Introduction
At present, about 3% of the human population is infected with hepatitis C virus (HCV). The first, acute stage of HCV infection is usually asymptomatic [
11]. However, only 15–30% of the infected eliminate the virus. The remaining patients develop chronic hepatitis C (CHC). Despite two decades of intensive studies, no obvious correlations have been found between the parameters describing CHC patients and the course of infection or the outcome of therapy [
1]. Instead, HCV genotype and virus load were confirmed to be important factors affecting the final result of antiviral treatment. Individuals infected with genotype 2 or 3 and those with a low level of viremia respond better to the therapy [
14,
18,
24]. This indicates that certain features of the virus, rather than those of the host, may actually be critical for CHC treatment.
Recently, it has become increasingly clear that one of the key factors making CHC therapy so difficult is the unusual genetic polymorphism of the virus [
8,
27]. The major cause of this polymorphism is the effective, but error-prone replication of the RNA genome [
8]. Consequently, HCV does not form a homogenous population but exists as a quasispecies [
8,
27]. The latter is defined as a pool of phylogenetically related but genetically slightly distinct variants present in a single infected organism [
3,
4]. All variants are subjected to continuous selection; therefore, only the most fit are able to survive and spread [
8]. Several regions of the HCV genome demonstrate an especially high level of variability. Among them, hypervariable region 1 (HVR1), encoding a 27-aa N-terminal fragment of the E2 glycoprotein, is thought to be the most heterogeneous one.
A number of attempts have been made in recent years to find a correlation between the complexity of HCV quasispecies and the outcome of CHC therapy in adults. The diversity of viral populations was usually evaluated based on two regions of the HCV genome: E1/E2 (including HVR1), and NS5A. As a result, it has been postulated that HCV quasispecies display decreased heterogeneity in patients who develop a sustained virological response [
5,
20,
25,
30]. In contrast, little is currently known about the structure of the HCV population in children. This issue raises many important questions, especially since the natural history and clinical features of HCV infection in children and adults are different. A milder, often asymptomatic liver disease with a normal or slightly elevated level of alanine aminotransferase activity impedes the diagnosis of HCV infection in childhood [
15]. Consequently, the percentage of CHC in pediatric patients is probably underestimated. In addition, available reports concerning HCV quasispecies in children are usually based on only a few cases. Some of these reports refer to viral populations in untreated children [
6,
12], the others to HCV quasispecies in patients coinfected with human immunodeficiency virus (HIV) [
2,
26].
In this paper, we present an analysis of HCV quasispecies isolated from 23 children with CHC, all infected with the same HCV subtype (1a). For each of them, the structure of the virus population was determined just before anti-HCV treatment and 2 weeks later.
Discussion
In spite of ample evidence that high genetic variability of RNA viruses affects host-virus interactions, the question concerning the correlation between particular parameters characterizing HCV quasispecies and the course of CHC or results of its treatment remains open. Considering earlier studies, we decided to focus our investigations on the 462-nt fragment of the E1/E2 region of the HCV genome, comprising HVR1. Previous analyses of the HCV quasispecies have usually involved several genotypes (often all six), although it was known that HCV-2 and -3 are eliminated more frequently than HCV-1 [
7]. The latter observation suggest that the results obtained for one genotype cannot be extrapolated to the others, and that is why our analysis involved a group of 23 children with CHC, all of whom were infected with HCV-1a and qualified for combined interferon–ribavirin therapy.
The collected data revealed an apparent correlation between the final effect of CHC therapy and the HVR1 polymorphism observed just before treatment (at T0). Accordingly, we distinguished two types of HCV quasispecies. The first type, formed by one dominating variant accompanied by a small number of closely related ones, was characteristic of the non-responding patients. The second type, composed of a large number of distantly associated equally ranked variants was typical for sustained responders. In both groups of patients, the level of HCV RNA varied over a wide range (Table
1), and we did not find a statistically significant correlation between the accumulation of viral RNA and HCV quasispecies variability or therapy outcome. Unfortunately, the amount of information concerning the correlation between the complexity of HCV quasispecies and the efficacy of anti-HCV treatment in children is still very limited. The existence of two patterns of HCV evolution in a group of 12 untreated neonates with HCV infection was reported previously by Farci et al
. [
6]. They observed that mono- or oligoclonal HCV populations were associated with a high level of ALT activity and hepatic injury, while heterogenous ones were associated with a lower ALT activity and mild or no liver damage [
6]. In the case of children with CHC, we did not observe such a correlation between ALT activity and HCV population structure. Interestingly, our findings are also inconsistent with previous observations made for adults with CHC undergoing anti-HCV therapy. In that study, it was suggested that homogeneity of HCV quasispecies at the baseline [
21,
29,
30] or dramatic reduction of HCV genetic diversity during the first weeks of the therapy [
5] is associated with viral clearance. In contrast, we found that patients with an almost homogenous HCV quasispecies at T0 did not eliminate the virus, while the ones with heterogeneous populations developed a sustained response. Similarly to Farci et al. [
5], we observed a reduction of HCV quasispecies diversity in some sustained responders. Taking into account differences in the course and manifestation of chronic hepatitis C in adults and children, one can speculate that the mechanism underlying resistance to anti-HCV treatment may also vary in these two groups of patients.
HVR1 is the main target of host antibodies and is the most variable region of the HCV polyprotein. Hence, the stability of this region reported here seems to be a surprising discovery. HVR1 homogeneity was reported earlier in immunocompromised children and adults [
10]. In the group we classified as non-responders, there were three patients who had previously undergone immunosuppressive treatment due to an oncological disease who were anti-HCV negative by fourth-generation ELISA (P2-05, P2-10, P2-28; Table
1). However, at T0, the level of HVR1 diversity in these children was similar to that in anti-HCV-positive non-responders. These observations suggest that the conservation of HVR1 in non-responders is not caused by defects in their humoral immune response.
An analysis undertaken by Gerotto and coworkers showed that, in children with CHC, the degree of HVR1 variability increases continuously in parallel with the natural development of the host immune system [
12]. As a consequence, a more constant HVR1 is expected in younger children. Since the mean age of non-responders was almost the same as that of sustained responders (11.1 years, range 8–14, vs. 11.8 years, range 8–16, respectively), the observed changes in HVR1 variability could not be explained by differences in immune system development. Statistical analysis also did not reveal any correlation between the patients’ age and the level of HVR1 variability within the examined group of patients or within two subgroups: NR and SR patients.
Only recently, a lack of HVR1 variability over a long period of time has been reported in two immunocompetent perinatally HCV-1a/c-infected children [
13]. Both of them were subjected to interferon alpha monotherapy, and both failed to eliminate the virus. In addition, Gismondi and coworkers detected almost complete conservation of the HVR1 amino acid sequence, accompanied by a strong negative selection operating at certain positions. A similar observation was made with our non-responders. Gismondi’s group proposed two possible explanations for this HVR1 homogeneity: immunotolerance and adaptation of the virus to the specific conditions within each host. The results presented here favor latter hypothesis. In addition to intra-quasispecies homogeneity, we identified the surprising phenomenon of inter-quasispecies conservation of HVR1 variants in children who failed to clear the virus. It is plausible that these variants demonstrate the highest fitness, resulting from optimal structural and/or functional properties. Accordingly, selection would favor their conservation. However, a hypothetical mechanism of immunotolerance in NR patients cannot be ruled out. It is probably a combination of the viral factors and the immunological aspects of a given patient that determines the HCV persistence. Our findings support earlier observations that variants that are inherently resistant to therapy may be present before its onset. It is also worth mentioning that non-responder-specific conservative HVR1 variants were never found in populations isolated from sustained responders. Based on the data provided by Gismondi et al. [
13], one can hypothesize that these variants remain unchanged throughout treatment. In contrast, in children demonstrating a sustained response, positive selection seems to accelerate viral evolution to generate a more fit population. The high variability in these quasispecies probably reflects an ongoing process of adaptation of suboptimal variants—apparently an unsuccessful one, at least in terms of viral survival, since these populations were eliminated during treatment.
At present, it would be highly tentative to speculate what specific features of the conserved HVR1 variants underlie their superiority. It is certainly beyond the scope of this study. Gismondi et al. predicted
in silico that the identified variants are antigenic. This suggests that the invariability of HVR1 is not due to the lack of an adequate humoral immune response [
13]. Recently, we have expressed the E2 gene in
E. coli. HVR1 of the resulting protein was identical to the variant most prevalent in cluster 1. During the preliminary experiments, we were able to detect the partially purified recombinant E2 protein by ELISA with homologous and heterologous patient sera [P.J., M.F., M.F., unpublished results]. Even if HVR1 alone was not targeted by the antibodies (obviously, this could not be concluded from the initial studies with patient sera), other regions of E2 were, calling into question the need to keep HVR1 unchanged. Consequently, it is presumably not immune evasion that causes HVR1 homogeneity. For this reason, to gain insight into the mechanisms driving HVR1 variability, attention should be focused on other processes involving the E2 protein during the host-virus interaction.
The same conclusion can be drawn from the distribution of codons subjected to positive selection in our sustained responders. Previously, two antigenic stretches (spanning residues 3-13 and 19-25) were identified within HVR1 [
22]. Together, they constitute the majority of HVR1. Surprisingly, our results demonstrate that positively selected codons prevail in the central, non-antigenic region (10–18), and these are probably involved in maintaining HVR1 conformation and interactions with other molecules [
22]. This observation indicates that adaptation to host proteins is at least as an important factor driving HVR1 evolution in children as diversification caused by host antibodies.
The results presented in this paper suggest that the degree of baseline HCV quasispecies diversity correlates with the outcome of chronic hepatitis C treatment in children. If confirmed in a larger group of patients in the future, this observation would be of great importance for further optimization of the anti-HCV therapy. We also provide another line of evidence that HVR1 diversification and evolution may be driven differently in pediatric patients than in adults. This matter should be considered during prospective clinical and virological studies.