Ab-coated magnetic Dynabeads (CD3/CD28 beads; Dynal Biotech/Invitrogen Life Technologies) at 4 × 10⁷ beads/ml and 1,000 IU/ml IL-2 for 7 days. Rapamycin at the final concentration of 1 nM was added to half of the cultures (Fig. 2). After 1 wk of expansion, cells were washed to remove IL-2 and the beads were removed with a magnetic particle concentrator (Miltenyi Biotec). The expanded cells were then transferred to wells of a 24-well plate and restimulated with CD3/CD28 beads and 1000 IU/ml IL-2. These cells were cultured in the presence or absence of 1 nM rapamycin for 1 wk (Fig. 2). Cells cultured in the absence of rapamycin are designated as “R0” and those cultured with 1 nM rapamycin as “R1.” After every week of expansion, cells sampled for assays were washed twice, the beads were removed, and the cell number was determined by direct counting to evaluate expansion. Cell aliquots sampled from each group (at least 200,000 cells) were maintained for 48 h in culture medium and low-dose IL-2 (50 IU/ml) in the absence of CD3/CD28 beads and rapamycin before assays.

To evaluate effects of rapamycin on expansion of fresh MACS-purified T cells, CD4⁺CD25⁺ and CD4⁺CD25⁻ cells were cultured ± rapamycin in the presence of 50 IU/ml IL-2 or PHA and irradiated syngeneic APC for 7 days and used for assays.

For RNA extraction, the RNase easy kit (Qiagen) was used according to the manufacturer’s recommendation. Extracted RNA was reverse transcribed using the SuperScript III First-Strand Synthesis System (Invitrogen Life Technologies). We performed quantitative PCR using the TaqMan ABI Prism 7000 Sequence Detection System (Applied Biosystems) on cDNA samples for Foxp3 expression normalized to the housekeeping gene 18s rRNA. The following primers for PCR were purchased from Applied Biosystems: forkhead/winged helix transcription factor (Foxp3), 5′-GCACCTTCCCAAATCCCAGT-3′ (sense) and 5′-GCAGGCAAGACAGTGGAAACC-3′ (antisense) in a 20-μl volume. Levels of 18s rRNA were quantified by using TaqMan PDAR Eukaryotic 18s Endogenous Controls (Applied Biosystems). The following cycling conditions were used: 10 min at 95°C, followed by 30 cycles of 15 s at 95°C, and 1 min at 60°C.

All CD4⁺ T cell samples were stained with anti-CD4, monocyte-discriminating Abs (anti-CD14, anti-CD32, and anti-CD116). Fresh as well as in vitro-expanded CD4⁺CD25⁺ (R0 vs R1) and CD4⁺CD25⁻ T cells (at least 200,000 cells/tube) were stained with the following mAbs for 15 min at 4°C to determine the “nTreg” surface marker profile: PE-conjugated anti-CD25, PerCP-labeled anti-CD4, allophycocyanin-conjugated anti-CD45RO, CCR7, and HLA-DR, FITC-conjugated anti- CD62L, Fas, GITR, CD45RA, and CD27. Criteria used for designating T cells as CD4⁺CD25^high were based on mean fluorescence intensity (MFI) as follows: with fresh PBMC when MFI was ≥20, and with isolated/cultured CD4⁺CD25⁺ T cells, which were highly enriched populations, when MFI was ≥500, as illustrated in Fig. 1, A and B. Appropriate isotype controls were included in all experiments.

Intracellular levels of IL-10 in fresh or in vitro-expanded R0 and R1 T cells were measured after a 16-h incubation with 25 ng/ml LPS (Sigma-Aldrich) in the presence of 1 μg/ml monensin (GolgiStop; BD Pharmingen). Intracellular staining was performed as previously described (27). Briefly, for detection of intracellular Foxp3, CTLA-4, and IL-10, samples were first incubated with mAbs against surface markers CD4 and CD25. After subsequent washes, cells were fixed with PBS containing 4% formaldehyde for 20 min at room temperature, washed once with PBS containing 0.5% BSA and 2 nM EDTA, permeabilized with PBS containing 0.5% (w/v) BSA and 0.2% (v/v) saponin, and stained with anti-CTLA-4-allophycocyanin, anti-Foxp3-FITC-labeled or anti-IL-10-PE-labeled mAb for 30 min at room temperature. Cells were further washed twice with PBS containing 0.5% BSA and 0.2% saponin, resuspended in FACS flow solution, and immediately analyzed by flow cytometry. Appropriate isotype controls were used in all experiments.

CD4⁺CD25⁺ and CD4⁺CD25⁻ T cells obtained by MACS isolation and R0 as well as R1 cells were assessed for annexin V (ANX V) binding. Samples were stained with FITC-conjugated ANX V (Molecular Probes/Invitrogen Life Technologies) and propidium iodide (PI; Molecular Probes), according to the instructions provided by the distributor.

Flow cytometry was performed using a FACScan flow cytometer (BD Biosciences) equipped with CellQuest software (BD Biosciences). The acquisition and analysis gates were restricted to the lymphocyte gate, as determined by characteristic forward and side scatter properties of lymphocytes. Analysis gates were restricted to the CD3⁺CD4⁺, CD4⁺CD25⁺, and CD4⁺CD25^high T cell subsets. Usually, 1 × 10⁵ lymphocytes were collected for analysis. For ANX V-stained samples the acquisition gates were restricted to the live (PI⁻) as well as the dead cell (PI⁺) populations present in each sample. Six parameters (forward scatter, side scatter, and four fluorescence channels) were used for list mode data analysis.

TCRV β-chain repertoire of fresh as well as in vitro-expanded R0 and R1 cells was determined by staining with a panel of Vβ chain-specific Abs (BD Pharmingen), followed by flow cytometry analysis.

CD4⁺ and CD8⁺ T cells enriched from PBMC using a miniMACS device were stimulated with HLA-A*0201-restricted Melan-A/MART-1_26–35- and HLA-DP*0401-restricted TT_947–967 peptides. To generate peptide-specific T cells, peptide-pulsed irradiated syngeneic APCs were cocultured with CD4⁺ or CD8⁺ T cells in the presence of IL-2 (25 IU/ml) and IL-7 (10 ng/ml) as described (27). To select for peptide-specific T cells in these cultures, flow cytometry was used to sort T cells stained by multimers or T cell-secreting cytokines in response to peptide stimulation. The isolated T cells were cloned and cultured as previously described (27).

Initially, suppression of proliferation mediated by MACS-purified CD4⁺CD25⁺ and CD4⁺CD25⁻ T cells before and after exposure to rapamycin was analyzed using an allogeneic system. Suppressive functions of these cells immediately after isolation, after 1 wk expansion (in the presence of 50 IU IL-2/ml or PHA and irradiated APC) or after 3 wk expansion (in the presence of CD3/CD28 beads and 1000 IU IL-2/ml) cocultured with fresh allogeneic responder CD4⁺CD25⁻ or CD8⁺CD25⁻ T cells was determined. For this purpose, the responder T cells were labeled with 0.5 μM CFSE (Molecular Probes) and stimulated with OKT3 (1 μg/ml) (American Type Culture Collection) and soluble anti-CD28 Ab (1 μg/ml; Miltenyi Biotec). Suppressor cells, either CD4⁺CD25⁺ or CD4⁺CD25⁻ cell fractions, were added at the suppressor:responder ratio of 1:1. Cells were incubated for 5 days in a CO₂ incubator at 37°C.

Next, Ag-specific suppression mediated by CD4⁺CD25⁺ R0 or R1 T cells that were expanded for 3 wk was analyzed in an autologous in vitro system. CD4⁺CD25⁻ TT_947–967 peptide-specific T cell line ZH-P1 and CD8⁺CD25⁻ Melan-A/MART1_27–35 peptide-specific T cell clone ZH- P7/2 were labeled with 0.5 μM CFSE (Molecular Probes) and incubated with either R0 or R1 T cells at the 1:1 ratio. Peptide-specific responder T cells were stimulated with the relevant peptide-pulsed (2 μg/ml) APC (autologous PBMCs) or with PHA (1 μg/ml).

All CFSE data were analyzed using the ModFit software provided by Verity Software House. The percentages of suppression were calculated based on the proliferation index of responder cells alone compared with the proliferation index of cultures containing responders and Tregs. The program determines the percent of cells within each peak and the sum of all peaks in the control culture is taken as 100% of proliferation and 0% of suppression.

For IFN-γ ELISPOT analysis, the CD4⁺CD25⁻ TT_947–967 peptide-specific T cell line ZH-P1 and CD8⁺CD25⁻ Melan-A_27–35 peptide-specific T cell clone ZH-P7/2 were stimulated either with the relevant peptide (TT_947–967, Melan-A/MART1_26–35, respectively) or irrelevant peptides (MAGE-A3_247–258, HIV-POL_476–484, respectively) using HLA-DP*0401⁺ and HLA-A*0201⁺ target cells, respectively. Twenty-four-hour ELISPOT assays were performed as previously described (27). Maximum IFN-γ release was determined by stimulation of responder cells with PMA/ ionomycin.

Differences between groups were assessed using the Student t test. Values of p were two tailed. Values of p < 0.05 were considered significant.

We first determined that CD4⁺CD25⁺ T cells comprised 2–5% of CD4⁺ T cells in the peripheral blood of 10 healthy individuals analyzed. The CD25^high T cells defined as described in Materials and Methods represented 0.5–1% of the total CD4⁺ T cell population, while the CD25^low T cells accounted for 2.5% of CD4⁺ T cells (Fig. 1,A). Next, CD4⁺CD25⁺ T cells were isolated using MACS on anti-CD4/anti-CD25 beads. The purity of positively selected CD4⁺CD25⁺ T cells was >96%. After MACS isolation, the CD4⁺CD25⁺ cell fraction contained both CD4⁺CD25^high and CD4⁺CD25^low cells (Fig. 1 B).

Fig. 2 is a schema designed for expansion of isolated CD4⁺CD25⁺ human T cells. It is based on our initial finding that these cells plated at the usual concentrations of ∼10⁵ cells/well in 96-well plates failed to proliferate even when IL-2, PHA, and APCs were added. We determined that plating CD4⁺CD25⁺ T cells at 10 or fewer cells per well in Terasaki plates allowed for their initial expansion to 15–20 × 10³ cells/well and their transfer to wells of 96-well plates 7 days later. To obtain sufficient cell numbers for subsequent culture ± rapamycin, CD4⁺CD25⁺ T cells remained in wells of a 96-well plate for 1 wk, yielding ∼100–150 × 10³ cells. At this point, the cells were divided into parallel R0 and R1 cultures. Viable cell counts performed at each transfer showed that the fold expansion was 50- to 100-fold for R0 as well as for R1 cultures at the end of the entire 3-wk expansion period.

The MACS-isolated cells were plated in the presence of 50 IU of IL-2 or in PHA and autologous feeder cells and assessed for phenotype after 1 wk of culture. In the presence of IL-2 alone, CD4⁺ T cells down-regulated surface expression of CD25 (Fig. 1,C). In contrast, most CD4⁺ T cells stimulated with PHA and autologous feeder cells retained CD25 expression (Fig. 1,D). These cultures were next incubated in the presence (R1) or absence (R0) of 1 nM rapamycin for 7 days. In IL-2 or PHA-containing R0 cultures, few CD4⁺ T cells were CD4⁺CD25^high and most were CD25⁻ (Fig. 1, E and F, left). In contrast, the parallel R1 cultures were enriched in CD4⁺CD25^high T cells (Fig. 1, E and F, right). The viable cell counts and ANX V-binding assays performed on day 7 indicated that most CD4⁺CD25⁻ T cells, which rapidly outgrow in the R0 cultures, are not sensitive to apoptosis (Fig. 1, G and H). In contrast, the R1 cultures largely contained proliferating CD4⁺CD25⁺ T cells and CD4⁺CD25⁻ T cells, representing a minority population, were highly sensitive to apoptosis (Fig. 1, G and H). These data are consistent with the preferential expansion of CD4⁺CD25⁺ T cells accompanied by the significant inhibition of CD4⁺CD25⁻ T cell expansion due to a high frequency of ANX V-binding cells in the R1 cultures (Fig. 1, G and H).

These initial experiments indicated that rapamycin-treated CD4⁺CD25⁺ cells could expand up to 1- to 2.5-fold in cultures supplemented with IL-2 or PHA and autologous APC. The optimal rapamycin concentration for expansion of nTreg cells was next determined. In contrast to experiments performed previously in mice (25), doses of rapamycin ≥10 nM considerably decreased proliferation of MACS-purified CD4⁺CD25⁺ T cells (data not shown). Conversely, lower doses of rapamycin (≤1 nM) together with CD3/CD28 beads and 1000 IU/ml IL-2 showed no difference in expansion rates for MACS-purified CD4⁺CD25⁺ and CD4⁺CD25⁻ T cells compared with expansion cultures without rapamycin (data not shown). A rapamycin concentration of 1 nM was determined to be optimal for expansion of CD4⁺CD25⁺ T cells and used for all additional experiments.

Fig. 3,A illustrates expansion of CD4⁺CD25⁻, CD4⁺CD25⁺, and CD4⁺CD25^high T cell populations in parallel R0 and R1 cultures during a 3-wk expansion period. The cultures were monitored weekly for cell numbers and the percentage of CD4⁺CD25^high T cells. MACS-purified CD4⁺CD25⁻ T cells proliferated readily in the absence of rapamycin (R0), reaching considerable fold expansion by week 3. However, CD4⁺CD25⁺ as well as CD4⁺CD25^high T cells were a minor component of R0 cultures. In contrast, while expansion of CD4⁺CD25⁻ and CD4⁺CD25⁺ T cells was delayed in the presence of rapamycin (R1), by week 3, vigorous proliferation of CD4⁺CD25⁺ and CD4⁺CD25^high T cells occurred, leading to their 50- to 100-fold expansion. CD4⁺CD25⁻ T cells were not proliferating in R1 cultures (Fig. 3,A), while a significant enrichment in CD4⁺CD25^high T cells was routinely observed (Fig. 3 B).

Previous experiments showed that rapamycin blocked proliferation of CD4⁺CD25⁻ T cells (25) and our initial data were in agreement with this possibility. To analyze whether rapamycin selectively induces apoptosis of CD4⁺CD25⁻ T cells, we examined ANX V binding to fresh and cultured CD4⁺CD25⁺ and CD4⁺CD25⁻ T cells after their exposure to rapamycin. As shown in Fig. 4,A, in normal fresh PBMCs exposed to rapamycin for 48 h, the CD4⁺CD25⁺ T cells bind little ANX V (0.1 ± 1.2%). In contrast, the CD4⁺CD25⁻ T cell subset showed significantly higher (p ≤ 0.00001) percentages of ANX V⁺ cells (97 ± 12%). Thus, fresh CD4⁺CD25⁺ nTreg are resistant, while CD4⁺CD25⁻ T cells are sensitive to rapamycin-induced apoptosis. Also, a significantly higher percentage of ANX V⁺ cells was observed in week 3 R1 cultures of CD4⁺CD25⁻ T cells (80 ± 13 vs 20 ± 15%) than in R0 cultures of CD4⁺CD25⁻ T cells (p < 0.0001), as shown in Fig. 4 B for a representative culture. In contrast, ANX V binding to CD4⁺CD25⁺ and CD4⁺CD25⁻ T cell subset was similar when these cells were cultured without rapamycin (23 ± 14 vs 20 ± 15%; p < 0.44; data not shown). Our results are consistent with the conclusion that rapamycin blocks proliferation of CD4⁺CD25⁻ T cells by inducing their apoptosis, while CD4⁺CD25⁺ Treg are resistant to apoptosis and continue to proliferate in the presence of 1 nM rapamycin.

Foxp3 has been considered to be a major phenotypic marker of Treg. Therefore, its expression was determined in fresh CD4⁺CD25⁺ and CD4⁺CD25⁻ T cells as well as after 3 wk of expansion in R0 and R1 cultures. Initially, mRNA for Foxp3 was measured, using Foxp3 specific real-time quantitative PCR (Fig. 5,A). Later, intracellular flow cytometry analysis of cells stained with Foxp3-specific Abs was used (Fig. 5,B). Neither fresh CD4⁺CD25⁻ R0 T cells nor fresh CD4⁺CD25⁻ R1 T cells expressed Foxp3 mRNA. In contrast, CD4⁺CD25⁺ T cells before and after exposure to rapamycin expressed Foxp3 mRNA (Fig. 5,A), and CD4⁺CD25⁺ R1 T cells expressed significantly higher levels of Foxp3 mRNA (p < 0.044) compared with CD4⁺CD25⁺ R0 T cells (Fig. 5,A). We have also studied Foxp3 protein expression by fresh and rapamycin-expanded CD4⁺CD25⁺ T cells by flow cytometry. As shown in Fig. 5 B, CD4⁺CD25⁺ T cells in 3-wk R0 cultures contained 10 ± 6.5% of Foxp3⁺ cells. In contrast, CD4⁺CD25⁺ T cells in 3-wk R1 cultures were all Foxp3 positive (98 ± 2.5%).

To establish the phenotype of T cells expanded in the presence of rapamycin relative to that of fresh CD4⁺CD25^high T cells, multicolor flow cytometry was performed. First, the phenotype of fresh PBMC before exposure to rapamycin was determined. The CD4⁺CD25^high T cell subset in fresh PBMC expressed higher levels of CD62L, GITR, CTLA-4, CD45RO, Fas, CCR7, and HLA-DR markers compared with the CD4⁺CD25^low subset or total CD4⁺CD25⁺ T cells (Fig. 5,C). R1 cells after 1 as well as 3 wk of expansion expressed significantly higher levels of the same markers present on the CD4⁺CD25^high subset in fresh PBMC (Fig. 5,C; p < 0.05). In R1 cultures, CD4⁺CD25^high cells, which represent the proliferating and expanded subset, are CD45RO⁺ (∼40%), CCR7⁺ (∼80%), and CD27⁻ (<10%), a phenotype resembling the central memory phenotype (28). The expression of all “nTreg” markers was higher on CD4⁺CD25⁺ T cells after 3 wk of expansion as compared with nTreg after 1 wk of expansion; however, the difference was not statistically significant (p ≤ 0.65) (Fig. 5 C). No significant difference in expression of the above-listed markers was observed in R1 cells expanded with <1 nM rapamycin vs R0 cells (data not shown).

To evaluate suppressive activity of R0 and R1 T cells after MACS isolation and 1 wk of expansion, cocultures (in the presence of IL-2 or PHA and irradiated APC) were set up with CFSE-labeled fresh allogeneic CD4⁺CD25⁻ T cells (ratio 1:1) responding to OKT3 and anti-CD28 Ab (Fig. 6,A). The mean ± SD percentage of inhibition was 43 ± 2.5% for fresh CD4⁺CD25⁺ T cells obtained after MACS isolation from healthy donors. No suppression was observed with CD4⁺CD25⁺ T cells after 1 wk of expansion in IL-2 alone or with CD4⁺CD25⁺ T cells after 1 wk of expansion with PHA and irradiated APC (Fig. 6,A). These cells tended to lose CD25 expression (see Fig. 1, C and D) and suppressor functions, possibly due to a rapid outgrowth of CD4⁺CD25⁻ populations (see Fig. 3,A). In contrast, as seen in Fig. 6 A, the mean ± SD percentages of inhibition obtained with CD4⁺CD25⁺ T cells from R1 cultures were: 62 ± 7.2% (fresh), 71 ± 3.5% (at 1 wk + IL-2), and 74 ± 6.5% (at 1 wk plus PHA/APC), respectively.

Furthermore, we analyzed the suppressive activity of R0 and R1 T cell cultures after 3 wk of expansion with CD3/CD28 beads and 1000 IU/ml IL-2 on proliferation of fresh allogeneic CD4⁺CD25⁻ and CD8⁺CD25⁻ T cells (ratio 1:1) responding to OKT3 and anti-CD28 Ab, using CFSE assays (Fig. 6,B). The mean ± SD percentage of inhibition obtained with cells of five healthy donors was 82 ± 4.5 and 98 ± 2.35%, respectively (Fig. 6,B), when R1 cells were tittered in at the 1:1 ratio. Much lower levels of suppression were observed (12 ± 4.5 and 16 ± 6.7%, respectively) when R0 T cells were added to the cocultures (Fig. 6 B).

To study Ag-specific suppression in an autologous in vitro system, syngeneic R0 and R1 T cells after 3 wk of expansion with CD3/CD28 beads and 1000 IU/ml IL-2 were added at 1:1 ratios to CD4⁺CD25⁻ T cell line ZH-P1 and CD8⁺CD25⁻ T cell clone ZH-7/2, which are specific for TT_947–967 peptide and Melan-A/MART1_26–35 peptide, respectively. For this purpose, the specific T cells were stimulated with either 1) PHA or 2) APC pulsed with the cognate peptides TT_947–967 and Melan-A/MART1_26–35, respectively. After 5 days, specific CD4⁺ and CD8⁺ T cell clones proliferated vigorously in response to cognate peptides in the absence of R1-expanded cells (Fig. 6,C). However, proliferation of CD4⁺ and CD8⁺CD25⁻ T cells after peptide stimulation was strongly inhibited upon coculture with syngeneic R1 T cells (Fig. 6,C) but not syngeneic R0 T cells. The mean ± SD percentage of inhibition obtained with cells of the same donor in three experiments was 85 ± 1.5% for CD4⁺ TT peptide-specific T cell line ZH-P1 (Fig. 6,C) and 58 ± 6% for CD8⁺CD25⁻ MART1-peptide specific T cell clone ZH-7/2 (Fig. 6 C). The mean percentages of inhibition of CD4⁺CD25⁻- and CD8⁺CD25⁻-specific T cells by syngeneic R0 or R1 T cells after stimulation with PHA peptide were similar to those obtained after peptide stimulation (data not shown).

To determine whether rapamycin induces suppressor T cells in CD4⁺CD25⁻ T cells, we analyzed inhibition of fresh allogeneic CD4⁺CD25⁻ T cell or CD8⁺CD25⁻ T cell proliferation in response to OKT3, anti-CD28 mAb, and IL-2 by CD4⁺CD25⁻ T cells after 7 days or 3 wk of expansion, respectively, in the presence or absence of rapamycin. CFSE-labeled responders stimulated with OKT and anti-CD28 Ab were cultured alone or coculture with CD4⁺CD25⁻ R0 or R1 cells at the ratio of 1:1 (Fig. 7). No evidence was obtained for induction of suppressor T cells in these cultures as indicated by low levels of suppressor activity in R1 cocultures (R0 cocultures not shown). The data indicated that rapamycin does not induce suppressor T cells in CD4⁺CD25⁻ populations during a 7-day or 3-wk incubation.

Next, we determined the suppressive activities of R1 T cells on cytokine secretion using IFN-γ ELISPOT assays. In contrast to R0 T cells, both syngeneic (Fig. 8) and allogeneic (data not shown) R1 T cells strongly inhibited IFN-γ secretion of TT_947–967 peptide-specific CD4⁺ T cells (99.1 ± 0.5 and 97.3 ± 1%, respectively; Fig. 8, top) and Melan-A/MART1_26–35 peptide-specific CD8⁺ T cells (98.7 ± 3 and 95.8 ± 0.6%, respectively; Fig. 8, bottom) upon peptide stimulation. Similar results were obtained upon stimulation with PHA (data not shown).

To distinguish nTreg cells from IL-10-secreting Treg cells (9), intracellular IL-10 expression was determined in both fresh CD4⁺CD25⁺ and in vitro-expanded R0 and R1 T cells after stimulation with LPS. Neither fresh CD4⁺CD25⁺ T cells nor R0 and R1 T cells after in vitro expansion expressed IL-10 as measured by flow cytometry (data not shown).

TCRVβ chain repertoire of R0 and R1 T cells was determined using Vβ chain-specific Abs in a flow cytometry-based assay. We found that R1 T cells after a 3-wk expansion displayed a polyclonal TCRVβ chain repertoire, comparable to CD4⁺CD25⁺ T cells of fresh PBMC and to 3 wk-expanded R0 T cells of the same donor (data not shown).

We demonstrate here that human CD4⁺CD25^highFoxp3⁺ nTreg cells upon stimulation via TCR and CD28 or by IL-2 undergo rapid and robust expansion when rapamycin is added. Purified CD4⁺CD25⁺ T cells exposed to rapamycin: 1) exhibit a strong proliferative capacity; 2) show a Foxp3⁺GITR⁺CTLA⁻4⁺ phenotype; 3) have a potential for homing to lymph node (LN) as indicated by CD62L and CCR7 expression; 4) mediate potent suppressive activities against syngeneic or allogeneic CD4⁺ and CD8⁺ T cells, T cell clones or populations with defined antigenic specificity as well as polyclonal T cells with various antigenic specificities. In contrast, purified CD4⁺CD25⁺ T cells cultured in the absence of rapamycin (R0) are mainly composed of nonsuppressor T cells with a CD25^low phenotype. The ex vivo use of rapamycin as outlined here should open the avenue for the in vitro generation of highly efficient nTreg cell populations in humans.

In the last few years, several strategies have been developed for long-term culture and polyclonal expansion of human nTreg cells (29, 30, 31). It has been demonstrated that the combined stimulation via TCR, CD28, and IL-2R is required to transiently induce human Treg proliferation and this strategy is especially effective with anti-CD3/CD28 beads or artificial APC (29, 30, 31). MACS-based purification has previously been used to obtain enriched CD4⁺CD25⁺ T cells for in vitro expansion. However, though straightforward, this method results in a considerable contamination with conventional CD4⁺CD25^low T cells. Consequently, the addition of high doses of IL-2 may promote an overgrowth of cells with modest or no suppressor activity. Thus, Battaglia et al. (25) showed that cultured murine CD4⁺CD25⁺ T cells not exposed to rapamycin lose their suppressive function. Others, using sorting for selection of only CD4⁺CD25^high cells, were able to generate suppressor cell lines from adult blood (32). However, as previously noted, even sorted populations may contain a mix of conventional and Treg cells (32), and relatively few CD25^high T cells are recovered by this technique. In contrast, MACS separation of human CD4⁺CD25⁺ T cells followed by a large-scale expansion in the presence of rapamycin yields a population of cells with a defined phenotype and strong suppressive function.

Our data suggest that in the presence of rapamycin (1 nM), human Treg subsets present in the peripheral blood survive and expand, while the other T cells (i.e., CD4⁺CD25^low) are inhibited from proliferation. We demonstrated that fresh or activated CD4⁺CD25⁻ T cells exposed to rapamycin (1 nM) in the presence of CD3/CD28 beads and 1000 IU/ml IL-2 do not proliferate, compared with CD4⁺CD25⁻ T cells cultured under the same conditions but in the absence of rapamycin. In contrast, rapamycin-exposed human CD4⁺CD25^high T cells vigorously expand for at least 3 wk in culture. Rapamycin also enhances PHA-stimulated expansion of Treg cells. We hypothesize that this expansion of CD4⁺CD25^high T cells results from rapamycin-driven selection of T cells at a unique state of activation, with rapamycin-induced apoptosis of CD4⁺CD25⁻ T cells. Sensitivity of these T cells to apoptosis, as indicated by ANX V binding, and resistance to apoptosis of fresh as well as rapamycin-expanded CD4⁺CD25^high T cells, suggest that the mechanism of rapamycin-based selection involves the death pathway. It is also noteworthy that rapamycin concentrations higher than 10 nM significantly decreased expansion rates of CD4⁺CD25^high T cells, while doses lower than 1 nM were not effective in inducing their expansion in our hands (data not shown). This indicates the existence of a threshold for rapamycin-mediated responses that could be related to the activation state of CD4⁺CD25⁺ T cells. Importantly, Treg cultured in the presence of IL-2 without rapamycin do not mediate suppression, while RT cells are highly suppressive (Fig. 6). In contrast to the findings reported by Valmori et al. (33), we demonstrate that the suppressive function of R1 T cells is maintained after rapamycin withdrawal (data not shown).

Phenotypic characteristics of RT cells are qualitatively similar to those of freshly isolated CD4⁺CD25^high T cells. We show here that rapamycin-expanded nTreg, like fresh nTreg, show high expression levels of CD25, Foxp3, CD62L, CD45RO, GITR, CTLA-4, CCR-7, Fas, and HLA-DR. In contrast, CD4⁺CD25⁺ T cells expanded in the absence of rapamycin express significantly lower levels of these markers. Furthermore, CD4⁺CD25⁻ and CD4⁺CD25^low T cells expanded in the presence of rapamycin showed low marker expression (Fig. 5 C). This observation suggests that rapamycin promotes survival of a subset of CD4⁺CD25^high T cells that constitutively express high levels of the markers characterizing the suppressor phenotype. Furthermore, RT cells displayed a broad TCRVβ usage similar to that of CD4⁺CD25⁻ from fresh PBMC. As this TCRVβ chain repertoire remained virtually unchanged after in vitro expansion in the presence of rapamycin, it appears that RT cells represent a polyclonal expansion of T cells originating from precursors present in the peripheral blood.

RT cells expressed significantly higher levels of CD95 (Apo-1/Fas) molecules on the surface as compared with R0 T cells. Upon TCR stimulation of CD4⁺CD25^high cells or in RT cells, we observed no significant activation-induced cell death in agreement with recent findings (34), suggesting that these T cells were resistant to apoptosis despite high levels of CD95 expression. In contrast, rapamycin markedly accelerated apoptosis in the CD4⁺CD25⁻ T cell cultures, as shown by ANX V binding, while expanded RT cells remained alive. Similar apoptosis-mediated depletion of activated T cells and, consequently, an increase in CD4⁺CD25⁺ T cells upon rapamycin treatment has been described in allotransplant and autoimmune animal models (21, 35).

It is believed that successful expansion of Treg is a prerequisite for future adoptive transfers of these cells in human diseases. In transplantation or autoimmunity, expansion and transfers of Treg might be beneficial, while in cancer, their attenuation could allow for recovery of antitumor immune responses. After 1–3 wk of expansion, T cells exposed to rapamycin (RT cells) represented a pure population of CD4⁺CD25^highFoxp3⁺ T cells (≥98%). Further phenotypic analysis showed that a considerable fraction expressed CD62L and CCR7, two receptors also present on T cells and some subsets of memory T cells and considered to be responsible for lymphocyte homing to secondary lymphoid organs (36). Indeed, accumulating data in animal models systems of autoimmunity or allotransplantation suggest that the LN is a critical site for tolerance induction (36, 37). The phenotype of human RT cells suggests that these cells could migrate to secondary lymphoid organs and to exert suppressive function in a site-specific manner.

Although our results indicate that rapamycin can be used for the expansion of Treg cells with a defined phenotype and suppressive functions from the peripheral blood of normal volunteers, in vitro expansion of Treg from blood of subjects with disease may not be equally feasible. In patients with autoimmune diseases, these cells might be less responsive or present in fewer numbers (38, 39). We are in the process of developing a protocol for a large-scale expansion of human Treg with rapamycin, and our preliminary data indicate that a 10⁵ expansion index after 3 wk of culture is achievable with peripheral blood-derived CD4⁺CD25⁺ T cells of normal donors. Also, it will be of interest to expand nTreg cell subsets isolated from cancer patients, especially patients with leukemia treated with hemopoietic stem cell transplantation. A possibility of therapy with expanded Treg to ameliorate graft-vs-host disease has been considered (40). In addition to therapeutic use, the availability of purified Treg in numbers sufficient for detailed phenotypic and functional analysis of these cells and understanding of their mode of action or Ag specificity is scientifically important.

In summary, we have shown that human polyclonal CD4⁺CD25⁺ nTreg cells capable of mediating potent suppression survive and readily proliferate upon exposure to rapamycin. Expanded CD4⁺CD25^high nTreg cells obtained from the peripheral circulation of normal donors show a “characteristic” suppressor phenotype and maintain expression of LN homing receptors. The data presented here may considerably accelerate the development of immunotherapeutic approaches for the treatment of autoimmune diseases or posttransplant alloreactions by the adoptive transfer of nTreg cells.

We thank Dr. Immanuel Luescher for multimer synthesis and Conny Schneider and Heidi Mattlin for technical assistance.

The authors have no financial conflict of interest.