Research ArticleGraft Vs Host Disease

Low-Dose Interleukin-2 Therapy Restores Regulatory T Cell Homeostasis in Patients with Chronic Graft-Versus-Host Disease

See allHide authors and affiliations

Science Translational Medicine  03 Apr 2013:
Vol. 5, Issue 179, pp. 179ra43
DOI: 10.1126/scitranslmed.3005265


CD4+Foxp3+ regulatory T cells (Tregs) play a central role in the maintenance of immune tolerance after allogeneic hematopoietic stem cell transplantation. We recently reported that daily administration of low-dose interleukin-2 (IL-2) induces selective expansion of functional Tregs and clinical improvement of chronic graft-versus-host disease (GVHD). To define the mechanisms of action of IL-2 therapy, we examined the immunologic effects of this treatment on homeostasis of CD4+ T cell subsets after transplant. We first demonstrated that chronic GVHD is characterized by constitutive phosphorylation of signal transducer and activator of transcription 5 (Stat5) in conventional CD4+ T cells (Tcons) associated with elevated amounts of IL-7 and IL-15 and relative functional deficiency of IL-2. IL-2 therapy resulted in the selective increase of Stat5 phosphorylation in Tregs and a decrease of phosphorylated Stat5 in Tcons. Over an 8-week period, IL-2 therapy induced a series of changes in Treg homeostasis, including increased proliferation, increased thymic export, and enhanced resistance to apoptosis. Low-dose IL-2 had minimal effects on Tcons. These findings define the mechanisms whereby low-dose IL-2 therapy restores the homeostasis of CD4+ T cell subsets and promotes the reestablishment of immune tolerance.


Allogeneic hematopoietic stem cell transplantation (HSCT) can provide curative therapy for various hematologic malignancies, bone marrow failure syndromes, and congenital immune deficiencies. Improvements in immunosuppressive therapy and supportive care have improved patient outcomes, but chronic graft-versus-host disease (GVHD) continues to be a major problem affecting long-term survivors (13). The clinical and laboratory manifestations of chronic GVHD resemble those of autoimmune diseases such as systemic lupus erythematous, Sjogren’s syndrome, and scleroderma (46). Both T and B cell responses play a role in the pathogenesis of chronic GVHD (79), suggesting that this syndrome reflects a general loss of immune tolerance.

CD4+Foxp3+ regulatory T cells (Tregs) contribute to the maintenance of peripheral tolerance, and patients with chronic GVHD have a relative deficiency of Tregs (1013). This deficiency appears to be a consequence of abnormalities in Treg homeostasis after HSCT wherein increased proliferation of Tregs is not sufficient to compensate for reduced thymic output and increased susceptibility to apoptosis (14). In contrast, homeostasis of conventional CD4 T cells (Tcons) does not appear to be impaired because the reconstitution of Tcons after HSCT is maintained through increased thymic production and reduced susceptibility to apoptosis compared to Tregs.

Cytokines that use the common γ chain play a central role in T cell homeostasis. In response to lymphopenia, interleukin-7 (IL-7) and IL-15 provide critical signals to drive T cell proliferation and survival (1518), and administration of IL-7 or IL-15 has been shown to increase different subsets of circulating T cells (1922). In contrast, IL-2 is a critical homeostatic cytokine for Tregs (2326). In animal models, neutralization of IL-2 results in Treg deficiency and autoimmunity, whereas administration of IL-2 can induce Treg expansion in vivo and prevent autoimmunity (25, 2731). These findings indicate that Tcon and Treg homeostases are regulated by distinct cytokines, and manipulation of the cytokine environment may influence the balance of these subsets in the specific settings of autoimmunity and GVHD.

Although T cell receptor (TCR) activation of effector T cells leads to rapid expression of IL-2 receptors, Tregs constitutively express high levels of high-affinity IL-2 receptors without activation. In recent clinical trials, administration of low-dose IL-2 has been shown to result in the selective expansion of Tregs and clinical improvement in symptoms of auto- and alloimmunity (32, 33). At our center, daily therapy with low-dose IL-2 for 8 weeks in patients with chronic GVHD led to a rapid increase in circulating Tregs, without a significant increase in CD4 Tcons or CD8 T cells (33). This was associated with clinical improvement in about 50% of patients, and no patients experienced GVHD progression. Here, we examined the role of homeostatic cytokines in chronic GVHD and the effects of IL-2 administration on the homeostasis of Tregs and Tcons. Patients with severe chronic GVHD had elevated levels of IL-7 and IL-15 associated with higher levels of phosphorylated signal transducer and activator of transcription 5 (pStat5) in Tcons than Tregs. This imbalance of Stat5 signaling was rapidly reversed in patients receiving low-dose IL-2, resulting in increased thymic generation and proliferation of Tregs and reduced susceptibility to apoptosis. These results demonstrate that daily administration of IL-2 at physiologic doses can restore Treg homeostasis in vivo and promote immune tolerance.


Differential effects of homeostatic cytokines on CD4 T cell subsets in vitro

The signals induced by IL-2, IL-7, and IL-15 interactions with their specific membrane receptors are mediated through the Janus kinase–Stat5 pathway (34). To compare the response of human CD4 T cell subsets to these cytokines, we purified CD4 Tregs and Tcons from healthy donors by cell sorting (Fig. 1A). Tregs isolated in this manner express high levels of Foxp3 (Fig. 1B) and effectively suppress proliferation of activated autologous Tcons (Fig. 1C). Purified Tregs and Tcons were stimulated by each cytokine for 15 min, and intracellular pStat5 was evaluated by flow cytometry. Results in Fig. 1D compare pStat5 levels in Tregs and Tcons after stimulation with IL-2, IL-7, and IL-15 over a 7-log range of cytokine concentrations. At high IL-2 concentrations (100 to 1000 IU/ml), Stat5 was activated in both Tregs and Tcons. However, at low IL-2 concentrations (1 to 10 IU/ml), Stat5 was activated only in Tregs. Different patterns of Stat5 activation were observed with IL-7 and IL-15. At low concentrations, IL-7 induced similar levels of Stat5 activation in Tregs and Tcons. However, at high concentrations, IL-7 preferentially activated Tcons. In contrast, IL-15 induced similar levels of Stat5 activation in Tregs and Tcons at all concentrations. These results indicate that human CD4 T cell subsets show distinct patterns of response to different homeostatic cytokines. The most selective effect on Tregs was noted with low concentrations of IL-2, and this was confirmed in functional assays comparing proliferation of Tregs and Tcons (Fig. 1E) after in vitro stimulation.

Fig. 1

Phenotypic and functional characterization of human Tregs and Tcons after in vitro cytokine activation. (A) Representative lymphocyte gate for identification of CD4 T cell subsets. Within the CD4 T cell gate, Tregs are identified as CD25med-highCD127low and Tcons are identified as CD25neg-lowCD127med-high. (B) Gated CD4 T cell subsets were examined for intracellular Foxp3 expression. Representative data are shown. Closed histogram represents isotype control. Blue and red histograms depict Tcons and Tregs, respectively. (C) Tregs (red) or Tcons (blue) isolated from peripheral blood were cultured with responder Tcons from the same donor and stimulated with irradiated allogeneic PBMCs for 5 days. Method for calculating percentage suppression of proliferation is described in Materials and Methods. Data are representative of five independent experiments. (D) Purified Tregs and Tcons were cultured for 15 min in various concentrations of IL-2, IL-7, and IL-15. The level of intracellular pStat5 was determined by flow cytometry. Cytokine dose-dependent phosphorylation of Stat5 in each CD4 T cell subset is shown. Data are representative of three independent experiments. MFI, mean fluorescence intensity. (E) Purified Tregs and Tcons were cultured in low (10 IU/ml) or high (100 IU/ml) concentrations of IL-2 for 5 days, and cell proliferation was measured by thymidine incorporation. Data are representative of three independent experiments.

Altered cytokine milieu in patients with severe chronic GVHD

To examine the role of homeostatic cytokines in vivo, we studied peripheral blood samples from 45 patients at a median of 3 years after transplantation (cohort 1, Table 1). Cohort 1 included 14 patients without chronic GVHD and 31 patients with chronic GVHD. In Fig. 2, A and B, patients with chronic GVHD have fewer lymphocytes and fewer CD4 T cells than healthy donors and HSCT patients without active GVHD (P < 0.0001 and P = 0.0005, respectively, Wilcoxon rank sum test). Plasma IL-2 levels were not significantly elevated after HSCT and were similar to the low levels present in healthy donors (Fig. 2C). Plasma IL-7 levels were significantly higher only in patients with severe chronic GVHD (P = 0.008, compared with healthy donors, Wilcoxon rank sum test). In contrast, IL-15 levels were higher in all HSCT patients compared with healthy donors, and there was no increase associated with severity of chronic GVHD (P < 0.001 for all comparisons with healthy donors). When the absolute numbers of circulating CD4 T cells were considered, IL-2/CD4, IL-7/CD4, and IL-15/CD4 T cell ratios were significantly elevated in patients with severe chronic GVHD compared to patients without GVHD (P < 0.01, P = 0.0002, and P = 0.0006, respectively, Wilcoxon rank sum test), but this is the most evident for the IL-7/CD4 T cell ratio (Fig. 2D).

Table 1

Patient characteristics. AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CML, chronic myelogenous leukemia; HL, Hodgkin’s lymphoma; ALL, acute lymphoblastic leukemia; MDS, myelodyplastic syndrome; NHL, non-Hodgkin’s lymphoma; PBSCs, peripheral blood stem cells; BM, bone marrow; MMF, mycophenolate mofetil.

View this table:
Fig. 2

Altered cytokine milieu and activation of intracellular Stat5 in chronic GVHD. (A and B) Lymphocyte counts and CD4 T cell counts in peripheral blood from healthy donors and chronic GVHD (cGVHD) patients (cohort 1). Box plots in each figure depict the 75th percentile, and median and 25th percentile values, and whiskers represent maximum and minimum values. NS, not significant. (C) Plasma concentrations of IL-2, IL-7, and IL-15 from healthy donors and HSCT patients. (D) Ratio of plasma cytokine concentration to CD4 T cell number for each patient group. (E) Expression of pStat5 in gated CD4 Tcons (blue) and Tregs (red) from healthy donors and patients with and without chronic GVHD. Representative panels are shown. (F) Expression of pStat5 in Tcon and Treg subsets from 25 healthy donors and 45 HSCT patients. The expression of pStat5 in Tcons was significantly greater than that in Tregs in patients with severe chronic GVHD (P < 0.005, Wilcoxon signed rank test), whereas pStat5 expression in other groups did not show significant differences between Tregs and Tcons. (G) Ratio of Treg-pStat5 MFI to Tcon-pStat5 MFI in healthy donors and transplant patients. Treg-pStat5/Tcon-pStat5 ratio in severe chronic GVHD was significantly lower than in patients without GVHD (P < 0.001, Wilcoxon rank sum test).

Activation of CD4 T cell subsets in vivo was also monitored by measuring pStat5. In healthy donors, levels of pStat5 were equivalent in Tregs and Tcons (Fig. 2, E and F). Similar results were obtained when pStat5 was measured in Tregs and Tcons using an alternate gating strategy based on expression of Foxp3 and CD25 to define these CD4 T cell subsets (fig. S1). As shown for representative patients in Fig. 2E and the entire cohort 1 in Fig. 2F, levels of pStat5 in Tregs and Tcons were not different in HSCT patients with no, mild, or moderate chronic GVHD. In contrast, pStat5 was selectively increased in Tcons in patients with severe chronic GVHD (P < 0.005, Wilcoxon signed rank test). This activation was specific for pStat5 because pStat1, pStat3, pStat4, and pStat6 were not significantly increased after HSCT (table S1). In Fig. 2G, Treg-pStat5/Tcon-pStat5 ratio steadily decreased as GVHD severity increased. We also evaluated whether pStat5 expression in CD4 T cell subsets was related to the time from transplantation and found no correlation between time from transplant and Tcon-pStat5, Treg-pStat5, and Treg-pStat5/Tcon-pStat5 (Spearman’s correlation coefficients are −0.23, −0.30, and −0.04, respectively). Together, these findings suggest a dominant effect of IL-7 and IL-15 on Tcon homeostasis as well as a relative functional deficiency of IL-2 on Treg homeostasis in patients with severe chronic GVHD.

Rapid correction of signaling imbalance between Tregs and Tcons during IL-2 therapy

To evaluate the effects of IL-2 on Stat5 signaling in Tregs and Tcons in vivo, we examined 14 patients with refractory chronic GVHD enrolled in a phase 1 clinical trial described previously (cohort 2, Table 1) (33). These patients received daily low-dose IL-2 for 8 weeks followed by a 4-week rest period without IL-2. One week after IL-2 therapy began, pStat5 significantly increased in Tregs, whereas pStat5 decreased in Tcons. This is shown for a representative patient in Fig. 3A and for all patients examined in Fig. 3B. This resulted in a significant increase in Treg-pStat5/Tcon-pStat5 ratio (P = 0.0002) 1 week after beginning daily low-dose IL-2 (Fig. 3C, Wilcoxon signed rank test). However, the high level of pStat5 in Tregs was not maintained with continued IL-2 treatment. After 2 weeks of daily IL-2 therapy, pStat5 levels were at their lowest levels in both Tregs and Tcons (Fig. 3D). Thereafter, pStat5 levels in both subsets gradually increased, but the ratio of Treg-pStat5/Tcon-pStat5 was maintained in the range observed in patients without GVHD (Fig. 3E). These findings indicate that exogenous low-dose IL-2 corrected the balance of Treg and Tcon activation in patients with chronic GVHD, and this was maintained for the duration of therapy.

Fig. 3

Selective activation of Tregs in patients with chronic GVHD receiving low-dose IL-2. (A) Expression of pStat5 in gated Tcons (blue) and Tregs (red) from patients before and after the start of IL-2 administration. A representative panel from a single patient is shown. (B) MFI of pStat5 in Tregs and Tcons was compared before and 1 week after starting IL-2 therapy. Treg-pStat5 MFI is significantly elevated after IL-2. Tcon-pStat5 was reduced after IL-2, but this change was not statistically significant. Median values are shown in red. (C) Ratio of Treg-pStat5 MFI/Tcon-pStat5 MFI was compared before and 1 week after starting IL-2 therapy. The ratio is significantly increased in all patients examined after IL-2 administration (P = 0.0002, Wilcoxon signed rank test). Median values are shown in red. (D) Changes of pStat5 MFI in Tregs and Tcons during IL-2 therapy. Median values from 13 patients are shown. (E) Ratio of Treg-pStat5 MFI/Tcon-pStat5 MFI during IL-2 therapy. Median values for MFI ratios were measured in 13 patients. Green range depicts the interquartile range of 14 patients without GVHD (Fig. 2E).

Functional effects of IL-2 therapy on Treg homeostasis

We also measured plasma IL-2 levels during therapy and correlated these levels with absolute numbers of circulating Tregs. In Fig. 4A, IL-2 levels rose sharply 1 week after starting treatment, but high levels were not maintained as the number of circulating Tregs increased. In contrast, IL-7 and IL-15 levels decreased after starting IL-2 (Fig. 4B). Changes in expression of membrane receptors for IL-2 and IL-7 were also monitored during IL-2 therapy (fig. S2, A and B). The expression of IL-2 receptor α (CD25) on Tregs increased after starting IL-2 therapy, and the expression was maintained at a high level during the entire treatment period. In contrast, expression of CD25 did not change in Tcons, and IL-7 receptor α (CD127) expression remained stable on both Tregs and Tcons during therapy. Ki-67 expression in each T cell subset was measured as a marker of cell proliferation. Results from a representative patient are shown in Fig. 4C, and for the entire cohort of patients in Fig. 4D. Before IL-2, Tregs showed higher levels of proliferation than Tcons (14, 35). After starting IL-2, Treg proliferation increased rapidly, peaking 1 week after treatment began. Ki-67 also increased in Tcons during this period but at a much lower level (fig. S3, A and B). Although patients continued to receive the same daily dose of IL-2, the initial high level of Treg proliferation rapidly returned to baseline levels as the absolute number of circulating Tregs increased and IL-2 levels decreased.

Fig. 4

Effects of low-dose IL-2 on Tregs and Tcons in vivo. (A) Concentrations of plasma IL-2 and absolute numbers of Tregs during IL-2 therapy. Median values for 14 patients are shown. (B) Concentrations of plasma IL-7 and IL-15 during IL-2 therapy. Median values for 14 patients are shown. (C) Representative flow cytometry histograms for identification of Ki-67+ proliferating cells in Treg and Tcon subsets. Percentage of Ki-67+ cells is shown for each histogram. (D) Percentage of Ki-67+ proliferating cells in Treg and Tcon subsets during IL-2 therapy (median values). *P < 0.001, Tregs versus Tcons, Wilcoxon signed rank test. (E) Recent thymic emigrants (RTEs) (CD45RA+CD31+) within Treg and Tcon subsets during IL-2 therapy. Median fold changes are shown for each subset. *P < 0.005, Tregs versus Tcons, Wilcoxon signed rank test. (F) Levels of intracellular Bcl-2 in Treg and Tcon subsets during IL-2 therapy. Median percent increase in MFI is shown. *P < 0.05, Tregs versus Tcons, Wilcoxon signed rank test.

Because IL-2 also has an essential role in Treg development (36), we examined thymic generation of Tregs during IL-2 therapy. Previous studies identified coexpression of CD45RA and CD31 as a marker of RTEs (37, 38). We confirmed this as a marker of RTE in both Tregs and Tcons by isolating CD45RA+CD31+ cells from each subset and quantified T cell receptor excision circle (TREC) levels by real-time polymerase chain reaction (PCR). Only CD4 T cell subsets that expressed both CD45RA and CD31 contained high levels of TREC (fig. S4, A and B). Whereas IL-2 treatment did not increase levels of Tcon-RTE, low-dose IL-2 significantly increased Treg-RTE (Fig. 4E, Wilcoxon signed rank test). Treg-RTE levels peaked 4 weeks after starting IL-2 and remained elevated 4 weeks after IL-2 was stopped.

IL-2 promotes cell survival by inhibiting the mitochondrial pathway of apoptosis, and antiapoptotic proteins Bcl-2 and Mcl-1 appear to play a dominant role in this effect (39). In Fig. 4F, Bcl-2 expression increased during IL-2 therapy, and this was significantly greater in Tregs compared to Tcons. To confirm that increased Bcl-2 enhanced resistance to apoptosis, we isolated Tregs and Tcons from patient samples and cultured them with or without anti-Fas antibody to induce apoptosis. In the example shown in Fig. 5A, Tregs became more resistant to both spontaneous and Fas-induced apoptosis during IL-2 therapy. In Fig. 5B, Tregs from healthy controls are more susceptible to both spontaneous and Fas-induced apoptosis than Tcons. Before starting IL-2, Tregs and Tcons from HSCT patients both exhibited higher levels of apoptosis compared to healthy donors, but this was much more evident in Tregs than Tcons. After IL-2 therapy, Tregs showed significantly decreased levels of apoptosis in the absence and presence of anti-Fas antibody. Apoptosis in Tcons did not change significantly during IL-2 therapy.

Fig. 5

Increased resistance of Tregs to apoptosis during low-dose IL-2. (A) Spontaneous and Fas-induced apoptosis in Tregs before and after IL-2 therapy. Tregs and Tcons were isolated by cell sorting and cultured separately for 6 hours with control medium or anti-Fas antibody. Apoptosis was assessed by annexin V/7-aminoactinomycin D (7-AAD) costaining. Representative histograms for Tregs are shown. (B) Spontaneous and Fas-induced apoptosis in Tcon and Treg subsets isolated from three healthy control donors (Ctrl) and eight patients with chronic GVHD before (Pre) and after (Post) IL-2 therapy. Median values in each group are shown in red (exact Wilcoxon rank sum test, Wilcoxon signed rank test).

Suppressive function of IL-2–expanded Tregs

To confirm that expanded Tregs maintain suppressive activity, we performed functional studies at various times during IL-2 treatment, including the initial proliferation phase and the later maintenance phase. As shown for a representative patient in Fig. 6A, Tregs suppressed the proliferation of carboxyfluorescein diacetate succinimidyl ester (CFSE)–labeled responder Tcons from the same patient. Tregs also markedly inhibited secretion of interferon-γ (IFN-γ) from activated autologous Tcons (Fig. 6B). Moreover, suppression of Tcon IFN-γ secretion was dose-dependent, and Treg suppression could be detected at low Treg/Tcon ratios (Fig. 6C). These data indicate that IL-2–expanded Tregs continue to maintain effective regulatory activity.

Fig. 6

Efficient suppressive function of IL-2–expanded Tregs in vitro. (A) IL-2–expanded Tregs suppress in vitro proliferation of stimulated Tcons from the same patient. Cells were harvested after 5 days, and the proliferation of responder Tcons was examined by CFSE dilution. Top panel: Unstimulated Tcons maintain high-level CFSE labeling in vitro. Middle panel: Anti–CD3/CD28-stimulated Tcons exhibit extensive CFSE dye dilution. Lower panel: Addition of Tregs at a 1:1 ratio inhibits proliferation of CD3/CD28-stimulated Tcons. A representative result from six independent experiments is shown. (B) Tregs from patients receiving IL-2 were cultured with Tcons from the same patient at a 1:1 ratio in the presence of anti-CD3 and anti-CD28 antibody. Suppressive activity was assessed by measuring the ability of Tregs to reduce secretion of IFN-γ by stimulated Tcons. IFN-γ in the culture supernatants was measured by ELISA. ND, not detected. Result of three independent experiments is shown (P < 0.04, exact Wilcoxon rank sum test). (C) Treg-mediated suppression of Tcon IFN-γ secretion was measured at various Treg/Tcon ratios. Result of three independent experiments is shown. *P < 0.05 versus no Treg setting, exact Wilcoxon rank sum test, Savage test.


CD4 Tregs play a critical role in the maintenance of immune tolerance, and relative Treg deficiency is thought to contribute to the pathogenesis of chronic GVHD. Because Treg and Tcon homeostases are critically dependent on specific cytokines, we hypothesized that an abnormal cytokine milieu contributed to the imbalance between Tregs and Tcons in patients with chronic GVHD. This was supported by our recent clinical trial demonstrating that daily administration of low-dose IL-2 results in the selective expansion of peripheral Tregs and improvement of clinical manifestations of chronic GVHD (33). Here, we directly examined homeostatic cytokine signaling in patients with chronic GVHD and the effects of IL-2 administration on Treg and Tcon homeostasis in vivo. These studies demonstrate that severe chronic GVHD is characterized by relatively high levels of IL-7 and IL-15 with selective activation of the Stat5 signaling pathway in Tcons compared to Tregs (34). Low-dose IL-2 rapidly corrected this imbalance by selectively activating Tregs in vivo. This led to profound changes in Treg homeostasis, including increased proliferation of peripheral Tregs, increased generation of thymic Tregs, and enhanced resistance to apoptosis. Together, low-dose IL-2 restored the homeostatic balance of Tregs and Tcons and promoted the reestablishment of immune tolerance.

To examine the differential role of IL-2, IL-7, and IL-15 in the regulation of human CD4 T cells, we monitored pStat5 in individual T cell subsets. This comparison was possible because the specific receptors for each of these cytokines use the common γ chain, and phosphorylation of Stat5 is a shared signaling pathway that mediates their cellular effects. Although all three cytokines induce rapid pStat5 in vitro, CD4 Treg and Tcon subsets respond very differently to each homeostatic cytokine. These differences are most evident at low physiologic concentrations, and, in particular, low doses of IL-2 had the most selective effect on Tregs. We subsequently measured plasma cytokine levels and intracellular pStat5 in CD4 T cell subsets in patients with chronic GVHD to evaluate the response of each subset to their endogenous cytokine milieu. In healthy donors and patients without GVHD, systemic levels of these homeostatic cytokines were similar, and neither Tregs nor Tcons were found to have high levels of constitutive pStat5. In contrast, the cytokine environment in patients with severe GVHD was characterized by elevated levels of IL-7 and IL-15, which primarily support Tcon homeostasis, and a relative functional deficiency of IL-2. This was confirmed by demonstrating increased levels of pStat5 in Tcons in these patients, and the ratio of Treg-pStat5/Tcon-pStat5 was inversely correlated with the severity of chronic GVHD. Because all patients with severe chronic GVHD also received immunosuppressive agents, it is possible that these therapies also influenced the constitutive activation on Stat5 in these patients. Nevertheless, these findings suggest that the altered endogenous cytokine environment contributes directly to the abnormal balance of Tregs and Tcons in chronic GVHD.

There are several possible causes for the altered cytokine milieu after HSCT. Patients often experience profound and prolonged periods of lymphopenia after HSCT, which stimulates production of IL-7 and IL-15 by stromal cells to promote lymphocyte recovery (15, 16, 18). Inflammation and tissue damage associated with active GVHD also promote production of IL-7 (1517, 40). Immunosuppressive agents also contribute to the altered cytokine environment after transplant. In contrast to IL-7 and IL-15, IL-2 is only produced by activated T cells. Calcineurin inhibitors selectively inhibit IL-2 production by activated T cells, and all patients in our study received immunosuppressive agents for GVHD prophylaxis after HSCT. This effect is amplified in patients who develop GVHD and continue to receive intensive immunosuppressive therapy for prolonged periods.

Although CD4 Tregs do not secrete IL-2, these cells constitutively express high levels of high-affinity receptors for IL-2 and are therefore primed to respond to low concentrations of exogenous cytokine (41, 42). CD4 Tcons constitutively express intermediate-affinity receptors (IL-2Rβ and IL-2Rγ) and therefore require higher concentrations of IL-2 for activation in the absence of TCR engagement (4345). The predicted selectivity of low concentrations of IL-2 on CD4 T cell subsets was confirmed in patients with chronic GVHD receiving IL-2. In these patients, low-dose IL-2 was sufficient to induce rapid and selective up-regulation of pStat5 in Tregs as well as the down-regulation of pStat5 in Tcons. After the balance of pStat5 activation in Tregs and Tcons was normalized, this balance was maintained for the duration of IL-2 therapy. Although there was no statistical correlation between pStat5 expression and clinical response to IL-2 therapy in this small number of patients, changes in levels of pStat5 in Tregs and Tcons may be useful as a functional biomarker to assess the effect of IL-2 therapy in future studies. Although the activation of pStat5 in Tregs is presumed to be a direct effect of IL-2, the mechanism responsible for suppression of pStat5 in Tcons is likely to be more complex. IL-2–activated Tregs may directly suppress activation of Tcons. Alternatively, lower levels of pStat5 in Tcons may reflect reduced levels of IL-7 and IL-15. Because expanded Tregs ameliorate inflammation in damaged tissues, this may result in the reduced production of IL-7. Resolution of lymphopenia after IL-2 therapy may also lead to the lower levels of IL-7 and IL-15.

Selective activation of Tregs by administration of low-dose IL-2 had profound effects on Treg homeostasis in vivo. These effects varied with the duration of IL-2 therapy. The initial effect was primarily on Treg proliferation, which peaked 1 week after starting IL-2. Treg proliferation appeared to be related to systemic levels of IL-2, which also peaked at this time and subsequently declined as the number of circulating Tregs rapidly increased. The high-affinity receptor expressed by Tregs rapidly binds exogenous IL-2, and increasing numbers of Tregs likely limit the levels of free ligand in vivo. Moreover, exogenous IL-2 has a short half-life in vivo that may limit the ability to maintain high systemic cytokine levels during continued treatment. Because systemic concentrations of IL-2 decreased and Treg proliferation returned to baseline levels, the number of peripheral Tregs stabilized and gradually decreased despite continued administration of IL-2. Because we were only able to monitor circulating levels of Tregs, we do not know whether tissue Tregs also declined or whether the decrease in circulating Tregs might actually reflect increased trafficking of Tregs to sites of inflammation.

IL-2–induced Treg proliferation was followed by a six- to sevenfold increase in thymic-derived Tregs. Thymic generation of Tregs peaked 4 to 6 weeks after starting IL-2, and this effect also waned by week 8. The reasons for this transient effect are unknown. Thymic Treg development may also be IL-2 concentration–dependent, and thus, this may simply reflect changes in systemic levels of IL-2 during prolonged therapy. It is also possible that the inability to sustain high levels of Treg thymic maturation may reflect intrinsic factors related to thymic function in adults. These might include limited numbers of T cell progenitors, limited functional capacity of involuted adult thymus, effects of other immunosuppressive therapies on thymic function, and autoregulatory pathways within the thymus that may limit the generation of excessive numbers of Tregs.

In patients with chronic GVHD, circulating Tregs have lower levels of Bcl-2 and are more susceptible to apoptosis compared to Tcons (14). Low-dose IL-2 induced increased Bcl-2 expression in Tregs, which peaked 8 weeks after starting therapy. Functional studies confirmed that this resulted in increased resistance to apoptosis in vitro. Notably, all of these cellular effects were highly selective for Tregs, and in vitro experiments confirmed the functional suppressive function of Tregs that had been expanded in vivo during IL-2 therapy. Although low-dose IL-2 had slight effects on Tcon homeostasis, the numbers of Tcons did not change during therapy, and importantly, this did not result in chronic GVHD progression in any patient.

Although the immunologic effects of low-dose IL-2 therapy were very consistent, our study was limited to the analysis of T cells in serial blood samples. We were not able to examine tissues affected by chronic GVHD, and further studies are needed to define the effects of IL-2 on chronic alloreactivity at sites of clinical disease. Indeed, expansion of Tregs in peripheral blood and modification of pStat5 levels occurred in all patients, but only 50% demonstrated clinical improvement. Further modifications of this approach will be needed to enhance the clinical efficacy of low-dose IL-2 therapy. These might include earlier intervention in the course of chronic GVHD before extensive tissue damage and fibrosis have developed. Adoptive therapy with highly purified donor Tregs can also be combined with low-dose IL-2 therapy. In this setting, low-dose IL-2 may be used to selectively expand infused Tregs in vivo, and this may reduce the need to expand large numbers of Tregs in vitro for adoptive therapy. Low-dose IL-2 may also be more effective if combined with immunosuppressive agents such as sirolimus (rapamycin) that would selectively suppress effector T cell function while sparing Tregs. It is also evident that most of the immunologic effects of IL-2 therapy only persisted as long as IL-2 treatment was continued. It is likely that long-term IL-2 treatment will be needed to maintain peripheral tolerance, and additional toxicities may be identified as further studies extend the duration of therapy.

The present study focused on allogeneic stem cell transplantation as a model for studying human CD4 Treg homeostasis. Autoimmune diseases have also been associated with elevated levels of IL-7 or IL-15 without increased IL-2, and improvement of disease manifestations has been reported after blockade of these cytokines (46, 47). These observations suggest that homeostatic imbalance of CD4 T subsets may also be involved in the pathogenesis of autoimmune diseases. Considering the restoration of normal Treg homeostasis during IL-2 therapy in patients with chronic GVHD, low-dose IL-2 may provide a new strategy for restoration of immune tolerance in other clinical settings characterized by Treg deficiency.

Materials and Methods

Patient characteristics

Laboratory studies were undertaken in 59 adult patients who underwent allogeneic HSCT at the Dana-Farber Cancer Institute and the Brigham and Women’s Hospital, Boston. All patients were enrolled in clinical research protocols approved by the Human Subjects Protection Committee of the Dana-Farber/Harvard Cancer Center. Written informed consent was obtained from each patient before sample collection, in accordance with the Declaration of Helsinki. Clinical characteristics of these patients are summarized in Table 1. All patients received peripheral blood stem cells (PBSCs) with standard immunosuppressive regimens for GVHD prophylaxis. No patients received T cell–depleted stem cell products or prophylactic therapy with anti-thymocyte immunoglobulin. Chronic GVHD status was categorized according to documented clinical exam and laboratory studies with National Institutes of Health (NIH) chronic GVHD consensus criteria.

Patient cohort 1. Plasma cytokines and pStat5 expression in CD4 T cell subsets were tested in 45 patients. Blood samples were collected at relatively late times after transplantation (median, 3 years). Chronic GVHD status, shown in Table 1, was determined at the time of sample collection. We also studied 25 age-matched healthy adults.

Patient cohort 2. Detailed analysis of CD4 T cell homeostasis was undertaken in 14 patients with chronic GVHD during low-dose IL-2 therapy. Recombinant IL-2 was administrated subcutaneously once daily, and all patients completed 8 weeks of treatment. Seven patients received IL-2 (3.0 × 105 IU/m2 per day), and seven received IL-2 (1.0 × 106 to 1.5 × 106 IU/m2 per day). Blood samples were obtained before and at 1, 2, 4, 6, 8, and 12 weeks after starting IL-2.

Flow cytometry

Peripheral blood mononuclear cells (PBMCs) were isolated from blood samples by density gradient centrifugation (Ficoll-Hypaque; GE Healthcare) and cryopreserved in 10% dimethyl sulfoxide before being analyzed. After thawing, PBMCs were first incubated with the following directly conjugated monoclonal antibodies for 20 min at 4°C: anti–CD4–Pacific Blue (clone RPA-T4; BD Biosciences), anti–CD25-PC7 (clone M-A251; BD Biosciences), anti–CD45RA-FITC (fluorescein isothiocyanate) (clone M-A251; Beckman Coulter), anti–CD31-APC (allophycocyanin) (clone WM59; eBioscience), and anti–CD127–APC–Alexa Fluor 750 (clone eBioRDR5; eBioscience). To detect intracellular Ki-67 and Bcl-2, we processed surface-stained PBMCs using Cytofix/Cytoperm buffer (BD Biosciences) and incubated them with phycoerythrin (PE)–conjugated anti–Ki-67 antibody (clone B56; BD Biosciences) or PE-conjugated anti–Bcl-2 antibody (clone Bcl-2/100; BD Biosciences) for 30 min at room temperature. To detect intracellular pStat proteins and Foxp3, we first incubated PBMCs with anti–CD4–Pacific Blue, anti–CD25-PC7, anti–CD45RA-FITC, and anti–CD127–APC–Alexa Fluor 750. Surface-stained PBMCs were processed with Fixation buffer and Fix/Perm buffer (BD Biosciences) or Fixation buffer and Permeabilization buffer (eBioscience), and incubated with the following APC-conjugated antibodies from BD Biosciences: pStat1 (clone pY701), pStat3 (clone pS727), pStat4 (clone pY693), pStat5 (clone pY694), and pStat6 (clone pY641) or Foxp3 (clone PCH101; eBioscience). CD4 Tcons are defined as CD4+CD25neg-lowCD127med-high and CD4 Tregs are defined as CD4+CD25med-highCD127low. Expression of Foxp3, pStat proteins, CD45RA, CD31, Ki-67, and Bcl-2 was determined for each T cell subset. Cell debris and doublets were excluded on the basis of side versus forward scatter. All cells were analyzed on a FACSCanto II (BD Biosciences) with FACSDiva (BD Biosciences) and FlowJo software (Tree Star).

Cell sorting

For in vitro functional assays, specific cell populations were isolated by cell sorting with FACSAria (BD Biosciences). CD4 Tregs and Tcons were sorted by the expression pattern of CD25 and CD127 (Fig. 1A). Sorted cell populations were confirmed to be more than 95% pure.

TREC analysis

Measurement of signal-joint TCR excision circle (sjTREC) DNA was performed according to a previously described protocol (11). Briefly, DNA was isolated from each Treg and Tcon subset with the QIAamp DNA Blood Mini Kit (Qiagen). Measurement of sjTREC DNA was performed by TaqMan real-time PCR with ABI 7300/7500 Real-Time PCR (Applied Biosystems).

In vitro suppression assays

In vitro suppression was evaluated with CFSE dilution to measure Tcon proliferation or enzyme-linked immunosorbent assay (ELISA) to measure IFN-γ secretion. In CFSE dilution assay, freshly sorted Tcons were labeled with CFSE (Invitrogen) according to the manufacturer’s directions. Briefly, cells were incubated at 37°C for 10 min with 5 μM CFSE. Staining was stopped by adding RPMI 1640 containing 10% fetal bovine serum at 4°C, followed by one wash in phosphate-buffered saline. CFSE-labeled responder Tcons were cultured with Tregs sorted from the same donor in the presence of anti-CD3 antibody (0.1 μg/ml) (clone OKT3; eBioscience) and anti-CD28 antibody (1 μg/ml) (clone L293; BD Biosciences) in 96-well round-bottom plates at a 1:1 ratio (1 × 104 each T cell subset per well). After 5 days, cells were harvested and incubated with anti–CD4–Pacific Blue (clone RPA-T4; BD Biosciences). Cell division analysis of CD4 Tcons was performed on a FACSCanto II (BD Biosciences). IFN-γ concentration in culture supernatant in each well was also measured by ELISA according to the manufacturer’s instructions (Pierce Biotechnology).

Fas-induced apoptosis assay

Tregs and Tcons were purified from thawed PBMCs by cell sorting and cultured separately with purified mouse anti-human CD95 antibody (5 μg/ml) (clone EOS9.1; BD Biosciences) or with control medium in 96-well round-bottom plates at a concentration of 1 × 104 T cells per well. Apoptosis induction was measured 6 hours after addition of anti-CD95. Cell death was assessed by annexin V/7-AAD costaining and forward to side scatter profiles.

Plasma cytokine measurement

Plasma IL-2, IL-7, and IL-15 concentrations were measured by ELISA according to the manufacturer’s instructions (Pierce Biotechnology and R&D Systems). Samples were obtained from patients after HSCT and cryopreserved in aliquots before being analyzed. We also studied plasma samples from 18 healthy adults.

Statistical analysis

Descriptive statistics was used for patient and transplant-related characteristics. Fisher’s exact test or a χ2 test was used for group comparisons for categorical variables in Table 1. The Wilcoxon rank sum test or the Kruskal-Wallis test was performed for two or more group comparisons for continuous variables. Nonparametric one-way analysis of variance (Savage test) was performed for analysis of functional suppression of Tcon IFN-γ secretion. The Wilcoxon signed rank test was used for difference of paired samples for continuous variables. Spearman’s rank test was used to describe the correlation between time from transplant and pStat5 expression. All tests were two-sided at the significance level of 0.05, and multiple comparisons were not adjusted.

Supplementary Materials

Fig. S1. Equivalent expression of pStat5 in CD4 T cell subsets from healthy donors.

Fig. S2. Effects of low-dose IL-2 therapy on the expression of cytokine receptors (CD25 and CD127) in T cell subsets.

Fig. S3. Effects of low-dose IL-2 on proliferation of Tcon subsets.

Fig. S4. Confirmation of CD45RA+CD31+ subsets of Tregs and Tcons as RTEs.

Table S1. Expression of pStat in T cell subsets in healthy donors and patients after HSCT.

References and Notes

  1. Acknowledgments: We thank J. Daley, S. Lazo-Kallanian, and R. Smith for excellent assistance with flow cytometric studies and D. Hearsey, G. Murga, and B. Selland for help obtaining clinical samples. Funding: This work was supported by NIH grants AI29530 and CA142106, the Jock and Bunny Adams Research and Education Endowment, and the Ted and Eileen Pasquarello Research Fund. J.K. is supported in part by a Dana-Farber Dunkin’ Donuts Rising Star award and an American Society of Blood and Marrow Transplantation/Pharmion New Investigator award. Author contributions: K.M. designed and performed the experiments and wrote the paper. J.K. designed and supervised the clinical trial and clinical data collection and edited the paper. H.T.K. designed the clinical trial, performed statistical analysis, and edited the paper. G.B., S.M., K.M., and Y.K. performed the experiments and edited the paper. C.C., V.T.H., E.P.A., P.A., J.H.A., and R.J.S. designed and carried out the clinical trial, analyzed the data, and edited the paper. B.R.B. reviewed the data and edited the paper. J.R. designed the clinical trial, supervised the laboratory studies, and edited the paper. Competing interests: The authors declared that they have no competing interests.
View Abstract

Stay Connected to Science Translational Medicine

Navigate This Article