Plasma chemokine levels correlate with the outcome of antiviral therapy in patients with hepatitis C

Abstract

Chronic infection with the hepatitis C virus (HCV) is associated with failures of T-cell-mediated immune clearance and with abnormal B-cell growth and activation. We examined the levels of chemokines that bind to CXC chemokine receptor 3 (CXCR3) to determine whether such chemokines might play a role in the failure of the immune system to clear HCV infection. Elevations in CXC ligand 9 (CXCL9), CXCL10, and CXCL11 were observed in all patients with HCV. CXCR3 expression was increased significantly on peripheral blood B lymphocytes, but not T lymphocytes, from individuals with HCV infection. Chemokine levels were measured in samples collected before, during, and after antiviral therapy from a group of 29 patients infected with HCV genotypes 1a (24 patients) and 1b (5 patients). Levels of CXCL10 and CXCL9 decreased following successful antiviral therapy; CXCL11 did not decline significantly during or in the first 6 months after therapy. The baseline level of CXCL10 (measured before the start of antiviral treatment) was greatest in patients with HCV who subsequently became nonresponders to therapy. These results suggest that plasma concentrations of immunoreactive CXCL10 may be a predictor of responsiveness or nonresponsiveness to antiviral therapy with pegylated interferon (IFN) with or without ribavirin. This observation has implications for understanding the pathogenesis of HCV infection.

Figure 1.

Plasma CXCL9, CXCL10, and CXCL11 levels are elevated in patients with current HCV infection. (A) CXCL10 was measured by cytometric bead array, and levels were confirmed by ELISA in control subjects (n = 58), sustained responders (SVR; n = 51), and patients with HCV (n = 82). (B) CXCL9 was measured by cytometric bead array, and levels were confirmed by ELISA in control subjects (n = 44), sustained responders (n = 50), and patients with HCV (n = 82). (C) CXCL11 was measured by ELISA in control subjects (n = 44), sustained responders (n = 47), and patients with HCV (n = 81). Lines across each column represent the median value for each set of measurements. Values of P at the top of each graph were calculated by the Kruskal-Wallis test. Dunn multiple comparison test was used as a posttest to compare each group of samples to each other group; NS indicates not significant; **P < .01; ***P < .001.

Figure 2.

Increased expression of CXCR3 on B cells from patients with HCV infection. (A) PBMCs from a healthy donor were stained for CD3, CD20, and CXCR3. T and B cells identified by staining for CD3 or for CD20, respectively, were analyzed for expression of CXCR3. Numbers in dot plots represent the percentage of gated cells within the region indicated. (B) PBMCs from a patient with HCV were stained for CD3, CD20, and CXCR3. T and B cells identified by staining for CD3 or CD20, respectively, were analyzed for expression of CXCR3. (C) Frequency of T lymphocytes and B lymphocytes expressing CXCR3 among healthy donors and patients with HCV. Patient samples came from chronically infected patients who had not received antiviral therapy for at least 6 months. Lines across each column represent the median value for each group. Values of P were calculated by the Mann-Whitney test. (D) Frequency of CD20⁺ B lymphocytes expressing CXCR3 plotted against the frequency of PBMCs expressing CD20 in the same sample for the 29 patients listed in Table 2. Samples were obtained 1 week before the start of antiviral therapy. The symbols used indicate the outcome of therapy (▪, no response; ▵, SVR; ×, transient response followed by breakthrough or relapse). The values are correlated for the patients as a group (Spearman r = –0.6407, P = .001) and for nonresponders to antiviral therapy (Spearman r = –0.8788, P = .001).

Figure 3.

Plasma CXCL10 levels in sustained responders versus nonresponders to antiviral therapy. Serial samples were obtained from patients described in Table 2. (A) CXCL10 levels measured 7 days before the start of therapy are lower in those who subsequently achieve an SVR (n = 11) than in those who subsequently have no reduction in HCV RNA during or after antiviral therapy (NR; n = 10). Lines across each column represent the median for each set of measurements. Value of P was calculated by the Mann-Whitney test. (B-C) CXCL10 levels were measured in plasma obtained 7 days before initiation of antiviral therapy, 29 days after initiation of therapy, and 24 weeks after completion of therapy. Results are plotted separately for patients who subsequently achieved an SVR (B) and for patients who did not respond to antiviral therapy (C). Value of P (comparing the levels of CXCL10 at different points during treatment) was calculated by the Friedman test, a nonparametric repeated measures analysis of variance (ANOVA).

Figure 4.

Plasma CXCL9 levels in sustained responders versus nonresponders to antiviral therapy. Serial samples were obtained from patients described in Table 2. (A) CXCL9 levels measured 7 days before the start of therapy are lower in those who subsequently achieve an SVR (n = 11) than in those who subsequently have no reduction in HCV RNA during or after antiviral therapy (NR; n = 10). Lines across each column represent the median for each set of measurements. Value of P was calculated by the Mann-Whitney test. (B-C) CXCL9 levels were measured in plasma obtained 7 days before initiation of antiviral therapy, 29 days after initiation of therapy, and 24 weeks after completion of therapy. Results are plotted separately for patients who subsequently achieved an SVR (B) and for patients who did not respond to antiviral therapy (C). Value of P (comparing the levels of CXCL9 at different points during treatment) was calculated by the Friedman test.

Figure 5.

Plasma CXCL11 levels in sustained responders versus nonresponders to antiviral therapy. Serial samples were obtained from patients described in Table 2. (A) CXCL11 levels measured 7 days before the start of therapy are lower in those who subsequently achieve an SVR (n = 11) than in those who subsequently have no reduction in HCV RNA during or after antiviral therapy (NR; n = 10). Lines across each column represent the median for each set of measurements. Value of P was calculated by the Mann-Whitney test. (B-C) CXCL11 levels were measured in plasma obtained 7 days before initiation of antiviral therapy, 29 days after initiation of therapy, and 24 weeks after completion of therapy. Results are plotted separately for patients who subsequently achieved an SVR (B) and for patients who did not respond to antiviral therapy (C). Value of P (comparing the levels of CXCL11 at different points during treatment) was calculated by the Friedman test.

Figure 6.

Plasma CXCL10 levels measured before start of therapy correlate with early responses to therapy. (A) CXCL10 levels measured 7 days before the start of therapy were lower in those whose serum viral load declined by > 2 log₁₀ during the first 29 days of therapy than in those whose serum viral load declined by < 1 log₁₀ during the first 29 days of therapy. (B) CXCL10 levels measured 7 days before the start of therapy were lower in patients in whom serum HCV RNA had declined by > 2 log₁₀ by 12 weeks of therapy than in those whose serum viral load declined by < 1 log₁₀ after 12 weeks of antiviral therapy. Measurements from all 29 patients are shown. Lines across each column represent the median for each set of measurements. Value of P was calculated by the Mann-Whitney test.

Figure 7.

Plasma CXCL10 levels measured before start of therapy correlate with long-term responses to therapy. (A) CXCL10 levels measured 7 days before the start of therapy were lower in patients in whom serum HCV RNA was undetectable at the end of antiviral therapy than in those who remained viremic at the end of antiviral therapy. (B) CXCL10 levels measured 7 days before the start of therapy were lower in patients in whom serum HCV RNA remained undetectable 24 weeks after completion of therapy than in those who were viremic 24 weeks after completion of antiviral therapy. Measurements from all 29 patients are shown. Lines across each column represent the median for each set of measurements. Value of P was calculated by the Mann-Whitney test.