Associations among clinical, immunological, and viral quasispecies measurements in advanced chronic hepatitis C


  • Conflict of interest: Nothing to report.


The relationships among host immune and viral factors and the severity of liver disease due to hepatitis C virus (HCV) are poorly understood. Previous studies have focused on individual components of the immune response to HCV, often in relatively small numbers of patients. We measured HCV-specific lymphoproliferation (LP), intrahepatic cytotoxic T lymphocyte (CTL), and neutralizing antibody (NA) responses and HCV quasispecies (QS) diversity and complexity in a large cohort of subjects with advanced liver fibrosis (Ishak stages 3-6) on entry into the HALT-C (Hepatitis C Antiviral Long-term Treatment against Cirrhosis) trial. We correlated LP, CTL, NA, and QS results with clinical characteristics, including serum alanine aminotransferase (ALT), HCV RNA level, HCV genotype, and hepatic histopathology. LP, CTL, and NA responses were detected in 37%, 22%, and 22% of subjects tested, respectively. The only association that was statistically significant was higher mean serum ALT values in patients with detectable HCV-specific CTL responses (P = .03). In conclusion, immune responses to HCV and viral diversity showed little relationship to clinical or histological features at a single time point in this selected population of patients with advanced chronic hepatitis C for whom prior interferon treatment had failed. (HEPATOLOGY 2005;41:617–625.)

Chronic infection with hepatitis C virus (HCV) is estimated to affect more than 3 million individuals in the United States and 170 to 200 million people globally.1 The course of liver disease related to chronic HCV infection is highly variable. Some individuals are infected for decades without clinically apparent liver disease, but liver injury and inflammation are detectable in most infected individuals, and overt liver disease develops in approximately 50%.2 Disease is slowly progressive in approximately 30% of individuals with chronic hepatitis C, leading to cirrhosis and liver failure, usually after 20 years or longer. As a consequence of these characteristics, chronic hepatitis C has become the most common indication for liver transplantation in the United States.1

The factors that determine the severity of liver disease in chronic HCV infection and their relative importance are not fully defined. Serum HCV RNA levels, as a proxy measurement of viral replication, correlate poorly with disease severity,3, 4 whereas conflicting data exist as to whether measurements of intrahepatic HCV replication correlate better with liver injury.3–9 Virus–host interactions are also thought to be a significant cause of liver injury. Persistent HCV infection occurs in the context of ongoing HCV-specific cellular and humoral immune responses. Although CD4+ and CD8+ T-cell responses in chronically infected individuals are less intense than those observed during acute HCV infection10–15 and are apparently inefficient at viral clearance,11 several small studies have reported associations between HCV-specific T-cell responses and liver injury or serum HCV RNA levels.16–18 Greater quasispecies (QS) diversification within the envelope hypervariable region (E2 HVR1) is thought to reflect immune selection and also has been associated with evolution to progressive disease.19

Previous studies of HCV-specific immune responses and viral QS evolution have included relatively small numbers of subjects. In particular, few studies heretofore have attempted to test associations among different arms of the immune response; in those rare instances in which this has been done, studies have involved few subjects.13 The Hepatitis C Antiviral Long-term Treatment against Cirrhosis (HALT-C) trial is a large, multicenter, randomized clinical trial to determine the efficacy of long-term therapy with pegylated interferon alpha (IFN-α) for the prevention of progression of liver disease in patients with chronic hepatitis C who did not achieve a sustained virological response to previous therapy.20 We measured HCV-specific lymphoproliferative (LP) responses, cytotoxic T lymphocyte (CTL) responses, E1- and E2-specific neutralizing antibody (NA) responses, and HCV E2 HVR1 QS diversity and complexity at entry into the HALT-C trial in a large cohort of enrolled subjects and correlated these findings with clinical characteristics and hepatic histopathology.


HCV, hepatitis C virus; QS, quasispecies; HVR1, hypervariable region 1; HALT-C, Hepatitis C Antiviral Long-term Treatment against Cirrhosis; IFN, interferon; LP, lymphoproliferation; CTL, cytotoxic T lymphocyte; NA, neutralizing antibody; IL-2, interleukin-2; PBMC, peripheral blood mononuclear cell; SOD, superoxide dismutase; SI, stimulation index; BLCL, B lymphoblastoid cell lines; VSV, vesicular stomatitis virus; HMR, heteroduplex mobility ratio; ALT, alanine aminotransferase.

Patients and Methods


The design of the HALT-C clinical trial has been published.20 In brief, criteria for enrollment were chronic HCV infection with detectable HCV RNA, unresponsive to previous IFN-α therapy with or without ribavirin, and with bridging fibrosis or cirrhosis on liver biopsy within 12 months of enrollment. The immunology and virology studies involved subjects enrolled at 4 of the 10 HALT-C trial sites (Saint Louis University, University of Massachusetts/University of Connecticut, University of Southern California, and University of Texas Southwestern) which were selected based on investigator interest. The study protocols were approved by the internal review boards of all participating institutions, and written informed consent was obtained from all subjects. Clinical and laboratory data were collected from all subjects with the use of standardized procedures. Serum HCV RNA levels were measured at a central laboratory by using the Roche COBAS Amplicor Monitor v.2.0 assay (Roche Molecular Systems, Inc., Branchburg, NJ), and liver histology was evaluated and graded by a central panel of pathologists, using the Ishak scoring system.


Peripheral blood (25 mL) was collected at baseline in ACD (acid citrate dextrose) anticoagulant Vacutainer tubes (Becton Dickinson, Franklin Lakes, NJ) and shipped overnight at ambient temperature for LP analysis. Serum obtained at baseline was frozen and stored at −70°C for NA and QS analyses.

To assess eligibility to participate in the trial, liver biopsies were performed at baseline in those subjects without a qualifying liver biopsy within the previous 12 months. When sufficient tissue was obtained, a section of biopsy tissue approximately 0.5 cm in length was placed into tissue culture medium containing 20 U/mL recombinant human interleukin-2 (IL-2) and shipped overnight at ambient temperature to one of the two CTL laboratories (see CTL Assays).

LP Assays.

Peripheral blood mononuclear cells (PBMCs) were isolated by centrifugation over Ficoll-Hypaque and used immediately. PBMCs (105 per well) were incubated at 37°C in quadruplicate wells with superoxide dismutase (SOD)-recombinant HCV protein antigens (courtesy of Dr. Michael Houghton, Chiron Corp., Emeryville, CA) at a concentration of 10 μg/mL in RPMI 1640 medium supplemented with 10% human serum (Gemini Bio-Products, Woodland, CA). Antigens included yeast-derived SOD (negative control), SOD-c22 (HCV aa 2-120), SOD-c100 (HCV aa 1569-1931), SOD-NS5 (HCV aa 2054-2995), and Escherichia coli–derived SOD-c33c (HCV 1192-1457). Phytohemagglutinin (PHA; 5 μg/mL, Sigma, St. Louis, MO), Candida albicans (20 μg/mL, Greer Laboratories, Lenoir, NC), and Tetanus toxoid (12 Lf/mL, Wyeth-Ayerst Laboratories, Marietta, PA) were also tested as positive controls. On day 5 of culture, 1 μCi 3H-thymidine was added to each well for an additional 16 to 18 hours before measuring 3H-thymidine incorporation into DNA. The stimulation index (SI) to individual HCV antigens was calculated as (mean cpm in presence of antigen) divided by (mean cpm in presence of appropriate control antigen). Assays were considered valid for analysis only if the SI value to one or more of the positive controls was greater than 3.

CTL Assays.

For logistical considerations related to the large cohort size, CTL assays were performed at two separate laboratories. Laboratory procedures were reviewed and standardized at the two laboratories at the initiation of the study, and comparative studies using separate fragments of liver tissue from the same five biopsies showed that assay results were equivalent (data not shown).

Autologous B lymphoblastoid cell lines (BLCLs) were prepared by infection of PBMCs with Epstein-Barr virus.21 BLCLs were infected with recombinant vaccinia viruses expressing regions spanning the entire HCV polyprotein—amino acids 1 to 339 (C-E1, vv9A), amino acids 347 to 906 (E2-NS2, vv1H), or amino acids 827 to 3011 (NS2-NS5, vv827)—or with recombinant vaccinia virus expressing β-galactosidase as a negative control, and incubated for 16 to 20 hours.

T-cell lines were derived from liver biopsy samples by expansion with anti-CD3 monoclonal antibody 12F6 as described.21–23 Briefly, liver biopsy fragments were incubated in culture medium in the presence of 1 μg/mL anti-CD3 and 100 U/mL recombinant human IL-2. When there was significant growth of lymphocytes extending out from the tissue (generally 2-4 weeks), the cells were restimulated at a density of approximately 2.5 × 106 cells/mL with anti-CD3 and recombinant human IL-2 in the presence of allogeneic γ-irradiated PBMCs. Seven to 21 days later, these effector T cells were incubated with vaccinia virus-infected, 51Cr-labeled BLCL target cells at effector-target (E/T) ratios of 100:1, 50:1, and 25:1. Percent specific target cell lysis was calculated as described.21–23

NA Assays.

Pseudotype vesicular stomatitis virus (VSV) virions were prepared by phenotypic mixing of HCV chimeric envelope glycoproteins (E1-G or E2-G) and incorporation into the VSVts045 temperature-sensitive mutant as described.24–27 The preparations of HCV/VSV pseudotype virions had a titer of 103 to 104 plaque-forming units (pfu)/mL by plaque assay in BHK cells at the permissive temperature (32°C); the titer was not affected by incubation with an antiserum to VSV G.

For the NA assay, pseudotype viruses (∼100 plaque-forming units) were incubated with serial dilutions of heat-inactivated test sera at 37°C for 1 hour. The virus-serum mixture (200 μL) was added to BHK cell monolayers in individual wells of a 6-well plate and incubated at 32°C for 1 hour. Cells were washed 2 times with Dulbecco's medium (DMEM) and overlaid with a 1:1 mixture of 1.6% agarose and 2× Dulbecco's medium containing a final concentration of 2% fetal bovine serum. An additional agar overlay containing 0.0005% neutral red was added after 24 hours for counting of plaques. A reduction of 50% or more in the number of plaques compared with the mean of at least 6 control sera from healthy, HCV-seronegative donors was the cutoff for a positive NA assay, and the NA titer was defined as the highest serum dilution showing NA activity. None of the study sera showing neutralization of HCV/VSV pseudotype virions exhibited detectable (>5%) neutralization of non-pseudotype VSVts045.

QS Assays.

HCV QS diversity and complexity were evaluated by clonal frequency analysis of the E2 HVR1 as described,28 except that the HVR1 was amplified by using Advantage HF-2 polymerase (Clontech, Palo Alto, CA). For each serum sample tested, 20 colonies of bacteria transformed with the cloned HVR1 polymerase chain reaction product, representing 20 individual QS clones, were analyzed. Colonies were picked directly into tubes for re-amplification of the second-round polymerase chain reaction product. Polymerase chain reaction products were visualized on ethidium bromide–stained agarose gels. For each sample, one HVR1 polymerase chain reaction product was selected, purified, and end-labeled with 32P-adenosine triphosphate and T4 polynucleotide kinase. This labeled probe was then hybridized directly to each of the 20 HVR1 QS clones derived from that serum sample for 2 hours at 55°C. Hybrids were separated on 6% non-denaturing polyacrylamide MDE gels, (Cambrex Biosciences, Walkersville, MD) and visualized by autoradiography.

QS complexity was determined by counting the total number of unique gel shift patterns in each set of 20 QS clones representing a single specimen. QS genetic diversity was determined by deriving the average heteroduplex mobility of all clones relative to the homoduplex probe control. A heteroduplex mobility ratio (HMR) was calculated by dividing the distance in millimeters from the origin of the gel to the heteroduplex by the distance in millimeters from the origin to the homoduplex control; HMR values closer to 1 indicate sequences that more closely match the probe, whereas lower HMR values reflect greater sequence divergence from the probe. The HMRs for all 20 QS clones tested were averaged to provide the final HMR value for that serum sample.


Statistical analyses were performed through the HALT-C Data Coordinating Center. LP, CTL, and NA assays were scored as positive or negative. LP assays were scored as positive if the SI value for any HCV antigen was greater than 3. CTL assays were scored as positive if the percent specific lysis of any HCV protein-expressing target cell was more than 10% above the percent specific lysis of the control target cells at one or more effector–target cell ratios. NA assays were scored as positive if 50% or more neutralization against either E1- or E2-pseudotyped virus was detected. Standard statistical methods (chi-square test, odds ratio, t test, Wilcoxon rank sum test, and correlation coefficient) were performed by using SAS release 8.2 (SAS Institute, Cary, NC). All analyses were two-tailed, with α = 0.05.


Characteristics of the Ancillary Study Population.

Four hundred thirty-one subjects were enrolled at the 4 participating study sites, of whom 341, 209, 243, and 103 subjects were studied for LP, CTL, NA, and QS, respectively. The principal reason for exclusion from the CTL study was lack of fresh liver tissue for analysis. NA and QS assays were performed on frozen specimens from consecutive subjects included in one or more of the other immunology studies; QS assays were only performed on subjects infected with HCV genotype 1. Among those subjects who had CTL and at least one other assay performed, the mean (±SD) interval between the liver biopsy (for CTL) and blood collection (for LP, NA, or QS) was 4.1 ± 3.9 weeks.

Demographic and clinical characteristics of the subjects enrolled in the clinical study at the 4 sites and the subjects enrolled in each laboratory study are shown in Table 1. Consistent with the characteristics of all subjects enrolled in the HALT-C trial, subjects included in the immunology and virology studies were predominantly male, most were infected with HCV genotype 1, and they had long-standing HCV infection, with mean duration of infection greater than 25 years. Subjects had not received IFN-α or ribavirin for at least 2.5 months (mean, 26.0 months) before study. Subjects included in each of the laboratory studies were comparable to the overall clinical study population except for a lower prevalence of cirrhosis; the differences in prevalence of cirrhosis were statistically significant for subjects included in the CTL and NA studies (Table 1), a result of the greater difficulty in obtaining sufficient liver tissue by needle biopsy in subjects with cirrhosis.

Table 1. Demographics Characteristics of the Immunology and Virology Study Subjects
CharacteristicAll Subjects at SitesLPCTLNAQS
  • NOTE. Columns show all subjects enrolled in the parent HALT-C study at the participating study sites and the subgroups that were studied by lymphocyte proliferation assays (LP), cytotoxic T lymphocyte assays (CTL), neutralizing antibody assays (NA), and quasispecies diversity and complexity assays (QS).

  • *

    We compared the characteristics of subjects studied by each assay with the remaining HALT-C subjects not studied by that assay using the nonparametric Wilcoxon rank sum test. P values are .38, .28, .32, and .02 for LP, CTL, NA, and QS, respectively.

  • P values for the comparison of subjects studied by each assay and those not studied are .100, .001, .017, and .228 for LP, CTL, NA, and QS, respectively.

Age (yr)50.1 ± 7.850.3 ± 7.950.6 ± 7.550.5 ± 8.150.2 ± 8.7
Sex (% male)7171717072
Race (% black)1112111212
BMI29.9 ± 5.430.0 ± 5.529.9 ± 5.829.7 ± 5.429.3 ± 4.5
Duration of infection (yr)28.3 ± 8.428.7 ± 8.228.6 ± 8.028.8 ± 8.427.8 ± 7.5
Time off therapy (mo)*25.8 ± 22.726.4 ± 23.224.3 ± 20.325.2 ± 23.121.8 ± 18.3
ALT (U/L)106.4 ± 74.3101.9 ± 66.8102.9 ± 71.0102.2 ± 69.796.4 ± 60.8
HCV RNA (log10)6.45 ± 0.556.45 ± 0.556.48 ± 0.476.46 ± 0.586.49 ± 0.52
Genotype 1 (%)87878888100
Cirrhosis (%)3231252827

HCV-Specific Immune Responses and Viral Diversity.

HCV-specific LP responses in the study population are shown in Fig. 1. Overall, 126 (37%) of 341 subjects tested had a positive response to at least one HCV antigen using the pre-established cutoff of SI greater than 3. Responses to c100 (HCV aa 1569-1931) were most frequent (24% positive), followed by responses to c22 (HCV aa 2-120, 16% positive) and NS5 (HCV aa 2054–2995, 6% positive). Seventeen subjects (5%) had positive responses to 3 or more HCV antigens, 27 (7%) had positive responses to 2 HCV antigens, and 82 (24%) had a response to only 1 of the HCV antigens tested.

Figure 1.

Proliferation responses to HCV antigens in PBMCs of study subjects. Freshly isolated PBMCs were incubated with the recombinant HCV protein antigens or a corresponding SOD negative control for 6 days. 3H-thymidine incorporation was measured, and the results are expressed as the stimulation index (SI) for each antigen. (A) The percentage of subjects with a positive proliferation response (highest SI for any HCV antigen above the threshold value) is shown as a function of the threshold SI. (B) The distribution of SI values to each of the antigens is shown. Boxes show interquartile range, with the median indicated by a line. Error bars show 5th and 95th percentiles, with outliers plotted individually. The threshold value of SI = 3 used for most analyses is shown by a horizontal line.

HCV-specific CTL responses in the study population are shown in Fig. 2. Overall, HCV-specific CTL responses using our criteria for statistical analysis were detected in 45 (22%) of the 209 liver specimens tested. Of the subjects who met criteria for HCV-specific CTL responses, 17 (9%) showed a response to target cells expressing C-E1, 25 (13%) showed a response to target cells expressing E2-NS2, and 29 (14%) showed a response to target cells expressing NS2-NS5; 19 subjects (9%) showed responses to more than one target cell expressing HCV proteins.

Figure 2.

HCV-specific cytolytic activity of liver-derived T-cell lines. Intrahepatic T lymphocytes were expanded in culture by stimulation with anti-CD3 and recombinant human IL-2 and tested for CTL activity against autologous Epstein-Barr virus–transformed B lymphoblastoid cells infected with vaccinia viruses expressing HCV C-E1, HCV E2-NS2, or HCV NS2-NS5 or a control (expressing β-galactosidase). (A) The percentage of subjects with a positive CTL response (highest % specific lysis of any HCV target cell at any effector-target ratio above the threshold value after subtraction of % lysis of control target cell) is shown as a function of the threshold percent specific lysis. (B-C) Illustrative data from two subjects demonstrate recognition of HCV E2-NS2 (B) and HCV NS2-NS5 (C).

HCV-specific NA responses were detected in 54 (22%) of the 243 subjects studied, using a 50% or more plaque reduction threshold. Of the subjects with positive NA responses, NA responses were detected to the E1 glycoprotein in only 5 subjects (2%), to the E2 glycoprotein in only 12 subjects (5%), and to both E1 and E2 glycoproteins in 37 subjects (15%). The highest NA titers were 1:20 in 44 subjects and 1:50 in 10 subjects; none of the subjects had NA titers greater than 1:50.

QS diversity scores ranged from 0.75 to 1.00, with a mean (±SD) of 0.97 ± 0.04 and a median (interquartile range) of 0.98 (0.95, 1.00). QS complexity scores ranged from 1 to 16, with a mean (±SD) of 5.22 ± 3.01 and a median (interquartile range) of 5 (3, 7).

Associations Among Immunological Responses and Correlations With Viral Diversity.

Of the 367 subjects who had one or more of the immunology or virology assays performed, 140, 88, and 71 subjects had data available from two, three, or four assays, respectively. The associations among the results of these immunological and virological studies are summarized in Table 2. There were no statistically significant associations between detectable responses for any pairwise comparisons of the three immunological assays, either for the study cohort as a whole (Table 2) or when the analysis was restricted to subjects infected with HCV genotype 1 (data not shown). Similarly, no significant associations were seen between QS complexity or diversity and results of any of the immunological assays; this remained true when we considered only CTL and NA responses to the E2 region (data not shown).

Table 2. Pairwise Comparisons Among Assays of HCV-Specific Immune Responses and Viral Quasispecies
  1. Abbreviations: LP, lymphocyte proliferation; CTL, cytotoxic T lymphocyte; NA, neutralizing antibody; QS, quasispecies.

CTL positive, N1329
CTL negative, N5892
Odds ratio (95% CI)0.71 (0.34, 1.48)  
NA positive, N20311020
NA negative, N661072783
Odds ratio (95% CI)1.05 (0.55, 1.98)1.54 (0.64, 3.69) 
QS, N584058197425
QS complexity, mean ± SD5.58 ± 3.295.16 ± 2.854.95 ± 3.175.86 ± 3.145.32 ± 3.215.31 ± 2.99
QS diversity, mean ± SD0.97 ± 0.040.96 ± 0.040.97 ± 0.040.96 ± 0.050.96 ± 0.030.96 ± 0.05

Associations With Clinical, Laboratory, and Histological Observations.

Table 3 summarizes the main results of the analysis of clinical and histological associations with the immunological and virological parameters measured. There were no significant associations with age, sex, body mass index, previous ribavirin therapy, alcohol use, diabetes mellitus, illicit drug use, neutrophil count, serum ferritin, or iron saturation (data not shown). Associations that were statistically significant at the P ≤ .05 level included a slightly higher frequency of detectable LP responses in black subjects (22 of 41, 54%) than in non-black subjects (104 of 300, 35%; P = .02) and a weak negative correlation between QS diversity and platelet counts (r = −0.31, P = .002).

Table 3. Associations of HCV-Specific Immune Responses and Viral Quasispecies With Clinical and Histological Data
GroupALTIshak Fibrosis ScoreIshak Inflammatory ScoreCirrhosisHCV RNA (log10)
  • Abbreviations: LP, lymphocyte proliferation; CTL, cytotoxic T lymphocyte; NA, neutralizing antibody; QS, quasispecies. Values shown are mean ± SD except where indicated.

  • *

    R values shown for correlations.

LP positive98.7 ± 66.53.7 ± 1.17.3 ± 2.125%6.48 ± 0.57
LP negative103.7 ± 67.13.9 ± 1.37.1 ± 2.234%6.44 ± 0.54
CI difference(−9.7, 19.8)(−0.1, 0.4)(−0.7, 0.3) (−0.2, 0.1)
CTL positive128.9 ± 94.33.8 ± 1.27.2 ± 1.827%6.45 ± 0.50
CTL negative95.7 ± 61.63.7 ± 1.27.3 ± 2.324%6.48 ± 0.47
CI difference(−56.4, −10.0)(−0.5, 0.3)(−0.6, 0.9) (−0.1, 0.2)
NA positive92.6 ± 52.74.0 ± 1.27.0 ± 2.132%6.44 ± 0.68
NA negative104.8 ± 73.83.7 ± 1.27.1 ± 2.226%6.46 ± 0.55
CI difference(−9.7, 32.6)(−0.6, 0.1)(−0.6, 0.7) (−0.2, 0.2)
QS complexity*−0.16−0.030.07−0.03−0.06
QS diversity*0.20−0.14−0.05−0.130.10

The frequency of HCV genotype 1 infection was not different in those with or without LP, CTL, or NA responses. Mean serum alanine aminotransferase (ALT) levels were significantly higher in subjects with positive CTL responses than in those without detectable responses (128.9 ± 94.3 vs. 95.7 ± 61.6 U/L, Fig. 3), but no similar associations were found with LP, NA, or QS results. Histological findings as assessed by the Ishak scoring system did not show any significant associations with LP, CTL, or NA immune responses or with QS measurements.

Figure 3.

Association of HCV-specific CTL responses of liver-derived T-cell lines with serum ALT levels. The distributions of ALT values in subjects with detectable CTL responses (CTL+, N = 50) and in those without detectable CTL responses (CTL, N = 183) are shown. Boxes show interquartile range, with the median indicated by a solid line. Error bars show 5th and 95th percentiles, with outliers plotted individually. The dashed lines represent the group means.


The HALT-C trial, involving a large number of subjects with detailed clinical information including liver biopsy, provided a unique opportunity to study the clinical implications of HCV-specific immune responses and viral QS characteristics in a large cohort. Our success at coordinating multiple specialized laboratory analyses, including assays of non-cryopreserved blood and tissue samples, in a significant fraction (∼30%) of the HALT-C cohort could therefore serve as a model for incorporating such studies into other large cohorts. The number of subjects in whom we tested for relationships among several immunological responses and among clinical, virological, and immunological parameters is considerably larger than that of any previously reported study.

Analysis of the HALT-C trial cohort has several limitations. This population is characterized by a long duration of infection, unresponsiveness to prior IFN-α therapy, and advanced hepatic fibrosis.20 Published data have suggested that immune responses are critical to achieving a sustained virological response to treatment,29 and, thus, selection for non-responders to previous treatment might have enriched this cohort for immunological nonresponders. We found a low frequency of HCV-specific immune responses measured by LP, CTL, and NA assays in our subjects; the intrahepatic CTL responses were lower than in our previous studies of patients with less severe disease.30 The clinical characteristics of HALT-C trial subjects noted above likely explain in part this low frequency of detectable HCV-specific immune responses. Furthermore, because immune responses were not measured at earlier times in these subjects, we cannot determine whether poor HCV-specific immune responses contributed to the establishment of chronic HCV infection, as has been suggested by studies of acute HCV infection,11, 30 or whether these poor HCV-specific immune responses are a result of immune exhaustion.31 In either case, we recognize that our findings may not apply to all categories of patients with HCV infection.

Specimen availability and logistical considerations led us to analyze proliferation responses of PBMCs to recombinant HCV antigens and cytotoxic activity of in vitro–expanded intrahepatic lymphocytes against autologous BLCL expressing HCV proteins as measures of HCV-specific cellular immunity. Both have limited sensitivity to detect HCV-specific immune responses. Other techniques, such as HLA-peptide tetramer staining and IFN-γ ELISPOT assay using overlapping peptides spanning the entire HCV genome, have been suggested as more sensitive and precise measures of HCV-specific T cells.11, 13, 32 Nevertheless, these approaches also have limitations and are complex and costly to apply to large cohorts. In addition, none of these measures of HCV-specific T-cell responses has yet shown greater clinical relevance than the assays we used. Similarly, although pseudotype VSV virions expressing HCV E1 or E2 alone do not reproduce the natural virion surface, there are no validated assays of HCV-specific NA responses given the absence of a robust tissue culture system. Conversely, the large size of the current cohort made it possible to make observations about clinical outcome even with relatively insensitive assays.

We failed to find significant associations of LP responses with either CTL or NA responses. These negative data suggest that CD4+ T-cell help may not be required for maintenance of HCV-specific CD8+ T cells or B cells late in infection after memory lymphocytes are established. A recent report in the chimpanzee model demonstrates the persistence of memory CD8+ T cells even after depletion of CD4+ T cells.33 Alternatively, CD4+ T-cell help may only be provided effectively by cells localized to the liver, which were not measured by the LP assay with PBMCs used in this study. The absence of associations between QS variables and HCV-specific immune responses suggests that selective pressure exerted by antibody and T-cell responses on the viral QS in these subjects with long-standing HCV infection is weak or absent. Consistent with this interpretation, HCV-specific T-cell responses detected in these subjects were predominantly directed at epitopes in the core and nonstructural proteins, which were not evaluated by QS analyses.

HCV-specific immune responses and HVR1 QS diversity had only limited associations with clinical and histological findings in this study cohort. None of the immune response or viral QS variables measured was associated with demographic characteristics of the patients or histological findings as measured by the Ishak scoring system. The lack of such associations was not an artifact of our approach to analysis of the data. Alternative methods of analysis, including using different thresholds to define positive and negative assay results, considering LP and CTL data as continuous variables, use of nonparametric statistical tests, and multivariate analyses, did not affect the main positive and negative findings of the study. A trend was seen toward a negative relationship between cirrhosis on liver biopsy and HCV-specific LP responses in PBMCs; however, this did not achieve statistical significance at the P = .05 level (Table 3). Strong HCV-specific LP responses are characteristic of acute HCV infection and decline during the transition to chronicity,10 so our findings could reflect a continuing decline in LP responses as disease progresses along the continuum from fibrosis to cirrhosis.

The positive associations found in the current study cohort confirm, in general, results from other smaller studies. Sugimoto et al.34 noted a higher frequency of LP responses to HCV antigens in blacks than in whites, and they and others have described racial differences in the likelihood of chronic infection and biochemical and histological measures of disease. The difference in LP responses between black and non-black subjects in our study remained significant even when the presence or absence of cirrhosis was considered in the analysis. An association between intrahepatic CTL responses and elevated serum ALT was also noted by Nelson et al.16 and supports the model of immune-mediated hepatocyte injury.35, 36 The association of CTL responses with higher ALT values but not with lower HCV RNA levels suggests that the rate of clearance of HCV-infected cells is not the principal determinant of steady-state HCV RNA levels during chronic infection. The failure of HCV-specific T-cell responses to effectively control viral replication may be a reflection of a failure of effector functions of HCV-specific T cells, as has been described.11, 37, 38

QS diversity showed a negative correlation with platelet counts and a positive correlation with serum ALT in our study cohort. Both thrombocytopenia and hepatocyte injury in chronic HCV infection have been attributed to viral and immunological mechanisms,39 and increased viral QS diversity has the potential to affect both components. Although these correlations were statistically significant, it should be noted that the effects were weak (r values of −.31 and .20). Therefore, these associations require confirmation in other study cohorts.

In conclusion, the current analysis considered clinical, laboratory, histological, immunological, and virological information gathered at a single point in time for each subject, and therefore represents only a snapshot in time during the long course of a chronic disease. If immune-mediated liver injury is cumulative, the liver histology would be expected to depend on both the level of HCV-specific immune responses and the duration of infection. Variation in the duration of infection in the study cohort would therefore constitute a confounding factor in our analysis; that we detected some immunological associations with disease despite this statistical bias likely strengthens our conclusions and necessitates cautious interpretation of the negative data. Longitudinal evaluation of disease progression and evolution of immune responses and viral QS in this and other cohorts should provide additional information critical to a more complete understanding of virus–host dynamics in chronic HCV infection.


This is publication number 3 from the HALT-C Trial Group. The authors would like to thank the following individuals who contributed to the laboratory studies: Megan Allison, Michael A. Austin, Michelle Gano, Qi He, M.D., Rachel B. Life, Keith Meyer, Kelly White, and Christina Yoshihara. The authors would also like to thank Leslye Johnson, Leonard Seeff, and James Everhart of the National Institutes of Health for support, encouragement and critical discussions, Michael Houghton of Chiron Corporation and Charles Rice of the Rockefeller University for generously providing recombinant HCV proteins and vaccinia viruses, and the HALT-C trial investigators, research staff, and participants. In addition to the authors and individuals noted above, the members of the HALT-C Trial Group who coordinated the collection of specimens and contributed to the execution of the clinical study are: University of Massachusetts Medical Center/University of Connecticut—Savant Mehta, M.D., Maureen Cormier, R.N., B.S.N., Michelle Kelley, A.N.P., Donna Giansiracusa, R.N., B.S.N., Dawn Bombard, R.N.; Saint Louis University School of Medicine—Debra King, R.N., Cherryl Korte, R.N., B.S.N., Judith Thompson, R.N., Patricia Osmack; University of Texas Southwestern Medical Center—William M. Lee, M.D., Rivka Elbein, R.N., Nicole Crowder, L.V.N.; University of Southern California—Maurizio Bonacini, M.D., Susan L. Milstein, R.N., Carol B. Jones, R.N.; National Institute of Diabetes & Digestive & Kidney Diseases, Division of Digestive Diseases and Nutrition—Patricia R. Robuck, Ph.D.; New England Research Institutes—Michael C. Doherty, M.S., Maggie McCarthy, M.C.I., M.P.H., Michael G. Burton-Williams, M.D., M.P.H.; Roche Laboratories, Inc.—Raymond S. Koff, M.D., Michael J. Brunda, Ph.D.