Regulatory T lymphocytes (Tregs) are regarded as key immunomodulators in immune-mediated disorders. Our data validated a protective role of Tregs in myocardial ischemia reperfusion injury (MIRI). Moreover, activated Tregs ameliorated MIRI via a CD39-dependent mechanism, representing a putative therapeutic strategy.
Immune system and inflammatory processes participate in the pathogenesis of MIRI. Tregs are generally regarded as key immunomodulators that maintain immune tolerance and counteract tissue damage in a variety of immune-mediated disorders. However, the role of Tregs in MIRI remains to be elucidated.
In the present study, we discovered that selective depletion of Tregs in the DEREG mouse model resulted in aggravated MIRI. In contrast adoptive transfer of in vitro activated Tregs could potentiate MIRI and this effect is associated with attenuated cardiomyocyte apoptosis via activation of the Akt and ERK1/2 pathways and reduced neutrophil infiltration via down-regulation of KC and LIX production through a CD39-dependent mechanism.
These findings provide an insight into the protective role of Tregs in MIRI. Moreover, in vitro-activated Tregs may represent a putative therapeutic strategy for MIRI in patients undergoing elective PCI.
Early reperfusion therapy is currently the most effective therapy to limit infarct size and to preserve cardiac function in patients with acute myocardial infarction (AMI). The restoration of blood flow however triggers a series of events called myocardial ischaemia/reperfusion injury (MIRI) that eventually culminates in additional cell death and increases infarct size . The pathogenesis of MIRI is multifactorial, with inflammation being a common feature. There are considerable data indicating the important roles of various elements of immune response in MIRI .
T-cells that mediate adaptive immune responses were long thought to be only spectators in MIRI; however, recent studies have contradicted this opinion . In the animal model of MIRI, Yang et al.  found that T-cells infiltrated the myocardium as early as 2 min after MIRI and antibody (Ab)-mediated depletion of cluster of differentiation (CD)4+ T-cells reduced the myocardial injury, implying that CD4+ T-cells were harmful in MIRI. CD4+ T-cells are a heterogeneous cell population that can be further subdivided into T-helper (Th) Th1, Th2 and Th17 and regulatory T-cells (Tregs) based on their surface markers and function. Different subsets of T-cells participate in the inflammatory process through distinct mechanisms. We and others have previously reported that interferon (IFN)-γ secreted by Th1 cells and interleukin (IL)-17A secreted by γδT-cells and Th17 cells promote the progression of MIRI [3,4].
Tregs are a subset of T-cells with an immunomodulatory function, commonly identified by their expression of CD4 and CD25 on the cell surface and the transcription factor Forkhead box P3 (Foxp3) in the nucleus . They are capable not only of modulating the adaptive response by preventing effector T-cell proliferation and function but also of suppressing the innate immune response by inhibiting the macrophage inflammatory phenotype and neutrophil function [6,7]. Tregs convey suppressive action through cell-to-cell contact-mediated inhibition of immune cells and/or secretion of soluble factors such as IL-10 and transforming growth factor (TGF)-β1 . Moreover, adenosine is a recently identified mediator of the Tregs' suppressive function. Tregs highly express CD39 and CD73, cell-surface enzymes that could sequentially degrade extracellular ATP and ADP to AMP and AMP to adenosine respectively . Previous publications have reported the involvement of Tregs in kidney, brain and liver ischaemia/reperfusion (I/R) injury (IRI) [10–13]. We have also demonstrated that Tregs modulate inflammatory responses, attenuate ventricular remodelling and subsequently improve cardiac function after myocardial infarction , but their functional role in acute tissue injury in MIRI has not been investigated thus far.
In the present study, we demonstrated that depletion of Tregs using ‘depletion of regulatory T-cell’ (DEREG) mouse model exacerbated murine MIRI, whereas adoptive transfer of in vitro-activated Tregs, which have a more suppressive phenotype than freshly isolated Tregs, attenuated murine MIRI. We also determined that this protection is CD39/adenosine-dependent and associated with suppression of cardiomyocyte apoptosis, up-regulation of reperfusion injury salvage kinase (RISK) and inhibition of neutrophil infiltration.
MATERIALS AND METHODS
Male C57BL/6 mice aged 8–10 weeks were purchased from Wuhan University. Foxp3–GFP knockin mice with a C57BL/6 background were kindly provided by Professor Vijay Kuchroo (Brigham and Women's Hospital, Harvard University) . The DEREG bacterial artificial chromsomes (BAC) transgenic mice were obtained from Jackson Laboratory. IL-10−/− mice with a C57BL/6 background (obtained from the Model Animal Research Center of Nanjing University) were hybridized with Foxp3–GFP knockin mice to generate IL-10−/− Foxp3–GFP knockin mice in our laboratory. CD39−/− Foxp3–GFP knockin mice with a C57BL/6 background were kindly provided by Professor Simon Robson (Beth Israel Deaconess Medical Center, Harvard University) . The mice were maintained on a chow diet in a 12-h light/12-h dark environment at 25°C in the Tongji Medical School Animal Care Facility, according to institutional guidelines. All animal studies were approved by the Animal Care and Utilization Committee of Huazhong University of Science and Technology. Anaesthetic procedures were used in full to ensure that the animals did not suffer unduly during and after the experimental procedure.
In vivo MIRI models
Surgical induction of myocardial I/R was performed as previously described [4,17]. Briefly, mice were anaesthetized with ketamine (50 mg/kg) and pentobarbital sodium (50 mg/kg), orally intubated and connected to a rodent ventilator. A left thoracotomy was performed. The left anterior descending (LAD) coronary artery was visualized and ligated using an 8-0 silk suture around fine PE-10 tubing with slip knot. Mice were subjected to 30 min of LAD ischaemia followed by varying periods of reperfusion. Sham-operated animals were subjected to the same surgical procedures except that the suture was passed under the LAD but was not tied.
Data are presented as the means ± S.E.M. or percentages in the figures and tables. Differences were evaluated using an unpaired Student's t test between two groups and one-way ANOVA for multiple comparisons. For the ranked data, Fisher's exact test was performed for the comparison between groups. All analyses were performed using SPSS 13.0 and statistical significance was set at P<0.05. More detailed description of the methods and materials is available in the Supplementary Material Online Data.
Myocardial infiltration of Tregs after MIRI
To investigate the kinetics of Tregs following MIRI, the population of Tregs in the heart and spleen tissue were analysed by flow cytometry. The percentage of CD4+CD25+Foxp3+ Tregs among CD4+ T-cells within the myocardium had a tendency to increase as early as 5 min following MIRI, but there was no significant difference at any time point, even 24 h, compared with sham-operated mice (Figures 1A and 1B). However, the absolute number of Tregs in the myocardium showed a significant increase, rising after 5 min, peaking at 6 h and remaining elevated at 24 h (Figure 1C). This indicated that Tregs were rapidly mobilized to the heart and then accumulated following reperfusion, which therefore implicated a possible role for Tregs in MIRI. Moreover, the frequency of Tregs in the spleen was similar between the MIRI group and the sham group at different time points post reperfusion (Figures 1A and 1B), suggesting that there was no overall increase in Tregs post reperfusion in mice.
The recruitment of Tregs to heart tissue following MIRI
Selective depletion of Treg in DEREG mice leads to deteriorated MIRI
To test whether endogenous Tregs play a role in MIRI, we first used an anti-mouse CD25 monoclonal antibody (mAb) (PC61), which could partially deplete Tregs. Consistent with previous data , the injection of PC61 resulted in depletion of half of the Foxp3+ cells and a more than 90% reduction in CD25+ cells in the blood and spleen (Supplementary Figure S1). However, despite of the successful reduction in CD4+Foxp3+ Tregs, the infarct size of the heart, the serum level of cTnT (cardiac troponin T) and the cardiac function evaluated by echocardiography were not significantly increased in PC61-treated mice compared with the isotype group (Supplementary Figure S2). Of note, although partially depleted of Tregs, anti-CD25 mAb did not result in exacerbated MIRI.
Compared with anti-CD25 mAb, DEREG mouse model allows specifically and totally depletion of Foxp3+ Tregs . Thus, we employed this more precise approach to study Treg function in MIRI. After diphtheria toxin (DT) administration, Foxp3-expressing Tregs were effectively depleted as previously reported . Heart infarct size and the serum level of cTnT at 24 h after MIRI were significantly increased in Treg-depleted DEREG mice as compared with DT-treated wild-type (WT) littermates or PBS-treated DEREG mice (Figures 2A and 2B). Furthermore, DT treatment in DEREG mice caused impaired cardiac function as revealed by ejection fraction (EF) and fractional shortening (FS; Figures 2C and 2D). Therefore, our data suggested that selective removal of total Foxp3+ Tregs in DEREG mice resulted in a poorer outcome after MIRI.
Depletion of Tregs in DEREG mice aggravates MIRI
Adoptive transfer of
in vitro-activated but not freshly isolated Tregs protects against MIRI
To further substantiate the role of Tregs in MIRI, we applied another widely used strategy: the adoptive transfer of Tregs. 1×106 CD4+GFP+ Tregs with purity above 95% (Figure 3A) were transferred into recipient mice prior to reperfusion. Although the successful engraftment of donor cells was confirmed by flow cytometry of recipient hearts (Figure 4E), mice were not protected from MIRI in terms of infarct size, cTnT and cardiac function detected by echocardiography (Figures 3B and 3D).
In vitro-activated Tregs attenuate MIRI, whereas freshly isolated Tregs do not
In vitro-activated Tregs produce more suppressive mediators and have a more powerful migration ability than do freshly isolated Tregs
Because previous reports have shown that both polyclonal and antigen specific stimulation of Tregs resulted in increased suppressive activity in vitro and in vivo [20–23], we planned to repeat experiments using in vitro-activated Tregs. Before that, we performed a set of experiments to investigate the enhanced suppressive activity of Tregs after in vitro activation in our system. It is well established that anti-inflammatory cytokines IL-10 and TGF-β1 and extracellular adenosine are considered functional mediators of Tregs  and accordingly we focused on these molecules in the following experiments. FACS-sorted CD4+GFP+ Tregs were activated with IL-2, anti-CD3 and anti-CD28 for 3 days. As shown in Figure 4(A), the mRNA levels of IL-10 and TGF-β1 of Tregs were significantly up-regulated by activation. Moreover, in vitro activation led to a significantly enhanced production of adenosine, whereas freshly isolated Tregs showed a mild capacity to degrade ATP to adenosine (Figure 4B).
In addition to enhanced suppressive mediators, increased expression of chemokine receptors and greater efficiency in the migration of Tregs after activation are also reported [24,25]. Therefore, we sought to corroborate the increased expression of chemokine receptors on Tregs after activation in our system. C-C chemokine receptor type 5 (CCR5) signalling is shown to mediate Treg recruitment in mouse MIRI . As anticipated, an elevation of CCR5 measured by real-time PCR (Figure 4C) was mediated by activation on Tregs. Moreover, we determined the CCR5 expression by flow cytometry and confirmed its up-regulation by activation (Figure 4D). Importantly, the adoptive transfer of in vitro-activated Tregs resulted in more Tregs migrating into the heart tissue than that of freshly isolated Tregs as detected by flow cytometry (Figure 4E). To directly monitor the migration of transferred Tregs, Foxp3–GFP+ Tregs in heart were further verified by fluorescence microscopy (Figure 4F).
After the validity of the activation was confirmed, 1×106 FACS-sorted Tregs that were pre-activated were adoptively transferred before reperfusion. As expected, mice treated with in vitro-activated Tregs exhibited attenuated MIRI. The infarct size was significantly smaller in mice that received activated Treg (Figure 3C). This was accompanied by lower serum cTnT (Figure 3B). In parallel, we observed improved cardiac function, as indicated by elevated EF and FS, in mice treated with activated Tregs after MIRI (Figure 3D). Taken together, with more suppressive mediator production and powerful migration ability, in vitro-activated Tregs improved the outcomes after MIRI, which corroborated the protective role of Tregs and the implicated therapeutic potential of Tregs.
Tregs protect against MIRI through a CD39-dependent mechanism that does not involve IL-10 or TGF-β1
As mentioned above, IL-10 and TGF-β1 are key mediators of Tregs in various experimental systems . Moreover, they both showed beneficial effects in previous MIRI [27,28]. To examine the potential roles of these molecules in our model, we compared the severity of injury between the adoptive transfer of IL-10−/− Tregs and WT Tregs and the adoptive transfer of WT Tregs in the presence of anti-TGF-β1 Ab or isotype Ab (all transferred Tregs were in vitro-activated). Surprisingly, IL-10−/− Tregs still had a protective effect on MIRI, although somewhat milder than that of WT Tregs (Supplementary Figure S3). Similarly, whereas WT Tregs were transferred in the presence of anti-TGF-β1 Ab or isotype Ab, they showed comparable protection from MIRI (Supplementary Figure S3). Taken together, the data from IL-10 knockout and TGF-β1 inhibition supported the notion that neither IL-10 nor TGF-β1 was a critical mediator of the Tregs' cardioprotective effect in MIRI.
Because IL-10 and TGF-β1 did not account for the injury-reducing effect of Tregs, other mechanism must be operative. It has been reported recently that adenosine is a significant mediator of Tregs function . Tregs highly express CD39 and CD73, cell-surface enzymes that generate adenosine from ATP . Because the importance of adenosine for Tregs function has been established in the context of inflammation in several disease models, we hypothesize that adenosine signalling may contribute to the protective effect of Tregs on MIRI. To test this hypothesis, we transferred CD39−/− Tregs, which have impaired adenosine production, as previously described , to recipient mice. Unlike WT Tregs, CD39−/− Tregs failed to protect the heart from MIRI. Specifically, CD39−/− Tregs were unable to reduce infarct size and cTnT (Figures 5A and 5B). Consistently, the cardiac function detected by echocardiography did not exhibit improvement in the CD39−/− Treg group when compared with the control group (Figures 5C and 5D). The above results indicated that the CD39 pathway was essential for mediating the suppressive effects of Tregs on MIRI.
Tregs protect against MIRI through a CD39-dependent mechanism
Tregs attenuated cardiomyocyte apoptosis and activated a pro-survival pathway involving Akt and ERK1/2
Cardiomyocyte apoptosis is a well-known participant in the pathophysiology of MIRI  and our previous data have shown that Tregs prevent cardiomyocyte apoptosis in rat myocardial infarction . Concerning the CD39-dependent protection of Tregs, we questioned whether this protection was achieved via regulating cardiomyocyte apoptosis. In vivo data demonstrated that WT Tregs transfer reduced apoptosis of cardiomyocytes, as revealed by terminal deoxynucleotidyltransferase mediated dUTP nick-end labeling (TUNEL) staining and caspase-3 activity, whereas this reduction was abolished by a CD39 deficiency, indicating the impaired function of CD39−/− Tregs (Figures 6A and 6B). Moreover, this effect was also apparent in vitro. Primary cultures of neonatal mice cardiomyocytes were incubated with pre-activated WT Tregs or CD39−/− Tregs in a co-culture system with stimulation of H2O2 for 4 h to analyse apoptosis. Pre-incubation with WT Tregs significantly decreased cardiomyocyte apoptosis, but CD39 deficiency partially blocked this anti-apoptosis effect of Tregs (Figure 6C). Moreover, pre-incubation with adenosine could also reduce cardiomyocyte apoptosis.
Tregs inhibit cardiomyocyte apoptosis and increase the phosphorylation of Akt and ERK1/2
The RISK pathway refers to a group of anti-apoptotic pro-survival kinase signalling cascades, including Akt and extracellular-signal-regulated kinase (ERK). Activating these pro-survival kinases at the time of reperfusion has been demonstrated to potentially attenuate reperfusion-induced cell death, thus conferring powerful cardioprotection . Next, we investigated the RISK relevance of the Tregs' protective effect. The phosphorylated levels and total amount of ERK1/2 and Akt in ischaemic heart tissue at different time points after reperfusion of the myocardium were determined by Western blot analysis. Consistent with previous findings [32,33], the phosphorylation of Akt and ERK1/2 increased significantly after reperfusion, with the peak of activation occurring at 30 min and decreasing at 3 h after reperfusion (Figure 6D). The WT Treg transfer further increased both Akt and ERK1/2 phosphorylation at 30 min of reperfusion. However, the further increased phosphorylation of ERK1/2 and Akt was not observed in CD39−/− Treg-transferred mice (Figure 6E). To demonstrate that the alterations observed in vivo reflect the direct response of cardiomyocytes to Tregs, we tested Akt and ERK1/2 signalling in cardiomyocytes after co-culturing with Tregs. The kinetics of activation of Akt and ERK1/2 following stimulation by H2O2 was first characterized by Western blot analysis (Figure 6F). The treatment of cardiomyocytes with WT Tregs significantly increased the phosphorylation of Akt and ERK1/2, whereas a CD39 deficiency significantly impaired this effect of Tregs. Moreover, such an effect of WT Tregs could be simulated by adenosine (Figure 6G). These results confirmed that Treg-activated pro-survival signalling of Akt and ERK1/2 conferred direct protection on cardiomyocytes.
Tregs inhibited neutrophil infiltration through down-regulation of the cardiac production of the chemokines KC and LIX
Neutrophil-mediated inflammatory injury plays an important role during myocardial I/R. Tregs have been shown to suppress inflammatory responses in various disease models and one of their major targets is inhibiting neutrophil infiltration . We then investigated whether Tregs impeded the recruitment of neutrophils in the ischaemic myocardium. Neutrophil infiltration was prominent at 3 h in the PBS group, as demonstrated by myelo-peroxidase (MPO) activity and FACS analysis. The transfer of Tregs from WT mice remarkably attenuated neutrophil infiltration, whereas the transfer of Tregs derived from CD39−/− mice had no such effect (Figures 7A and 7B). Because the migration of neutrophils is regulated by the cysteine-X-cysteine glutamic acid–leucine-arginine+ (CXC ELR+) chemokines cytokine-induced neutrophil chemoattractant (KC), macrophage inflammatory protein (MIP)-2 and lipopolysaccharide (LPS)-induced CXC chemokine (LIX), we speculated that the manipulation of Tregs might alter chemokine production in response to myocardial I/R. Indeed, quantitative PCR performed at the site of ischaemic myocardium revealed that WT Tregs, but not CD39−/− Tregs, blunted the rise of the neutrophil-attracting chemokines KC and LIX, whereas the expression of MIP-2 was not significantly affected (Figure 7C). All of the above results suggested that Tregs inhibited local neutrophil chemoattractant production, thereby reducing neutrophil infiltration.
Tregs-reduced cardiac neutrophil recruitment and migration
We also tested this effect of Tregs on chemokine production and neutrophil migration in ex vivo cultures. Under stimulation with H2O2, cardiomyocytes were incubated with or without Tregs. After 24 h, the amounts of KC and LIX in the supernatants and the migration of neutrophil toward conditioned supernatants were determined. The production of KC and LIX by cardiomyocytes and neutrophil migration were significantly reduced in the presence of Tregs. However, this effect was reversed by CD39 deficiency. Moreover, pre-incubation with adenosine achieved comparable effects with WT Tregs (Figures 7D and 7E), indicating the pivotal role of CD39/adenosine in Tregs.
The frequency of circulating Tregs decreased, whereas the expression of CD39 by Tregs increased, after primary PCI in patients with AMI
After the above investigation of Tregs in an animal model was conducted, we sought to determine the regulation of Tregs in the acute reperfusion of ST-elevation myocardial infarction (STEMI) in humans. Our study focused on 15 patients undergoing primary percutaneous coronary intervention (PCI). Blood samples were taken before re-opening the infarct-related artery, during predilation and 1 h, 6 h, 12 h, 24 h and 72 h after stent implantation. Fifteen age-matched individuals with angiographically normal arteries were enrolled as the control group. We found that the percentage of Tregs (defined as CD4+CD25+CD127low) in the CD4+ T-cell population did not show significant differences between the control group and the baseline of the STEMI group. However, compared with the baseline, Tregs decreased after reperfusion of the STEMI group, which began at pre-dilation. A significant reduction occurred at 1 h, reaching a minimum at 6 h and this reduction lasted for at least 72 h after PCI (Figure 8B), which indicated a deficit of peripheral Tregs during the reperfusion phase of STEMI.
Dynamic changes of Tregs before and after primary PCI in patients experiencing AMI
Because CD39 has been proven to be the key regulator of Tregs from the above results, we next analysed the fraction of CD39+ cells within the Tregs population and the fraction of CD39-expressing Tregs within the CD4+ T-cell population. Surprisingly, the fraction of CD39+ Tregs (% of CD4+ T-cells) was not decreased along with the proportion of total Tregs at each time point after PCI (Figure 8C). By contrast, the relative frequency of CD39+ cells within the Tregs population increased after reperfusion and reached statistical significance at 12 h and 24 h following PCI (Figure 8D).
MIRI is considered an inflammatory condition that is characterized by immune responses, which makes immune cells and mediators potential therapeutic targets [2,3]. Our results qualify that Tregs provide protective effect in MIRI and activated Treg transfer is a promising cardioprotective treatment. Specifically, we first detected a dynamic infiltration of endogenous Tregs in the myocardium after MIRI. Selectively, depletion of Tregs by the use of DEREG mice led to substantially deteriorated MIRI, whereas anti-CD25 mAb administration did not potentiate MIRI. The adoptive transfer of in vitro-activated but not freshly isolated Tregs significantly protected mice from MIRI. An additional mechanistic study demonstrated that CD39 is responsible for exogenous Tregs-afforded cardioprotection and this effect was achieved by regulating cardiomyocyte apoptosis, pro-survival pathway ERK1/2 and Akt and neutrophil infiltration.
Treg-mediated suppression occurs at peripheral sites during immune reactions, in draining lymph nodes or in both  and the former is considered more important . We observed a rapid infiltration of Tregs in myocardium as early as 5 min post reperfusion, which was consistent with previous reports that T-cells extravasated into the myocardium within 2 min after the initiation of reperfusion . Moreover, Treg accumulation progressed between 5 min and 24 h after reperfusion compared with sham, as indicated by the results from absolute cell counts, further suggesting a recruitment of Tregs to myocardium which is driven by the process of MIRI. However, we did not evaluate the activation status of these cells. A proportion of recently activated Tregs (CD69+) exist in vivo even at steady state , suggesting the possibility that these recently activated Tregs could infiltrate in heart following MIRI. In addition, Tregs could be activated in lymph nodes, in peripheral tissues or in both [24,38]; hence we assumed that Tregs could be activated for full suppressive function during MIRI, although the activation dynamics and sites of Tregs need to be elucidated in the future. Besides, we found Tregs in sham-operated heart tissue, which agrees with a recent observation that there are resident immune cells, including Tregs, within the unstressed heart .
Treg infiltration into myocardial tissue post reperfusion indicates a possible role for endogenous Tregs in this process. To address this hypothesis, we used two methods to deplete Tregs. First, anti-CD25 mAb (PC61), which could partially deplete Tregs, had no effect on MIRI. It may attribute to the fact that CD25 is also up-regulated on activated T-cells and the existence of a CD25-negative Treg subpopulation, which made the depletion of Tregs by anti-CD25 Ab less specific and thorough. Additionally, anti-CD25 Ab was reported to deplete an inactive population of Tregs that expressed a low level of CD69 . Therefore, the remaining Tregs with an activated phenotype may be sufficient to confer protection, which resulted in the negative results for anti-CD25 Ab. Secondly, we used the DEREG mouse model, which allowed depletion of Foxp3+ cells with very high specificity and effectiveness and had been employed in several recent studies to delineate the functional role of Treg in vivo [19,40,41], to exclude the technical limitation of anti-CD25 Ab. Our data demonstrated that Treg ablation in DEREG mice resulted in increased infarct size and impaired cardiac function. A recent report and our previous study have confirmed that Tregs improved healing and ameliorate remodelling at a later stage in a permanent myocardial infarct model [14,40]. In the present study, using a transient ischaemia and reperfusion model, we concluded that Tregs also offered protection in acute myocardial injury at early stage.
The adoptive transfer of Tregs not only represents a promising strategy to elucidate the role of Tregs but also is an attractive therapy to treat or prevent a number of experimental autoimmune diseases and graft-versus-host disease (GVHD) and to facilitate solid-organ transplantation . Moreover, the direct infusion of Tregs has been shown to have the potential to alleviate IRI of the liver, kidney and brain [10,12,13]. However, our results revealed that freshly isolated Tregs had no beneficial effect on MIRI. Previous studies have provided evidence that in vitro-activated Tregs are more suppressive than freshly isolated Tregs [20,21]. In a model of polymicrobial sepsis, the adoptive transfer of in vitro-activated Tregs increased bacterial clearance and improved survival, whereas freshly isolated Tregs had little effect . Similarly, in vitro-activated Tregs can ameliorate ongoing chronic GVHD, but freshly isolated donor naive Tregs showed no protective effect . In line with these observations, the current study demonstrated that treatment with in vitro-activated Tregs protected mice from MIRI. The mechanisms related to the therapeutic effect of in vitro-activated Tregs on MIRI were further investigated. Consistent with previous reports [44,45], in vitro stimulation of Tregs not only augmented the secretion of soluble suppressive mediators (IL-10 and TGF-β1), but also enhanced their ability to generate adenosine. Thus, the cardioprotection afforded by activated but not naive Tregs can be partly attributed to enhanced suppressive activity by in vitro stimulation. We observed an increased migration of activated Tregs to the myocardium after being transferred into MIRI mice and a higher expression of CCR5 on Tregs after activation. Of note, CCR5 signalling was shown to mediate Treg recruitment in MIRI . We therefore speculated that up-regulation of the migration may also be the underlying mechanism for the increased therapeutic effect of activated Tregs. Finally, although Tregs could be efficiently activated at the site of tissue inflammation [24,35], in vivo activation of freshly isolated Tregs requires sequential migration from the blood to the target tissue and to draining lymph nodes to differentiate and fully execute their suppressive function , which is certainly a time-consuming process. The delay in activation may negate their potential for cardioprotection because the first few minutes of myocardial reperfusion is a critical time window for cardioprotection . Any interventions instituted after the first few minutes of reperfusion would become ineffective in cardioprotection . Taken together, the data suggest that in vitro-activated Tregs, with their enhanced suppressive activity, increased migration ability and fast-working potential, suppress injury post reperfusion, which in turn confer cardioprotection, represent a promising cell therapy for MIRI.
The mechanisms underlying the suppressive effects of Tregs in vivo appear to be complex. In considering the candidates for suppressive mediators that are responsible for Treg-afforded cardioprotection, we noted that IL-10 and TGF-β1 were both implicated as being protective in MIRI [27,28]. However, our data showed that neither IL-10 nor TGF-β1 is likely to be a direct mediator of exogenous Treg-afforded cardioprotection. Adenosine has been recognized as a soluble mediator of suppression conveyed by Tregs in a colitis model in 2006 . Tregs from CD39-null mice show impaired suppressive properties in vitro and fail to block allograft rejection and the ear-swelling response in contact hypersensitivity reactions in vivo [9,47]. Moreover, in kidney IRI, CD73-deficiency led to inhibition of Treg function, indicating the critical role of adenosine in the Treg-mediated suppression of kidney IRI . Besides, adenosine is a well-established cardioprotective agent  and it has been applied as an adjunct to PCI, showing an 11% reduction in the size of myocardial infarcts in clinical studies [49,50]. In our study, we found that Tregs that lack CD39 no longer protected the heart from reperfusion injury, suggesting that CD39 was vital for the suppressive activity of Tregs in MIRI. Of note, CD73 is also critical for Tregs to produce adenosine; whether CD73 is required for the protective effect of Tregs in MIRI remains to be elucidated. Nevertheless, we cannot exclude a contribution from other mechanisms of Treg-mediated suppression, such as cell–cell contact through CTLA-4 (cytotoxic T-lymphocyte-associated protein 4) or cytolysis through the secretion of granzymes.
In our recent study, we showed that Treg transfer reduced cardiomyocyte apoptosis on day 28 after rat myocardial infarction and that Tregs directly protect neonatal rat cardiomyocytes against LPS-induced apoptosis in vitro through cell–cell contact and IL-10 . We, in the present study, extend these findings by demonstrating that Tregs attenuated cardiomyocyte apoptosis after 3 h of reperfusion in vivo and under circumstances of oxidative stress in vitro and confirmed CD39 as a novel protective mechanism for cardiomyocytes used by Tregs. It has been reported that adenosine protects cardiomyocytes from the apoptosis caused by hypoxia/re-oxygenation through ERK1/2 and Akt pathways . In addition, ERK1/2 and Akt belong to a group of pro-survival protein kinases named RISK, which afford powerful cardioprotection when activated specifically at the time of myocardial reperfusion. Our results indicated that the adoptive transfer of WT Tregs strengthened the activation of ERK1/2 and Akt, whereas this strengthening was not observed for the adoptive transfer of CD39−/− Tregs. The results were further validated by our experiments in vitro. Therefore, our study reaches the conclusion that cardiomyocyte apoptosis is a direct target for Tregs cardioprotection against myocardial I/R, with CD39/adenosine serving as a novel and critical mediator and activation of RISK as an underlying mechanism.
Proinflammatory cytokine and chemokine production and leucocytes, especially neutrophil infiltration, are indispensable parts of the inflammatory cascade during MIRI. Previous reports about cerebral  and renal  I/R demonstrated that Treg-conferred protection was associated with the inhibition of neutrophil accumulation and inflammatory mediator production. We showed previously that Tregs restricted the infiltration of inflammatory cells (including neutrophils, macrophages and T-lymphocytes) and reduced proinflammatory cytokines [including tumour necrosis factor a (TNF-α) and IL-1β] in the infarcted rat hearts . We, in the present study, further extend these findings by demonstrating that Treg-inhibited neutrophil infiltration and neutrophil-specific chemokine KC and LIX production. In addition to the decreased chemokine production by cardiomyocytes detected in our study, Tregs could also reduce neutrophil infiltration through other mechanisms that we could not rule out, such as increased neutrophil apoptosis  or decreased neutrophil chemotaxis and adhesion activity. Moreover, these effects of Tregs could be abrogated by CD39 deficiency and simulated by adenosine. Adenosine, as an endogenous anti-inflammatory agent, plays a central role in the regulation of inflammatory responses and in suppressing inflammatory tissue destruction . Thus, CD39 is an appropriate suppressor mechanism for Tregs to prevent neutrophil infiltration and chemokine production.
Several clinical studies have investigated circulating Tregs in patients with AMI [54–56], but little attention has been paid to the reperfusion phase following PCI. We identified a decrease in Tregs and a relative increase in CD39+ cells within the Treg population in peripheral blood following PCI, suggesting a possible role for this subset during the early phase of myocardial reperfusion after PCI. However, these are preliminary data, which need to be established in a large cohort of patients.
In conclusion, we report that Tregs confer cardioprotection against MIRI and that this effect is associated with attenuated cardiomyocyte apoptosis via activation of the Akt and ERK1/2 pathways and reduced neutrophil infiltration via down-regulation of KC and LIX production. Furthermore, we characterized a CD39-dependent mechanism whereby Tregs exert cardioprotection. Our study suggests that Treg adoptive transfer may represent an efficient and innovative cell-based therapy for MIRI. Given the course of in vitro activation, this strategy may be applied to patients undergoing elective PCI, although further preclinical study is warranted.
area at risk
acute myocardial infarction
bacterial artificial Chromsomes
C-C chemokine receptor type 5
cluster of differentiation
cardiac troponin T
depletion of regulatory T-cell
Forkhead box P3
cytokine-induced neutrophil chemoattractant
left anterior descending
lipopolysaccharide-induced CXC chemokine
macrophage inflammatory protein
myocardial ischaemia/reperfusion injury
percutaneous coronary intervention
reperfusion injury salvage kinase
ST-elevation myocardial infarction
transforming growth factor
terminal deoxynucleotidyltransferase mediated dUTP nick-end labeling
Ni Xia and Jiao Jiao performed most of the experiments together with Bing-Jie Lv, Yu-Zhi Lu, Xiao-Bo Mao and Ke-Jing Wang, and drafted the manuscript. Ting-Ting Tang, Xin Tu and Hong Xiao analysed the data and contributed to the study design. Zheng-Feng Zhu and Shao-Fang Nie performed blood collection from patients and flow cytometry analysis. Qing Wang, Guo-Ping Shi and Yu-Hua Liao were involved in study supervision and manuscript revision. Xiang Cheng designed the experiments, supervised and funded the study and revised the manuscript.
We thank Professor Vijay Kuchroo (Brigham and Women's Hospital, Harvard University) for Foxp3–GFP knockin mice and Professor Simon Robson (Beth Israel Deaconess Medical Center, Harvard University) for CD39−/− Foxp3–GFP knockin mice.
This work was supported by grants from the National Basic Research Program of China [grant numbers 2013CB531103 and 2012CB517805 (to X.C.)]; and the National Natural Science Foundation of China [grant numbers 81222002 and 91339118 (to X.C.), 81400364 (to N.X.), 81200177 (to T.-T.T.) and 81470596 (to H.X.)].
These authors contributed equally to this work.