Rapamycin Monotherapy in Patients With Type 1 Diabetes Modifies CD4⁺CD25⁺FOXP3⁺ Regulatory T-Cells

Patients with long-lasting type 1 diabetes (>5 years) who had reduced awareness of hypoglycemia, brittle diabetes, or progressive complications, despite optimization of insulin therapy, were candidates for solitary islet transplant at the Telethon-Juvenile Diabetes Research Foundation Center for Beta Cell Replacement, San Raffaele Scientific Institute, Milan. In this study, six patients received rapamycin treatment at 0.1 mg · kg⁻¹ · day⁻¹ (Table 1). Each patient is identifiable by a specific symbol, which can be followed throughout the manuscript (e.g., patient Hsr-066-ITA-rp06 is recognizable by the ○ symbol). Peripheral blood was obtained before and during rapamycin treatment (before receiving islet transplantation) after informed consent and ethics committee approval. Five normal donors of similar age and sex to the patients were recruited through the courtesy of Centro Trasfusionale, San Raffaele Scientific Institute, Milan, and donated peripheral blood after informed consent and ethics committee approval.

Peripheral blood mononuclear cells (PBMCs) were isolated over a Ficoll-Hypaque (Amersham Pharmacia Biotech Europe, Uppsala, Sweden) density gradient centrifugation from sodium-heparinized venous blood samples and washed twice in PBS (Cambrex-Biowhittake, Walkersville, MD). All samples were frozen in FCS (Cambrex-Biowhittaker) containing 10% dimethyl sulfoxide (Sigma-Aldrich, St. Louis, MO) in a controlled-rate automated freezing device to −80°C and then stored in liquid nitrogen. All of the experiments were performed on thawed cells. Samples from the same type 1 diabetic patient collected before and during rapamycin treatment were tested simultaneously in parallel to cells from one normal donor. Each patient was tested alongside a separate normal donor.

Thawed PBMCs were stained for surface antigens with the following monoclonal antibodies (mAbs) all purchased from BD Pharmingen (San Diego, CA): anti-CD4 PerCP (peridinin chlorophyll protein) (clone SK3), anti-CD25 allophycocyanin (clone 2A3), anti-CD62L phycoerytrin (PE) (clone SK11), anti-CD127 PE (clone M21), anti-CTLA4 biotin (clone BN13), anti-CD45RO fluorescein isothiocyanate (FITC) (clone UCHL1), and anti–HLA-DR PE (clone TU36) mAbs. Intracellular staining for human FOXP3 was performed using the anti FOXP3-Alexa 488 mAb (clone 259D; BioLegend, San Diego, CA) according to the manufacturer's instructions. At least 20,000 events were acquired from each sample on a BD FACScalibur and analyzed with FCS Express V3 software.

Thawed PBMCs were stained with anti-CD4 FITC (clone SK3; BD Pharmingen) and anti-CD25 PE (clone 2A3; BD Pharmingen). CD4⁺CD25^bright (top 1–2%) and CD4⁺CD25⁻ fractions were sorted with a fluorescence-activated cell sorter (FACS; BD FACSvantage) (for gating strategy, see Fig. 2). Sorted CD4⁺CD25⁻ and CD4⁺CD25^bright T-cells had a purity of 96–98% both from type 1 diabetic patients (before and during rapamycin treatment) and from normal donors. Sorting CD4⁺CD127^lowCD25^bright T-cells provided a population of nTregs with suppressive ability similar to those sorted as CD4⁺CD25^bright cells, contrary to what has previously been published (Fig. 1) (6). We therefore concluded that there was no need to sort nTregs based on the expression of CD127.

FACS-sorted CD4⁺CD25⁻ T-cells were plated at 1 × 10⁴ cells per well (in 200 μl X-vivo 15 [Cambrex-Biowhittaker] supplemented with 10% human serum AB [Sigma], 100 units/ml penicillin, and 100 units/ml streptomycin) in 96 round-bottom plates and stimulated with anti CD3/CD28-coupled beads (one bead per six cells) (Invitrogen-Dynal, Oslo, Norway). FACS-sorted CD4⁺CD25^bright T-cells were added at a 1:1 ratio (CD4⁺CD25⁻:CD4⁺CD25^bright). T-cell proliferation was assessed at day 4, after addition of [³H]thymidine for the last 18 h of culture (1 μCi per well) (Amersham, Buckingham, U.K.). Suppressive activity of CD4⁺CD25^bright T-cells was measured as inhibition of cell proliferation of CD4⁺CD25⁻ T-cells compared with proliferation of CD4⁺CD25⁻ T-cells stimulated in the absence of nTregs. Historical data from our laboratory demonstrate that repeated measures of suppression using normal donor cells is relatively consistent (data not shown). In one patient, the suppression assay was performed using the CFSE dilution assay. Briefly, 1 × 10⁵ FACS-sorted CD4⁺CD25⁻ T-cells were stained with CFSE (Molecular Probes, Eugene, OR) as described previously (18). Following the same protocol as for CFSE staining, 1 × 10⁵ FACS-sorted CD4⁺CD25^high T-cells were first stained with SNARF (Molecular Probes) and were subsequently mixed with an equal number of CFSE⁺CD4⁺CD25⁻ T-cells in round-bottom 96-well plates precoated with 10 μg/ml anti-CD3 and 1 μg/ml soluble anti-CD28 mAbs (BD Biosciences). Seven days later, the cells were collected and analyzed by FACS. The proportion of CFSE⁺ (FL-1) T-cells proliferating in vitro was calculated by gating on lymphocytes plus alive cells (TOPRO⁻ FL-4) (Molecular Probes) and by excluding SNARF⁺ (FL-2) cells. The number of gated cells (events) in a given cycle (division: n) was divided by 2 raised to power n, to calculate the percentage of original precursor cells from which they arose. The sum of original precursors from division 1–6 represents the number of precursors cells that proliferated. The percentage of CFSE⁺ divided cells was calculated by ([no. of precursors that proliferated_1–6/no. of total precursors_0–6] × 100) (18). The percentage of CD4⁺ T-cells CFSE⁺ cells divided in the presence of CD4⁺CD25^high T-cells was compared with the percentage of CD4⁺ CFSE⁺ divided T-cells in the absence of nTregs.

FACS-sorted CD4⁺CD25⁻ and CD4⁺CD25^bright T-cells were plated at 1 × 10⁴ cells per well (in 200 μl X-vivo 15 supplemented with 10% human serum AB, 100 units/ml penicillin, and 100 units/ml streptomycin) in 96 round-bottom plates and stimulated with anti CD3/CD28-coupled beads (one bead per six cells) (Invitrogen-Dynal). Culture supernatant was collected 3 days after activation and frozen at −80°C. Detection of IL-2, interferon-γ (IFN-γ), tumor necrosis factor-α (TNF-α), IL-4, IL-5, and IL-10 was performed using a cytometric bead array kit (BD Pharmingen) according to the manufacturer's instructions.

Comparisons between patients and normal donors were performed using Student's t test. Comparisons between samples collected before and samples collected during rapamycin treatment were performed using Student's paired t test. In all cases, two-tailed P < 0.05 was considered significant. Analyses were performed using the Prism V4.03 software (GraphPad, San Diego, CA).

To define whether rapamycin monotherapy modifies number and phenotype of circulating nTregs, PBMCs isolated from type 1 diabetic patients before and during rapamycin treatment were thawed and tested for the expression of regulatory-cell markers (6,25,26). The percentage of circulating CD4⁺ T-cells was not altered by rapamycin treatment (Fig. 2A). Similarly, the percentage of CD25^bright T-cells, within the CD4⁺ T-cell subset, did not change during rapamycin monotherapy compared with that before treatment (Fig. 2B). Markers of regulatory T-cells (namely, FOXP3, CD127, CTLA-4, HLA-DR, CD45RO, and CD62L) were also similarly expressed by CD4⁺CD25^bright T-cells isolated before and during rapamycin monotherapy (Fig. 2C).

To determine whether rapamycin therapy modifies the functional features of peripheral CD4⁺CD25⁻ effector T-cells and CD4⁺CD25^bright T-cells (which should be enriched for nTregs), their proliferative capacity, cytokine production profile, and suppressive function were tested in vitro. The two T-cell subsets were first purified from type 1 diabetic patients before and during rapamycin monotherapy and from age-matched normal donors by flow cytometry (Fig. 3A). Purified CD4⁺CD25^bright T-cells were all FOXP3⁺ (data not shown). FACS-sorted T-cells were activated in vitro by T-cell receptor–mediated polyclonal stimulation. Type 1 diabetic CD4⁺CD25⁻ effector T-cells proliferated to a similar extent to those isolated from normal donors, irrespective of rapamycin therapy (Fig. 3B). CD4⁺CD25^bright T-cells from type 1 diabetic subjects were anergic both before and during rapamycin treatment, as were those isolated from normal donors (Fig. 3B), and their proliferative capacity was restored on addition of exogenous IL-2 (data not shown). These data show that rapamycin does not alter the proliferative capacity of effector T-cells isolated from type 1 diabetic patients. In addition, the anergic phenotype of CD4⁺CD25^bright T-cells (i.e., [H³]thymidine incorporation ≤1,000 cpm upon activation) purified from both type 1 diabetic patients (before and during rapamycin treatment) and normal donors suggests that the sorted T-cells comprise mainly nTregs that are not contaminated with activated CD4⁺CD25⁺ effector T-cells with high proliferative capacity.

To ascertain whether rapamycin monotherapy alters the cytokine production profile of CD4⁺CD25⁻ effector T-cells and CD4⁺CD25^bright nTregs, the sorted CD4⁺ T-cell subsets were activated polyclonally, and supernatants were collected for cytokine measurements. The ability of type 1 diabetes CD4⁺CD25⁻ effector T-cells to produce both Th1 (i.e., IL-2, IFN-γ, and TNF-α) and Th2 (i.e., IL-4, IL-5, and IL-10) cytokines was similar before and during rapamycin monotherapy, and the cytokine levels were comparable with those of CD4⁺CD25⁻ T-cells isolated from normal donors (Fig. 4). Sorted CD4⁺CD25^bright T-cells, due to their anergic state, did not produce significant levels of any of the tested cytokines, irrespective of rapamycin treatment (data not shown). Overall, effector CD4⁺CD25⁻ T-cells isolated from type 1 diabetic patients are functionally similar before and during rapamycin monotherapy, in terms of proliferative ability and cytokine production profile, and are not different from those isolated from normal donors. The same holds true for CD4⁺CD25^bright nTregs.

Finally, the ability of sorted CD4⁺CD25^bright nTregs to suppress proliferation of CD4⁺CD25⁻ effector T-cells was tested in vitro. CD4⁺CD25^bright nTregs isolated from type 1 diabetic subjects during rapamycin monotherapy suppressed syngeneic effector T-cells significantly more than those isolated before treatment (Fig. 5A, compare pretreatment samples and samples during rapamycin treatment). The average level of nTreg suppression in type 1 diabetic patients before rapamycin therapy was significantly reduced in comparison with that of normal donors (Fig. 5A, compare pretreatment samples and normal donors). On the contrary, type 1 diabetes nTregs isolated during rapamycin therapy suppressed to the same levels of normal donor nTregs (Fig. 5A, compare patients during rapamycin treatment and normal donors). Type 1 diabetes nTregs isolated during rapamycin treatment had an increased suppressive capacity compared with those isolated before treatment also when normal donor CD4⁺CD25⁻ effector T-cells were used as responder cells (Fig. 5B). These data indicate that the increased suppressive ability observed during rapamycin monotherapy is due to an improved capacity of type 1 diabetes nTregs to suppress proliferation of CD4⁺CD25⁻ effector T-cells rather than to an increased susceptibility of type 1 diabetes CD4⁺CD25⁻ effector T-cells to be suppressed in vitro on in vivo exposure to rapamycin. The observation that CD4⁺CD25^bright nTregs isolated from normal donors suppressed type 1 diabetes CD4⁺CD25⁻ effector T-cells isolated before and during rapamycin treatment to a similar extent (Fig. 5C) further supports the hypothesis that rapamycin monotherapy directly improves nTregs suppressive activity rather than modifying the susceptibility of effector T-cells to be suppressed.

We have shown that rapamycin monotherapy in long-term type 1 diabetic patients does not alter the frequency and total number of circulating CD4⁺CD25^brightFOXP3⁺CD127^low/neg nTregs. Similarly, the ability of rapamycin-exposed nTregs to proliferate and produce cytokines in vitro is unaffected by the therapy. Interestingly, peripheral nTregs isolated from type 1 diabetic patients during rapamycin treatment have an intrinsic improved capacity to suppress proliferation of syngeneic and allogeneic CD4⁺CD25⁻ effector T-cells, compared with that before treatment. Rapamycin therapy, therefore, refits nTreg suppressive activity, but it does not directly change effector T-cells.

Rapamycin administration to pre-diabetic NOD mice leads to expansion of nTregs that accumulate in the pancreatic lymph nodes and block diabetes development (19). Similarly, C57BL/6 mice receiving rapamycin for 14 days show an enhanced ratio between nTregs and CD4⁺ T-cells in the spleen and thymus (27). In humans, the frequency of circulating nTregs in renal transplant recipients is preserved under rapamycin treatment, whereas it is significantly decreased during therapy with calcineurin inhibitors (28). Similarly, we observed that rapamycin monotherapy preserves the number of circulating nTregs. In humans, a large expansion of peripheral nTregs has been observed only in renal transplant recipients receiving rapamycin as maintenance immunosuppressive therapy after profound T-cell depletion with Campath-1H (22). The absence of such expansion in patients receiving the calcineurin inhibitor cyclosporine A (22) suggests that acute lymphopenia and calcineurin-mediated signaling are important prerequisites for peripheral nTreg expansion in humans.

Rapamycin treatment causes G1 arrest in a variety of cell types, including T-cells (29). However, we demonstrated that, in contrast to CD4⁺CD25⁻ effector T-cells, human nTregs are resistant to the antiproliferative effect of rapamycin in vitro (20). This antiproliferative effect is operational as long as rapamycin is present in the culture, because restimulation of rapamycin-exposed CD4⁺CD25⁻ effector T-cells in the absence of rapamcyin leads to normal T-cell proliferation (Fig. 2B; 20).

The existence of CD4⁺CD25⁺ nTregs with defective suppressor function in type 1 diabetic patients has been demonstrated in some studies (9–11), whereas others showed that type 1 diabetes CD4⁺CD25⁺ Tregs are as suppressive as those of normal donors (12). Among the factors that could account for the discrepancies in these studies is the different purity of the tested Tregs. Studies have used FACS-sorted CD4⁺CD25^bright T-cells (12) or bead-purified CD4⁺CD25⁺ T-cells (9–11), and it is possible that a defect in CD4⁺CD25⁺ nTregs was due to confounding effects from contaminating T effector cells when bead-purified T-cells were used. To reduce this risk, we FACS-sorted both CD4⁺CD25^bright T-cells (falling in the top 1–2% of CD25⁺ cells) and CD4⁺CD25⁻ effector T-cells from type 1 diabetic patients and normal donors. Functional data, in terms of cell proliferation and cytokine production profile of both CD4⁺CD25⁻ and CD4⁺CD25^bright T-cells, prove that in the current study, we are comparing similar T-cell subsets isolated from type 1 diabetic subjects before and during rapamycin treatment and from normal donors. Using these highly purified T-cells, we show that the suppressive function of nTregs freshly isolated from type 1 diabetic patients before rapamycin therapy is significantly reduced compared with that of normal donors. Although the number of patients was low, some patients had nTregs that were completely absent in suppressive function as determined by current protocols. Moreover, quantitative suppression by nTregs isolated from normal donors was relatively similar between subjects, whereas it varied markedly between type 1 diabetic patients. Thus, type 1 diabetic subjects appear to be variable in their freshly isolated peripheral nTreg function.

Increased and restored suppressive function of nTregs was observed in type 1 diabetic patients undergoing rapamycin monotherapy. These results are in accordance to our in vitro data, which demonstrated that peripheral CD4⁺ T-cells isolated from type 1 diabetic subjects and expanded in vitro with rapamycin have the same suppressive ability as those isolated from normal donors (20). The improved nTreg activity during rapamycin therapy is unlikely due to differences in numbers of bona fide nTregs (i.e., FOXP3⁺CD127^low/neg) in the type 1 diabetic CD4⁺CD25^bright T-cells isolated during rapamycin treatment compared with that before treatment because they are equally present, irrespective of rapamycin therapy. FOXP3 can be upregulated in T-cells on activation, and this might influence their suppressive function (30,31). We tested whether nTregs isolated during rapamycin treatment differ in their ability to upregulate FOXP3 on in vitro stimulation. Increased FOXP3 expression in RAPA-nTregs on in vitro activation was observed in only one of six patients (data not shown), indicating that this was unlikely to explain the rapamycin-associated increased nTreg suppression. Other explanations include heterogeneity of the Treg subsets in the sorted CD4⁺CD25^bright T-cells before and during rapamycin treatment. It has been shown that rapamycin allows the in vitro generation of inducible human CD4⁺ Tregs that are CD25^bright (21,32), and it is possible that rapamycin treatment leads to an increased representation of this subset in the subsequently purified T-cells. Alternatively, rapamycin monotherapy might modulate some molecule/s crucial for nTreg function. Gene expression profiling of T-cells exposed to rapamycin in vivo is currently ongoing to address this question.

Overall, these data show that rapamycin can refit nTreg function in vivo in long-term type 1 diabetic patients and can therefore influence the consequent islet transplant outcome. However, the current use of anti-CD25 mAb therapy (daclizumab) may abrogate this tolerogenic effect of rapamcyin. We observed that type 1 diabetic patients undergoing islet transplantation and receiving daclizumab have a massive depletion of circulating CD4⁺CD25⁺FOXP3⁺ T-cells even months after treatment (P.M. and M.B., unpublished observation). At present, one of the major challenges in transplantation is therefore to carefully design new combinational therapies so as not to abrogate the observed positive effects of rapamycin on nTreg function.

This work was supported by the Italian Telethon Foundation and Juvenile Diabetes Research Foundation Grants JT-01 and GJT04014.

We thank Alessandra Ferraro (San Raffaele Telethon Institute for Gene Therapy) for performing the suppressive experiments comparing CD4⁺CD25^brightCD127⁻ and CD4⁺CD25^bright nTregs.