Impaired Thymic Export and Apoptosis Contribute to Regulatory T-Cell Defects in Patients with Chronic Heart Failure

Objective Animal studies suggest that regulatory T (Treg) cells play a beneficial role in ventricular remodeling and our previous data have demonstrated defects of Treg cells in patients with chronic heart failure (CHF). However, the mechanisms behind Treg-cell defects remained unknown. We here sought to elucidate the mechanism of Treg-cell defects in CHF patients. Methods and Results We performed flow cytometry analysis and demonstrated reduced numbers of peripheral blood CD4+CD25+FOXP3+CD45RO−CD45RA+ naïve Treg (nTreg) cells and CD4+CD25+FOXP3+CD45RO+CD45RA− memory Treg (mTreg) cells in CHF patients as compared with non-CHF controls. Moreover, the nTreg/mTreg ratio (p<0.01), CD4+CD25+FOXP3+CD45RO− CD45RA+CD31+ recent thymic emigrant Treg cell (RTE-Treg) frequency (p<0.01), and T-cell receptor excision circle levels in Treg cells (p<0.01) were lower in CHF patients than in non-CHF controls. Combined annexin-V and 7-AAD staining showed that peripheral Treg cells from CHF patients exhibited increased spontaneous apoptosis and were more prone to interleukin (IL)-2 deprivation- and CD95 ligand-mediated apoptosis than those from non-CHF individuals. Furthermore, analyses by both flow cytometry and real-time polymerase chain reaction showed that Treg-cell frequency in the mediastinal lymph nodes or Foxp3 expression in hearts of CHF patients was no higher than that of the non-CHF controls. Conclusion Our data suggested that the Treg-cell defects of CHF patients were likely caused by decreased thymic output of nascent Treg cells and increased susceptibility to apoptosis in the periphery.


Introduction
Chronic heart failure (CHF) is regarded as a state of chronic inflammation with elevated T-cell activation and inflammatory cytokine production in the circulatory system [1,2]. However, the pathogenic mechanisms responsible for this abnormal immune activation remain unknown. T reg cells represent a unique lineage of T cells that play an essential role in the modulation of immune responses and the control of potentially harmful immune activations because of their immunoregulatory and immunosup-pressive characteristics [3]. Among the several types of T reg cells that have been defined, one particular subset that constitutively expresses CD4, CD25 and the transcription factor Foxp3 has received much attention. Alterations in CD4 + CD25 + Foxp3 + T regcell number or function is directly associated with the pathogenesis of several common human diseases, including acute coronary syndrome (ACS) [4,5], multiple sclerosis [6], type 1 diabetes [7], and rheumatoid arthritis [8]. Adoptive transfer of purified T reg cells suppresses immune injury and improves recovery in animal disease models [9][10][11][12].
Adverse ventricular remodeling occurs upon acute and chronic injury regardless of etiology, and it is related to poor prognosis of patients with heart failure [13]. There is compelling evidence that inflammatory mechanisms contribute to the process of adverse ventricular remodeling [14]. In animal models of heart failure, previous studies demonstrated that T reg cells could be a target of heart failure therapeutics because CCR5-mediated T reg -cell recruitment in the infarcted heart [15] and adoptively transferred T reg cells [16] provided protection from adverse cardiac remodeling by preventing expansion of inflammation and fibrosis after adoptive transfer. In a previous publication, we found that circulating T reg cells were reduced and their function was altered in CHF patients, regardless of etiology, suggesting that the defects in T reg cells are responsible for the aberrant chronic immune activation in CHF patients [17]. It is believed that the understanding of mechanisms underlying T reg -cell defects in CHF patients is of great significance, especially with respect to therapy through T reg -cell manipulation. In the present study, we attempt to explore the mechanisms that might account for the T reg -cell defects in CHF patients by studying T reg -cell production, survival, and tissue reallocation in these patients.

Reduced nT reg -, mT reg -and RTE-T reg -cell frequency in CHF patients
To determine the number of total T reg cells and T reg subsets, PBMCs were obtained from 52 CHF patients and 43 age-matched non-CHF controls followed by 6-color flow cytometric analysis. Basic clinical characteristics of the study population are summarized in Table 1. Within the naïve CD4 + CD45RA + CD45RO 2 (R1 in Figure 1A) or memory CD4 + CD45RA 2 CD45RO + (R2 in Figure 1A) T cells, a small subpopulation of cells with high expression of both CD25 and Foxp3 could be readily detected. mT reg cells were characterized as CD4 + CD25 + Foxp3 + CD45RA 2 CD45 RO + cells (R3 in Figure 1B, upper panel) and nT reg cells were characterized as CD4 + CD25 + Foxp3 + CD45RA + CD45RO 2 cells (R4 in Figure 1B, lower panel). nT reg cells exhibited a lower expression of CD25 as compared with mT reg cells (mean fluorescent intensity, MFI: nT reg vs. mT reg : 24.165.4 vs. 54.668.7, p,0.01). RTE-T reg cells were identified as CD31 coexpressing nT reg cells ( Figure 1C). The proportion of T reg cells in total CD4 + T cells was significantly decreased in CHF patients when compared with non-CHF subjects ( Figure 1D). The percentages of nT reg and mT reg cells within CD4 + T cells were also significantly lower in CHF patients than in age-matched non-CHF subjects (non-CHF vs. CHF patients: nT reg : 1.1760.41% vs. 0.5960.31%, p,0.01; mT reg : 3.2760.92% vs. 2.0260.65%, p,0.01; Figure 1D). CHF patients showed a significantly lower nT reg /mT reg ratio (non-CHF vs. CHF patients: 36610% vs. 30612%, p,0.05; Figure 1E). Furthermore, we observed that proportions of RTE-T reg cells in the total T reg -cell population in CHF patients were significantly reduced when compared to agematched, non-CHF controls, suggesting that thymic production of T reg cells was impaired in CHF patients (non-CHF vs. CHF: 4.976.34% vs. 3.0060.97%, p,0.01; Figure 1F). However, no difference in total T reg , nT reg , mT reg , and RTE-T reg cells between IHF or NIHF patients was observed ( Figure 1D-F). Similar results were obtained when we compared the absolute numbers of total T reg , nT reg , mT reg , and RTE-T reg cells between CHF patients and non-CHF controls ( Table 2).
Consistent with our previous report [17], we observed that total T reg number was negatively correlated with NT-proBNP in CHF patients. Furthermore, the present study also found that NT-proBNP and nT reg or mT reg numbers were negatively correlated (Table 3).

Decreased intracellular TREC levels in T reg cells from CHF patients
TREC is a marker for nascent thymic T cells [18]. We studied intracellular levels of TRECs in T reg cells isolated from 25 CHF patients and 15 age-matched non-CHF subjects using quantitative real-time PCR. Flow cytometry was used to determine the purity of T reg cells after cell sorting (Figure 2A, left panel). The TREC content in T reg cells was significantly lower in CHF than that in non-CHF patients (non-CHF vs. CHF patients: 1.3860.49610 3 / 10 6 cells vs. 0.6060.32610 3 /10 6 cells, p,0.01; Figure 2A, right panel). There was no significant difference in T reg -cell TREC content between IHF and NIHF patients. Spearman's correlation test revealed a positive association between T reg -cell TREC level and RTE-T reg cell proportion in both CHF patients and non-CHF controls (r = 0.75, p,0.001; Figure 2B).

Increased spontaneous and IL-2 deprivation/Fasmediated apoptosis in T reg cells from CHF patients
Increased apoptosis and decreased survival could be a mechanism of T reg -cell defects in CHF patients. Because of the fixation and permeabilization procedures used for detecting T reg cells using Foxp3 antibodies by FACS, we detected apoptotic T reg cells with antibodies against CD127, a newly identified T reg surface marker that correlates well with Foxp3 [19]. In both CHF patients and non-CHF controls, CD4 + CD25 + Foxp3 + T reg cells were correlated with CD4 + CD25 + CD127 low/2 T reg cells (r = 0.91, p,0.001, Figure 3). When we gated the CD4 + CD25 + CD127 low/2 cells, we found that when this cell population was derived from CHF patients, irrespective of the etiology, there was a significantly higher percentage of apoptotic annexin V + 7-AAD 2 cells than when derived from non-CHF controls (non-CHF vs. CHF patients: 8.7963.37% vs. 14.7864.08%, p,0.01; Figure 4A).
Enhanced apoptosis often correlates with altered expression of apoptosis-associated genes. We compared the levels of antiapoptotic gene Bcl-2 and pro-apoptotic gene Bak expression between CD4 + CD25 + CD127 low/2 T reg cells isolated from CHF patients and non-CHF controls. Significantly lower Bcl-2 expression (p,0.01) and higher Bak expression (p,0.01) were observed in T reg cells from CHF patients when compared with those from non-CHF controls ( Figure 4B). IL-2 is essential for the development, function and homeostasis of T reg cells [20]. However, human T reg cells do not produce this cytokine and therefore are susceptible to IL-2 deprivation, which leads to T reg -cell apoptosis [21]. T reg cells from CHF patients and non-CHF controls might exhibit different susceptibilities to IL-2 deprivation. To test this hypothesis, we incubated T reg cells from the different patient populations with anti-human IL-2 monoclonal antibodies for 3 days. T reg cells from CHF patients were more sensitive to IL-2 deprivation-induced apoptosis when compared with T reg cells from non-CHF subjects (non-CHF vs. CHF: 22.3564.12% vs. 33.2665.89%, p,0.01; Figure 5A).
T reg -cell apoptosis could also be induced by the interactions between death receptor CD95 with the CD95 ligand (CD95L) [22]. Human T reg cells constitutively express these death receptors and are thus highly sensitive to CD95-CD95L-mediated apoptosis [23]. Increased apoptosis in T reg cells from CHF patients suggests that T reg cells from these patients express high levels of CD95 and/or are more sensitive to CD95L. To test this hypothesis, we compared the expression level of CD95 and sensitivity toward CD95L-triggered apoptosis in T reg cells from CHF patients and non-CHF controls. CD95 expression on T reg cells from CHF patients was significantly higher than on T reg cells from non-CHF controls (non-CHF vs. CHF: 73.7868.12% vs. 84.3066.67%, p,0.01; Figure 5B). CD95L induced apoptosis of T reg cells from CHF patients in a dose-and time-dependent manner ( Figure 5C). CD95L initiated T reg -cell apoptosis in 3 hrs, but apoptosis reached a peak after 12 hrs of induction. When incubated with 100 ng/ml of CD95L for 12 hrs, T reg cells prepared from CHF patients showed higher percentages of cells undergoing CD95L-induced apoptosis than in non-CHF subjects (non-CHF vs. CHF patients: 19.4366.87% vs. 36.52612.03%, p,0.01; Figure 5D). These observations could explain the increased T reg -cell apoptosis in CHF patients ( Figure 4A). Furthermore, we detected significantly higher plasma levels of soluble CD95L in CHF patients than in non-CHF controls (non-CHF vs. CHF patients: 77.2865.26% vs. 101.2265.06%, p,0.01; Figure 5E). Among the CHF subgroups IHF and NIHF, we did not detect any differences in either IL-2 deprivation-or CD95-mediated T reg -cell apoptosis ( Figure 5A/ 5D). Plasma CD95L levels were also similar between CHF, IHF, and NIHF patients ( Figure 5E). Taken together, these findings suggest that T reg cells from CHF patients were more prone to apoptosis and that IL-2 and CD95/CD95L might be involved in regulation of T reg -cell survival.

T reg cells accumulate neither in mediastinal lymph nodes nor in failing hearts
One possible explanation for reduced T reg -cell number in CHF patients is the reallocation of these cells to the lymph nodes or disease-affected organs. We compared the proportion of CD4 + CD25 + Foxp3 + T reg cells to total CD4 + T cells in the mediastinal lymph nodes from CHF patients and non-CHF controls. Mediastinal lymph node T reg cells from CHF patients were significantly fewer than from non-CHF controls ( Figure 6A/ 6B). Total lymphocyte Foxp3 mRNA levels were also significantly lower in CHF, IHF and NIHF patients than in non-CHF controls ( Figure 6C). To examine whether T reg cell accumulation in the heart was different between CHF and non-CHF controls, Foxp3 RT-PCR was performed on biopsied cardiac samples. No difference was found between failing hearts and hearts from donors, although Foxp3 levels were low in all tested heart samples ( Figure 6C).

Discussion
As the final common pathway of many cardiovascular diseases, CHF is a complex multi-step disorder and several mechanisms participate in its pathogenesis. There is compelling evidence that inflammation and autoimmune abnormalities play an important role in the progression of heart failure [1,24,25]. Various autoantibodies, which are directed against different cardiac antigens, such as cardiac myosin, cardiac troponin I, cardiolipin, beta 1 -adrenergic and M2 muscarinic receptors can be detected in the serum of patients with NIHF or IHF [26][27][28][29]. These autoantibodies can lead to cardiac injury, and they correlate with the deterioration of cardiac function. Other autoimmune abnormalities include infiltration of T cells in endomyocardial biopsies from patients with idiopathic dilated cardiomyopathy (DCM). Additionally, the transfer of peripheral blood lymphocytes from DCM patients to severe combined immunodeficiency (SCID) mice leads to ventricular remodeling [30]. In animal models, lymphocytes from rats with IHF can recognize and kill normal neonatal rat cardiac myocytes in vitro [31] and lead to autoimmune myocarditis in vivo after adoptive transfer [32].
T reg cells play a key role in the control of inflammation and autoimmune responses, and altered T reg cells predispose patients for uncontrolled immune activation or autoimmunity [3]. CHF patients were previously reported to have impaired T reg -cell number and function, but the precise mechanism behind this defect remains largely unknown [17]. In this study, we showed that reduced T reg cell number and function in CHF patients might be explained by impaired T reg -cell thymic output and increased apoptosis of these cell populations.
Like other T cells, T reg cells develop in the thymus [33]. A small fraction of T reg cells with a naïve CD45RA + CD45ROsurface profile (nT reg ) has recently been detected in the circulation. However, this nT reg subset declines with age, as does thymic output and other naïve T cells [34]. By contrast, the majority of circulating T reg cells appear as a mature population with a memory CD45RA 2 CD45RO + phenotype; these mT reg cells are stable throughout the life span, and the levels of mT reg cells increase during aging [35,36]. nT reg cells could represent the de novo generation of thymic lymphocytes, so the assessment of nT reg Table 2. Absolute number of T reg , nT reg , mT reg and RTE-T reg in the study population.
CD4 + T cells (10 6 /L) T reg (10 6 /L) nT reg (10 6 /L) mT reg (10 6 /L) RTE-T reg (  cells is used to evaluate thymic T reg -cell production. In this study, we provided evidence that, in addition to decreased percentages of nT reg and mT reg cells, a shift from nT reg cells toward mT reg cells was evidenced by a reduced nT reg /mT reg cell ratio in CHF patients. This result indicated the possibility that impaired thymic export contributes to T reg cell defects in this patient population. However, nT reg cells can proliferate after thymic output while retaining their naïve phenotype [37]. CD31 has been used as a direct marker of thymic output and enabled the discrimination of recent thymic emigrant (RTE) T reg cells from peripherally expanded nT reg cells [38]. Thus, the assessment of nT reg cells co-expressing CD31 (RTE-T reg ) is now used to evaluate the thymic output of T reg cells. The significant reduction of peripheral RTE-T reg cell content in CHF patients, when compared to the non-CHF controls, suggests a reduction of thymic T reg -cell output during the development of heart failure. An alternative approach to determine impaired T reg -cell thymic output in CHF patients was to assess intracellular concentration of TRECs in purified T reg cells. TRECs are generated during the process of T-cell receptor rearrangement in T-cell differentiation and do not duplicate during mitosis. TRECs are diluted out during homeostatic or antigen-driven T-cell proliferation in the periphery [18]. Therefore, TRECs are enriched in the newly synthesized and exported T-cell pool. nT reg cells, especially RTE-T reg cells, have higher frequencies of TRECs as compared with mT reg cells [38]. TREC content reduction in total T reg cells from CHF patients further supported our hypothesis that the Treg-cell output in the thymus of a CHF patient is functionally altered. Hass et al. recently reported that T reg cells from patients with and without multiple sclerosis showed different activities in suppressing T-effector cells. However, such differences disappeared after depleting the RTE-T reg cells, indicating a crucial role of RTE-T reg cells in the functional properties of the entire T reg population [38]. Thus, impaired thymus export of T reg cells could be associated not only with the number but also with the functional defect of T reg cells in CHF patients. Over the course of multiple sclerosis, for example, patients appear capable of amplifying mT reg -cell subpopulations to compensate for impaired thymic production of T reg cells [39]. In the case of CHF, in contrast, the homeostatic control of T reg cells seems to be disturbed. Both nT reg and mT reg cells were reduced in CHF patients ( Figure 1D). The homeostasis of T reg cells is maintained by a balance between T reg -cell generation and depletion. Apoptosis-induced alteration of T reg -cell levels has been associated with several diseases. For example, intrathyroidal CD4 + CD25 + T reg cells from patients with autoimmune thyroid diseases were prone to apoptosis, which led to a local T reg -cell reduction [40]. In contrast, patients with metastatic epithelial cancer demonstrated a significantly elevated proportion of peripheral T reg cells, and these cells were protected from apoptosis [41]. Apoptosis not only reduces the number of T reg cells, but also reduces their functions. By using T-effector cell suppression assays, T reg -cell apoptosis was closely associated with their capacity to inhibit T-effector cell proliferation [42]. In patients with type 1 diabetes, an increase in apoptosis was correlated with a decline in the suppressive potential of T reg cells [43]. As suggested by these studies, high sensitivity to IL-2 deprivation or FasL-induced apoptosis may contribute in part to the defect of T reg cells in CHF patients. T reg cells from CHF patients were more susceptible to apoptosis following IL-2 deprivation. Upon antigen activation, T cells induce the expression of CD95, a member of the tumor necrosis factor receptor/nerve growth factor receptor superfamily that induces apoptosis by binding to CD95L and subsequently activating   caspase [44]. In the present study, we demonstrated that T reg cells in CHF patients had higher CD95 expression levels and were more sensitive to CD95/CD95L-mediated apoptosis than those in non-CHF subjects. Indeed, we also detected concurrent increases in serum soluble CD95L levels in CHF patients, consistent with prior observations [45]. These findings strongly suggest that the CD95/CD95L pathway is an important regulator of T reg -cell apoptosis in CHF patients. After release from the thymus, T reg cells circulate continuously from blood to lymphoid tissues. In disease conditions, the expression of chemokine receptors, such as CCR4 and CCR8, on T reg cells allows their migration and recruitment to the site of inflammation [46]. In several human diseases, T reg cells preferentially accumulate at lymphoid tissues or sites of affected organs [47,48]. Therefore, it is possible that decreases in peripheral T reg cells in CHF patients are caused by T reg -cell reallocation rather than an overall decrease. To investigate this possibility, we compared the T reg -cell numbers in the mediastinal lymph nodes or Foxp3 expression in cardiac biopsies between CHF patients and non-CHF controls. The results revealed that T reg -cell frequency in the mediastinal lymph nodes or Foxp3 expression in hearts of CHF patients was no higher than that of the non-CHF controls. However, this possibility could not be excluded due to the very small sample number. In addition to generation in the thymus, T reg cells can also be converted from activated effector or memory CD4 + CD25 2 T cells in the periphery [49]. Peripherally converted T reg cells and thymusgenerated T reg cells demonstrate a similar phenotype and suppressive functions. It is possible that such peripheral T-cell phenotype conversion was altered in CHF patients. This hypothesis merits further investigation.
TNF-a is central in the inflammatory cytokines response in CHF and play a role in the pathogenesis and clinical progression of the disease [50]. IL-10, an anti-inflammatory cytokine, may offer protection against TNF-a and an improvement in cardiac function in CHF has been associated with an increase in IL-10 [51] or a decrease in TNF-a/IL-10 ratio [52]. Our data indicated that T reg frequency was negatively correlated with serum level of TNF-a or the TNF-a/IL-10 ratio ( Figure S1). In both our previous study [17] and the present study, we observed that total T reg number was significantly negatively correlated NT-proBNP which is considered as the most sensitive index of cardiac dysfunction in CHF patients. Based on these observations, we may speculate that T reg cells provided protection for the failing heart and defects in T reg cells is involved in the deterioration of cardiac function in CHF patients. However, the direct effect of T reg cells on cardiac dysfunction still needs to be studied in animal model.
IL-10 and TGF-b1 have been identified as the main effector cytokines of T reg cells [53]. We investigated the hypothesis that impaired T reg -cell function was associated with the decreased expression of these two cytokines. Disappointedly, we failed to observe a decrease in the expression of either IL-10 or TGF-b1 in CHF patients ( Figure S2).
To conclude, our study revealed that both impaired export from the thymus and enhanced apoptosis can account for impaired T reg -cell number and function in CHF patients, offering a mechanistic explanation for the phenotypes and providing possible novel targets for CHF therapy through T reg -cell manipulation.

Subjects
samples were collected from 52 CHF patients (31 men and 21 women, 44613 years old) and 43 non-CHF controls (28 men and 15 women, 42612 years old). Peripheral blood mononuclear cells (PBMCs) were prepared by Ficoll density gradient centrifugation (Sigma, USA). Plasma was obtained after centrifugation and stored at 280uC. CHF diagnoses were based on clinical history, physical examination, echocardiography, chest X-ray, electrocardiography and levels of N-terminal pro-brain natriuretic peptide (NT-proBNP), according to available guidelines pertaining to CHF. Patients were classified as having non-ischemic heart failure (NIHF) (n = 32, 17 men and 15 women) if they had no history of myocardial infarction and did not have significant coronary artery stenosis upon coronary angiography. Patients were considered to have ischemic heart failure (IHF) (n = 20, 14 men and 6 women) if the coronary angiography presented significant coronary artery disease (.50% stenosis in more than one major epicardial coronary artery) or the patients had a history of myocardial infarction or previous revascularization. Patients were excluded (1) if they were currently treated with anti-inflammatory drugs, such as non-steroidal anti-inflammatory drugs and steroids, (2) if they had collagen disease, thromboembolism, disseminated intravascular coagulation, advanced liver disease, renal failure, malignant disease, other inflammatory disease (such as septicemia, pneumonia), valvular heart disease, or atrial fibrillation, or (3) if they had pacemakers. Patients with higher serum cholesterol than the target values after risk stratification [54], who received statin therapy within 3 months, were also excluded. Mediastinal lymph nodes  [55] and left ventricular biopsies were obtained from six CHF patients (three patients with dilated cardiomyopathy who underwent cardiac transplantation and three patients with coronary heart disease who underwent the combined bypass surgery and left ventricular aneurysm resection) and three controls (heart graft donors without cardiomyopathy who died in car accidents).

Ethics statement
The investigation conforms to the principles outlined in the Declaration of Helsinki. The trial was approved by the ethics committee of Tongji Medical College of Huazhong University of Science and Technology and patients and controls provided written informed consent.
3. Naïve T reg (nT reg ), memory T reg (mT reg ) and recent thymic emigrant-T reg (RTE-T reg ) cells in the circulation A 6-color flow cytometry analysis was performed to determine levels of nT reg , mT reg and RTE-T reg in the circulation. PBMCs were stained with surface antibodies for APC/Cy7 anti-human CD4, PE anti-human CD25, FITC anti-human CD45RA, Percp/ Cy5.5 anti-human CD45RO and PE/Cy7 anti-human CD31 (Biolegend) for 30 min at 4uC. After surface staining, cells were fixed, permeabilized, and stained with APC anti-human Foxp3, according to the manufacturer's instructions (eBioscience, USA). Antibody isotype controls were performed to ensure antibody specificity. Stained cells were analyzed by flow cytometry with FACS Aria (BD Biosciences, USA).

T reg 2cell isolation
A two-step selection using a CD4 + CD25 + CD127 dim/2 Regulatory T cell Isolation Kit (Miltenyi Biotec, Germany) was used to isolate T reg cells according to the manufacturer's instructions. Briefly, non-CD4 + and CD127 high cells were magnetically labeled with a cocktail of biotin-conjugated antibodies and anti-biotin microbeads and subsequently depleted by negative selection. Preenriched CD4 + T cells were then labeled with anti-CD25 microbeads, and CD4 + CD25 + CD127 dim/2 T reg cells were isolated by positive selection. FACS was used to confirm the purity (.90%) of the isolated T reg cells.

Soluble CD95 ligand (sCD95L) ELISA detection
Human FasL/TNFSF6 Quantikine ELISA Kit (R&D Systems, USA) was used to determine the plasma sCD95L levels. The minimal detectable concentration was 2.66 pg/ml, and intra-assay and inter-assay coefficients of variation were ,10%. All samples were measured in duplicate.

T reg -cell detection in mediastinal lymph node
Mediastinal lymph nodes were minced and filtered through a cell strainer to create a single cell suspension preparation. Lymphocytes were isolated using Ficoll-Hypaque, stained with specific antibodies for CD4, CD25 and Foxp3, and subjected to FACS analysis. The number of T reg cells in the lymph nodes was quantified by flow cytometry. Percentages of CD4 + CD25 + Foxp3 + T reg cells in the mediastinal lymph nodes were determined in six CHF patients (three with idiopathic cardiomyopathy and three with ischemic cardiomyopathy) and three controls without cardiomyopathy. c. Comparison of Foxp3 expression in the mediastinal lymph nodes and hearts of CHF and non-CHF controls. *p,0.05 vs. non-CHF controls. doi:10.1371/journal.pone.0024272.g006 9. Real-time PCR Total RNA was extracted using Trizol lysis buffer (Invitrogen, USA), and cDNA was prepared using the Revertra AceH kit (Toyobo, Japan). Expression of target genes (Bcl-2 and Bak in purified T reg and Foxp3 cells) in heart tissues and lymphocytes isolated from mediastinal lymph nodes was quantified using the SYBR Green Master Mix (Takara, Japan) on an ABI Prism 7900 Sequence Detection system (Applied Biosystems, USA). Primer pairs were as follows: Bcl-2: F, 5-tacctgaaccggcacctg-3, R, 5-gccgtacagttccacaaagg-3; Bak: F, 5-cctgccctctgcttctgag-3, R, 5-ctgctgatggcggtaaaaa-3; Foxp3: F, 59-gaaacagcacattcccagagttc-39, R, 59-atggcccagcggatgag-39 GAPDH: F, 59-ccacatcgctcagacaccat-39, R, 59-ggcaacaatatccactttaccagagt- 39 For each sample, the mRNA expression level was normalized to that of GAPDH. The mean of duplicate measurements was normalized and expressed as a ratio of target mRNA copies to GAPDH mRNA copies.

Tumor necrosis factor (TNF)-a and IL-10 ELISA detection
Commercial ELISA Kits (Invitrogen, USA) were used to determine the plasma TNF-a and IL-10 levels. The minimal detectable concentrations were 0.5 pg/ml and 0.78 pg/ml for TNF-a and IL-10 respectively. All samples were measured in duplicate.

Statistical analysis
Values are expressed as means 6 standard deviation (SD) or percentage in the text and figures. For variables with normal distribution and homogeneity of variance, independent t-test or one-way analysis of variance (ANOVA) was used to test differences among two or more groups. For skewed variables, either nonparametric Kruskal-Wallis H test or Mann-Whitney U test were used for analyses. For the ranked data, Pearson's chi-square test or Fisher's exact test were used for the comparison between multiple groups. Spearman's correlation analysis was performed to determine correlation between the variables. In all cases, twotailed, p,0.05 was considered significant. Figure S1 Correlation analysis between T reg frequency and plasma levels of cytokines in CHF patients (n = 20). (TIF) Figure S2 Comparison of intracellular IL-10 and TGF-b1 in CD4 + CD25 + CD127 low T reg between CHF patients (n = 10) and healthy controls (n = 10).