Intrahepatic virus-specific IL-10-producing CD8 T cells prevent liver damage during chronic hepatitis C virus infection

Authors


  • Potential conflict of interest: Nothing to report.

Abstract

CD8 T cell killing of hepatitis C virus (HCV)-infected hepatocytes is thought to contribute to liver damage during chronic HCV infection, whereas the participation of HCV-nonspecific immune cells is unclear. To visualize the spatial relationship of HCV-specific CD8 T cells with parenchymal target cells, and to examine their local functional activity in relation to hepatocellular necrosis and fibrosis, we used HLA tetramers and confocal microscopy in biopsies from 23 HLA-A2 or HLA-B7 patients with chronic HCV infection. Intrahepatic tetramer+ (HCV-specific) CD8 T cells protected from hepatic necroinflammatory disease activity, independently of age, gender, viral load, and viral genotype. Indeed, tetramer+ cells were scattered in the liver within regions of weak fibrosis (low laminin expression) and low hepatocellular apoptosis (TUNEL method), and expressed IL-10 but not IFNγ. By contrast, tetramer-negative CD8 T cells were associated with active necroinflammatory liver disease, colocalized with strong laminin expression and hepatocellular apoptosis, and expressed more frequently IFNγ than IL-10. Overall, liver regions harboring HCV-specific CD8 T cells tended to be healthier than areas containing only inflammatory cells of undefined specificity. In conclusion, HCV-specific IL-10-producing CD8 T cells, although not cytotoxic and unable to control viral replication, can attenuate hepatocellular necrosis, liver fibrosis, and inflammation mediated by bystander T cells, and may thus represent antigen-induced regulatory CD8 T cells. Therapeutic modulation of the intrahepatic balance between specific and bystander CD8 T cells might be beneficial in patients with chronic hepatitis C. (HEPATOLOGY 2006;44:1607–1616.)

Hepatitis C virus (HCV) is a leading cause of liver disease worldwide and is the main indication for liver transplantation in western countries (for review see Poynard et al.1). Despite a substantial virus-specific cellular immune response2, 3 up to 70% of patients become chronically infected, owing to rapid mutational escape and host immune impairment. During the course of persistent infection, continuous intrahepatic inflammation (hepatitis) maintains a cycle of hepatocyte destruction and regeneration that often ends in cirrhosis and hepatocellular carcinoma. Because HCV is hepatotropic but not cytopathic, it has been assumed that the liver damage it causes is due to recognition and destruction of HCV-infected hepatocytes by cytotoxic T lymphocytes (CTL), even though virus-specific CTL responses ultimately fail to prevent progression of HCV disease. Furthermore, the possible pathophysiological contribution of liver-infiltrating inflammatory cells is poorly understood: relevant studies have been hampered by the lack of a small-animal model of HCV infection, and by difficulties in studying the human liver in situ. Most immunological studies in this setting have thus focused on peripheral blood cells, which may not accurately reflect the composition and function of the hepatic infiltrate. The low frequency of HCV-specific CD8 T cells in the circulation of chronically infected patients4–8 has been attributed to their sequestration in the liver. This is based on work on intrahepatic lymphocytes isolated from liver biopsy material and studied either after in vitro amplification9–11 or directly ex vivo.12–14 However, the relationship between virus-specific CTL and disease severity is controversial.15, 16 Furthermore, there are no data on the distribution of virus-specific cells within the infected liver. As a result, the contribution of HCV antigen-specific mechanisms to chronic hepatocellular injury is largely unknown.2, 3

The in situ tetramer technique can be used to directly visualize virus-specific CD8 T cells in infected tissues, to determine their spatial relationship with parenchymal target cells, and to examine their local functional activity by comparison with other inflammatory cells.17–19 Using this technique, we describe here a population of intrahepatic HCV-specific interleukin (IL)-10-positive CD8 T cells, the presence of which is associated with protection from severe necroinflammatory liver disease during chronic HCV infection. These data suggest that immunotherapy based on virus-specific regulatory CD8 T cells might be beneficial to limit liver damage.

Abbreviations

HCV, hepatitis C virus; CTL, cytotoxic T lymphocytes; IL, interleukin; IFNγ, interferon gamma; PBMC, peripheral blood mononuclear cells; ALT, alanine aminotransferase.

Patients and Methods

Patients.

We studied 23 patients with chronic HCV infection, of whom 15 were HLA-A2 and 8 were HLA-B7. None of the patients had received interferon (IFN)-α or ribavirin during the previous 6 months. Their characteristics are shown in Table 1. Written informed consent was obtained from each patient and the study protocol was approved by the ethics committee of Hôpital La Pitié (Paris, France).

Table 1. Clinical Parameters of Patients With Chronic HCV Infection
PatientAgeGenderActivityaFibrosisaALTHCV RNAHCVEpitopesb
GradeStage(UI/mL)Log10 (kUI/mL)GenotypeIHLPBMC
  1. Abbreviations: F, female; M, male; ALT, alanine aminotransferase; ND, not determined; a, METAVIR score; b, number of epitopes recognized (of 7 for HLA-A2 and of 1 for HLA-B7); IHL, intrahepatic lymphocytes.

A2-142M12770162
A2-247M11496.381a40
A2-383M12644.761b20
A2-467F12205.581b40
A2-562F11425.671b54
A2-636M11995.6106
A2-771F11105.31b40
A2-849M21104.161b30
A2-952F21882.183a51
A2-1050F24783.433a20
A2-1148F11476.662a02
A2-1232F11102.78430
A2-1345M124125.61b00
A2-1435F21252.771b12
A2-1539M12240ND31
B7-158M241725.66410
B7-258F11265.461a11
B7-329F00745.091b11
B7-447M11496.381a10
B7-570F04145.99210
B7-649M12104.161b00
B7-752F21882.183a00
B7-850F24783.433a00

Peripheral blood mononuclear cells (PBMC) were collected, separated by Lymphoprep gradient centrifugation (Biowest, Oslo, Norway) and cryopreserved for further analysis. The same day, all the patients underwent percutaneous needle biopsy of the liver. Part of the sample was processed for diagnostic evaluation and part was frozen for later in situ tetramer staining. Histological findings were graded with the METAVIR activity score (necrosis and inflammation) and the METAVIR fibrosis score.20 Serum HCV RNA levels and HCV genotypes were determined by using the Versant HCV RNA 3.0 (bDNA) (Bayer Diagnostics, Berkeley, CA) and INNO-LiPa HCV II (Innogenetics, Gent, Belgium) tests, respectively. Controls consisted of 10 healthy blood donors (PBMCs only) and 20 HLA-A2- and B7-negative chronically HCV-infected patients (PBMCs and liver biopsies).

MHC Class I Tetramers, Antibodies and Flow Cytometry.

HLA-A2 and -B7 tetramers were produced as described in Altman et al.21 Peptides were synthesized by NeoSystems (Strasbourg, France). The tetramers included HLA-A2 tetramers loaded with HCV Core35-45 (YLLPRRGPRL) and Core132-141 (DLMGYIPLV), and HCV non-structural protein NS3-11073-1081 (CINGVCWTV), NS3-21406-1415 (KLVALGINAV), NS4B1807-1816 (LLFNILGGWV), NS5B2594-2602 (ALYDVVTKL), NS5B2727-2736 (GLQDCTMLV), and HLA-B7 tetramer loaded with Core41-49 (GPRLGVRAT). HLA-A2 and HLA-B7-restricted pp65 CMV tetramers were used as controls. Each tetramer was titered individually on polyclonally expanded intrahepatic lymphocytes and used at the optimal concentration (2.5-5.0 μg/mL). For staining, 106 PBMC were incubated at 37°C for 30 minutes in the dark with one APC-labeled and one PE-labeled tetramer, then with CD8-FITC (Caltag Laboratories, Burlingame, CA) and 7-AAD (BD Pharmingen, San Diego, CA) for 20 minutes at 4°C. Data were collected on a FACSCalibur flow cytometer within one hour after staining, and analyzed using Cell Quest software (Becton Dickinson, San Diego, CA). Lymphocytes were gated according to forward/side scatter profiles, then CD8high cells were selected and 7-AAD was used to exclude dead cells. Results were considered positive above a cutoff value of 0.01% (99th percentile of staining in 10 HLA-A2+ healthy controls).

In situ Tetramer Staining.

Liver biopsy samples were immersed in Tissue-Tek OCT (Bayer Corporation, Pittsburgh, PA) and quick frozen in liquid nitrogen. Sections (5 μm) were cut and fixed in acetone for 10 minutes at −20°C. Nonspecific binding sites were blocked for 30 minutes at room temperature with 5% bovine serum albumen (BSA) in phosphate-buffered solution (PBS), followed by the Biotin-Blocking System (Dako, Glostrup, Denmark) and washing with PBS. Sections were first stained with Cy3-labeled CMV-specific tetramer for 4 hours at room temperature, then overnight with FITC-labeled HCV-specific tetramer loaded with the relevant epitope (5-10 μg/mL diluted in 2% fetal calf serum in PBS). Sections were washed 3 times for 10 minutes in PBS with 0.1% Triton X-100 (Sigma, St. Louis, MO), and incubated for 90 minutes on ice with mouse anti-human CD8 (clone DK25, Dako) diluted 1/100. After 3 washing steps with PBS, sections were incubated for 45 minutes at room temperature with Cy5-labeled goat anti-mouse (Jackson Laboratories, Bar Harbor, ME) diluted 1/50. Slides were mounted in Vectashield medium (Vector Laboratories, Burlingame, CA) and analyzed using a confocal laser scanning microscope with photomultipliers (TCS SPS AOBS model, Leica, Mannheim, Germany). Cells were analyzed in a minimum of 50 Z stacks per liver section by an operator who was unaware of the clinical data. Images were acquired using a 63x/1.43 numerical aperture oil objective and Leica confocal software TCS SP2 (V2.61.1537), and were analyzed using ImageJ software version 1.36b (http://rsb.info.nih.gov/ij). To increase the specificity of the assay, in each section the Cy3-labeled CMV-specific tetramer was used as internal control, to correct for patient/tissue-dependent background staining due to inter-individual differences. Each experiment also included a biopsy sample from an HLA-A2-negative/HLA-B7-negative HCV-infected patient as additional negative control, to correct for tetramer-dependent background staining. Data are presented after background correction.

Hepatocellular apoptosis was measured by using the TUNEL (TdT-mediated dUTP nick-end labeling) TMR Red Apoptosis kit (Roche Molecular Biochemicals, Mannheim, Germany). Unfixed sections were first stained with a mixture of FITC-labeled HLA-A2 or B7 tetramers (10 μg/mL total) for 4 hours at room temperature, followed by incubation with anti-CD8 and Cy5-labeled goat anti-mouse antibodies. Sections were then fixed with 10% formaldehyde, permeabilized with 0.1% Triton X-100, 0.1% citrate sodium buffer and incubated with the TUNEL reaction mixture containing terminal deoxynucleotidyl transferase for 45 minutes at 37°C. The number of positive cells was expressed relative to the number of CD8 T cells in the same Z stack. The negative control was generated by substituting PBS for TdT enzyme.

For IFNγ, IL-10 and extracellular matrix staining, acetone-fixed slides were stained with a mixture of HLA-A2 tetramers, followed by rabbit anti-CD8 (SP16, NeoMarkers Labvision, Fremont, CA) and mouse anti-IFNγ (BD Pharmingen), rat anti-IL10 (Biolegend, San Diego, CA) or mouse anti-laminin (Novocastra, Newcastle upon Tyne, UK) antibody for 90 minutes on ice, followed by Cy3-labeled goat anti-rabbit and Cy5-labeled goat anti-mouse or Cy5-labeled donkey anti-rat. Laminin expression was quantified as a percentage of the volume stained per Z stack.

Statistical Analysis.

The non parametric Mann-Whitney rank sum test was used to assess differences between groups. Associations between variables were identified with Fisher's exact test. Statistical significance was assumed when P (two-tailed) was less than .05. Multiple regression was used to test whether correlations between CD8 responses and liver disease activity were independent from other risk factors. Analysis of variance (ANOVA) was used to assess the relationship between the presence of tetramer+ cells and local liver fibrosis or hepatocellular apoptosis.

Results

Liver Homing of HCV-Specific CD8 T Cells.

Using a panel of HLA-A2 and HLA-B7 tetramers specific for eight epitopes of structural and non-structural HCV proteins, we observed very low frequencies of HCV-specific CD8 T cells in the peripheral blood of the 23 chronically infected patients, in keeping with previous studies.13, 14, 22 HCV-tetramer+ cells were detected in the peripheral blood of six HLA-A2+ (40%) and two HLA-B7+ (25%) patients, at frequencies ranging from 0.05% to 2% of total CD8 T cells (Supplementary Fig. 1; Supplementary material is available at: http://interscience.wiley.com/jpages/0270-9139/suppmat/index.html). The median magnitude of HCV-specific CD8 T cell responses, calculated as the sum of HCV-tetramer+ cells frequencies per patient (with values below the cutoff considered as zero), was 0.01% (range 0.0 to 2.3%). Peripheral HCV-specific CD8 T cells targeted a mean of 1.2 epitope per patient.

We then analyzed liver biopsy specimens obtained on the same day as peripheral blood samples, by means of tetramer staining and confocal laser scanning microscopy. CD8+ and tetramer+ cells were counted in a minimum of 50 Z sections per patient and per tetramer, and the percentage of tetramer+ cells among all CD8+ cells was calculated. We observed two distinct distribution patterns of CD8 T cells in the liver, consisting of extensive tetramer-negative CD8 cell infiltrates, and sparse tetramer+ cells scattered among few CD8 T cells (Fig. 1; Supplementary Fig. 2). This localization of HCV-specific cells outside areas of large CD8 infiltrates was independent of the HCV epitope. HCV-tetramer+ cells were more frequent in the liver than in peripheral blood: they were observed in the liver of 12/15 (80%) HLA-A2+ patients and 5/8 (62.5%) HLA-B7+ patients. The median magnitude of the intrahepatic HCV-specific CD8 T cell response was 2.14% (range 0.00-11.43%, P = .0027 versus peripheral tetramer+ cells). NS31073 and NS31406-specific CD8 cells were the most frequent tetramer+ cells (NS31073: mean 1.06% of all liver CD8 T cells, versus 0.18% in the periphery, P = .012; NS31406: mean 1.29% versus 0.14%, P = .025), while NS5B2596 tetramer+ cells were rarely observed (Fig. 2). This enrichment of tetramer+ lymphocytes in the liver relative to the periphery ranged from 0.7-fold for Core132 to 9.6-fold for NS31406.

Figure 1.

In situ tetramer staining of HCV-specific CD8 T cells in liver samples from chronically HCV-infected patients. (A) Images of single channels are shown on the left (CD8- red; HCV tetramer-green; CMV tetramer-violet; differential interference contrast (DIC)-gray scale). A larger 3-channel merged image is shown on the right. Double-stained HCV-specific CD8 T cells appear in yellow. Use of CMV tetramer as internal control significantly increased the specificity of the method. Bar = 10 μm. Compared to large tetramer-negative CD8 cell infiltrates (B), clusters of tetramer+ CD8 lymphocytes are located within regions of mild CD8 infiltration (C).

Figure 2.

The frequency and magnitude of HCV-specific CD8 cells are higher in the liver than in peripheral blood. Each line joining a pair of circles representing peripheral blood (PBMC) and liver-infiltrating lymphocytes (LIL) represents a single patient. Statistical analysis was based on Wilcoxon's test for paired data, including values below the cutoff (negative results shown arbitrarily as 0%).

Taken together, these results indicate that the frequency and magnitude of HCV-specific CD8 T cell responses are higher in the liver than in the periphery of patients with chronic hepatitis C.

The Magnitude of Intrahepatic HCV-Specific Responses Correlates Negatively With Hepatic Disease Activity.

We then examined whether intrahepatic HCV-specific CD8 T cells contribute to the liver damage observed during chronic infection. Unexpectedly, the magnitude of intrahepatic HCV-specific responses correlated negatively with METAVIR necroinflammatory activity score (r = −0.6, P = .006). The frequency of intrahepatic tetramer+ cells was significantly higher in patients with mild hepatitis, i.e., those with activity scores <2, than in patients with severe hepatitis, i.e., those with activity score =2 (median: 3.58% and 0.01% of CD8 T cells, respectively, P = .0035) (Fig. 3A), as confirmed by regression analysis (P = .009) (Fig. 3B). Importantly, this effect was independent of other progression factors such as age, gender, viral load, and the viral genotype, as shown by logistic regression analysis (OR= 0.07, 95% confidence interval = 0.005-0.086, P = .038). The NS31406 epitope was particularly strongly associated with protection against severe chronic hepatitis, as none of the patients with NS31406-specific intrahepatic CD8 T cells had high activity scores (P = .044 vs. other patients). There was no correlation between the presence or magnitude of intrahepatic HCV-specific CD8 responses and alanine aminotransferase (ALT) levels or HCV RNA load.

Figure 3.

Intrahepatic, but not peripheral, HCV-specific CD8 cells are associated with a low disease activity score. (A) The sum of tetramer+ CD8 cells per patient was analyzed according to the METAVIR activity score. Each circle represents the sum of HCV-specific CD8 T cells (empty circles: LIL; black circles: PBMC). Values below the cutoff (arbitrarily 0%) are shown on the zero line. Results are given on a log10scale. Statistical analysis used the Mann-Whitney test. (B) Regression analysis of the relationship between intrahepatic HCV tetramer+ cells and activity score (P = .009) (Each circle represents one patient. (C) The total number of intrahepatic CD8 T cells correlated with the METAVIR fibrosis score (linear regression analysis, P = .034).

By contrast, the total number of intrahepatic CD8 T cells was significantly higher in patients with necroinflammatory activity score =2 than in patients with lower scores (mean CD8 cell counts determined in 50 Z sections, 763 and 343, respectively, P = .001). This number also correlated with the ALT level (r = 0.6 Spearman test, P = .003). Taken together, these results suggest that some liver-infiltrating CD8 T cells specific for HCV epitopes play a beneficial role in chronic HCV infection by preventing progression to severe necroinflammatory liver disease independently from the control of viral replication, while CD8 T cell infiltrates appear to participate actively in liver damage.

HCV-Specific CD8 Cells Are Located in Liver Areas With Low Hepatocellular Apoptosis and Low Liver Fibrosis.

To investigate the potential cytotoxicity of intrahepatic CD8 cells, we assessed hepatocyte apoptosis by using a TUNEL assay. Little TUNEL labeling was seen in the vicinity of tetramer+ CD8 T cells. By contrast, the number of TUNEL-positive cells was markedly increased in areas showing abundant CD8 T cell infiltration (Fig. 4A-B; Supplementary Fig. 3). The number of CD8 T cells correlated strongly with the number of TUNEL+ cells present in the same Z stack (r = 0.7, P < .0001) (Fig. 4C). This correlation was strengthened when stacks containing tetramer+ cells were excluded (r = 0.79, P < .0001), but was not significant when tetramer+ cells were considered. The absolute number of TUNEL+ cells was lower in stacks containing tetramer+ cells than in stacks containing only tetramer-negative CD8 lymphocytes (median 4 and 8 TUNEL+ cells, respectively; P = .0006, Mann-Whitney test). Regression analysis confirmed this beneficial effect of HCV tetramer+ cells on hepatocellular apoptosis (P = .001, Fig. 4D).

Figure 4.

Intrahepatic HCV-specific CD8 cells are associated with low hepatocellular apoptosis. (A-B) Representative TUNEL assay on a liver section from a chronically infected patient (HCV tetramer: green; CD8: red; TUNEL: blue) Hepatocellular apoptosis is observed in the vicinity of a large CD8 infiltrate (A), but not in close vicinity of HCV tetramer+ (yellow) cells (B). (C) The number of CD8 cells correlated with the number of TUNEL+ cells present in the same Z stack (r = 0.7, Spearman test, P < .0001). (D) Bivariate linear regression analysis of the relationship between absence (0, black dots) or presence (1, open dots) of HCV tetramer+ cells and hepatocellular apoptosis (number of TUNEL+ cells in a liver region) (P = .001, dashed line) Apoptosis was less important in stacks containing tetramer+ cells than in stacks containing only tetramer-negative CD8 cells. Each dot represents the number of cells counted in one stack. Horizontal lines represent the median values of TUNEL+ cells.

To determine the possible relationship between CD8 T cell infiltration and liver fibrosis, we co-stained liver biopsy material with an antibody against laminin, a major extracellular matrix component whose abundance reflects the degree of liver fibrosis.23 We found that tetramer+ cells were scattered within regions of low laminin expression while large infiltrates of tetramer-negative CD8 T cells colocalized with strong laminin expression (Fig. 5A-B; Supplementary Fig. 4). The total number of CD8 T cells correlated strongly both with the METAVIR fibrosis score (linear regression, P = .034) (Fig. 3C) and with laminin expression in the corresponding liver region (r = 0.69, P < .0001) (Fig. 5C). In addition, laminin expression was stronger in regions containing only tetramer-negative CD8 cells than in regions that also contained tetramer+ cells (median volume, 3.65% and 1.32%, respectively, P = .0025) (Fig. 5D).

Figure 5.

Intrahepatic HCV-specific CD8 T cells are associated with low liver fibrosis. Staining of tetramer (green) and CD8 (red) was combined with staining for laminin (cyan). Large CD8 infiltrates colocalized with areas of strong laminin expression (A), while HCV-tetramer+ cells were scattered within regions of low laminin expression (B). (C) Laminin expression (expressed as a log10 percentage of the volume stained per Z stack, to normalize the distribution), correlated with the number of tetramer-negative CD8 cells in the corresponding liver region (r = 0.69 Spearman test, P < .0001). (D) Bivariate linear regression analysis (dashed line) of the relationship between absence (0) or presence (1) of HCV tetramer+ cells and liver fibrosis (log10 laminin expression) (P = .003). Each dot represents one stack. Horizontal lines represent the median values of laminin expression.

These results suggest that some HCV-specific CD8 T cells can attenuate the deleterious effect of CD8 T cell infiltrates, which contribute directly to liver damage by killing hepatocytes and promoting fibrosis.

Intrahepatic HCV-Specific CD8 Cells Produce IL-10 But Not IFNγ.

We then tested intrahepatic CD8 T cells for their expression of IL-10 and IFNγ. A majority of HCV-specific CD8 T cells stained positively for intracellular IL-10, while only a small fraction of tetramer-negative CD8 T cells expressed this regulatory cytokine (median percentage, 58% and 9.5%, respectively, P = .015) (Fig. 6). By contrast, variable percentage of tetramer-negative CD8 T cells stained positively for the proinflammatory cytokine IFNγ, while no IFNγ-expressing tetramer+ HCV-specific CD8 T cells were observed.

Figure 6.

HCV-specific CD8 T cells produce IL-10 Staining of tetramer (green) and CD8 (red) was combined with staining for IL-10 (cyan). Triple-stained HCV-specific CD8 T cells appear in light pink.

Together, these results strongly support the protective role of HCV-specific IL-10 producing CD8 T cells, which act at least in part by antagonizing the detrimental effects of liver-infiltrating CD8 T cells.

Discussion

Human and chimpanzee studies have shown that the outcome of acute HCV infection is determined by the vigor and quality of the antiviral T cell response,6, 24 and that a weak initial response may lead to viral persistence. Once chronic infection is established, the host is obliged to compromise between active virus eradication by lysis of infected hepatocytes and the need to protect a vital organ. It is not known whether liver damage due to hepatocyte killing is proportional to the number of virus-specific cytotoxic T cells or to the number of other liver-infiltrating immune cells. In other words, we do not know whether virus-specific intrahepatic T cells are beneficial or deleterious during chronic HCV infection. Here we used HLA tetramers to identify CD8 T cells specific for some HCV epitopes directly within liver biopsy material and to examine their relationship with liver histopathology. Of note, these HCV epitopes do not represent the whole set of potentially immunogenic HCV epitopes, and therefore, CD8 T cell responses to these epitopes do not predict the overall response in an individual.25 However, our results provide direct evidence that at least some virus-specific intrahepatic CD8 T cells are beneficial in chronic hepatitis C.

First, we found that both the presence and the abundance of tetramer+ T cells in the liver of chronically infected patients correlated negatively with disease activity. These cells were found in patients with very different circulating viral loads, suggesting that they have little impact on viral replication. CD8 T cells have two different antiviral mechanisms — killing of infected host cells, and secretion of cytokines such as IFNγ that directly inhibit viral replication. We used the TUNEL assay to identify apoptotic cells and thus to estimate hepatocyte killing. HCV-tetramer+ cells were sparsely scattered among non-apoptotic liver cells, indicating that they do not exert their protective effect via direct cytotoxicity of HCV-infected cells. Our observation that HCV-tetramer+ cells did not produce IFNγ, as previously observed,11 also helps to explain their minor role in virus control. Thus, it appears that some virus-specific CD8 T cells, which persist in the liver during chronic infection, do not eliminate HCV-infected hepatocytes, but are nonetheless able to control the disease, as their presence is associated with significantly milder hepatitis.

On the other hand, large CD8 T cell infiltrates, the specificity of which remains to be determined, were associated with hepatocellular apoptosis, liver fibrosis, and overall disease activity. The importance of T cell recruitment in hepatic pathogenesis has been attributed to the release of IFNγ, which can be produced by HCV-specific CD4 T cells.26, 27 Contrary to HCV-tetramer+ CD8 T cells, all of which were IFNγ-negative, we found that a substantial proportion of tetramer-negative CD8 cells expressed IFNγ which should therefore be seen not only as an antiviral cytokine, but also as a proinflammatory cytokine that promotes T cell recruitment.28, 29 Of course we cannot rule out that some tetramer-negative CD8 T cells may be specific for other non-tested HCV epitopes. In addition, areas of dense CD8 infiltrates might be the center of HCV infection and subsequent antigen-induced cell death of HCV-specific CD8 T cells, explaining the absence of tetramer+ cells in these areas. These possibilities, however, do not contradict the main finding of the present study, i.e., the protective effect of some HCV-specific CD8 cells against liver damage. We found that HCV-specific CD8 cells, most of which expressed the regulatory cytokine IL-10, antagonized the deleterious effect of chronic inflammation on apoptosis and liver fibrosis. Liver immunopathology was milder in regions harboring HCV tetramer+ cells than in regions containing only tetramer-negative cells. Accapezzato et al.14 recently described liver-expanded IL-10-expressing HCV-specific CD8 T cells with in vitro suppression function in chronic HCV infection. These cells might thus represent the CD8 counterparts of antigen-driven IL10-producing type 1 CD4 regulatory (TR1) cells,30 which have been reported to suppress antigen-specific immune responses via cell-to-cell interactions and/or production of IL-10.31 IL-10 has broad immunomodulatory properties: it inhibits antigen-specific activation, proliferation and production of cytokines by T cells, and plays a key role in blocking proinflammatory cytokine production by various cell types.32 These combined effects may explain the attenuation of necroinflammatory liver disease in regions where IL-10+ CD8 T cells are present. In addition, IL-10 has specific effects on liver fibrosis. Persistent activation of hepatic stellate cells, the main fibrinogenic cell type in the damaged liver,33 leads to progressive fibrosis during chronic infection.34 An antifibrotic activity of IL-10 has been suggested by studies of cultured hepatic stellate cells35 and by animal models of liver injury, including exogenous IL-10 administration,36 IL-10 knockout mice37 and transgenic IL-10 expression in liver.38 Therefore, attenuation of liver fibrosis in regions containing IL-10-expressing CD8 T cells may be directly related to the antifibrotic IL-10 activity.

The protective action of intrahepatic HCV-specific CD8 T cells in terms of hepatocyte necrosis could be conceivably due their failure to recognize infected target cells owing to viral antigenic variability. Indeed, the epitopes used to construct our tetramers were derived from the prototype sequence of genotype 1a, and some of our patients were infected by different genotypes. However, the presence of tetramer+ cells in our patients was independent of the HCV genotype, supporting the notion of cross-genotype CD8 T cell recognition.11, 15 HCV can also escape CTL recognition by mutation of target epitopes,3, 39 an early event, which remains fixed thereafter,6, 40 and determines viral persistence.41 Moreover, despite their ability to clear one HCV strain, patients may be reinfected with a heterologous strain that can then persist.42 For these different reasons, it will be important to analyze the autologous HCV sequence in our patients. The question arises as to why some HCV-specific protective/regulatory CD8 cells are generated in some patients but not in others. Host factors, such as HLA and cytokine/chemokine-receptor gene polymorphisms, have been shown to play a role in the severity of HCV disease. Progression to severe hepatitis would appear to result from an imbalance between regulatory T cells, effector cells, and inflammatory cells: the protective response is usually quantitatively inadequate in a large organ such as the liver, as rapid production of HCV virions overwhelms the bystander response.43 Adoptive immunotherapy might thus be an interesting adjunctive treatment for patients with chronic progressive HCV infection.

In conclusion, by analyzing intrahepatic HCV-specific CD8 T cells in patients at different stages of chronic hepatitis, we found a strong and broad-based virus-specific IL-10+ regulatory CD8 T cell response in liver areas where inflammation, necrosis, and fibrosis were controlled. HCV-specific CD8 cells, particularly those specific for the HLA-A2-restricted epitope NS3-1406-1415, appeared to prevent chronic liver injury mediated by other effector cells of undefined specificity and to keep inflammation at a low level. These results point to the curative potential of such virus-specific CD8 T cells. By modulating the balance between a weak protective CD8 T cell response and an excessive proinflammatory reaction, one might be able to bolster the antiviral response without inducing excessive tissue damage.

Acknowledgements

This work was supported by an ANRS grant (HC EP11). MA was supported by ANRS and INSERM, and DS was supported by Association pour la Recherche contre le Cancer (ARC). We thank Annette Lesot and Nicole Vignot for technical support, and Patricia Przednowed, Edith Audran and Christine Deshaye for HLA typing.

Ancillary