September 22, 2010

Concerns About HCV Genotyping: Hepatitis C Virus (HCV) Genotype 1 Subtype Identification in New HCV Drug Development and Future Clinical Practice

Download the PDF here

PLoS ONE 4(12): e8209. doi:10.1371/journal.pone.0008209

Published December 8, 2009

"INNO-LiPA HCV 2.0 currently is the best available commercial assay for HCV genotype 1 subtype identification and should be used in clinical trials and practice.....

Trugene HCV Genotyping Kit and INNO-LiPA HCV 1.0, failed to correctly identify HCV subtype 1a in 22.8% and 29.5% of cases, and HCV subtype 1b in 9.5% and 8.7% of cases, respectively....The second-generation line probe assay is currently the best commercial assay for determination of HCV genotype 1 subtypes 1a and 1b. It can therefore be used locally in clinical trials to identify the HCV subtype and stratify the patients at inclusion, as well as to interpret efficacy and resistance data. When reporting final data, direct sequence analysis of the NS5B region and/or another coding region (for instance the region encoding the antiviral drug target HCV protein) should always be performed as it may identify mistyping or mis-subtyping with commercial assays, especially in the case of rare subtypes."

Stephane Chevaliez1,2, Magali Bouvier-Alias1,2, Rozenn Brillet2, Jean-Michel Pawlotsky1,2*

1 French National Reference Center for Viral Hepatitis B, C and delta, Department of Virology, H™pital Henri Mondor, Universite Paris 12, Creteil, France, 2 INSERM U955, Creteil, France Methods based on the sole analysis of the 5'NCR, namely Trugene HCV Genotyping Kit and INNO-LiPA HCV 1.0, failed to correctly identify HCV subtype 1a in 22.8% and 29.5% of cases, and HCV subtype 1b in 9.5% and 8.7% of cases, respectively (Table 1)...... The results clearly show that, although they are by far the most widely used techniques in new HCV drug development trials, genotyping techniques based on the sole analysis of the 5'NCR should be avoided, as they mistype approximately 25% and 10% of HCV subtype 1a and 1b strains, respectively......INNO-LiPA HCV 2.0 displays the same 5'NCR oligonucleotide probes as INNO-LiPA HCV 1.0, plus core-encoded oligonucleotide probes aimed at better discriminating between HCV subtypes 1a and 1b. With INNO-LiPA HCV 2.0, subtype identification was corrected in 64 of the 70 subtypes 1a that were incorrectly typed with INNO-LiPA HCV 1.0. Five samples could not be PCR-amplified in the core-coding region and the result was not interpretable with INNO-LiPA HCV 2.0 in the remaining case (Table 1). INNO-LiPA HCV 2.0 also corrected subtype identification in 13 of 23 subtypes 1b that were incorrectly typed with INNO-LiPA HCV 1.0. Eight samples could not be PCR-amplified in the core-coding region and the result was not interpretable with INNO-LiPA HCV 2.0 in the remaining two cases (Table 1). Overall, the second-generation line probe assay correctly classified 97.5% of subtype 1a and 96.2% of subtype 1b strains. When only samples that could be PCR-amplified with the assay procedure were taken into account, correct subtype determination was achieved in 99.6% and 99.2% of cases, respectively (Table 1)The real-time PCR-based assay targeting both the 5'NCR and the NS5B region, Abbott RealTime HCV Genotype II assay, correctly identified 93.2% of subtype 1a and 88.9% of subtype 1b strains. Only 2 HCV subtype 1b samples could not be PCR-amplified with this method (Table 1)......Novel assays have been recently developed that aim at better discriminating among the different HCV genotype 1 subtypes and between genotypes 1 and 6. Abbott RealTime HCV Genotype II assay is a real-time PCR method using several sets of genotype- and subtype-specific primers and probes located in both the 5'NCR and the NS5B-coding region. As shown in Table 1, adding a second target region for analysis led to substantially improving HCV genotype 1 subtype identification compared to methods targeting the sole 5'NCR. However, in contrast with a previous report [33], we found that this assay failed to correctly identify HCV genotype 1 subtype in approximately 10% of cases"

Table 1. Ability of the different molecular methods tested in this study to correctly identify HCV subtypes 1a and 1b in a series of 500 patients infected by one or the other of these subtypes.

Abstract

Background

With the development of new specific inhibitors of hepatitis C virus (HCV) enzymes and functions that may yield different antiviral responses and resistance profiles according to the HCV subtype, correct HCV genotype 1 subtype identification is mandatory in clinical trials for stratification and interpretation purposes and will likely become necessary in future clinical practice. The goal of this study was to identify the appropriate molecular tool(s) for accurate HCV genotype 1 subtype determination.

Methodology/Principal Findings

A large cohort of 500 treatment-naïve patients eligible for HCV drug trials and infected with either subtype 1a or 1b was studied. Methods based on the sole analysis of the 5' non-coding region (5'NCR) by sequence analysis or reverse hybridization failed to correctly identify HCV subtype 1a in 22.8%-29.5% of cases, and HCV subtype 1b in 9.5%-8.7% of cases. Natural polymorphisms at positions 107, 204 and/or 243 were responsible for mis-subtyping with these methods. A real-time PCR method using genotype- and subtype-specific primers and probes located in both the 5'NCR and the NS5B-coding region failed to correctly identify HCV genotype 1 subtype in approximately 10% of cases. The second-generation line probe assay, a reverse hybridization assay that uses probes targeting both the 5'NCR and core-coding region, correctly identified HCV subtypes 1a and 1b in more than 99% of cases.

Conclusions/Significance

In the context of new HCV drug development, HCV genotyping methods based on the exclusive analysis of the 5'NCR should be avoided. The second-generation line probe assay is currently the best commercial assay for determination of HCV genotype 1 subtypes 1a and 1b in clinical trials and practice.

Funding: The Trugene HCV 5'NC Genotyping kits and the INNO-LiPA HCV kits were kindly provided by Siemens Medical Solutions Diagnostics. The Abbott RealTime HCV Genotype II kits were kindly provided by Abbott Molecular. This work is part of the activity of the VIRGIL European Network of Excellence on Antiviral Drug Resistance supported by a grant (LSHM-CT-2004-503359) from the Priority 1 "Life Sciences, Genomics and Biotechnology for Health" program in the 6th Framework Program of the European Union. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Introduction

Over 170 million individuals are infected with hepatitis C virus (HCV) worldwide. Phylogenetic analyses have shown that HCV strains can be classified into at least 6 major genotypes (numbered 1 to 6), and a large number of subtypes within each genotype [1]. Genotype 1 is by far the most frequent genotype in chronically infected patients worldwide, with subtypes 1a and 1b representing the vast majority of circulating strains [2], [3], [4].

Current treatment of chronic hepatitis C is based on the combination of pegylated interferon (IFN)-α and ribavirin [5]. This treatment fails to eradicate infection in 50%-60% of patients infected with HCV genotype 1 and approximately 20% of those infected with HCV genotypes 2 and 3 [6], [7], [8]. Thus the need for more efficacious therapies is urgent, especially for patients infected with HCV genotype 1. A number of novel antiviral molecules currently are in preclinical or clinical development [9]. The most advanced ones are specific inhibitors of viral enzymes and functions involved in the HCV life cycle. Molecules that have reached clinical development include inhibitors of the nonstructural (NS) 3/4A serine protease and inhibitors of HCV replication that belong to different categories: nucleoside/nucleotide analogue and non-nucleoside inhibitors of the HCV RNA-dependent RNA polymerase (RdRp), NS5A inhibitors and cyclophilin inhibitors [9]. These agents have shown potent antiviral efficacy when used alone, and encouraging results have been recently published showing that HCV clearance can be achieved in approximately 70% of cases when a potent NS3/4A inhibitor is used in combination with pegylated IFN-α and ribavirin [10], [11], [12].

HCV genotype 1 is generally considered as a homogeneous group. There are however biological differences between the different subtypes of HCV genotype 1, which are related to differences in their nucleotide and amino acid sequences. Importantly, differences between subtype 1a and 1b (by far the most frequently encountered genotype 1 subtypes in clinical practice) include different efficacies of antiviral drugs and different resistance profiles to such drugs. Indeed, several HCV inhibitors appear to have selective activity against different HCV genotype 1 subtypes, both in vitro and in vivo. Differences have been observed in vitro with NS3/4A protease inhibitors, non-nucleoside inhibitors of HCV RdRp and NS5A inhibitors [13], [14], [15], [16], [17]. For instance, BILB 1941, a non-nucleoside inhibitor of HCV RdRp, has been shown to have better antiviral efficacy in patients infected with HCV subtype 1b than in those infected with HCV subtype 1a, a finding reflecting in vitro experiments [13].

A major issue that limits the efficacy of direct acting antiviral therapies for HCV is the selection by these drugs of resistant variants upon administration [18]. Recent studies with NS3/4A protease inhibitors have shown that the genetic barrier and resistance profiles substantially differ between the different genotype 1 subtypes. For instance, the Arg to Lys substitution at position 155 of the NS3 protease (R155K) is usually selected in subtype 1a replicons treated with telaprevir, but not in subtype 1b replicons [19]. The reason is that only one nucleotide substitution is needed relative to the subtype 1a sequence to generate this variant, whereas two substitutions are needed relative to the 1b sequence (codon usage bias). Overall, natural polymorphisms at positions R155 and V36 are frequent in subtype 1a, but rare in subtype 1b where substitutions at position A156 are preferentially selected in vitro [19]. This is reflected in vivo by the different resistance profiles in patients infected by HCV subtypes 1a and 1b. In the former, the V36 and R155 substitutions represent the backbone of resistance, whereas in the latter resistance is less frequent as it is preferentially associated with substitutions at position A156 that are associated with a decreased fitness of the variants [19], [20], [21]. Similarly, important differences in the resistance profiles have been described in vitro with HCV-796, a non-nucleoside inhibitor of HCV RdRp. The C316Y amino acid substitution has been reported to be selected in both subtype 1a and 1b replicon cells. However, in genotype 1a replicons, the C316Y substitution has low replication capacity that must be compensated for by additional "compensatory" substitutions, including L392F or M414T, resulting in an increase in replication levels of at least 10-fold [19]. A higher genetic barrier to resistance to HCV-796 and related compounds is therefore expected in patients infected with HCV subtype 1a than 1b. In vivo, HCV-796 monotherapy was however shown to select subtype 1a variants with a single C316Y substitution, whereas the C316Y substitution was associated with a number of additional substitutions in subtype 1b patients [22].

As a result of these findings, correct identification of HCV subtypes 1a and 1b is crucial in clinical trials assessing new HCV drugs in order to correctly stratify and interpret efficacy and resistance data. It may also become important in future clinical practice, as tailoring treatment schedules with HCV inhibitors to HCV genotype 1 subtype might become necessary. A variety of molecular methods can be used to identify the HCV genotype and subtype both in clinical trials and practice. Commercial assays have been developed, most of them targeting the 5' noncoding region (5'NCR) of the HCV genome, although this region is the most conserved one. These methods have been shown to differentiate well the different HCV genotypes (1 to 6), except genotype 1 from genotype 6, a rare HCV genotype in the Western world [23], [24]. The goal of our study was to assess the ability of molecular methods targeting the 5'NCR to correctly identify the HCV genotype 1 subtype in patients eligible for clinical trials, and to identify the best method for this purpose.

Results

Hepatitis C Virus Genotype and Subtype Determination by Phylogenetic Analysis of a Portion of the NS5B Gene

Direct sequence analysis of a sufficiently long portion of the NS5B gene followed by phylogenetic analysis is the reference method for identification of HCV genotype and subtype [1], [25]. It was used to identify the HCV genotype and subtype in 516 treatment-naïve patients included in a multicenter clinical trial assessing different schedules of pegylated IFN-α2a and ribavirin [26]. All of these patients were thought to be infected with HCV genotype 1 at inclusion based on local assessment. In fact, 6 patients were infected with genotype 6, including 2 with subtype 6e, one with subtype 6o, one with subtype 6p, one with subtype 6q and one with subtype 6r. These 6 samples were not considered for further analysis in the present study. The remaining 510 patients were confirmed to be infected with HCV genotype 1: 237 of them (46.5%) were infected with HCV subtype 1a and 263 (51.6%) with subtype 1b (Figure 1). As shown in Figure 1, HCV subtype 1a strains segregated into two distinct clades, that were termed 1a clade I (n = 83, 35.0%) and 1a clade II (n = 154, 65.0%). Eight patients (1.6%) were infected with another HCV genotype 1 subtype, including 4 patients with subtype 1d, 2 with subtype 1e, one with subtype 1i, and one with subtype 1l. The remaining 2 patients (0.3%) were infected with genotype 1 but the subtype could not be determined. The ability of the different molecular methods to correctly identify HCV subtypes 1a and 1b was then tested on the 237 and 263 samples containing HCV subtypes 1a and 1b, respectively.

INNO-LiPA HCV 2.0 displays the same 5'NCR oligonucleotide probes as INNO-LiPA HCV 1.0, plus core-encoded oligonucleotide probes aimed at better discriminating between HCV subtypes 1a and 1b. With INNO-LiPA HCV 2.0, subtype identification was corrected in 64 of the 70 subtypes 1a that were incorrectly typed with INNO-LiPA HCV 1.0. Five samples could not be PCR-amplified in the core-coding region and the result was not interpretable with INNO-LiPA HCV 2.0 in the remaining case (Table 1). INNO-LiPA HCV 2.0 also corrected subtype identification in 13 of 23 subtypes 1b that were incorrectly typed with INNO-LiPA HCV 1.0. Eight samples could not be PCR-amplified in the core-coding region and the result was not interpretable with INNO-LiPA HCV 2.0 in the remaining two cases (Table 1). Overall, the second-generation line probe assay correctly classified 97.5% of subtype 1a and 96.2% of subtype 1b strains. When only samples that could be PCR-amplified with the assay procedure were taken into account, correct subtype determination was achieved in 99.6% and 99.2% of cases, respectively (Table 1).

The real-time PCR-based assay targeting both the 5'NCR and the NS5B region, Abbott RealTime HCV Genotype II assay, correctly identified 93.2% of subtype 1a and 88.9% of subtype 1b strains. Only 2 HCV subtype 1b samples could not be PCR-amplified with this method (Table 1).

5'NCR Sequence Analysis in Misclassified Subtype 1a Strains

Among the HCV subtype 1a strains, 47 were misclassified as subtype 1b by Trugene HCV Genotyping Kit and/or INNO-LiPA HCV 1.0, including 33 that were misclassified by both assays, 7 that were misclassified by Trugene HCV Genotyping Kit only, and 7 that were misclassified by INNO-LiPA HCV 1.0 only (Table 2). Figure 2 shows an alignment of their 5'NCR sequences relative to the consensus sequences of the correctly classified strains (including subtype 1a clade I, subtype 1a clade II and subtype 1b). As shown in Figure 2, misclassification of subtype 1a strains into subtype 1b in one or both assays was related to the presence of natural polymorphisms at nucleotide positions 204 and 243, both of which are located within the sequence of an INNO-LiPA HCV 1.0 probe. At position 243, A is the most frequent nucleotide in HCV subtype 1a, in both subtype 1a clade I and clade II. Substitution into a G, the most frequent nucleotide at position 243 in subtype 1b, was found in all cases that were misclassified as subtype 1b by Trugene HCV Genotyping Kit and/or INNO-LiPA HCV 1.0 (Figure 2). At position 204, A is the most frequent nucleotide for subtype 1a clade I, whereas C is the most frequent nucleotide for subtype 1a clade II, and C or T are the most frequent nucleotides for subtype 1b. In spite of the presence of a G at position 243, the presence of an A at position 204 allowed correct identification of subtype 1a with Trugene HCV Genotyping Kit but not with INNO-LiPA HCV 1.0 (Figure 2). The usual presence of a C at position 204 in subtype 1a clade II explains why misclassifications were far more frequent with this clade than with subtype 1a clade I.

Among the 12 subtype 1a strains that were classified as genotype 1, indeterminate subtype with Trugene HCV Genotyping Kit, one had a G and 5 had mixed A and G populations at position 243. Two additional patients with an A at position 243 had a C at position 248. In the remaining 4 cases, no explanation was found in the 5'NCR sequence for the failure to identify the HCV subtype (data not shown). Among the 25 subtype 1a strains that were classified as genotype 1, indeterminate subtype with INNO-LiPA HCV 1.0 (including 6 with the same profile in Trugene HCV Genotyping Kit), 4 had a G and 4 had mixed A and G populations at position 243. Three additional patients with an A at position 243 had a C at position 248 (C only in two of them, a mixture of C and T in one). In the 14 remaining cases, no explanation was found in the 5'NCR sequence for the failure to identify the HCV subtype (data not shown).

5'NCR Sequence Analysis in Misclassified Subtype 1b Strains

Among HCV subtype 1b strains, 8 were misclassified as subtype 1a by Trugene HCV Genotyping Kit and/or INNO-LiPA HCV 1.0, including 3 that were misclassified by both assays, 4 that were misclassified by Trugene HCV Genotyping Kit only, and 1 that was misclassified by INNO-LiPA HCV 1.0 only (Table 2). Figure 3 shows an alignment of their 5'NCR sequences relative to the consensus sequences of the correctly classified subtype 1a and subtype 1b strains. As shown in Figure 3, and as for misclassified subtype 1a strains discussed above, misclassification of subtype 1b strains into subtype 1a was related to the presence of natural polymorphisms at positions 204 and 243. At position 243, G is the most frequent nucleotide in HCV subtype 1b. Substitution into an A, the most frequent nucleotide at position 243 in subtype 1a, was found in all cases that were misclassified as subtype 1a by both Trugene HCV Genotyping Kit and INNO-LiPA HCV 1.0 and by INNO-LiPA HCV 1.0 only, but not in those that were misclassified by Trugene HCV Genotyping Kit only (Figure 3). In the latter, it is the presence of an A at position 204 instead of a C or a T that was responsible for misclassification in all but one case (Figure 3).

Among the 11 subtype 1b strains that were classified as genotype 1, indeterminate subtype with Trugene HCV Genotyping Kit, one had an A at position 243. In the remaining cases, no explanation was found in the 5'NCR sequence for the failure to identify the HCV subtype (data not shown). Among the 15 subtype 1b strains that were classified as genotype 1, indeterminate subtype with INNO-LiPA HCV 1.0 (none of which were classified as indeterminate in Trugene HCV Genotyping Kit), one had an A and one harbored mixed A and G populations at position 243. Both of them had a C at position 248 (C only in one of them and a mixture of C and T in the other one). In the remaining 13 cases, no explanation was found in the 5'NCR sequence for the failure to identify the HCV subtype (data not shown).

Incorrect Subtyping with Abbott RealTime HCV Genotype II Assay, that Targets Both the 5'NCR and NS5B Region

Among the HCV subtype 1a strains, 16 were incorrectly classified by Abbott RealTime HCV Genotype II assay (Table 1): 2 were misclassified as subtype 1b, 12 were classified as genotype 1, indeterminate subtype, one was identified as a mixed 1a/1b infection, and one gave an indeterminate result. In one case, PCR amplification failed, and in one case, not enough serum volume was available for testing.

Among the HCV subtype 1b strains, 27 were incorrectly classified by Abbott RealTime HCV Genotype II assay (Table 1): 3 were misclassified as subtype 1a, 18 were classified as genotype 1, indeterminate subtype, 5 were identified as a mixed 1a/1b infection, and one gave an indeterminate result. In 2 cases, PCR amplification failed, and in one case, not enough serum volume was available for testing.

Source

No comments:

Post a Comment