An immunomodulatory role for CD4+CD25+ regulatory T lymphocytes in hepatitis C virus infection

Authors


Abstract

The CD4+CD25+ regulatory T lymphocytes have been implicated in suppressing T cell immune responses. Our aim was to characterize the frequency, phenotype, function, and specificity of CD4+CD25+ T cells in hepatitis C virus (HCV) infection. Peripheral CD4+CD25+ cells from recovered (n = 15), chronic infected (n = 30), and normal control (n = 15) subjects were analyzed ex vivo for quantitation, phenotype, and effect on HCV-specific interferon gamma production and proliferation. CD4+CD25+ specificity was determined by intracellular cytokine staining for interleukin 10 (IL-10). A higher proportion of CD4+CD25+ were found in chronic infection (mean, 3.02%) when compared with recovered (1.64%, P = .001) and normal controls (2.27%, P = .02). CD4+CD25+ cells display CD45ROhigh, CD45RAlow, CD28high, CD62Lhigh, and CD95high phenotype. HCV-specific interferon gamma activity was enhanced in peripheral blood mononuclear cells depleted of CD4+CD25+ and suppressed in peripheral blood mononuclear cells enriched with CD4+CD25+. Depletion of CD4+CD25+ cells also enhanced HCV-specific CD4+ and CD8+ T cell proliferation. Cytokine analysis suggested CD4+CD25+ cells secrete transforming growth factor beta (TGF-β1) and IL-10. The inhibitory role for TGF-β1 was confirmed by anti–TGF-β1. Transwell studies showed CD4+CD25+ mediated suppression to be dose dependent and requiring cell contact. CD4+CD25+ cells showed HCV-specificity through IL-10 production, with a frequency ranging from 1.9% to 5.3%. A positive correlation was detected between CD4+CD25+ T cell frequency and HCV RNA titer, whereas an inverse relation was found with liver inflammatory activity. In conclusion, CD4+CD25+ T lymphocytes constitute a highly differentiated population and appear to play a role in viral persistence by suppressing HCV-specific T cell responses in a cell–cell contact manner. (HEPATOLOGY 2004;40:1062–1071.)

After hepatitis C virus (HCV) infection, interaction between the innate and adaptive immune response plays a pivotal role in perpetuation or clearance of HCV infection. After exposure to HCV, the quantity and breadth of interferon (IFN)-γ HCV specific CD8+ T cells has been directly associated with eradication in those patients with self-limited infection versus patients with a chronic course.1 Furthermore, the differentiation of naïve CD4+ T cell into HCV-specific inflammatory CD4+ T cell (TH1) or helper CD4+ T cell (TH2) with their respective, disparate cytokine profiles is also an integral part in determining the outcome of HCV infection.2 The TH1 proinflammatory cytokine profile (IFN-γ, tumor necrosis factor alpha [TNF-α]) promote an antiviral immune response targeted at cytosolic and intracellular antigens by stimulating cytotoxic CD8+ T cells and natural killer activation. The TH2 anti-inflammatory cytokines (interleukin 10 [IL-10], transforming growth factor beta [TGF-β]) downregulate the TH1 response. A strong TH1 response has been observed in those who clear acute HCV infection, whereas a blunted TH1 response with a more pronounced TH2 response is seen in those with chronicity.3

Recently, much attention has focused on regulatory T cells (Tregs) and their contribution to disease states. The immune system employs naturally arising CD4+ regulatory (or suppressor) T cells, which are engaged in the maintenance of self-tolerance by suppressing the activation and expansion of self-reactive lymphocytes that may cause autoimmune disease.4 A controlled balance between initiation and downregulation of the immune response is vital in maintenance of immune homeostasis, because dysfunctional immune resolution may lead to chronic inflammation and/or autoimmunity.5 Most regulatory cells constitutively express CD25 (the IL-2 receptor alpha-chain) in the physiological state. This Treg population, which constitutes 5% to 10% of peripheral CD4+ T cells in humans, has a broad repertoire that recognizes various self and nonself antigens.6 Evidence has accumulated that removal or reduction of this population can also enhance immune responses against infectious microbes.7–14 More recently, Tregs in humans appear to suppress proliferation to self antigen as well as to foreign antigens.15

Because a hallmark of chronic HCV infection is impaired HCV-specific effector T cell responses,16 it is intriguing to hypothesize that Tregs may play a role with long-term persistence of HCV infection. In this study, we show that Tregs are a highly differentiated population that suppress HCV-specific CD4+ and CD8+ T cell responses in a dose-dependent, cell contact manner involving TGF-β1 and IL-10 and may play a role in viral persistence.

Abbreviations

HCV, hepatitis C virus; IFN, interferon; TNF-α, tumor necrosis factor alpha; IL-10, interleukin-10; TGF β1, transforming growth factor beta 1; Tregs, CD4+CD25+ regulatory T lymphocytes; PBMC, peripheral blood mononuclear cells; DC, dendritic cells; PHA, phytohemagglutinin antigen; HIV, human immunodeficiency virus; TT, tetanus toxoid.

Patients and Methods

Patients and Controls.

Sixty individuals were recruited using a protocol approved by the internal review board at the University of Florida. The study subjects included spontaneous recovered (n = 15), chronic HCV infection (n = 30), and healthy controls (n = 15). The recovered samples were from individuals with anti-HCV antibody, a history of acute HCV exposure with resolution, and negative HCV RNA by the transcription-mediated amplification qualitative assay.17 The chronic group represents treatment-naïve subjects, all of whom had a liver biopsy, HCV RNA level, and serum alanine aminotransferase. In addition, 5 patients with chronic infection undergoing liver transplantation were used to evaluate compartmentalization of CD4+CD25+ cells.

Peripheral Blood Mononuclear Cells, Lymph Nodes, and Liver Preparation.

Peripheral blood mononuclear cells (PBMC) were isolated from fresh heparinized blood as described previously.18 Liver and perihepatic lymph nodes obtained at the time of transplantation were processed with collagenase treatment. Tissue samples were dissected into 1-mm3 pieces, added to cRPMI 1640 medium and collagenase type IV at 100 μL/mL, and the mixture incubated at 37°C for 60 minutes. The supernatant was removed and diluted in cRPMI 1640 medium, centrifuged at 125g for 10 minutes.

HCV Peptide, Tetramers, and Antigen.

HLA-A0201–restricted peptides NS3 (CINGVCWTV) and NS5B (GLQDCTMLV) synthesized at ICBR Protein Chemistry (Gainesville, FL), and HCV core antigen peptide (aa 2-122; Biodesign, Kennebunk, ME) were purified and processed as described previously.18 PE-labeled HLA-A2 tetramer complexes were synthesized as described by Altman et al.19 HCV antigens were obtained from commercial sources (Chiron, Emeryville, CA): Core, NS3, NS4, NS4/5, and NS5. These antigens were pooled at a final concentration of 10 μg/mL for use in some assays.

Generation of Dendritic Cells.

Dendritic cells (DCs) were generated from PBMC as described by Romani et al.20

Preparation of PBMC, CD8+, CD4+, CD4+CD25+ and CD4+CD25.

The effect of CD4+CD25+ cells on HCV-specific IFN-γ responses was tested with 3 sets of PBMC: (1) PBMC; (2) PBMC depleted of CD4+CD25+ (PBMC – CD4+CD25+); and (3) PBMC supplemented with CD4+CD25+ (PBMC + CD4+CD25+). The PBMC – CD4+CD25+ subset was created by mixing CD4-depleted PBMC, obtained by negative CD4+ cell selection from whole PBMC (20–30 × 106) with CD4 microbeads (Miltenyi Biotec, Auburn, CA), and the CD4+CD25 fraction, obtained by depleting the negatively selected CD4+ cell fraction of CD25+ cells (%CD4+CD25+ cells < 10%), using positive selection beads (Miltenyi). The PBMC + CD4+CD25+ subset was established by adding the positively selected CD25+ cells from the negatively selected CD4+ cell fraction to PBMC at a 10:1 ratio (0.2 × 106 PBMC: 0.02 × 106 CD4+CD25+). The final %CD4+CD25+ cells/well in the PBMC + CD4+CD25+ subset ranged between 8% and 18%.

The effect of CD4+CD25+ cells on HCV-specific T cell proliferation was examined by using CD4+, CD4+CD25, and negatively selected CD8+ cells (Miltenyi) as effectors, when cocultured with CD4+CD25+ cells in CFSE and [3H]thymidine assays. Purity of each fraction was > 90% by flow cytometry.

CFSE Experiments.

CD8+ cells were incubated with 10 μmol/L CFSE (Molecular Probes, Eugene, OR) for 10 minutes at 37°C. Labeled CD8+ cells were used as effectors in mixing culture experiments. Four groups of mixing experiments were analyzed to determine the suppressive effect of Tregs on HLA-A2–restricted HCV-cytotoxic T lymphocytes generated from pure CD8+ cells as described previously.18 CFSE-labeled CD8+ cells were cocultured with DCs at a ratio of 20:1, 10 μg/mL peptide, and 10 ng/mL IL-7 (B&D) as well as 1 of 3 fractions: (1)CD4+CD25+; (2) CD4+CD25; or (3) CD4+ cells at a 10:1 ratio (1 × 106 CD8:0.1 × 106 CD4 fraction) . On day 7, cells were stained with PE-labeled HLA-2 tetramer complexes and analyzed by flow cytometry.

ELISpot and Enzyme-Linked Immunosorbent Assay.

An established ELISpot was used to evaluate IFN-γ HCV-specific responses in the 3 PBMC subsets.18 Enzyme-linked immunosorbent assays were used to measure TGF-β1 (Promega, Madison, WI) and IL-10 (Endogen, Woburn, MA) in 48-hour culture supernatants. Neutralizing experiments were performed by adding 2 μg monoclonal anti-human TGF-β1 (R&D Systems, Minneapolis, Minnesota) to 1 of the duplicate sets of the 3 PBMC subsets.

Proliferation and Transwell Studies.

The [3H]thymidine protocol was performed as described previously.18 In this assay, CD4+CD25 cells served as effectors. They were cocultured with DCs in the presence of pooled HCV antigens (10 μg/mL), phytohemagglutinin antigen (PHA; 5 μg/mL), human immunodeficiency virus (HIV) antigens (10 μg/mL), or in the absence of antigens. Treg-mediated suppression of HCV-specific CD4 proliferation was examined by mixing effectors with CD4+CD25+ at a 10:1 (2 × 105 effector: 0.2 × 105 Treg) and 2:1 (2 × 105 effector:1 × 105 Treg) ratio. The mean counts per minute (cpm) were calculated as mean cpm × 103 after subtracting the cpm obtained in the absence of antigen.

Transwell experiments were conducted to determine whether CD4+CD25+-mediated suppression of CD4+CD25 effectors requires cell contact. CD4+CD25+ cells (2 × 105) were stimulated in the upper chamber with autologous DCs (20:1 ratio) and pooled HCV antigens, and CD4+CD25 effectors (2 × 105) were stimulated in the lower chamber with DCs (20:1 ratio) and pooled HCV antigens. After 72 hours, [3H]thymidine was added to measure the responses of CD4+CD25 effectors in the outer chambers.

IL-10 Intracellular Cytokine Staining.

Intracellular cytokine staining was performed by use of the BD Cytofix/Cytoperm kit (BD Biosciences, San Diego, CA). Briefly, 5 × 105 CD4+CD25+ cells, and DCs were incubated at a ratio of 10:1 with medium, 10 μg/mL HIV antigens, or 10 μg/mL HCV pooled antigens at 37°C for 2 days. After adding the transport inhibitor BD Golgistop at 0.67 μL/mL, the culture was incubated at 37°C for 5 hours. After blocking with 10% human antibody serum, cells were labeled with CD4–FITC, CD25-cychrome and permeabilized. Intracellular staining was performed with PE-labeled IL-10 antibodies or isotype control. CD4+ and CD4+CD25 cells cultured with DCs in medium or with pooled HCV antigens also served as controls.

Flow Cytometry.

The following monoclonal antibodies were used: anti-CD4, anti-CD25, anti-CD8, anti-CD45RO, anti-CD45RA, anti-CD27, anti-CD28, anti-CD95, anti-CD69, anti-CD38, anti-CCR5, anti-CD62L, and anti-Ki67 (BD PharMingen). Staining was performed with FITC, PE, Cychrome, and APC-coupled antibodies. Acquisitions and data analysis were performed with a FACSCalibur (BD Biosciences) and CellQuest software.

Statistical Analysis.

The Mann-Whitney U test and linear regression analysis using the Spearman correlation coefficient were performed with the SPSS 10.1 program. All P values are 2-tailed.

Results

CD4+CD25+ Frequency and Distribution.

Prevalence of CD4+CD25+ cells in HCV infection was defined ex vivo from PBMC in chronic infection, self-limited infection, and normal controls by flow cytometry. A significantly higher proportion of CD4+CD25+ cells were found in those with chronic infection (median, 2.98%; mean, 3.02%; SE ± 0.11) when compared with recovered (1.68%, 1.64% ± 0.17, P = .001) and normal controls (2.30%, 2.27% ± 0.22, P = .02), suggesting a relationship with viral persistence (Fig. 1). Additionally, a lower prevalence of CD4+CD25+ cells was found in those with spontaneous resolution when compared with normal controls (P = .04). The distribution of CD4+CD25+ in PBMC, lymph node, and liver tissue was determined in 5 patients with chronic infection. The frequency of CD4+CD25+ cells was higher in the periphery (median, 5.97%; mean, 6.59% ± 0.93), followed by perihepatic lymph nodes (1.55%, 1.56% ± 0.50); they were least abundant in liver parenchyma (0.69%, 0.84% ± 0.23) (Fig. 2).

Figure 1.

An ex vivo comparison of peripheral blood CD4+CD25+ T cell frequency in spontaneous resolution (n = 15), chronic infection (n = 30), and normal control (n = 15) subjects. Data are expressed as box plots, in which the horizontal lines illustrate the 25th, 50th, and 75th percentiles of the frequencies of the CD4+CD25+ T lymphocytes as assessed by flow cytometry. The vertical lines represent the 10th and 90th percentiles, and circles denote values outside these percentiles. The P values were calculated by using the nonparametric Mann-Whitney U test.

Figure 2.

Comparison of CD4+CD25+ T cell frequency in the liver, perihepatic lymph node, and peripheral blood of subjects with chronic infection at the time of liver transplantation (n = 5). Data are expressed as box plots, in which the horizontal lines illustrate the 25th, 50th, and 75th percentiles of the frequencies of the CD4+CD25+ T lymphocytes as assessed by flow cytometry. The vertical lines represent the 10th and 90th percentiles, and circles denote values outside these percentiles. The P values were calculated by using the nonparametric Mann-Whitney U test.

CD4+CD25+ Cells Display CD45ROhigh, CD45RAlow, CD27high, CD28high, CD62Lhigh, CD95high, CD69low, and CCR5low Phenotype.

Cell surface phenotypes of CD4+CD25+ in spontaneous resolution, chronic infection, and normal controls were compared by phenotype, level of activation, differentiation, memory, and homing marker with flow cytometry (Table 1). CD4+CD25+ cells display CD45ROhigh, CD45RAlow, CD27high, CD28high, CD62Lhigh, CD95high, CD69low and CCR5low phenotype in peripheral blood. No significant phenotypic differences were found when comparing recovered, chronic infected, and controls. In addition, the phenotypes of the CD4+CD25+ lymphocytes in peripheral blood, lymph node, and liver tissue from 5 patients with cirrhosis from chronic infection showed no compartmental differences (data not shown).

Table 1. Phenotypic Expression Analysis of Circulating CD4+CD25+ T Cell in Spontaneous Recovered, Chronic Infected, and Normal Control Subjects as Determined by 4-Color Flow Cytometry
CD4+CD25+ PhenotypeNormal ControlsSpontaneous ResolutionChronic InfectionNote
  1. NOTE. PBMCs from recovered, chronic, and control subjects were stained with fluorochrome-labeled CD4, CD25, CD45RO, CD45RA, CD27, CD28, CD95, CD69, CCR5, and CD62L surface molecules. Isotype antibodies served as controls. No significant phenotypic differences are noted when comparing the 3 groups.

  2. Abbreviation: NS, not significant.

CD45RO+/CD45RA+9.089.169.71NS
CD45RO+/CD45RA−67.4767.6068.11NS
CD45RO−/CD45RA+20.6021.6119.74NS
CD45RO−/CD45RA−3.091.642.44NS
CD27+/CD28+69.4775.2370.72NS
CD27+/CD28−13.5316.5215.11NS
CD27−/CD28+15.517.4312.75NS
CD27−/CD28−1.490.821.42NS
CD95+/CD69+0.651.761.27NS
CD95+/CD69−69.5476.5171.10NS
CD95−/CD69+0.670.610.98NS
CD95−/CD69−29.1421.1326.65NS
CCR5+/CD62L+3.221.562.47NS
CCR5+/CD62L−2.001.751.76NS
CCR5−/CD62L+72.6873.5383.71NS
CCR5−/CD62L−22.1024.1513.06NS

Effect of CD4+CD25+ on HCV-specific CD4+ T Cell Immune Responses.

To demonstrate the effect of Tregs on HCV-specific CD4+ responses, PBMC depleted or enriched with CD4+CD25+ was analyzed by IFN-γ ELISpot and compared with whole PBMC responses in subjects with chronic infection and spontaneous recovery. As shown in Fig. 3, depletion of CD4+CD25+ increased HCV-specific IFN-γ activity in chronic infection in response to core (mean, 108 vs. 84, P < .01), NS3 (37 vs. 26, P < .01), NS4 (29 vs. 9, P < .01), NS4/5 (100 vs. 64, P < .01), and NS5 (40 vs. 18, P < .05); and in recovered subjects in response to NS3 (645 vs. 295, P < .01), NS4/5 (591 vs. 468, P < .05) and NS5 (328 vs. 113, P < .05). Addition of CD4+CD25+ to PBMC at a 1:10 ratio (≈10% of CD4+CD25+) reversed the mean number of HCV-specific IFN-γ spots below baseline in response to HCV antigens. Overall, HCV-specific IFN-γ responses were enhanced by depletion of CD4+CD25+ and inhibited by the addition of these regulatory T lymphocytes.

Figure 3.

CD4+CD25+ T cells suppress HCV-specific IFN-γ secretion. HCV-specific IFN-γ responses were assessed with the IFN-γ ELISpot using 3 PBMC cultures (PBMC, PBM − CD4+CD25+; PBMC + CD4+CD25+) that were mixed with HCV antigens, positive controls (PHA, TT), or negative controls (medium, HIV) in subjects with (A) chronic infection (n = 30) and (B) spontaneous recovery (n = 15). The magnitude of the bars represent the mean number of IFN-γ spots per 2 × 105 mononuclear cells from the 3 PBMC cultures after subtracting the number of IFN-γ spots from their negative control wells. Although depletion of CD4+CD25+ T cells (PBMC − CD4+CD25+ T cells) enhances HCV-specific IFN-γ responses, supplementation of whole PBMC with CD4+CD25+ T cells (PBMC + CD4+CD25+ T cells) suppresses HCV-specific IFN-γ responses when compared with whole PBMC. *P < .05 and **P < .001 calculated with the Mann Whitney U test. PBMC, peripheral blood mononuclear cells; PHA, phytohemagglutinin antigen; HIV, human immunodeficiency virus; TT, tetanus toxoid.

The regulatory properties of CD4+CD25+ T cells were also examined with 2 fractions created from total CD4+, CD4+CD25+, and CD4+CD25, by [3H]thymidine proliferation in chronic infection (n = 30). In the first series of tests, CD4+CD25 cells were used as effectors and cultured with autologous DCs and pooled HCV antigens (Fig. 4). CD4+CD25 cells were then mixed with CD4+CD25+ cells at indicated ratios (10:1, 2:1) and stimulated with pooled HCV antigens and DCs. CD4+CD25+ cells significantly inhibited HCV-specific CD4+CD25 proliferation when added at a 10:1 (mean 18 × 103 cpm vs. 30, P < .05) and 2:1 ratio (7.5 vs. 30, P < .01) when compared with the control CD4+CD25 cell culture in the absence of Treg cells. The degree of suppression increased with CD4+CD25+ cell number added to the coculture, suggesting a dose-dependent relationship.

Figure 4.

CD4+CD25+ T cell immunosuppression of HCV-specific CD4+ T cell proliferation is dose dependent. CD4+CD25 T cells from subjects with chronic infection (n = 30) were used as effector cells to examine HCV-specific CD4+ T cell proliferation as measured by 3H-thymidine incorporation in response to pooled HCV antigens in coculture with CD4+CD25+ T cells at various ratios (10:0, 10:1, and 2:1). Positive (PHA) and negative controls (HIV antigens) are included. Results are expressed as mean counts per minute (cpm × 103) after subtracting the cpm in the absence of antigen. CD4+CD25+ T cells suppress HCV-specific CD4+CD25 proliferation in a dose-dependent manner. To determine whether the suppression was cytokine mediated and/or cell-contact requiring, transwell studies were performed. Suppression of HCV-specific CD4+CD25 proliferation was abolished after separating the 2 cell fractions (CD4+CD25 and CD4+CD25+) with a transwell insert, suggesting that suppression is cell–cell contact dependent. *P < .05, **P < .01 by the Mann-Whitney U test. PHA, phytohemagglutinin antigen; HIV, human immunodeficiency virus; HCV, hepatitis C virus.

Effect of CD4+CD25+ on HCV-Specific CD8+ T Cell Immune Responses.

Two HLA-A2–restricted-peptides (NS3 and NS5B) and autologous DCs were used to generate cytotoxic T lymphocytes from pure CD8 cells of HLA-A2+ chronic patients. The CFSE dye was used to assess the proliferative capacity of HCV-specific CD8+ T cells.21 CD8+ T cells were stained with CFSE, and HCV-specific CD8+ T cells were quantified by determining the CFSE content of tetramer-positive cells after 7 days of in vitro stimulation with peptide and coculture DCs as well as CD4+ T cell fractions (CD4+, CD4+CD25, or CD4+CD25+). The results of a representative patient with chronic infection are shown in Fig. 5. After 7 days of coculture, NS3- and NS5-specific CD8+ T cell proliferation are 16.0% and 16.4% in the presence of a normal CD4 cell population. Addition of CD4+CD25+ cells decreased CD8+ T cell NS3-specific proliferation to 10.7% and NS5B-specific proliferation to 12.1%. Coculturing with CD4+CD25 cells increased CD8+ T cell NS3-specific proliferation to 27.2% and NS5B-specific proliferation to 28.8%. No NS3 or NS4 tetramer-positive proliferation was observed with HIV peptide (0.5%-1.2% CFSE signal in the left upper quadrant) when used as a negative control.

Figure 5.

CD4+CD25+ T cells suppress HCV-specific CD8+ T cell proliferation. CFSE assay for a representative patient with chronic infection. CD4+CD25+ T cells suppress proliferation of HCV-specific tetramer+ CTLs. Two HLA-A2–restricted peptides (NS3 and NS5B) were used to generate CTLs from pure CD8+ cells of HLA-A2+ subjects. HCV-specific CD8+ T cells proliferation were evaluated with CFSE, anti-CD8, and PE-HLA-A2 tetramers. (A) CFSE-labeled tetramer+ CTLs as detected after a 7-day culture; with unlabeled CD4+ T cells, with unlabeled CD4+CD25 T cells and with unlabeled CD4+CD25+ T cells. CD8+ T cells were cocultured with the various CD4 fractions at 10:1 ratio (1 × 106 CD8:0.1 × 106 CD4 fraction).The CFSE signal in the left upper quadrant represents the percentage of HCV-specific CTLs that have proliferated during the 7-day culture. Dot plots were generated by gating for lymphocytes. (B) Histogram plot showing the CFSE signal intensity of the CD8 T cells in the presence of CD4+, CD4+CD25, and CD4+CD25+ T cells when using the NS3 and NS4 tetramers.

CD4+CD25+ Secrete TGF-β1 and IL-10.

To evaluate for cytokine release from Tregs, ELISA was performed for TGF-β1 and IL-10 by using the supernatants from the 3 PBMC cultures (PBMC, PBMC − CD4+CD25+, and PBMC + CD4+CD25+) in chronically infected subjects (n = 30). As shown in Fig. 6, cytokine analysis showed lower levels of TGF-β1 and IL-10 in the supernatants of PBMC depleted of CD4+CD25+ versus baseline PBMC (for TGF-β1: core 75 vs. 173, P < .05; NS4/5 139 vs. 272, P < .01; NS5 114 vs. 270, P < .01; PHA 264 vs. 289, HIV 18 vs. 19; for IL-10: core 38 vs. 57, P < .01; NS3 37 vs. 59, P < .05; NS4/5 30 vs. 44, P < .05, NS5 23 vs. 38 P < .01; PHA 60 vs. 75, HIV 6 vs. 8). In addition, significantly higher levels of both cytokines were noted in PBMC enriched with CD4+CD25+ on comparison with their respective whole PBMC. These results suggest that CD4+CD25+ regulatory T cells secrete significant amounts of TGF-β1 and IL-10.

Figure 6.

CD4+CD25+ T cells in chronic infection (n = 30) produce TGF-β1 and IL-10. Enzyme-linked immunosorbent assay cytokine analysis for TGF-β1 and IL-10 was performed on culture supernatants from the 3 PBMC cultures representing whole PBMC, PBMC depleted of CD4+CD25+ cells (PBMC − CD4+CD25+), and PBMC supplemented with CD4+CD25+ cells (PBMC + CD4+CD25+). Bars represent the difference in cytokine production after HCV antigen and negative control antigen stimulation. (A) Lower levels of TGF-β1 are found in CD4+CD25+ depleted PBMC cultures and higher levels noted in PBMC cultures supplemented with CD4+CD25+ T cells in comparison with whole PBMC cultures in subjects with chronic infection. (B) IL-10 levels are also lower in PBMC − CD4+CD25+ T cell cultures and higher in PBMC + CD4+CD25+ T cell cultures when compared with whole PBMC. *P < .05 and **P < .01 calculated using the Mann-Whitney U test. TGF-β, transforming growth factor beta; PBMC, peripheral blood mononuclear cells; IL, interleukin.

Neutralization of CD4+CD25+ Suppression with Anti-TGF-β1.

To determine whether TGF-β1 was an immunoregulatory cytokine involved in the suppressive effect by CD4+CD25+ T cells, neutralizing experiments using anti-TGF-β1 antibody were performed in the chronically infected subjects (n = 15). The effect of introducing anti-TGF-β1 on the mean number of HCV-specific IFN-γ spots was assessed by using HCV-stimulated PBMC, PBMC - CD4+CD25+, and PBMC + CD4+CD25+ cell cultures and compared with their respective subsets in the absence of anti–TGF-β1 (Fig. 7). Addition of anti–TGF-β1 increased HCV-specific IFN-γ responses in the stimulated PBMC (29 vs. 15, P < .01) and PBMC + CD4+CD25+ (35 vs. 13, P < .01) cultures when compared with stimulated subsets in the absence of anti–TGF-β1. Although addition of anti–TGF-β1 abrogated CD4+CD25+-mediated suppression in these cultures, no reversal was found with an isotype antibody (P > .87), confirming that TGF-β1 was a critical mediator of this suppressive effect. Introduction of anti–TGF-β1 into the mononuclear cells depleted of CD4+CD25+ did not elicit any significant change in HCV-specific IFN-γ response when compared with the PBMC − CD4+CD25+ culture in the absence of anti–TGF-β1 (P > .65)

Figure 7.

TGF-β1 neutralization reverses CD4+CD25+–mediated suppression in chronic infection (n = 15). HCV-specific IFN-γ response in the IFN-γ ELISpot by whole PBMC, PBMC − CD4+CD25+, and PBMC + CD4+CD25+ cultures in response to anti-TGF-β1 antibody alone (diagonal striped column), pooled HCV antigens in the absence of a blocking antibody (white columns), presence of an isotype antibody (horizontal striped column), and after introduction of anti-TGF-β1 antibody (black column) in chronic infection. Introduction of anti-TGF-β1 antibody significantly enhances HCV-specific IFN-γ responses in PBMC and PBMC + CD4+CD25 cultures when compared with those observed in PBMC and PBMC + CD4+CD25+ cultures in the absence of a neutralizing antibody. HCV-specific response is not increased by anti–TGF-β1 antibody in PBMC cultures depleted of CD4+CD25+. **P < .01 by the Mann-Whitney U test. HCV, hepatitis C virus; IFN-γ interferon gamma; PBMC, peripheral blood mononuclear cells.

CD4+CD25+-Mediated Suppression Is Cell-Contact Dependent.

Having established that CD4+CD25+ T cells mediate their suppressive effect through TGF-β1 secretion, we then investigated whether such a suppression requires cell contact between CD4+CD25+ cells and their targets in chronic infection. Using a transwell culture system, CD4+CD25+ cells in the inner well and CD4+CD25 targets cells in the outer well were both cultured with pooled HCV antigens and autologous DCs. Suppressor function of the CD4+CD25+ was measured after a 72-hour incubation by assessing CD4+CD25 [3H]thymidine incorporation. Whereas coculture of CD4+CD25 and CD4+CD25+ cells at increasing ratios (10:1, 2:1) resulted in significant suppression of HCV-specific CD4+CD25 proliferation, CD4+CD25+ cells failed to suppress through a membrane (Fig. 4). This finding suggests that CD4+CD25+ T cell immunosuppression is cell–cell contact dependent.

CD4+CD25+ Respond to HCV Antigens via IL-10 Secretion.

To evaluate for viral specificity, we performed intracellular cytokine staining for IL-10 on CD4+CD25+ T cells (>95% purity by flow) that were cocultured with autologous DCs and either pooled HCV antigens or HIV antigens (negative control). Intracellular cytokine staining results showed HCV-specific IL-10 production that ranged between 1.9% and 5.3% compared with no spontaneous, PHA-, or HIV-stimulated response in the 5 chronically infected patients evaluated. An example of the specificity of the HCV-specific IL-10 production for a representative patient is shown in Fig 8. The absence of detectable IL-10 by intracellular cytokine staining in the total CD4+ and CD4+CD25 fractions support the existence of CD4+CD25+ cells responding directly to HCV antigens by IL-10 production. The failure of PHA alone to stimulate the CD4+CD25+ IL-10 production further illustrates its HCV specificity and is in agreement with studies classifying Treg as unresponsive (anergic) to nonspecific stimulation.22, 23

Figure 8.

A minority of CD4+CD25+ regulatory T cells in chronic infection exhibit HCV-specificity. Pure CD4+CD25+ T cells cocultured with autologous monocyte-derived dendritic cells produce IL-10 in response to pooled HCV antigens by intracellular cytokine staining. (A) Intracellular cytokine staining of CD4+CD25+ regulatory T cells for IL-10 in a subject with chronic infection. Pure CD4+CD25+ T cells cultured with PHA, HIV antigens and in the absence of antigen showed no IL-10 production highlighting their HCV-specificity. (B) Coculture of CD4+ and CD4+CD25+ fractions with pooled HCV antigens did not elicit IL-10 production. IL, interleukin; PE,; PHA, phytohemagglutinin antigen; HIV, human immunodeficiency virus; HCV, hepatitis C virus.

Correlation With Clinical Parameters.

Analysis of chronic infected patients (n = 30) showed no correlation between CD4+CD25+ T cell frequency and ALT level and fibrosis scores (Fig. 9). However, there was a positive correlation between CD4+CD25+ Treg frequency and HCV RNA titer (R = 0.457, P = .011) and an inverse relationship with histologic inflammatory scores (R = −0.67, P = .001) by the nonparametric Spearman rank.

Figure 9.

Correlation analysis between CD4+CD25+ regulatory T cell frequency in chronic infection and (A) serum alanine aminotransferase level, (B) HCV RNA viral load, (C) histological inflammatory grade, and (D) fibrosis score. A significant positive correlation is found between CD4+CD25+ regulatory T cell frequency and HCV RNA titer (R = 0.457, P = .011) , and a negative correlation is observed with inflammatory scores (R = −0.67, P = .001) by the nonparametric Spearman rank. ALT, alanine aminotransferase; HCV, hepatitis C virus; bDNA, branched DNA; HAI, histological activity index.

Discussion

In this study, we provide an extensive analysis of the frequency, phenotype, and function of CD4+CD25+ T lymphocytes in HCV infection. We postulate that CD4+CD25+ T cells play a role in suppressing CD4+ and CD8+ T cell responses to HCV, and thereby promote HCV persistence. In support of this hypothesis, we found a significantly higher proportion of CD4+CD25+ T cells in chronically infected patients compared with spontaneously recovered and normal controls. This finding also raises the possibility that the steady-state level of regulatory CD4+CD25+ cells before infection may predetermine and/or predict whether individuals will clear or develop chronic infection. Furthermore, our results indicate that CD4+CD25+ T cells suppress IFN-γ secretion and proliferation of HCV-specific T cells in a dose-dependent and cell-contact fashion. Surprisingly, a small number of CD4+CD25+ T cells can respond directly to HCV antigens through cytokine production. The demonstration of “suppressive” HCV antigens is novel and implies a viral-specific Treg response. These data implicate the involvement of CD4+CD25+ T cells in maintenance of chronic HCV infection.

The distribution of CD4+CD25+ T cells suggest that these cells accumulate mainly in peripheral blood and not the actual site of infection, the liver. Typically, HCV-specific CD4+ and CD8+ T cells are compartmentalized to the liver and not the periphery.24, 25 Thus, we had anticipated that these peripherally induced regulatory T cells would be concentrated at the site of inflammation or antigen presentation. This discrepancy is difficult to explain in the context of our current understanding of HCV and regulatory T cell homeostasis and suggests a potential interfering role for HCV. It is possible that an intrahepatic deficiency in this regulatory cell leads to progressive liver injury. An important limitation of the data is that the population sampled represents a selected group with heavily fibrosed liver tissue obtained at the time of liver transplantation. Such distorted liver architectures are well known to have lower numbers of intrahepatic lymphocytes, which may contribute to the lower percentage of CD4+CD25+ T cells in the liver parenchyma of this group.

Phenotypic analysis of CD4+CD25+ T cells showed that that they are CD27high, CD28high CD45ROhigh, CD45RAlow CD62Lhigh CD95high, CD69low and CCR5low. This suggests that CD4+CD25+ T cells are an early antigen-primed memory T cell population and susceptible to apoptosis. In T cell activation and clonal expansion, the costimulatory receptor CD28 and its closest structural relative, the de-activating receptor cytotoxic T lymphocyte antigen-4 (CTLA-4), are of central importance.26 CTLA-4 expression is primarily restricted to CD4+CD25+ T cells, where it is constitutively expressed, serving the functional role of the primary costimulatory signal.27 Indeed, blockade studies of CTLA-4 abrogating the ability of CD4+CD25+ T cells to suppress T cell–mediated immune responses reveal its critical role in regulating T cell responses.28 Lin et al. found that the response of regulatory T cells to conventional costimulation was ineffective in comparison with conventional T cells, but readily expanded in vitro with “superagonistic” CD28-specific monoclonal antibody without the need for T cell receptor engagement.29 Therefore, costimulatory receptors CD28 and CD27 expression on CD4+CD25+ regulatory T cells is probably not a reflection of function as observed on general CD4+ T cells. Most of the CD4+CD25+ T cells express the CD45RO isoform, which is characteristic of memory T cells, and few express CD45RA, a marker for naïve T cells. The upregulation of CD95, a Fas apoptotic receptor or late activation marker, and CD62L, a lymph node homing marker, support that they are a differentiated population. Lastly, we found that very few cells express CCR5, which is preferentially expressed on lymphocytes with a TH1 cytokine profile.

Current data suggest that once Tregs are activated, they inhibit both CD4+ and CD8+ T cell responses in an antigen-nonspecific manner.30, 31 Our results indicate that CD4+CD25+ cells secrete IL-10 and TGF-β1, and that anti–TGF-β1 reverses their suppressive effect. This confirms an inhibitory role for TGF-β1. Furthermore, our findings suggest CD4+CD25+ T cell–mediated suppression of HCV-specific T cell response is dose-dependent and requires cell contact. Recent mechanistic investigations have shed some light on this juxtaposition requirement by finding expression of the cell surface form of TGF-β1 by regulatory CD4+CD25+ T cell and expression of the TGF-β receptor by target CD4 and CD8 T cells.32 Importantly, a minority of Treg display HCV specificity to IL-10 secretion, which may play a regulatory role in promoting viral persistence. This is consistent with the findings of Belkaid et al., who detected a key role for IL-10 in their model.7 Their use of IL-10–deficient mice or blockade of IL-10 resulted in a sterilizing cure of Leishmania infection. This indicates that IL-10 secreted by CD4+CD25+ regulatory T cells may attenuate the function of macrophages to kill the intracellular parasite. IL-10 has also been shown to inhibit HCV-specific immunity when administered exogenously in patients with chronic HCV.18 Although the observed percentage of IL-10 producing Tregs is low, this finding may provide a novel mechanism used by HCV to evade the immune system and have a significant implication in the arena of vaccine development.

Expression of CD25 has been used as a useful marker of CD4+CD25+ regulatory T cells. An important caveat is that this immunophenotype CD4+CD25+ is also observed on activated CD4+ T cells, including those that are pathogen specific. This represents the major limitation of this study. The expression of CD25 occurs early during T cell stimulation, and its up-regulation has been used as a sensitive technique to detect HCV-specific T cells during in vitro HCV antigen stimulation.33 Thus, CD25 is a poor marker of Treg cells after T cell activation and is best used on a naïve T cell repertoire in an ex vivo fashion before any T cell stimulation as performed in this study and others.34 Although the surface expression of other molecules such as the TNF family molecule glucocorticoid-induced TNF receptor and CTLA-4 may serve as markers for Treg cells, an exclusive marker to identify this population of T cells remains elusive. Recently, foxp3 was reported potentially as an exclusive marker for CD4+CD25+ regulatory T cells in mice, but its relevance in human CD4+CD25+ Treg cells is not clear.23 Until an exclusive marker is found to identify CD4+CD25+ regulatory T cells, the best marker for their presence is demonstrating their suppressive properties on CD4 and CD8 T cell proliferation and cytokine production.

In conclusion, our results indicate that CD4+CD25+ T cells can respond directly to HCV antigens and suppress the HCV-specific immune response in a dose-dependent cell contact manner, which may promote HCV survival within the host. The higher frequency of CD4+CD25+ regulatory T cell in chronicity when compared with recovered subjects and its positive correlation with HCV viral load further supports its association with HCV persistence. Current investigation is underway to clarify the specificity of the antigenic determinants and the mechanism of their action. It is likely that antiviral vaccine candidates would want to avoid epitopes that lead to strong regulatory T cell stimulation. Conversely, the finding that CD4+CD25+ T cell frequency is inversely related to histological inflammatory activity underscores the potential role in suppressing pathological T cell–mediated immunological responses. Future investigation needs to be considered on how to manipulate the function of these regulatory T cells in the clinical setting. Possible manipulation strategies to enhance immune responses may include altering CD4+CD25+ Treg cell activity by means of cell depletion, CTLA-4 blockade, glucocorticoid-induced TNF receptor signaling, or blocking effector cytokine function.35

Ancillary