SEARCH

SEARCH BY CITATION

Keywords:

  • Interferon-α;
  • MCP-1;
  • necroinflammation;
  • ribavirin;
  • β-chemokines

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Clin Microbiol Infect 2011; 17: 204–209

Abstract

The role of chemokines in chronic hepatitis C virus (HCV) infection is not fully understood. The present study aimed to characterize the baseline serum concentrations and the initial β-chemokine response to treatment with interferon-α and ribavirin with respect to the final clinical outcome of virological response to treatment. Serum concentrations of alanine aminotransferase (ALT) and of the CC subfamily chemokines [macrophage inflammatory protein (MIP)-1α, MIP-1β, monocyte chemoattractant protein (MCP)-1 and the regulated on activation, normal T expressed and secreted (RANTES) chemokine] were measured in patients with chronic HCV infection and in healthy individuals. Necroinflammation and fibrosis were scored in liver biopsies. Treatment outcomes were classified as with or without a sustained virological response after a full-course treatment according to the genotypes. The main treatment group consisted of 72 patients with chronic hepatitis C, whereas 24-h blood samples were available for 42 patients. Increased baseline levels of all CC chemokines were found in the two responder groups compared to the healthy controls, although significant levels were reached only for MIP-1α and MCP-1. No correlation was observed between chemokine levels and serum ALT levels, any histological necroinflammatory parameters, or the fibrosis grade. After 24 h of treatment, increases in MIP-1α, MIP-1β and RANTES levels were exclusively observed in the group with sustained virological response. MCP-1 was also significantly increased after 24 h in both responder groups, although no differences were observed between the two responder groups. In conclusion, an early MIP-1α, MIP-1β, and RANTES response may predict a sustained response to virological treatment.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

The understanding of successful virus elimination mediated by the immune response to hepatitis C virus (HCV) infection is far from settled. Both viral and host factors are determinants for the spontaneous elimination or persistence of the virus. In HCV infection, some 50–90% of the patients develop chronic disease [1]. There is evidence that unspecific CD8 cells and a subset of antigen-specific CD4 cells deviating T-helper cells in a TH1 profile (interferon (IFN) and tumour necrosis factor-α) play a pivotal role in the immune response to the HCV [2].

A rapid and strong host T cell response plays a major role in effective clearance of the HCV virus. Most likely, the tendency towards chronic infection represents a lack of a strong and specific immune response to viral antigens as well as an ineffective rapid clearance of the HCV [3,4].

The recruitment of effector cells from the blood stream to the liver is dependent upon local chemokines and their receptors. Chemokines constitute a large family of small proteins consisting of four main subfamilies (CXC, CC, C, CX3C) mediating their effects on a family of seven-transmembrane domain G-protein coupled receptors [5]. The chemokines are responsible for the leukocyte migration by creating a chemical gradient from the vascular endothelium to the infected cells [6] and by selective expressions of different chemokine receptors on the leukocytes [7].

There is increasing evidence for a role of chemokines as inflammatory mediators in HCV infection [8].

The CC chemokines macrophage inflammatory protein (MIP)-1α (CCL3), MIP-1β (CCL4) and the regulated on activation, normal T expressed and secreted (RANTES) chemokine (CCL5) are expressed by the portal vessel endothelium and recruit macrophages and lymphocytes into the liver [6,9]. These chemokines bind to their corresponding receptors, CCR5, on lymphocytes with type 1 cytokine secretion [9,10]. Theoretically, both the chemokine response to HCV and the availability of the corresponding receptors on the lymphocytes are obligatory for a strong immunological response to the virus. Genetically determined loss of CCR5 expression has been linked to chronic hepatitis [11]. Most studies, although not all, have shown increased intrahepatic and/or blood levels, as well as increased expression, of the CXC and CC families of chemokines in chronic HCV infection [8]. However, the exact role of chemokines both in spontaneous remission and in response to immunomodulatory agents such as IFN-α is so far poorly understood. According to the few studies available, high baseline serum levels of MIP-3α [12] are associated with a positive prognostic response, whereas high levels of IFN-γ-inducible protein 10 are associated with a negative prognostic response to treatment [13,14]. Moreover, increase in CXCR3-expressing CD8+ cells during treatment has been associated with achievement of virus control [15]. However, the dynamics of the apparent critical initial rapid CC chemokine response with emphasis on clearance of virus is so far not well investigated.

Therefore, in the present study, we describe the blood concentrations of MIP-α, MIP-β, monocyte chemoattractant protein (MCP)-1 and RANTES and their relationship to liver enzyme levels, microbiological and histological parameters at baseline, and then describe the association between the initial (24 h) chemokine response and the final clinical outcome of the treatment of chronic HCV infection.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Patients

Seventy-two patients with chronic HCV infection were included. The patients were admitted to the University Hospital of North Norway, Tromsø, and Nordland Hospital, Bodø, Norway, for diagnosis and treatment between 1998 and 2005. Blood tests were made for alanine aminotransferase (ALT), chemokines and HCV at admission (all 72 patients) and after 24 h (42 patients).

The inclusion criteria for treatment were elevated ALT, positive HCV serology and viraemia as detected by PCR. Patients with concomitant active hepatitis B virus infection and patients with severe diseases were excluded.

Patients included before 2001 received IFN-α-2b (Introna; Schering Plough AS, Oslo, Norway), 3 million international units thrice weekly (monotherapy). After 2001, treatment was a combination of pegylated IFN-α-2b (PegIntron, Schering Plough), 1.5 μg/kg once weekly, and ribavirin (Rebetol, Schering Plough), 600–1200 mg/day (dual therapy), for either 24 weeks (genotype 2 or 3) or 48 weeks (genotype 1 or 4). The treatment effect was classified as sustained virological response (SVR; negative HCV PCR 24 weeks after cessation of treatment) and non-SVR [comprising two subgroups: relapse (negative/positive HCV PCR at the end of treatment and positive test after an additional 24 weeks, respectively) and nonresponse (positive HCV PCR after 12 weeks of treatment)]. The treatment responses among the 72 patients were SVR in 39 and non-SVR in 33 patients (with 17 cases of relapse and 16 of nonresponse). Liver biopsies were performed on a regular basis before the start of treatment. The modified hepatic activity index (HAI) grading necroinflammatory scores (0–18) was determined and fibrosis grading (0–6) was performed according to Ishak et al. [16].

In Table 1, the baseline characteristics and clinical profiles are presented for the 72 patients (monotherapy, n = 40; dual therapy, n = 32) included in the study. In 12 patients (all in the monotherapy group), genotype determination was not performed as a result of the combination of the unavailability of the method at the time of inclusion and the lack of blood for testing at a later time point. Histological examinations of liver biopsies were performed in 65 patients. In seven patients, biopsy specimens were not available, either because of patient refusal to participate or because of a regular, missed biopsy. Regarding the 24-h analyses, only blood samples from 42 patients were available [SVR, n = 23; non-SVR, n = 19 (relapse, n = 12; nonresponse, n = 7)].

Table 1.   Clinical characteristics at baseline of patients with chronic hepatitis C virus infection grouped according to response to treatment
 AllSustained responseNonsustained response
  1. Data are presented as the mean (95% CI) (for ALT geometric mean) or n (%).

  2. ALT, aminotransferase; HAI, hepatic activity index.

n7239 (54%)33 (46%)
Male/female42/3022/1715/18
Age (years)38 (36–40)36 (34–38)41 (36–45)
ALT (U/L) (n = 72)99 (86–114)111 (91–135)87 (72–106)
HAI index (n = 65)9.0 (8.3–9.8)8.4 (7.4–9.3)10.0 (8.9–11.0)
Fibrosis score (n = 65)1.5 (1.1–1.8)1.4 (1.0–1.8)1.6 (1.0–2.2)
Genotype 1:2:3:4 (n = 60)26:4:27:36:4:16:120:0:11:2

Analysis of chemokines

Serum from blood samples was obtained after centrifugation at 4°C and subsequently kept at −27°C. Measurement of the CC chemokines MIP-1α, MIP-1β, MCP-1 and RANTES was performed with commercial ELISA kits (Quantikine, human chemokines; R&D Systems Europe Ltd, Oxon, UK). Blood samples from five healthy volunteers were used as controls.

Approvals

All participants provided their written informed consent. The Regional Committee for Medical and Health Research Ethics approved of the study, and permissions for storage of biological material and the research database were granted from the Norwegian Ministry of Health and the Norwegian Social Science Data Services (NSD), respectively.

Statistical analysis

Unless otherwise stated, data are presented as the mean (95% CI). For log-linear variables (chemokine and ALT measurements), geometric means were used. For further analyses, logarithmic transformation was used when appropriate. Ordinal variables (HAI, fibrosis score) were analyzed using nonparametric statistics (Wilcoxon test). Correlation statistics were performed by Pearson correlation. Categorical data were analyzed using a chi-square test. Analysis of basal chemokine levels was performed by one-way analysis of variance followed by a simple contrast with healthy controls as reference. The 0–24-h differences showed normal distribution, and nontransformed data could be entered into t-test analysis. Chemokine levels were analyzed comparatively between groups using Student’s t-test for independent observations and, within groups, one-sample t-tests were used. None of the analyses violated the assumption of normal distribution or the assumption of equality of variances.

Linear and logistic regression analyses were performed to test for confounding effects of sex, age, type of treatment, ALT levels, genotype and histological data.

Receiver-operating characteristics (ROC) analysis was performed with 0–24-h differences in chemokine levels as continuous variable and SVR/non-SVR as the state variable.

All analyses were performed using spss software, version 16.0 (SPSS Inc., Chicago, IL, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

Clinical characteristics at baseline

In the entire group of 72 patients, no significant differences were detected in ALT levels, HAI scores or fibrosis scores across the treatment groups with and without SVR (Table 1) or when further subgrouping according to SVR and the two non-SVR subgroups (relapsers and nonresponders) (data not shown). As expected, there was more SVR in the genotype 2 + 3 group than in the genotype 1 + 4 group (p 0.008, chi-square). Moreover, logistic regression analysis showed that genotype 1 + 4 (p 0.006), but not age, sex or ALT levels, were independent predictors of treatment response (SVR vs. non-SVR).

Baseline chemokines in HCV patients and in healthy controls

There were increases in all chemokines in the SVR group and the non-SVR group (relapsers and nonresponders) compared to healthy controls, but only MIP-1α and MCP-1 levels were at significant levels (Table 2). No significant differences were observed between the various chemokine levels in the various responder groups (data not shown). When comparing the levels of each of the four chemokines, linear regression analyses showed no significant effects of factors such as ALTlevels, sex, age, type of treatment, and treatment response on the chemokine levels. Moreover, no significant correlations could be detected between histological and microbiological parameters vs. chemokine levels (data not shown).

Table 2.   Serum chemokine concentrations at baseline in patients with chronic hepatitis C virus infection grouped according clinical outcome
 nMIP-1α (pg/mL)MIP-1β (pg/mL)MCP-1 (pg/mL)RANTES (ng/mL)
  1. Data are presented as geometric mean (95% CI).

  2. *Significant difference from healthy controls by simple contrast.

  3. MCP, monocyte chemoattractant protein; MIP, macrophage inflammatory protein; RANTES, regulated on activation, normal T expressed and secreted.

  4. NS, not significant.

Sustained response397 (3–14)*77 (56–107)239 (200–286)*33 (27–41)
Relapse1711 (6–22)*79 (54–115)225 (179–284)*31 (22–43)
Nonresponse165 (1–20)87 (63–120)181 (83–398)*31 (21–47)
Healthy controls51 (0.8–2.0)64 (33–126)61 (51–73)21 (17–26)
p (analysis of variance) 0.003NS<0.0005NS

Early chemokine changes during initiation of hepatitis C treatment

In the group of 42 patients (16 with monotherapy and 26 with dual therapy) with 24-h blood samples available, significant increases from baseline (within groups) were exclusively observed in the SVR group for MIP-1α and MIP-1β, and at a nonsignificant level for RANTES (Fig. 1). MCP-1 levels were significantly increased in both responder groups (Fig. 1), whereas, in the non-SVR group, the significant increase was only observed in the subgroup of relapsers (data not shown). Moreover, the mean differences between the SVR and non-SVR groups at 24 h were significant (between groups) with respect to MIP-1α, MIP-1β and RANTES, but not MCP-1 (Fig. 1).

image

Figure 1.  Early serum chemokine responses to the treatment of chronic hepatitis C virus infection. Values are geometric means (95% CI for geometric mean); p values are generated by a t-test of 0–24 h within group (long underline) and between group (short underline) differences.

Download figure to PowerPoint

These observed differences in response were also observed after adjustment for sex, age, genotype and treatment protocol in a linear regression analysis.

ROC analysis

To determine whether the early chemokine response could predict SVR, an ROC analysis was performed. Only MIP-1α and MIP-1β had significant areas (Fig. 2). A cut-off of 0 pg/mL 0–24 h increase in serum MIP-1α would yield a sensitivity of approximately 0.8 and a specificity of 0.5 for virological response. A cut-off of 0 pg/mL 0–24 h increase in serum MIP-1β would yield a sensitivity of approximately 0.8 and a specificity of 0.4 for SVR.

image

Figure 2.  Receiver-operator characteristics for the macrophage inflammatory protein (MIP)-1α and MIP-1β increase at 24 h after the initiation of interferon-α therapy. The basis of this analysis is a continuous variable (chemokine value) and an actual state variable [sustained virological response (SVR )/no SVR]. A choice of cut-off will define values as being positive or negative, and a comparison of this value with the actual state gives the estimate of sensitivity and specificity. Different cut-off values then define the blue curve in the plot. Both observed areas are significant and cut-off values can be deduced by inspecting the plot coordinates and their corresponding chemokine values (not shown).

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

In the present study, we have shown that baseline serum levels of the CC chemokines MIP-1α, MIP-1β, MCP-1 and RANTES were increased compared to healthy controls. There were no significant relationships between the serum levels of the chemokines and liver function tests, microbiological and histological parameters. Of the various chemokines studied, a rapid (24 h) response exclusively associated with a sustained virological response (SVR) was observed only for MIP-1α, MIP-1β and RANTES. MCP-1 was increased in both the SVR and non-SVR group, although no differences between the two responder groups were observed.

In agreement with several previous studies, the serum levels of CC chemokines were increased in patients with chronic HCV infection compared to healthy controls [8]. Increased intrahepatic expression has also been reported for CC chemokines [17,18]. Furthermore, in the present study, serum levels of the CC chemokines were not associated with the grade of necroinflammatory activity or the fibrosis grade. In the literature, there are conflicting results concerning this topic; in one study, serum MIP-3α and ALT correlated significantly [12]; in another study, intrahepatic RANTES expression correlated well with the histological activity [17], whereas, in yet another study, the expression of RANTES or MIP-1β in the liver did not correlate with necroinflammation activity scores [18]. Most likely, during a chronic HCV infection, there is an inflammatory circle: an activated chemokine response is responsible for recruitment of inflammatory cells, which in turn activate further production of cytokines and chemokines. This inflammatory circle is in turn responsible for the chronic necroinflammation and the development of cirrhosis [13]. However, the exact contribution of chemokines and especially the CC chemokines to this inflammatory circle is unclear and needs further studies.

The conventional treatment of chronic HCV infection consists of a combination of the immunomodulatory and antiviral agents IFN-α and ribavirin. Several studies have shown that a rapid viral clearance from the general circulation is associated with a high rate of sustained virological response [19]. It has been proposed that the natural immune response to the HCV has to be rapid, strong and complete with the various epitopes to prevent virus escape and subsequent chronic inflammation [20]. Therefore, the present study was designed to study early (24 h) responses of the CC chemokines during the initiation of antiviral treatment. Significant increases exclusively associated with a sustained virological response were observed for MIP-1α and MIP-1β. For RANTES, the responses at 24 h were less pronounced but were at significant levels between the two groups.

It is of interest to note that MCP-1 levels were also increased at 24 h of treatment in the sustained response group and in the relapser group, but not in the nonresponder group. Our data indicate that, of the various chemokines mediating the immune response during the treatment, MIP-1α, MIP-1β and RANTES, in contrast to MCP-1, apparently play an important role in inducing an effective clearance of the HCV.

Interestingly, the present data indicate a possible role of early chemokine response as a predictor of an SVR response. As can be seen form the ROC curves in Fig. 2, the assays cannot yield both high sensitivity and specificity at the same time because this would require a steep sigmoid curve shape with an area approaching 1. The clinical use of such a test would be to decide whether a patient should discontinue treatment, assuming the patient will not respond to treatment. Consequently, the cut-off values were set for high sensitivity of response to give the patient the benefit of doubt.

The present study indicates that various chemokines play a role in mediating an effective immunomodulatory role during treatment of HCV infection. However, the results obtained must be interpretated with great caution. First, the number of patients studied was low and two types of treatment with IFN-α were used in the patients studied. Second, the study was not primarily designed to test independent 24-h chemokine responses to the final clinical outcome. A future study should be prospective and properly stratified for potential confounding factors such as sex, genotypes and grade of liver cirrhosis. In this setting, a possible role of early chemokine response as a clinical predictor of sustained virological response could be better evaluated.

In conclusion, serum levels of CC chemokines are elevated in chronic HCV infection but not associated with the grade of necroinflammation. An early response of MIP-1α, MIP-1β and RANTES to treatment may predict an effective HCV clearance.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

We are grateful for technical support from I. Christiansen and O. S. Moen at the Laboratory of Gastroenterology, Institute of Clinical Medicine, University of Tromsø, Tromsø, Norway.

Transparency Declaration

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References

This study was supported by grants from the Foundation of Gastroenterolgy, University Hospital of North Norway, Tromsø, from the University of Tromsø and from Schering Plough, Oslo, Norway. The authors declare no conflict of interest in relation to this study.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Transparency Declaration
  9. References