T-cell response relative to genotype and ethnicity during antiviral therapy for chronic hepatitis C

Authors

  • David E. Kaplan,

    1. Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia
    Search for more papers by this author
  • Kazushi Sugimoto,

    1. Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia
    Search for more papers by this author
  • Fusao Ikeda,

    1. Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia
    Search for more papers by this author
  • Jason Stadanlick,

    1. Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia
    Search for more papers by this author
  • Mary Valiga,

    1. Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia
    Search for more papers by this author
  • Kirti Shetty,

    1. Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia
    2. Division of Gastroenterology, Department of Medicine, Philadelphia VA Medical Center, Philadelphia, PA
    3. Division of Gastroenterology, Department of Medicine, Georgetown University, Washington, DC
    Search for more papers by this author
  • K. Rajender Reddy,

    1. Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia
    Search for more papers by this author
  • Kyong-Mi Chang

    Corresponding author
    1. Division of Gastroenterology, Department of Medicine, University of Pennsylvania, Philadelphia
    2. Division of Gastroenterology, Department of Medicine, Philadelphia VA Medical Center, Philadelphia, PA
    • Division of Gastroenterology, Department of Medicine, University of Pennsylvania & Philadelphia VAMC, A212 Medical Research, Philadelphia VAMC, University and Woodland Avenues, Philadelphia, PA 19104
    Search for more papers by this author
    • fax: 215-823-4394


  • Potential conflict of interest: Nothing to report.

Abstract

Viral genotype and host ethnicity are important predictors of viral clearance during antiviral therapy for chronic hepatitis C virus (HCV) infection. Based on the role of T cells in natural HCV clearance, we hypothesized that T cells may contribute to the genotypic and ethnic difference in treatment outcome. To test this hypothesis, T-cell response to HCV antigens (core, nonstructural NS3/4 and NS5) and control phytohemagglutinin (PHA) was monitored prospectively and was correlated with virological outcome in 41 patients chronically infected with HCV (27 genotype 1, 14 genotype 2 or 3; 19 black persons, 22 white persons) undergoing combined interferon alfa and ribavirin therapy. Interestingly, in patients with genotype 2 or 3 infection, enhanced virological response coincided with a greater T-cell response to HCV NS3/4 antigen at baseline (50% vs. 15%; P = .026) that augmented further during therapy (29% vs. 4%; P = .035) compared with genotype 1-infected patients. However, HCV-specific T-cell response remained weak in genotype 1-infected patients regardless of virological outcome or ethnicity. Furthermore, virological outcome was associated with a suppressed baseline proliferative response to phytohemagglutinin (P < .03) that increased during therapy (P < .003) independent of ethnicity or genotype. In conclusion, HCV-specific T-cell response was associated with HCV genotype but not with therapeutic clearance of HCV infection. The association between treatment outcome and phytohemagglutinin response suggests more global and antigen-nonspecific mechanisms for therapeutic HCV clearance. (HEPATOLOGY 2005;41:1365–1375.)

T cells play a critical role in natural hepatitis C virus (HCV) clearance. Indeed, natural HCV clearance is associated with a vigorous and multispecific virus-specific T-cell response, contrasting with a weak, focused, or dysfunctional response in chronic evolution.1–8 In established chronic infection, HCV can be cleared therapeutically in approximately 50% of patients with the current standard of therapy that combines pegylated interferon alfa (IFN-α) and ribavirin.9–11

Although the determinants of therapeutic HCV clearance are not precisely defined, both host and viral factors contribute to the therapeutic outcome. One key factor is the viral genotype, as shown by the cure rate of 46% in genotype 1-infected patients compared with 80% in genotype 2- or 3-infected patients after standard therapy.9–11 Host ethnicity also influences treatment outcome, with significant IFN-α resistance in black persons relative to white persons.12–15 Early virological response (EVR), defined as at least a 2 log drop in serum HCV titer within the first 12 weeks of therapy, also predicts the ultimate virological outcome,16 suggesting that the determinants of treatment outcome are already present early in the course of antiviral therapy.

Based on the key role for T cells in natural HCV clearance and many potential pleiotropic effects of IFN-α, it has been suggested that IFN-α–based antiviral therapy promotes HCV clearance by augmenting the HCV-specific T-cell response in some, although not all, studies.17–22 Along this line, the apparent IFN-α resistance in black patients or some genotype 1-infected patients may be the result of persistent T-cell dysfunction that is unresolved during IFN-α–based antiviral therapy. In this study, we examined these hypotheses prospectively in a cohort of ethnically and genotypically defined United States veterans undergoing IFN-α plus ribavirin therapy for chronic HCV infection. Our results show that HCV-specific T-cell response during antiviral therapy is differentially augmented relative to genotype but not necessarily to viral clearance or ethnicity. Furthermore, treatment response was associated with lymphoproliferative response to a nonspecific mitogen phytohemagglutinin (PHA), suggesting a more global, antigen-nonspecific immune mechanism in viral clearance during combined IFN-α plus ribavirin therapy.

Abbreviations

HCV, hepatitis C virus; IFN-α, interferon alfa; IFN-γ, interferon gamma; EVR, early virological response; NS, nonstructural; PHA, phytohemagglutinin; SVR, sustained virological response; C1, genotype 1; C2, genotype 2/3; PBMC, peripheral blood mononuclear cell; SFU, spot-forming unit.

Patients and Methods

Patients.

Patients were recruited from the Gastroenterology Clinics at the Philadelphia Veterans Administration Hospital and the Clinical Research Center at the University of Pennsylvania from 2001 through 2003 under institutional review board–approved protocols after patients gave informed consent. Treatment-eligible patients with chronic hepatitis C were enrolled prospectively before starting IFN-α plus ribavirin therapy. Due to evolving clinical practice, the type of IFN-α included IFN-α2b, pegylated IFN-α2b, or pegylated IFN-α2a with expected treatment duration of 48 weeks for genotype 1 and 24 weeks for genotype 2 or 3 infection.11 All patients were assessed for baseline demographic, clinical, and virological parameters, including HCV viremia by Roche COBAS qualitative or quantitative reverse-transcriptase polymerase chain reaction assay (Roche Diagnostics, Branchburg, NJ) and HCV genotype by InnoLIPA (Innogenetics, Gent, Belgium). Ethnicity was defined based on patient self-report and medical records. All patients were assessed for HCV-specific CD4 T-cell response to recombinant HCV antigens at baseline, within the first month (2-4 weeks); at 12, 24, and 48 weeks after treatment initiation; and at 4 to 6 months after treatment for genotype 1-infected patients. A subset of patients also was examined for T-cell IFN-γ response to HCV peptides based on available lymphocytes. Although blood draw was limited in some patients because of treatment-induced anemia and early treatment cessation limiting follow-up, patients completing therapy were examined at 94% of protocol time points.

Forty-one patients were included in the analysis based on completing at least 12 consecutive weeks of therapy to assess for EVR, defined as at least a 2 log-fold reduction in HCV titer by Roche reverse-transcriptase polymerase chain reaction at 12 weeks. Patients were divided based on genotype into groups C1 (genotype 1) and C2 (genotypes 2 and 3), with further subdivision by EVR. End-of-treatment response (ETR) and sustained virological response (SVR) were defined as negative HCV RNA by reverse-transcriptase polymerase chain reaction at treatment completion and 6 months after completion, respectively. Controls included 23 spontaneously recovered (HCV antibody positive/HCV RNA negative) individuals without prior antiviral therapy (group R) and 20 healthy HCV-seronegative individuals without a history of HCV exposure (group N). Exclusion criteria included a history of human immunodeficiency virus or HBV coinfection, immunosuppressive therapy, autoimmune hepatitis, primary biliary cirrhosis, and conditions precluding research blood donation.

Recombinant HCV Proteins.

Recombinant genotype 1a–derived HCV core, NS3/4, and NS5 and control superoxide dismutase proteins were provided by Dr. Michael Houghton (Chiron Corp., Emeryville, CA).17, 23, 24 We previously showed that proliferation and IFN-γ responses to these proteins are based on CD4 T cells.4, 7

Overlapping HCV Peptides.

Three hundred sixty-one overlapping 15-mer peptides (offset by six amino acid residues) spanning the entire HCV core and NS3-NS5 proteins based on HCV-H sequence (genotype 1a) were commercially synthesized (Genemed, San Francisco, CA) and mixed into 12 separate pools each containing 20 to 31 consecutive peptides or 36 pools of 10 to 11 consecutive peptides as previously described.6, 7 These peptides were immunogenic for CD4 and CD8 T cells.6, 7

Peripheral Blood Mononuclear Cells.

Peripheral blood mononuclear cells (PBMCs) were isolated from blood and were used directly to assess HCV-specific T-cell response.6, 7 HCV-specific CD8 T-cell response was assessed using CD4-depleted PBMC T cells by negative selection as previously described.6, 7

HCV-Specific CD4 Proliferative T-Cell Response.

CD4 proliferation assay was performed as previously described.4, 6, 7 Briefly, PBMC (2 × 105 cells/well) were stimulated for 7 days with HCV core, NS3/4, NS5, and control superoxide dismutase proteins (10 μg/mL) and were harvested after 16 hours of 3H-thymidine incorporation (1 μCi/well; Dupont NEN, Boston, MA). As a positive control for general T-cell responsiveness, PBMCs were stimulated with a T-cell mitogen PHA at 2 μg/mL and were harvested on day 4. The results were expressed as a stimulation index with mean counts per minute in stimulated wells divided by the mean counts per minute in control wells. A positive response was defined by cutoff values derived from 20 healthy controls (>mean stimulation index + 2 SD): core (2.8), NS3/4 (2.1), and NS5 (3.8) as previously described.7 Significant HCV-specific T-cell augmentation during therapy was defined as at least two positive responses during therapy that were at least 50% more than the baseline response.

IFN-γ Elispot Assay.

IFN-γ Elispot assay was performed using 2 × 105 PBMC/well in triplicates with and without HCV and control antigens as described previously.7 HCV-specific CD4 T-cell IFN-γ response (or type 1 helper T-cell response) in whole PBMCs was examined using recombinant HCV and control proteins (10 μg/mL). HCV-specific total T-cell IFN-γ response in PBMCs or CD8 T-cell IFN-γ response in CD4-depleted PBMCs were examined using overlapping HCV-derived 15-mers (5 μmol/L/peptide) with positive controls including PHA (2 μg/mL), tetanus toxoid (0.5 μg/mL), and Candida albicans (20 μg/mL).6, 7 IFN-γ spot-forming units (SFUs) were counted using an Elispot reader (Hitech Instruments, Media, PA). HCV-specific type 1 T-cell frequency was calculated by subtracting the mean IFN-γ SFUs in negative control wells from mean SFUs in antigen-stimulated wells and was expressed as HCV-specific IFN-γ SFU/106 PBMCs. HCV-specific CD8 T-cell IFN-γ response in CD4-depleted PBMCs was calculated by correcting for relative CD8 enrichment after CD4 depletion.6, 7 A positive response was defined by cutoff values derived from 20 healthy controls (>mean + 2 SD). Significant HCV-specific T-cell augmentation during therapy was defined as at least two positive responses during therapy that were at least 50% more than the baseline response. Results from each HCV antigen were summed for total HCV-specific type 1 helper T-cell response.

Statistical Analysis.

The median values for clinical and immunological parameters were compared using the nonparametric Kruskal-Wallis ANOVA, the Wilcoxon rank sum test, or the Mann-Whitney U test. The frequency of positive responses was compared using the chi-square test or Fisher's exact test based on sample size. Spearman rank correlation was used for bivariate correlation of variables. Multivariate regression was performed using JMP 5.1 (SAS Institute Inc., Cary, NC). A P value less than .05 was considered significant.

Results

Characteristics of Patients Undergoing IFN-α Plus Ribavirin Therapy.

Twenty-seven genotype 1-infected (group C1) and 14 genotype 2- or 3-infected (group C2) patients were included in the analysis based on completion of at least 12 weeks of antiviral therapy. C1 patients were separated further by early virological response to EVR-negative (EVR; n = 11) and EVR-positive (EVR+; n = 16) subgroups. All 14 C2 patients were EVR+ (Table 1). Thus, more group C2 patients were EVR+ than group C1 patients (100% vs. 59%; P = .016), consistent with the known genotypic difference in therapeutic response.9–11

Table 1. Baseline Patient Characteristics of IFNα-Treated Patients
 Chronic Patient Subgroups
All (n = 41)Genotype 1 (C1)Genotype 2,3 (C2)P Value*
EVR (n = 11)EVR+ (n = 16)EVR+ (n = 14)
  • Abbreviations: ALT, alanine aminotransferase; INR, International Normalized Ratio; PEG, pegylated.

  • *

    P value comparing C1, C1+, and C2+ subgroups by Kruskal-Wallis for median values and by chi-square analysis for frequency.

  • Biopsy results available in 28 of 41 patients.

  • Fibrosis score based on modified histologic activity index.

  • §

    P value derived by 3 × 3 × 3 chi-square analysis for frequency.

Demographic parameters     
 Age (yr)535354540.31
 % Black46736370.0012
 % Alcohol history765575930.34
 % Antiviral treatment naive806488860.25
Median clinical parameters     
 ALT (IU/mL)544758500.61
 Albumin (g/dL)4.14.14.34.20.91
 Total bilirubin (mg/dL)0.80.70.80.90.46
 INR1.01.01.01.00.14
 Platelets (×103/mm3)2172392351840.02
 HCV RNA titer (IU/mL)825,000850,000693,500840,0000.45
 % >850,000 IU/mL465538500.64
Liver fibrosis/cirrhosis     
 % with fibrosis grade >346 (13/28)29 (2/7)55 (5/11)70 (6/10)0.44
Antiviral therapy (all including ribavirin)     
 Standard IFN-α2b3102 
 PEG IFN-α2b2671090.58§
 PEG IFN-α2a12363 

Furthermore, there were fewer black patients in the C2 group compared with the C1 subgroups with and without EVR (7% C2 vs. 73% C1 EVR vs. 63% C1 EVR+; P = .0012) consistent with the known genotype 1 predominance among black persons.7, 25, 26 With the exception of platelet count, the subgroups shared similar baseline characteristics, including alcohol use, previous antiviral therapy, liver-associated laboratory measures, HCV RNA titer, histological results, and types of IFN-α administered (Table 1).

As shown in Tables 2 and 3, 27 patients (66%) completed the full course of therapy (48 weeks for C1, 24 weeks for C2), including 21 patients who were HCV RNA negative at the end of therapy (i.e., ETR+) and 18 patients who remained HCV RNA negative 6 months thereafter (i.e., SVR+). As expected, there were more ETR+ and SVR+ patients in the C2 group than the C1 group. As for ethnicity, more of the white patients were SVR+ than the black patients in our cohorts (82% vs. 40%; P = .039; Table 3). However, this was largely because of a higher frequency of non-genotype 1 infection among white patients with no apparent ethnic disparity among C1 patients (%SVR+; 44% of black patients vs. 43% of white patients; P = .99; Table 3). These genotypically and ethnically defined patients with diverse treatment outcomes provided the basis for our study.

Table 2. Virological Outcome Relative to Genotype and EVR
 Genotype and EVR Subgroups
All (n = 41)Genotype 1 (C1)Genotype 2,3 (C2)P Value* C1 vs. C2
EVR (n = 11)EVR+ (n = 16)EVR+ (n = 14)AllEVR+
  • Abbreviation: NA, not applicable.

  • *

    P value was calculated by chi-square or Fisher's exact test as appropriate; P < .05 in bold.

  • One C2 patient completed only 21 weeks of an expected 24 weeks of therapy but remained HCV RNA–negative by reverse-transcriptase polymerase chain reaction at weeks 24 and 48. Thus, the number of ETR+ and SVR+ patients in the C2 group was more than number of treatment completed patients.

No. EVR+30 (73%)0 (0%)16 (100%)14 (100%).007NA
No. treatment completed27 (87%)5 (45%)11 (69%)11 (79%).30.69
 No. ETR+21 (78%)0 (0%)9 (81%)12 (100%).022.22
 No. SVR+18 (67%)0 (0%)6 (54%)12 (100%).0009.013
Table 3. Virological Outcome Relative to Genotype and Ethnicity
 Ethnicity and Genotype Subgroups
AllP Value*Genotype 1 (C1)P Value*Genotype 2,3 (C2)P Value*
Black (n = 19)White (n = 22)Black (n = 18)White (n = 9)Black (n = 1)White (n = 13)
  • Abbreviation: NS, not significant.

  • *

    P value was calculated by chi-square or Fisher's exact test as appropriate. P < .05 in bold.

% EVR+58860.0855670.69100100NS
% Treatment completed53770.1150780.2310085NS
 % EVR+70820.6467430.61100100NS
 % SVR+40820.03944430.99100100NS

Baseline HCV-Specific T-Cell Response Does Not Predict Early or Sustained Virological Response.

As expected, the baseline HCV-specific CD4 T-cell response in the chronic subgroups was significantly weaker, particularly for the nonstructural NS antigens when compared with spontaneously HCV-recovered patients (Fig. 1; Table 4). Interestingly, among the chronic patients, CD4 T-cell responses to the NS antigens were detected more frequently in C2 than in C1 patients (50% vs. 15% for NS3/4; P = .026; Table 4). Although this initially suggested that the baseline T-cell responsiveness may contribute to the greater treatment response in C2 patients, only 2 of 16 C1 EVR+ patients responded to the NS antigens (Table 4), suggesting an effect associated with HCV genotype rather than treatment response. Along this line, HCV-specific CD4 T-cell response was not associated significantly with early or sustained virological responses to therapy when all of the chronic patients were considered together (Table 4). Furthermore, HCV-specific T-cell IFN-γ response was equally weak in all three subgroups regardless of EVR, contrasting with the vigorous response maintained in spontaneously HCV-recovered patients (Fig. 2A-B). Thus, pretreatment HCV-specific T-cell response did not predict viral clearance during IFN-α–based antiviral therapy for established chronic HCV infection in our study.

Figure 1.

Baseline hepatitis C virus (HCV)-specific CD4 proliferative T-cell responses. CD4 proliferative responses in stimulation index (SI) to HCV core, nonstructural (NS)3/4, and NS5 antigens are shown for 11 genotype 1 (C1) early virological response–negative (EVR), 16 C1 EVR-positive (EVR+), 14 genotype 2 (C2) EVR+, and 23 recovered (R) patients. Median values are indicated by horizontal black bars. Cutoffs for positive responses to the HCV antigens were derived from 20 healthy controls (2 SD + mean SI) and are indicated as horizontal dotted lines. As expected, T-cell responses to the NS antigens are significantly greater in the recovered compared with the chronic patients (NS3/4, 8.7 vs. 1.4; P < .0001; NS5, 2.0 vs, 1.1; P = .04). Among the chronic subgroups, the C2 EVR+ subgroup showed greater proliferative response to NS3/4 than the C1 subgroups (2.1 vs. 1.2; P = .015), whereas no significant differences were noted between C1 EVR+ and C1 EVR− subgroups. T-cell response to the NS5 antigen also differed significantly when all three chronic subgroups were compared with a Kruskal-Wallis ANOVA (P = .049). P values were calculated by the nonparametric Mann-Whitney U test. P values < .05 were considered significant and are highlighted in bold.

Table 4. Baseline HCV-Specific T-Cell Response Relative to Genotype and Virological Response
 % Responders*
CoreNS3/4NS5
  • *

    Percentage of responses in CD4 proliferation assays at baseline by for group in which the stimulation index (SI) was greater than cutoffs based on the mean plus 2 SD for 20 normal controls. P values, derived from chi-square analysis, are bolded if <.05.

  • Excludes EVR+ patients who did not complete therapy. SVR, therefore, includes EVR+ who completed therapy (but were SVR) and EVR.

Comparison of all subgroups   
 All (n = 41)27275
 C1 EVR (n = 11)27180
 C1 EVR+ (n = 16)25130
 C2 EVR+ (n = 14)295014
 Recovered437835
 Normal (n = 20)555
 Recovered vs. normal (P value).26<.0001.003
Comparison by genotype   
 C1 (n = 27)26150
 C2 (n = 14)295014
 P value1.0.026.11
Comparison by EVR   
 EVR (n = 11)27180
 EVR+ (n = 30)27307
 P value1.0.691.0
Comparison by SVR   
 SVR33200
 SVR+22336
 P value.70.461.0
Figure 2.

Baseline hepatitis C virus (HCV)-specific T-cell interferon gamma (IFN-γ) response. The bar graphs show the median T-cell IFN-γ responses (in IFN-γ spot-forming units [SFUs]/106 peripheral blood mononuclear cells [PBMCs]) to HCV core (black bars), nonstructural (NS)3/4 (gray bars), and NS (white bars) in the genotype 1 (C1) early virological response–negative (EVR), C1 EVR-positive (EVR+), and genotype 2 (C2) EVR+ as well as the recovered (R) and normal control (N) participants. (A) CD4 T-cell IFN-γ response to recombinant HCV core, NS3/4, and NS5 antigens shows significant differences between the chronic and recovered groups (C < R) but not the C1 EVR, C1 EVR+, and C2 EVR+ subgroups. (B) Total T-cell IFN-γ to overlapping 15mer peptides derived from HCV core, NS3/4, and NS5 antigens shows significant difference between the chronic and recovered subgroups (C < R) but not the chronic subgroups. P values were calculated using a Kruskal-Wallis ANOVA and are shown in bold if significant (P ≤ .05).

HCV-Specific CD4 T-Cell Response During Antiviral Therapy Is Associated with the Viral Genotype But Not HCV Clearance.

We then examined the evolution of HCV-specific T-cell response during antiviral therapy. As shown in Fig. 3, significant increases in HCV-specific T-cell response was observed in only a minority of patients based on our criteria described in Patients and Methods (at least 50% more than baseline at two or more time points). For example, CD4 proliferative T-cell response to HCV core was augmented in 9% to 25% of patients without a significant difference between all three subgroups (P = .58; Fig. 3A). T-cell responses to the NS antigens increased minimally in C1 patients with and without EVR (0% vs. 9%), suggesting a lack of correlation between virological outcome and HCV-specific T-cell response during antiviral therapy. Conversely, sustained virus suppression in the SVR+ patients did not result in increased HCV-specific T-cell response, contrasting with the vigorous responses maintained with natural HCV clearance. Indeed, although most genotype 1–infected SVR+ patients (red lines, Fig. 3) showed little to no augmentation in HCV-specific T-cell response during or after antiviral therapy, transient augmentations were seen in some SVR-negative (SVR) patients (black lines, Fig. 3). Notably, C2 patients tended to display augmented T-cell response particularly for the NS3/4 antigen compared with C1 patients (43% C2 vs. 4% C1; P = .004). However, this difference persisted even when only EVR+ patients were considered (C2 EVR+ 43% vs. C1 EVR+ 6%; P = .031), suggesting an effect associated with genotype rather than virological response. As for HCV-specific type 1 helper T-cell response, it did not differ significantly between groups at baseline (Fig. 2A-B) or during therapy, despite an increased tendency toward augmentation among the C2 patients (Fig. 3B). Concurrent analysis for HCV-specific CD4 response, CD8 T-cell IFN-γ response, or both using overlapping peptide pools in a subset of patients also showed no significant difference (data not shown). These results suggest that HCV-specific T-cell augmentation during antiviral therapy is associated with the infecting HCV genotype but not with virological outcome. Furthermore, they suggest that defective HCV-specific T-cell proliferation and IFN-γ production in chronic HCV infection is not necessarily reversed with viral clearance after IFN-α–based therapy.

Figure 3.

Evolution of hepatitis C virus (HCV)-specific CD4 T-cell response in chronically HCV-infected patients during antiviral therapy relative to HCV genotype and virological response. (A) CD4 proliferative T-cell responses (stimulation index [SI]) to HCV core, nonstructural (NS)3/4, and NS5 antigens during and after antiviral therapy for each patient in the genotype 1 (C1) early virological response–negative (EVR), C1 EVR-positive (EVR+), and genotype 2 (C2) EVR+ subgroups are displayed as line graphs before therapy (time 0) and at several time points during and after therapy: at the first month (2-4 weeks), between 3 and 6 months, at 12 months (end of therapy for C1 subgroups and 6 months after treatment completion for C2 subgroup), and at 18 months (6 months after therapy for C1 subgroup). A higher value was used if two data points were available for analysis within each time range. The percentage of patients with significant augmentations in HCV-specific T-cell responses (defined as 50% or more of baseline at two or more time points during therapy) is indicated on the right side of each graph. Only a minority of patients displayed augmented HCV-specific T-cell response (0%-43%), with a significant difference between the subgroups for NS3/4 (C2 > C1 subgroups, P = .004) but not core (P = .99) or NS5 (P = .11). Similarly, HCV-specific T-cell augmentation occurred in only minority of sustained virological response–positive (SVR+) patients (red lines). (B) HCV-specific type 1 helper T-cell response (IFN-γ + spot-forming units [SFUs]/106 peripheral blood mononuclear cells [PBMCs]) during and after antiviral therapy is shown summing the response to recombinant HCV core and NS3-NS5 antigens. A higher value was used if two data points were available for analysis within each time range. Similar to the proliferative response, C2 EVR+ patients showed augmented IFN-γ responses more frequently than patients in the C1 subgroups (50% vs. 16%; P = .026). T-cell responses in SVR+ patients are indicated in red. There was no significant association between augmented HCV-specific T-cell IFN-γ response and SVR or EVR.

Host Ethnicity Is Not Associated with Treatment-Induced Augmentation of HCV-Specific T-Cell Response.

Because host ethnicity can influence treatment outcome and T-cell responses in HCV infection,7, 12–15, 25, 26 we examined the potential impact of ethnicity in HCV-specific T-cell response relative to treatment outcome, focusing on the C1 subgroup with 18 black and 9 white patients. Although the number of patients is low, there was no apparent ethnic difference in baseline HCV-specific T-cell responsiveness relative to early virological response (Fig. 4A). During antiviral therapy, augmentation in HCV-specific T-cell proliferative and IFN-γ responses occurred in equally small subsets of black and white patients, mostly for the HCV core (17% vs. 22%) rather than the NS antigens (0%-6%; Fig. 4B-C). In both ethnic groups, these augmentations occurred in SVR+ (red lines, Fig. 4B-C) and SVR (black lines, Fig. 4B-C) patients, suggesting that ethnic difference in treatment response is not based on HCV-specific T-cell response.

Figure 4.

Hepatitis C virus (HCV)-specific T-cell response relative to ethnicity and virological response to therapy. (A) Percent HCV-specific CD4 proliferative T-cell responders at baseline among genotype 1–infected black and white patients with and without early virological response (EVR). Both ethnic groups showed no association between CD4 T-cell response and any of the HCV antigens and EVR. (B) HCV-specific CD4 T-cell proliferation (stimulation index [SI]) during therapy did not show a distinct ethnic difference between black and white patients. The frequency for patients with significant augmentations in HCV-specific CD4 T-cell response (defined as at least two positive responses that were at least 50% more than the baseline response and indicated on the left upper corner of each graph) was no different between the ethnic groups. A higher value was used if two data points were available for analysis within each time range. (C) HCV-specific CD4 T-cell interferon gamma (IFN-γ) response during therapy. The frequency of subjects with significant augmentations in HCV-specific CD4 T-cell response (defined as at least two positive responses that were at least 50% more than the baseline response and indicated on the left upper corner of each graph) was no different between the ethnic groups. Sustained virological response–positive (SVR+) patients are indicated by red lines. A higher value was used if two data points were available for analysis within each time range. Dotted horizontal lines indicate the cutoffs for a positive responses derived from normal controls. SFU, spot-forming unit; PBMC, peripheral blood mononuclear cell; NS, nonstructural.

Early Virological Response Correlates with Weak Baseline Proliferative Response to T-Cell Mitogen PHA That Is Rapidly Augmented with Therapy.

Because HCV-specific T-cell response generally was suppressed without consistent augmentation during therapeutic viral clearance, we looked for a more global T-cell dysfunction associated with treatment outcome using a nonspecific T-cell mitogen, PHA. T-cell response to PHA was no different among the chronic, recovered, and normal control participants at baseline (Fig. 5A), consistent with our previous report.7 Interestingly, the baseline PHA response was significantly greater in C1 EVR than in the EVR+ subgroups (Fig. 5B). Indeed, patients with pretreatment PHA response less than the average for normal control participants (stimulation index, 272) were more likely to be EVR+ (86% vs. 46%; P = .008). In multivariate regression analysis controlling for various baseline characteristics (HCV genotype, age, ethnicity, body mass index, alanine aminotransferase, bilirubin, albumin, International Normalized Ratio (INR), platelet count, HCV RNA titer, prior antiviral therapy, and PHA response), baseline PHA response was the only significant predictor of early virological response (P = .036). Interestingly, PHA response tended to increase within 1 month of antiviral therapy in C1 EVR+ and C2 EVR+ patients but not in C1 EVR patients (Fig. 5C). The increased on-treatment PHA response from baseline was significant for EVR+ as well as SVR+ patients, but not EVR (and presumably SVR) patients (Fig. 5D). Patients with augmented PHA response of 50% more than the baseline within the first month of therapy were more likely to be EVR+ (89% vs. 63%; P = .046) and SVR+ (73% vs. 37%; P = .032, excluding EVR+ patients who discontinued therapy) than those without PHA augmentation. The baseline PHA response covaried with ethnicity (mean stimulation index, 265 for black patients vs. 111 for white patients; P = .014), although lower PHA response was seen in both EVR+ black and white patients (Fig. 5E). The difference in PHA response did not correlate with the circulating frequency of total T cells or CD4+CD25+ T cells (data not shown), suggesting that the association between PHA responsiveness and treatment outcome is independent of total T cell or CD25+ regulatory T-cell frequency.

Figure 5.

Phytohemagglutinin (PHA) response and the outcome of antiviral therapy. Proliferative responses to PHA in peripheral blood mononuclear cells (PBMCs) are shown in stimulation index (SI) with associated P values by Wilcoxon rank sum test (P ≤ .05 in bold). (A) PHA responses at baseline in the 23 recovered (R), 41 chronic (C), and 20 normal control (N) patients are shown as scatter plots. Horizontal black lines indicate the median values without a significant difference between the subgroups. (B) PHA responses at baseline in 11 genotype 1 (C1) early virological response–negative (EVR), 16 C1 early virological response–positive (EVR+), and 14 genotype 2 (C2) EVR+ patients are shown as scatter plots. Horizontal black lines indicate the median values (significantly lower in both EVR+ subgroups compared with the C1 EVR subgroup). (C) Evolution of PHA responses in median SI before and after starting therapy is shown at various time points for 11 C1 EVR, 16 C1 EVR+, and 14 C2 EVR+ subgroups, demonstrating that the baseline difference between the EVR+ subgroups and C1 EVR subgroup is lost at 1 month of therapy. A higher value was used if two data points were available for analysis within each time range. (D) Median PHA responses at baseline and on therapy relative to EVR and sustained virological response (SVR) are shown as bar graphs for 11 EVR, 30 EVR+, and 18 SVR+) patients, highlighting the augmented responses with therapy in EVR+ and SVR+ but not EVR patients. This difference persisted when C2 patients were excluded, suggesting a genotype-independent effect (P = .012 for month 1; P = .014 for months 3-6; data not shown). (E) Ethnicity and the trend in PHA response relative to virological outcome. There was no difference in the pattern of PHA responsiveness between black and white patients among EVR+ and EVR groups.

Discussion

IFN-α is a type 1 interferon that induces numerous gene products with antiviral activity as well as enhances cell-mediated immunity with increased HLA expression and natural killer cell activity.27, 28 Although the direct antiviral activity of IFN-α against HCV has been demonstrated in vitro in the HCV replicon system,29, 30 it has been suggested that augmentation of HCV-specific T-cell response by IFN-α in vivo can contribute to viral clearance during therapy in some although not all studies.17–22 Furthermore, the nucleoside analog ribavirin may promote a type 1 helper T-cell response that favors viral clearance.19, 31, 32 Because HCV genotype and host ethnicity (at least in the United States) are important determinants of treatment outcome,12, 13, 25 we examined the role of HCV-specific T-cell response in virological response to combined IFN-α plus ribavirin therapy relative HCV genotype and ethnicity in patients with established chronic HCV infection.

Compared with the robust and broad responses maintained in patients with natural HCV clearance, HCV-specific T-cell response was weak and focused in patients with chronic HCV infection with a baseline genotypic difference (genotype 2 or 3 > 1) that was accentuated further during therapy. Because most patients with genotype 2 or 3 infection achieved sustained virological response, this initially suggested an association between HCV-specific T-cell responsiveness and treatment response. However, in genotype 1-infected patients, both early and sustained viral clearance occurred with little to no increases in HCV-specific T-cell response. These results suggested that IFN-α–based antiviral therapy can augment HCV-specific T-cell response with an apparent association with genotype. However, augmentation in HCV-specific T-cell response was neither sufficient nor necessary for viral clearance.

Our findings differ from several previous reports demonstrating an association between sustained increase in HCV-specific T-cell response and virological response to antiviral therapy in chronically HCV-infected patients.18, 19 However, because the influence of viral genotype was not examined specifically in these studies, genotypic differences in HCV-specific T-cell response could have contributed to the discrepant results. Our results are more consistent with recent reports in the setting of acute hepatitis C in which virus-specific T-cell response was not augmented during antiviral therapy.21, 22

The nature of the genotypic difference in HCV-specific T-cell response is not clear. Because most genotype 2- or 3-infected patients were SVR+, we could not directly compare the level of HCV-specific T-cell augmentation relative to SVR in these patients. However, because HCV-specific T-cell response was augmented in only one third of SVR+ genotype 2- or 3-infected patients with antiviral therapy, T-cell augmentation apparently was not necessary for viral clearance in most cases. It remains an open question if T-cell augmentation may be relevant for early viral kinetics in genotype 2 or 3 infection. Another consideration is that HCV-specific T-cell response was detected using viral antigens derived from a prototype genotype 1 sequence in most studies, including ours. T cells specific for genotype 1-derived antigens in patients without active genotype 1 infection could reflect a memory response to a previously cleared virus rather than the circulating virus.33 If this is correct, genotype 1-specific T cells in patients with genotype 2 or 3 infection may be nonspecific for the circulating virus and therefore ineffective in viral clearance, even after immune augmentation by IFN-α plus ribavirin therapy.

The ethnic disparity in resistance to IFN-α–based antiviral therapy was not immunologically based in our study. Indeed, we previously reported that HCV-specific T-cell response was more readily detectable in black than in white patients, albeit dysfunctional in coordinated antigen-specific proliferation and IFN-γ production.7 However, the lack of enhanced HCV-specific T-cell proliferation or IFN-γ response in patients with virological response suggested that effector T-cell dysfunction is not resolved with antiviral therapy or sustained viral clearance. The general lack of immune augmentation in our patients also was consistent with the generally decreased biochemical liver disease in HCV-infected patients undergoing antiviral therapy. It also contrasts with the immune reconstitution and biochemical flares observed in HBV-infected patients undergoing antiviral therapy.34–37 Because the outcome of IFN-α plus ribavirin therapy is independent of HCV-specific T-cell response, therapeutic augmentation of HCV-specific T-cell response could be a useful adjuvant treatment method, particularly in patients resistant to IFN-α plus ribavirin therapy.

An unexpected finding in our study was the association between treatment response and lymphoproliferative response to PHA, a T-cell mitogen often used to control for overall T-cell function and cell viability.3, 7, 38–40 The reason for this association is not clear. It was not associated with disparities in standard clinical, demographic, or virological parameters. The association was not the result of cell viability differences, because all assays were performed with freshly isolated lymphocytes on the same day of blood draw. PHA responsiveness also could be genetically based, although this was not assessed directly in our study. It was linked to but not dependent on ethnicity, because EVR+ patients showed low baseline PHA responses that rose promptly with antiviral therapy regardless of ethnicity. Although unresponsiveness to PHA may be associated with the frequency and activation of the circulating T cells (e.g., in simian/human immunodeficiency virus),41 there was no significant difference in circulating total T-cell or regulatory CD4+CD25+ T-cell frequencies in our patient subgroups. PHA response also can be a function of accessory cells such as monocytes and dendritic cells as well as their secreted factors (e.g., interleukin-6).42, 43 Along this line, HCV gene products (e.g., core and NS3) can activate inflammatory pathways in monocytes via the toll-like receptor44 or can suppress effector T-cell differentiation by inhibiting interleukin-2 production (e.g., core),45, 46 whereas dendritic cells in HCV-infected patients may have impaired differentiation and allostimulatory capacity.47, 48 Although further studies are needed to determine the underlying mechanism for our observation, the PHA response before and during therapy may identify a subset of patients in whom IFN-α plus ribavirin therapy can be more effective.

It is also important to acknowledge some of the potential limitations in our study. First, we did not assess of the evolution of intrahepatic T-cell response. Thus, the absence of apparent T-cell augmentation in our patients who cleared HCV may reflect compartmentalization to the liver, where they may be more effective. Nonetheless, virus-specific T-cell response has been detected readily in the peripheral blood of patients during natural HCV clearance, suggesting that peripheral T-cell responses can be associated with viral outcome. Second, we used early rather than sustained virological response to define the patient subgroups because more than half of the EVR patients did not complete therapy. However, because 99% of EVR patients are also SVR after a full course of therapy,16 it is not unreasonable to assume that our EVR patients define a group of treatment nonresponders. Third, most assays were performed prospectively using freshly isolated lymphocytes at the time of sample collection, and thus were subject to potential variables in experimental condition. Although such variables may be reduced by performing all assays simultaneously using cryopreserved cells, cryopreserved cells can be less proliferative and subject to variable cell viability during freezing and thawing. Therefore, given the generally weak HCV-specific T-cell response in chronic HCV infection, we used freshly isolated cells to maximize the chance of detecting a positive HCV-specific T-cell response while avoiding falsely negative responses. Finally, because the patients in our study were monitored only for 6 months after treatment completion, we cannot exclude the possibility that HCV-specific T-cell response will improve after a longer duration of viral clearance.

In conclusion, clearance of chronic HCV infection with combined IFN-α plus ribavirin therapy was not associated with increased HCV-specific T-cell response, contrasting with the robust T-cell response in natural HCV clearance. The evolution of HCV-specific T-cell response during antiviral therapy was associated with genotype but not ethnicity. Remarkably, HCV clearance with antiviral therapy correlated with proliferative response to a nonspecific T-cell mitogen, PHA, suggesting a more global immune mechanism of HCV persistence and treatment response that warrants further investigation.

Acknowledgements

The authors thank Dr. Michael Houghton at Chiron Corporation for the generous provision of the recombinant HCV antigens; and Barbara Rensman, NP, and Ms. Rose Keevill, Pharm D, for patient recruitment as well as the study patients who participated in this study at the Philadelphia Veterans Affairs Medical Center clinics, the Hospital of University of Pennsylvania, and the NIH-funded Clinical Research Center within the University of Pennsylvania.

Ancillary