Intrahepatic CD8+ T-cell failure during chronic hepatitis C virus infection


  • Potential conflict of interest: Nothing to report.


The precise mechanisms responsible for the failure of intrahepatic hepatitis C virus (HCV)-specific CD8+ T cells to control the virus during persistent infection have not been fully defined. We therefore studied the CD8+ T-cell response in 27 HLA-A2–positive patients using four previously well-defined HLA-A2–restricted HCV epitopes. The corresponding HCV sequences were determined in several patients and compared with the intrahepatic HCV-specific CD8+ T-cell response. The results of the study indicate: (1) intrahepatic HCV-specific CD8+ T cells are present in the majority of patients with chronic HCV infection and overlap significantly with the response present in the peripheral blood. (2) A large fraction of intrahepatic HCV-specific CD8+ T cells are impaired in their ability to secrete interferon γ (IFN-γ). This dysfunction is specific for HCV-specific CD8+ T cells, since intrahepatic Flu-specific CD8+ T cells readily secrete this cytokine. (3) T-cell selection of epitope variants may have occurred in some patients. However, it is not an inevitable consequence of a functional virus-specific CD8+ T-cell response, since several patients with IFN-γ–producing CD8+ T-cell responses harbored HCV sequences identical or cross-reactive with the prototype sequence. (4) The failure of intrahepatic virus–specific CD8+ T cells to sufficiently control the virus occurs despite the presence of virus-specific CD4+ T cells at the site of disease. In conclusion, different mechanisms contribute to the failure of intrahepatic CD8+ T cells to eliminate HCV infection, despite their persistence and accumulation in the liver. (HEPATOLOGY 2005;42:828–837.)

There is emerging consensus that cellular immune responses play an important role in the immunopathogenesis of hepatitis C virus (HCV) infection.1, 2 Indeed, the appearance of functional HCV-specific T-cell responses is kinetically associated with control of viremia in the acute phase of infection.3–7 Studies in chimpanzees revealed that virus-specific T cells accumulate in the liver and that this coincides with liver disease and viral clearance.8, 9 The central role of virus-specific CD4+ and CD8+ T cells in HCV clearance has further been recently demonstrated by cell depletion studies in chimpanzees.10, 11

HCV-specific T cells are also detectable in chronic hepatitis, although at a lower frequency compared with acute resolving infection. HCV-specific CD8+ T-cell lines have also been generated from the human liver.12–15 In two of these studies, cytotoxic CD8+ T-cell lines could be established from approximately half of the chronically infected livers.12, 15 The long-term survival of HCV-specific CD8+ T cells in the liver has also been demonstrated by major histocompatibility complex (MHC) class I tetramer analysis. While no tetramer-positive cells were found in the liver in two studies,16, 17 other studies have detected tetramer-positive CD8+ T cells in livers at frequencies often exceeding 1% to 2% of intraheatic CD8+ T cells,18–20 indicating that these cells accumulate at the primary site of infection.

The mechanisms by which HCV is able to evade the virus-specific T-cell response that is present in the infected liver are still only poorly understood. Several different mechanisms could explain the failure of the intrahepatic HCV-specific CD8+ T-cell response.2 For example, several groups have suggested that functional alterations of HCV-specific T cells may be associated with persistent infection, including weak interferon γ (IFN-γ) production, impaired proliferation, and an immature differentiation state.21–24 However, these studies have been performed with peripheral blood lymphocytes. Thus, it was not previously known whether a lack of functional and mature T cells in the blood is a result of compartmentalization of these cells into the liver. A recent study revealed that intrahepatic HCV-specific CD8+ T cells are often impaired in their ability to secrete IFN-γ but frequently produced considerable amounts of interleukin 10 (IL-10), suggesting an immunoregulatory role of these T cells.18

HCV persistence may also be facilitated by viral evolution over the course of infection, enabling escape from prominent CD8+ T-cell responses.25 Indeed, studies in acutely infected patients and chimpanzees have revealed that viral escape from CD8+ T-cell responses may contribute to the development of viral persistence.26–29 These escape mutations occur early in infection and then remain fixed in the individual HCV genomes26 but can revert in the absence of the restricting HLA allele when transmitted to another subject with a different HLA type.29

Although a great deal has been learned about the role of the virus-specific T-cell response in recent years, very little information is currently available about the efficiency with which virus-specific T cells home to the liver, how well they function once they arrive, how well the intrahepatic T-cell response reflects the T-cell response present in the peripheral blood, and to what extent different mechanisms, such as T-cell dysfunction and viral escape, contribute to the failure of the intrahepatic HCV-specific T-cell response to control HCV infection. This study was performed to address these important issues.


HCV, hepatitis C virus; MHC, major histocompatibility complex; IFN, interferon; IL, interleukin; PCR, polymerase chain reaction; PBMC, peripheral blood mononuclear cell; FACS, fluorescence-activated cell sorting; RT, reverse transcription

Patients and Methods


Twenty-seven HLA-A2–positive patients with chronic HCV infection (HCV antibody– and HCV PCR–positive) were enlisted in the study. Blood samples were drawn for serological, virological, and immunological analyses, and the medical history of each subject was recorded. The patient characteristics are summarized in Table 1. The study protocol was approved by the local ethics committee.

Table 1. Study Population
PtAgeSexALT (U/L)HCV GenotypeViral Load (×106)Histology (METAVIR)
  1. NOTE. The table presents the characteristics of 27 HLA-A2–positive patients with chronic HCV infection. Normal ALT < 35 U/L, viral load in copies/mL. ALT, alanine aminotransferase.


HCV Assays.

HCV antibody testing was performed using the VITROS Anti-HCV assay (Ortho-Clinical Diagnostics, Neckargemünd, Germany). HCV RNA polymerase chain reaction (PCR) was performed using the Cobas Amplicor (Roche, Basel, Switzerland). HCV genotypes were identified by InnoLIPA (Innogenetics, Ghent, Belgium) and viral load was determined using the HCV RNA 3.0 assay (bDNA) (Bayer Diagnostics, München, Germany). All tests were performed at the Institute of Virology, University of Freiburg.

Peripheral Blood Mononuclear Cells.

EDTA anticoagulated blood (50-70 mL) was obtained from each patient, and peripheral blood mononuclear cells (PBMCs) were immediately isolated on Ficoll-Paque PLUS density gradient (Amersham, Oslo, Norway), washed three times in Dulbecco's balanced salt solution (DBSS; Gibco, Eggenstein, Germany), and either analyzed immediately or cryopreserved in media containing 80% fetal calf serum (Biochrom, Berlin, Germany), 10% dimethyl sulfoxide (Sigma-Aldrich, St. Louis, MO), and 10% RPMI 1640 (Gibco).

Liver Biopsy.

Liver tissue was obtained by ultrasound-guided needle biopsy. One fragment was put into RPMI medium (Gibco) containing 10% fetal calf serum; one was fixed in 10% zinc formalin for histological examination. Classification of the histological changes was performed according to the METAVIR score at the Institute of Pathology, University of Freiburg.

Intrahepatic T Cells.

Isolation of liver-infiltrating lymphocytes was performed exactly as previously described.9 Briefly, homogenized cell suspensions were incubated with magnetic beads coupled to anti-CD8/anti-CD4 antibodies, and bound cells were isolated using a particle magnetic concentrator. The purity of the T-cell subset, which was confirmed by fluorescence-activated cell sorting (FACS) analysis, was always greater than 95%. The isolated intrahepatic T cells were then plated into separate wells in 24-well plates (Greiner, Essen, Germany) in 1 mL 10% human AB serum, 100 U/mL IL-2 (Hoffmann–La Roche, Basel, Switzerland), 0.04 μg/mL anti-human CD3 monoclonal antibody (Immunotech, Marseilles, France), and 2 × 106 irradiated autologous PBMCs as feeder cells.

Synthetic Peptides and HLA-A2 Tetramers.

HCV-derived peptides (genotype 1a) and an influenza (Flu)-derived peptide (Table 2) previously shown to be HLA-A2–restricted epitopes were synthesized and used as previously described.7 For the analysis of HCV-specific CD4+ T-cell responses, overlapping peptides derived from HCV strain H77 (genotype 1a) spanning 18 amino acids and overlapping by 11 amino acids were obtained from the NIH AIDS Research and Reference Reagent Program (Bethesda, MD). HLA-A2 tetramers corresponding to the HCV peptides were provided by the National Tetramer Core Facility at Emory University (Atlanta, GA). The HLA-A2 tetramer corresponding to the Flu peptide was produced by ProImmune, Oxford, UK.

Table 2. HLA-A2–Restricted Epitopes
VirusAA SequenceAA Position
  1. NOTE. Part of this table has been reproduced from The Journal of Experimental Medicine, 2001;194:1395–1406 by copyright permission of The Rockefeller University Press.

  2. Abbreviations: AA, amino acid; HCV, hepatitis C virus.

InfluenzaGILGFVFTLMatrix 58


Anti-CD8 FITC, anti-CD8 PE, anti–IFN-γ PE, anti–IL-2 PE, anti–IL-4 PE, isotype FITC, isotype PE, and isotype Cy7-PE antibodies were obtained from BD Pharmingen (Heidelberg, Germany). All antibodies were used for immunostaining and FACS analysis according to the manufacturers' instructions.

HCV-Specific CD8+ T-Cell Enumeration.

Quantification of HCV-specific CD8+ cells was performed using HLA-A2 tetramers containing the peptides shown in Table 2 exactly as previously described.7 The detection limit was 0.03% of CD8+ cells for HCV and Flu tetramer. The detection limit was determined as the background signal plus 3 SDs after staining intrahepatic lymphocytes and PBMCs from HLA-A2+/HCV-negative blood donors (n = 10) and from HLA-A2–negative patients with chronic hepatitis C (n = 5).

Intracellular IFN-γ Staining.

This procedure was performed exactly as previously described.7 Briefly, lymphocytes (PBMCs: 0.5 × 106 cells/well; cytotoxic T lymphocytes and intrahepatic lymphocytes: 0.2–0.3 × 106 cells/well) were stimulated with the peptides (10 μg/mL) in duplicate wells in the presence of recombinant human interleukin-2 (Hoffmann-La Roche) and Brefeldin A (BD PharMingen, Heidelberg, Germany) and stained after 5 hours of incubation with antibodies to CD8 and after permeabilization with antibodies to IFN-γ (BD PharMingen). One well was stimulated with 10 ng/mL phorbol myristate acetate (Sigma-Aldrich) and 200 ng/mL ionomycin to serve as positive control for IFN-γ staining. FACS analysis was performed using FACSCalibur flow cytometer and analyzed with CELLQuest software (Becton Dickinson, San Diego, CA). The frequency of cytokine-positive CD8+ T cells was defined as the difference between the frequency detected in peptide-stimulated and unstimulated cells.


ELISPOT assays were performed exactly as previously described30 with duplicate cultures of 2.5 × 105 polyclonally expanded CD4+ T cells. Four hundred forty-one HCV peptides (final concentration, 10 μg/mL) were grouped into pools consisting of 10 peptides each in a matrix array, such that each individual peptide could be found in 2 pools only, allowing the identification of putative responses. Antigen-specific spot-forming cells (SFCs) in the presence of antigen versus SFCs in media alone were quantitated with an ELISPOT Reader (Biosys, Karben, Germany). Average background responses were 5 SFCs/2.5 × 105 cells. A response was scored positive if it was more than 5-fold above the background response and the mean response +2 SD in healthy, anti-HCV–negative control persons.

Antigen-Specific Cell Proliferation.

PBMCs (4 × 106) were resuspended in 1 mL RPMI (Gibco) containing 10% human AB serum, 1% streptomycin/penicillin, and 1.5% Hepes buffer (1 mol/L); stimulated with 10 μg/mL synthetic HCV or influenza peptide and 0.5 μg/mL anti-human CD28 (BD PharMingen); and cultured in a 24-well plate (Greiner). The cultures were restimulated on day 7 with autologous irradiated (3,000 rad) PBMCs (1 × 106 cells/well) and peptide (10 μg/mL), and on days 3 and 10 with 20 U/mL recombinant human interleukin-2 (Hoffmann-La Roche) in fresh media.

DNA Amplification and Sequencing.

RNA was extracted from sera with the QIAamp Viral RNA Mini Kit (Qiagen, Valencia, CA) and subjected to reverse transcription (ThermoScript Reverse Transcriptase, Invitrogen, Karlsruhe, Germany) and two rounds of amplification (Expand High Fidelity PCR System, Roche). The following combination of primers was used: (1) amplification of the fragment of the core region for all HCV types and subtypes. Reverse transcription (RT) was done with the primer p874a (5′-AAGAAGATAGARAARGAGCAACC). PCR I was run with primers p417s (5′-GGCGGYGGNCAGATCGTTGG) and p874a. PCR II was performed with primers p439s (5′-GAGTWTACBTGYTGCCGCGCAG) and p874as. (2) Amplification of the fragments of HCV genotype 1a. NS3 fragment: RT–primer 5317as (5′-CGCCAACGAGCACCCAGG); PCR I–primers p3276s (5′-ATGGAGACCAAGCTCATCACG) and p5317as; PCR II–primers p3314s (5′-CGCGTGCGGTGACATCATC) and p4680as (5′-CAGTCTATCACAGAGTCGAAGTC). NS5B fragment: RT–primer SV75as (5′-GGAGTCAAARCAGCGGGTATC); PCR I–primers SV73s (5′-CGTAGCCCACATCAACTCCG) and SV75as; PCR II–primers SV72s (5′-GCTCGTCTCATCGTGTTCC) and SV74as (5′-CGCTGTCCTGGTGAGTATTGG).

(3) Amplification of the fragments of HCV genotype 1b. NS3 fragment: RT–primer p5511as (5′-CTGCTTGAATTGCTCGGCGAG); PCR I–primers p3399s (5′-GGGCAGGGGTGGCGRCTCC) and p5511as; PCRge II–primers p3404s (5′-AGGGGTGGCGRCTCCTY-GCG) and p5371as (5′-CACAATGACCACRCTGCCYG). NS5B fragment: RT–primer SV70as (5′-GGGGCCAAGTCACAACATTG); PCR I–primers SV69s (5′-CAACCAGAGAARGGAGGCC) and SV70as; PCR II–primers SV68s (5′-CTCGCCTTATCGTATTCCCAG) and SV71as (5′-TTCACCAGGAACTCGACYCG).

The amplified DNA was gel purified with the QIAquick Gel Extraction kit (Qiagen) and subjected to direct sequencing in both directions (Dye Terminator DNA Sequencing kit, Perkin-Elmer, Boston, MA) in order to establish the predominant sequence. The genotype determination of HCV isolates was performed as described elsewhere.31

Binding Assays.

Peptide binding to HLA-A2 was measured exactly as previously described.32 Briefly, various doses of the test peptides (ranging from 100 μmol/L to 1 nmol/L) were coincubated with 0.5 nmol/L radiolabeled HBVc18-27 (FLPSDYFPSV) peptide and HLA-A2.1 heavy chain and β2-microglobulin for 2 days at room temperature in the presence of a mixture of protease inhibitors. Percentage of MHC-bound radioactivity was determined by gel filtration, and the concentration required to inhibit 50% of the radiolabeled peptide binding (IC50), was calculated for each peptide.

Statistical Analysis.

Linear regression analysis using Spearman's correlation coefficient was performed.


Frequency of Intrahepatic HCV-Specific CD8+ T Cells in Chronic HCV Infection.

First, we determined the frequency of HCV-specific CD8+ T cells in polyclonally expanded intrahepatic CD8+ T cells from 27 HLA-A2–positive patients with chronic HCV infection (Table 1) using 4 HLA-A2 tetramers (Table 2). Of note, this expansion method has been successfully used to detect intrahepatic HCV-specific CD8+ T cells in chimpanzees.8–10 In this study, intrahepatic HCV-specific CD8+ T-cell responses were detectable in 17 (63%) of 27 patients and targeted an average of 1.3 epitopes (range, 0–3), primarily within the nonstructural region of the genome (Table 3). It is also important to note that 12 (52.2%) of 23 tested patients had Flu-specific CD8+ T-cell responses (Table 3).

Table 3. Frequency of Peripheral and Intrahepatic HCV- and Flu-Specific CD8+ T Cells
PtPBMCs % TetramerIHLs % Tetramer
  • NOTE. Table shows a summary of tetramer-positive CD8+ T-cell responses in the peripheral blood and the liver, expressed as %. All positive responses are shown in bold. Abbreviations: HCV, hepatitis C virus; Flu, influenza; PBMC, peripheral blood mononuclear cell; IHL, intrahepatic lymphocyte; ND, not determined.

  • *

    Additional responses in CTL lines.

% Responses033.329.625.985.77.451.944.425.952.2

Overlap Between the Peripheral and Intrahepatic CD8+ T-Cell Response.

Peripheral CD8+ T lymphocytes purified from PBMCs obtained at the same time and expanded polyclonally in the same manner as the intrahepatic lymphocytes were also studied to compare the frequency of HCV-specific CD8+ T cells in peripheral blood and in the liver. In addition, the peripheral CD8+ T-cell response was also tested directly ex vivo. Importantly, the frequency of HCV-specific CD8+ T cells was comparable in both cell populations: PBMCs ex vivo versus polyclonally expanded PBMCs (data not shown). Peripheral HCV-specific CD8+ T-cell responses were detectable in 16 (59.3%) of 27 patients, targeting an average of 0.9 epitopes (range, 0–3). All HCV-specific T cells targeted the nonstructural region. The frequency of HCV-specific CD8+ T cells was low, with mean frequencies ranging from 0.087% against the NS52594 epitope to 0.142% against the NS31073 epitope. Thus, the frequency of peripheral HCV-specific CD8+ T-cell responses was significantly lower compared with the intrahepatic compartment. In contrast, Flu-specific CD8+ T cells were detectable in the peripheral blood of most patients (18 [85.7%] of 21 tested patients) and had a higher mean frequency (0.44%) compared with the intrahepatic compartment.

The T-cell epitope repertoire overlapped significantly between both compartments. Indeed, from the total of 35 intrahepatic HCV tetramer–positive CD8+ T-cell responses, 24 also were detectable directly ex vivo in the peripheral blood and an additional 5 peptide-specific CD8+ T-cell responses became detectable after two rounds of peptide-specific stimulation, as shown by the asterisks in Table 3. A few HCV-specific CD8+ T-cell responses (5 [14.3%] of 35 intrahepatic CD8+ T-cell responses) were preferentially localized in the liver because these responses could not be expanded from the blood, even after several rounds of stimulation with the cognate peptide (Table 3, patients 16, 18, 22, 23, and 27).

Function of Intrahepatic CD8+ T Cells in Chronic HCV Infection.

Next, we determined the ability of intrahepatic HCV-specific CD8+ T cells to produce IFN-γ after stimulation with the corresponding peptides. As shown in Fig. 1B, only a minority of intrahepatic tetramer–positive, HCV-specific CD8+ T cells secreted IFN-γ after specific stimulation with the cognate peptide, suggesting that a large fraction of CD8+ T cells are impaired in their ability to secrete IFN-γ at the site of disease. Notably, this dysfunction was observed despite IL-2 supplementation of these cells in vitro. Representative results from 4 patients are shown in Fig. 1A. In contrast to the dysfunction of most HCV-specific CD8+ T cells, Flu-specific CD8+ T cells produced IFN-γ in all but 2 patients (Fig. 1), indicating that dysfunction is not a general phenomenon of intrahepatic virus–specific CD8+ T cells but rather is characteristic of HCV-specific cells. In order to determine whether the inability of HCV-specific CD8+ T cells to produce IFN-γ may reflect a skewing to a different cytokine profile we also tested intrahepatic CD8+ T-cell responses for the ability to secrete IL-2 and IL-4 in selected cases: in no case did we observe HCV-specific or Flu-specific CD8+ T cells producing these cytokines (data not shown). Further, the impaired ability of HCV-specific CD8+ T cells to produce IFN-γ was not specific to the intrahepatic compartment, because the same phenotype was observed in the peripheral blood (data not shown).22, 23 Of note, peripheral and intrahepatic HCV-specific T cells expressed CCR7, as has been recently suggested.24 Finally, it is also important to note that we never observed peptide-specific IFN-γ production in the absence of tetramer response in any of the 27 tested patients.

Figure 1.

(A) Representative dot blots of indicated patients showing tetramer-positive CD8+ T-cell responses and corresponding IFN-γ responses. (B) Black bars indicate the percentage of intrahepatic tetramer-positive CD8+ T cells, and gray bars represent the percentage of IFN-γ positive CD8+ T cells of all tested patients. (C) Intrahepatic tetramer-specific CD8+ T-cell responses (x-axis) in correlation with their corresponding IFN-γ secretion (y-axis) are shown. IFN-γ, interferon γ; tet, tetramer.

Sequence Heterogeneity of the Viral Epitopes and Correlation With the Intrahepatic Virus–Specific T-Cell Response.

Next, the amino acid sequence of the four tested epitopes was deduced by viral nucleotide sequence analysis in 13 patients infected with genotype 1a or 1b and correlated with the corresponding peptide-specific intrahepatic CD8+ T-cell response. The peptides used in this study were derived from genotype 1a and were thus identical with the prototype sequence of genotype 1a with the exception of the NS52594 epitope, since this peptide was initially synthesized with the sequence of its original description.3 However, it is important to note that functional analysis revealed cross-immunogenicity between the NS52594 peptide and synthesized NS52594 peptides with the prototype 1a and 1b sequences and also between the NS31073 peptide and a synthesized peptide with the prototype 1b sequence (data not shown). In contrast, no cross-immunogenicity was observed between the NS31406 genotype 1b and the genotype 1a–derived NS31406 peptide.

As shown in Table 4, 17 of 28 tetramer-positive CD8+ T-cell responses that did not produce IFN-γ were observed. Importantly, we detected an identical or cross-immunogenical viral sequence in 10 cases. A total of 11 intrahepatic HCV-specific CD8+ T-cell responses were detected that were able to secrete IFN-γ after stimulation with the peptide. Notably, 6 of the corresponding autologous viral sequences did not display sequence variations, and 1 patient (patient 11, NS31073 epitope) had a cross-immunogenic sequence, indicating that the emergence of T-cell escape mutations is not an inevitable consequence of a functional intrahepatic CD8+ T-cell response. In 5 patients with functional intrahepatic CD8+ T-cell responses, sequence variations were detectable, although at least 1 was again cross-immunogenic (patient 11, NS31073). To test the hypothesis that some of these variant peptides are potentially cytotoxic T-lymphocyte escape variants, the variant NS31406 peptides present in patients 26 and 27 were synthesized and tested for CD8+ T-cell cross-reactivity and immunogenicity. As shown in Fig. 2A, no recognition of the respective NS31406 variant peptides was observed in both patients across a range of peptide concentrations, compatible with cytotoxic T-lymphocyte escape at these positions. The binding affinity of these variant peptides, however, was unaffected (Fig. 2B). Finally, we also tested all genotype 1b–infected patients with the corresponding genotype 1b NS31406 peptide and did not find any significant IFN-γ production (data not shown). The detection of NS31406-specific CD8+ T cells in these patients by using genotype 1a peptide–derived HLA-A2 tetramers suggests that tetramer binding can still occur in the absence of peptide-specific IFN-γ production, as has recently been described.7 However, it is also possible that the T-cell response against the prototype 1a peptide may present a residuum from a previous exposure, which is especially likely in the case of patient 23, who also showed genotype 1a peptide–specific IFN-γ production.

Table 4. Viral Peptide Sequence and Intrahepatic Prototype HCV Peptide–Specific CD8+ T-Cell Response
  1. NOTE. Table shows comparison of prototype 1a and 1b sequences with the corresponding sequence detected in the patients and correlation with epitope-specific CD8+ T-cell responses (tetramer and IFN). Boxes indicate T-cell responses tested in the presence of prototype and variant peptide (Fig. 2). Positive responses are shown in bold.

inline image
Figure 2.

(A) Cytotoxic T lymphocyte (CTL) lines stimulated with prototype peptide were tested for intracellular IFN-γ production against prototype peptide and variant peptide (Table 4) at decreasing concentrations. (B) Binding affinity of the genotype 1a and 1b prototype NS3 1406 and the two variant peptides tested as shown in panel A.

Peripheral and Intrahepatic HCV-Specific CD4+ T-Cell Responses.

To determine the HCV-specific CD4+ T-cell response, polyclonally expanded peripheral and intrahepatic CD4+ cells of 10 patients were comprehensively screened in a matrix ELISPOT using overlapping peptides. As shown in Fig. 3, peripheral HCV-specific CD4+ T-cell responses were only detectable in 3 patients. In contrast, intrahepatic HCV-specific CD4+ T-cell responses were detectable in 8 patients, although these responses were rather weak in most cases. The strongest intrahepatic CD4+ T-cell response was observed in patient 17, who also mounted an intrahepatic IFN-γ–producing CD8+ T-cell response (Table 4).

Figure 3.

Frequency of HCV-specific CD4+ T cells in polyclonally expanded peripheral (A) or intrahepatic (B) CD4+ cells in response to overlapping HCV peptides covering the entire HCV polyprotein. Responses are shown as antigen-specific, IFN-γ spot forming cells (SFCs) per 2.5 × 105 cells. Putative epitope-specific responses within a specific HCV protein are added together, and only positive responses are displayed. IHL, intrahepatic lymphocyte; IFN-γ, interferon γ; PBMC, peripheral blood mononuclear cell.


In this study we used the tetramer technique in combination with functional cytokine assays to analyze and compare the frequency and cytokine production capacity of HLA-A2–restricted intrahepatic and peripheral HCV-specific CD8+ T cells at the single-cell level. This approach certainly underestimates the total pool of HCV-specific CD8+ T cells present in the infected liver. Thus, this study was not performed to comprehensively analyze the CD8+ T-cell response against all possible viral epitopes but rather to simultaneously determine the frequency and function of virus-specific CD8+ T cells in the intrahepatic compartment.

The first finding of our study was that 17 (63%) of 27 patients with chronic HCV infection had intrahepatic HCV-specific CD8+ T-cell responses against at least one of the four epitopes. These results are in agreement with previous tetramer analyses that demonstrated the presence of HCV-specific CD8+ T cells in the liver,18–20 supporting the notion that HCV can persist despite the long-term survival of virus-specific CD8+ T cells at the site of infection. Importantly, the magnitude and multispecificity of the HCV-specific CD8+ T-cell response present in the liver did not correlate significantly with viral titers or the activity of liver disease, suggesting that these HLA-A2–restricted virus-specific cells fail to sufficiently control the virus.

How can the failure of the intrahepatic HCV-specific CD8+ T-cell response be explained? One explanation would be the impaired effector functions of HCV-specific CD8+ T cells at the site of disease. Although the exact effector mechanisms used by the intrahepatic HCV-specific T cells to control the virus still remain to be determined, recent work suggests that IFN-γ might perform noncytolytic effector functions during HCV infection, similar to its ability to control hepatitis B virus replication in transgenic mice and acutely infected chimpanzees.33 Indeed, the recently observed correlation between the intrahepatic induction of IFN-γ and HCV-specific CD8+ T-cell responses in acutely infected chimpanzees that clear the virus9, 34 and the fact that IFN-γ efficiently inhibits the replication of an HCV replicon in vitro35 support this hypothesis. Therefore, it is important to note that the majority of intrahepatic HCV-specific CD8+ T cells in our cohort of patients with chronic HCV infection were unable to sufficiently secrete IFN-γ after peptide-specific stimulation (Fig. 1). Interestingly, the impairment to secrete IFN-γ is specific for HCV-specific CD8+ T cells, since Flu-specific CD8+ T cells present in the same HCV-infected livers readily secreted IFN-γ after peptide-specific stimulation. These results extend previous observations that have described impaired effector functions of HCV-specific CD8+ T cells in the peripheral blood21–24 and show that a lack of functional and mature T cells in the blood is not a result of compartmentalization of these cells to the liver.

The simultaneous analysis of the viral nucleotide sequences of the four epitopes and the intrahepatic HCV-specific CD8+ T-cell response in 13 patients gave important insights into the host–virus interactions in chronic hepatitis C. First, the observed impaired ability of HCV-specific CD8+ T cells to produce IFN-γ is not solely due to a mismatch of the synthetic peptides used for functional analysis with the sequences encoded by the infecting virus, since both sequences were identical or cross-immunogenic in most patients (Table 4). Second, T-cell selection of epitope variants is not an inevitable consequence of a functional virus-specific CD8+ T-cell response, since patients with IFN-γ–producing CD8+ T-cell responses often harbored HCV sequences identical or cross-reactive with the prototype sequence (Table 4), as has been previously shown for the NS31073 epitope.36 Thus, our results indicate that epitopes sometimes do not undergo mutation, even when they are targeted by virus-specific CD8+ T cells, as has been shown in HIV infection.37 Sequence variations that were consistent with the selection of escape mutations were primarily observed in the NS31406 epitope, supporting previous studies that described viral escape mutations in the presence of HCV-specific CD8+ T-cell responses in patients with chronic HCV infection.38 It is important to note, however, that without documentation of the sequence of the infecting virus, viral escape cannot be formally proven. Third, the lack of recognition of epitopes in several patients cannot be explained by sequence variation alone, since several patients harbored viral sequences identical or cross-reactive with the prototype sequence but still did not mount an HCV-specific CD8+ T-cell response (Table 4). These results support the concept that HCV has the ability to interfere with the induction of primary immune responses by antigen-presenting cells39 or that some of these responses may have been primed early but have been exhausted or deleted in the course of infection.4, 7

It is tempting to speculate that the different mechanisms of CD8+ T-cell failure are a direct result of inadequate virus-specific CD4+ T-cell help that is common in persistent HCV infection.2, 40, 41 Of note, we were able to detect IFN-γ–producing CD4+ T cells in the livers of most tested patients (Fig. 3). Our results support the notion that HCV-specific CD4+ T-cell responses are rather weak during chronic infection but do accumulate in the intrahepatic compartment.42–44 Other mechanisms that could explain the CD8+ T cell failure include a dysfunction of CD4+ T cells (e.g., IL-2 secretion),41 suppressive role of regulatory T cells,45–48 unresponsiveness due to high antigen levels, suppression by viral factors, and influence by the innate immune system.2

In summary, our results may have some important therapeutic implications, since successful therapeutic vaccination protocols may depend not only on the induction of HCV-specific CD8+ T cells, but also on the maintenance of their function in vivo, especially in the liver. In addition, knowledge about the failure of epitope-specific CD8+ T-cell responses will be crucial for the optimal design of new T-cell–based immunotherapeutic strategies.


The authors thank Ms. Natalja Nazarova and Scott Southwood for excellent technical assistance, Dr. Frank Chisari for continuous support and helpful advice throughout this study, and Dr. Joerg Timm for critically reading the manuscript. The HLA-A2 tetramers were obtained from the NIH tetramer facility (Rockville, MD).