- Email: [email protected]
Genotyping of hepatitis C virus—comparison of three assays
Journal of Clinical Virology 27 (2003) 276 /285 www.elsevier.com/locate/jcv
Genotyping of hepatitis C virus* comparison of three assays /
Alexander C. Haushofer a,*, Jo¨rg Berg b, Rene´ Hauer c, Doris Trubert-Exinger a, Herbert G. Stekel b, Harald H. Kessler d a
Institute of Laboratory Medicine, General Hospital St. Po¨lten, Propst Fu¨hrer-Strasse 4, A-3100 St. Po¨lten, Austria b Institute of Laboratory Medicine, General Hospital Linz, Krankenhausstrasse 9, A-4020 Linz, Austria c Institute of Laboratory Medicine, Municipal Hospital Lainz, Wolkersbergenstrasse 1, A-1130 Vienna, Austria d Institute of Hygiene, Karl-Franzens-University, Universitaetsplatz 4, A-8010 Graz, Austria Received 29 March 2002; received in revised form 6 August 2002; accepted 7 September 2002
Abstract Background: Genotyping of hepatitis C virus (HCV) is clinically relevant to epidemiology, prognosis, and therapeutical management of HCV infection. Objectives: Accuracy and specificity of three assays for HCV genotyping/subtyping were determined. The TruGene HCV 5?NC Genotyping Kit (TruGene), which is a direct sequencing test and two assays based on reversed hybridization, Inno-LiPA HCV II assay and ViennaLab HCV Strip Assay, were compared. Amplification products generated by the Cobas AmplicorTM HCV Test were used. Study design: A total of 100 consecutive HCV RNA positive samples derived from patients with chronic hepatitis C were examined for their genotypes/subtypes by the three assays. Results: Identification of genotypes and subtypes by the TruGene assay as reference test for the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay or Inno-LiPA HCV II assay as reference test for the TruGene and the ViennaLab HCV Strip Assay showed similar results for overall accuracies (TruGene as reference test for Inno-LiPA HCV II and ViennaLab HCV Strip Assay, genotypes/subtypes: 100%/95.5% and 97%/92%; Inno-LiPA HCV II as reference test for TruGene and ViennaLab HCV Strip Assay, genotypes/subtypes: 99%/85.9% and 97%/87.9%) and specificities (TruGene as reference test for Inno-LiPA HCV II and ViennaLab HCV Strip Assay, genotypes/subtypes: 100%/97.8% and 99%/97.7%; Inno-LiPA HCV II as reference test for TruGene and ViennaLab HCV Strip Assay, genotypes/subtypes: 100%/99.4% and 99.7%/98%). Conclusions: The three assays were found to be reliable for the detection and discrimination of all HCV genotypes common in Europe and in North America and to be suitable for the routine diagnostic laboratory. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Hepatitis C virus; Genotype; Sequencing; Hybridization assay; Genotyping assay
1. Introduction * Corresponding author. Tel.: /43-2742-300-3750; fax: / 43-2742-300-2710. E-mail address: [email protected] (A.C. Haushofer).
Hepatitis C virus (HCV) is a small, enveloped human flavivirus with a positive strand ribonucleic acid (RNA) genome of approximately 9400 bases
1386-6532/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 6 5 3 2 ( 0 2 ) 0 0 1 8 3 - X
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
in length. The genome comprises one large open reading frame flanked by highly conserved untranslated regions (UTRs) at the 5? and 3?termini. Nucleotide sequences of various HCV isolates have been examined worldwide, and a substantial nucleotide sequence variation was found throughout the entire HCV genome (Choo et al., 1991). The genetic variation of HCV is thought to reflect sequence errors that are predominantly introduced during virus replication due to a non-proofreading RNA-dependent RNA polymerase. Many different but closely related HCV variants are generated within one HCV infected individual (Bendinelli et al., 1999; Davis, 1999). HCV isolates show four levels of genetic variations: types, subtypes, isolates, and quasispecies (Bukh et al., 1995). Up to date, 11 genotypes including more than 90 subtypes have been identified (Maertens et al., 1997). All HCV isolates separate into phylogenetically related clusters called subtypes. Subtypes can be classified into several major types that show sequence similarities of 65 /75% of the total genome (Leon et al., 2002). HCV genotypes may differ from each other by as much as 33% of their nucleotide sequence distributed over the entire viral genome. HCV genotypes were originally identified by a comparison of published HCV sequences (Okamoto et al., 1992). Despite a high degree of sequence conservation in the 5?UTRs, genotype-specific differences do exist in this region. Based on the sequence variations in the 5?UTR and the NS5 region Simmonds et al. proposed a classification of HCV types and subtypes, that is currently most often used (Simmonds et al., 1994). Genotyping and subtyping of HCV is relevant to the epidemiology of HCV, vaccine development, clinical management, and assessment of the risk-benefit ratio of therapeutic measures against chronic HCV infection (McHutchison et al., 1998; Poynard et al., 1998, 2000). Therapeutical management of chronic HCV infection based on the HCV genotype has been suggested in the Consensus Statement of the European Association of the Study of the Liver EASL Consensus Panel (EASL Consensus Panel, 1999). With regard to chronic hepatitis C, HCV genotyping has shown to be of predictive value,
277
which was found independent of viral load, age, ethiology, duration of disease, and histological findings (Chemello et al., 1995a; Nousbaum et al., 1995). Furthermore, differences between genotypes in their response to treatment were observed in recent studies in which a combined treatment with interferon-alpha and ribavirin was administered (Chemello et al., 1995b; McHutchison et al., 1998; Poynard et al., 1998). The HCV genotype can be determined by nucleotide sequencing of the entire genome followed by composition of a phylogenetic tree, which is presently the ‘gold standard’ for the detection and identification of the various HCV genotypes and subtypes (Lau et al., 1995). This approach, however, is cumbersome and regarded as impractical for routine clinical laboratory settings (Ross et al., 2000). More convenient methods focus on the amplification of defined regions of the HCV genome by reverse transcription (RT)-PCR followed by nucleotide sequencing, RT-PCR with genotype-specific primers, RT-PCR followed by restriction fragment length polymorphism, and RT-PCR with universal primers followed by subtype-specific hybridization (Simmonds et al., 1993; Widell et al., 1994; Davidson et al., 1995; Stuyver et al., 1996). The TruGene HCV 5?NC Genotyping Kit (Visible Genetics, Toronto, Ontario) has been introduced recently. This assay is based on a nucleotide sequencing reaction of RT-PCR products from the 5?UTR. The RT-PCR products can conveniently be retrieved from the Cobas AmplicorTM HCV Test (Roche Molecular Systems, Pleasanton, CA) or the Cobas AmplicorTM HCV Monitor Test (Roche). Those RT-PCR products are also used in genotyping assays that are based on subtype-specific hybridization, the Inno-LiPA HCV II (Innogenetics N.V., Gent, Belgium) assay and the ViennaLab HCV Strip Assay (ViennaLab Labordiagnostika GmbH, Vienna, Austria). The latter two assays have especially been designed for routine laboratory settings. The aim of the present study was to compare three different HCV genotyping/subtyping assays with regard to accuracy and specificity. Furthermore, performance of these assays in a diagnostic routine laboratory was evaluated.
278
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
2. Materials and methods 2.1. Samples A total of 100 consecutive HCV RNA positive samples derived from patients with chronic hepatitis C, analyzed at the Molecular Diagnostics Laboratory, Institute of Hygiene, Karl-FranzensUniversity Graz, were studied. All samples were negative for HIV-1 antibodies and HBsAg. Amplification products were generated by RTPCR with the qualitative Cobas AmplicorTM HCV Test (Roche Molecular Systems). Aliquots of the amplification products were stored at /70 8C until further examination. Amplification products were analyzed with all three genotyping assays. 2.2. HCV genotyping/subtyping For the TruGeneTM HCV 5?NC Genotyping Kit (Visible Genetics), amplification products were neutralized and purified with the High Pure PCR Product Purification Kit (Roche Molecular Biochemicals, Mannheim, Germany). The sequencing procedure was performed according to the manufacturer?s instructions employing the CLIPTM sequencing technique (Visible Genetics). The CLIP technique allows both directions of the amplification product to be sequenced simultaneously within the same tube using two different dye-labeled primers for each of the four sequencing reactions. Electrophoresis and data analysis were carried out with the automated OpenGeneTM DNA sequencing system (Visible Genetics) that were combined with the automated Long-Read TowerTM for the separation of the sequenced fragments. Data were acquired with the GeneLibrarianTM module of the GeneObjectsTM (Visible Genetics) software. Forward and reverse sequence data were combined and compared with a HCV sequence library. Thereafter, the sequences of the HCV 5?UTR were subjected to a phylogenetic analysis using the PHYLIP software package, version 3.5c (Forns and Bukh, 1998). Distances between pairs of sequences were estimated with the DNA-DIST program. Phylogenetic trees were constructed by the unweighted pair group method using arithmetic
averages on the previous sets of pair-wise distance according to Ross et al. (2000). For the correct identification of a certain HCV subtype, the sequence has to agree for more than 90% with a corresponding sequence of the HCV sequence library according to the guideline for the interpretation of subtypes in the ISO 9001 certified molecular diagnostics laboratory. The Inno-LiPA HCV II assay (Innogenetics N.V.) and the ViennaLab HCV Strip Assay (ViennaLab Labordiagnostika GmbH) were performed according to the manufacturer’s instructions. These assays are based on the reversehybridization principle of RT-PCR products derived from the 5?UTR of the HCV genome. Briefly, biotinylated RT-PCR products are hybridized to a selection of specific oligonucleotide probes immobilized on membrane strips. The labeled RT-PCR products will only hybridize to a probe (or line) matching the sequence of the isolate, allowing stringent discrimination at the subtype level. The high specificity is obtained by using very stringent hybridization conditions. The presence of probe/amplicon-biotin hybrids is revealed by an alkaline phosphatase-labeled streptavidin conjugate by converting a chromogen gaining a color reaction. The Inno-LiPA HCV II strip is based on 21 lines of probes, the ViennaLab HCV Strip Assay on 18 lines of probes. The genotype/subtype is found by comparing the pattern of positive (colored) lines of the hybridization assays with the special interpretations charts enclosed to the package insert by the companies. 2.3. Statistical analysis The TruGene assay or the Inno-LiPA HCV II were alternatively used as reference test for the evaluation of concordance and for the calculation of accuracy and specificity. The accuracy was defined as the number of correctly determined genotypes or subtypes from the TruGene assay or the Inno-LiPA HCV II or the ViennaLab HCV Strip Assay divided by the total number of genotypes/subtypes determined firstly by the TruGene assay and secondly by the Inno-LiPA HCV II as reference tests. Specificity was defined as the number of correctly identified negative results
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
obtained with the TruGene assay or the InnoLiPA HCV II or the ViennaLab HCV Strip Assay for a certain genotype/subtype divided by this number plus false positive results determined firstly by the TruGene assay and secondly by Inno-LiPA HCV II as reference tests.
3. Results 3.1. Detection of genotypes All samples genotyped by the three assays showed 100 distinct results with the TruGene assay, 101 results (one sample with a double infection showed two genotypes) with the InnoLiPA HCV II assay, and 99 distinct results (plus one non-interpretable genotype) with the ViennaLab HCV Strip Assay (Table 1). With the TruGene assay as reference test, overall accuracies of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 100 and 97%, respectively. With the Inno-LiPA HCV II as reference test, overall accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 99 and 97%, respectively. With the TruGene assay as reference test, overall specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were
279
100 and, 99% respectively. With the Inno-LiPA HCV II assay as reference test, the overall specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 100 and 99.7%, respectively. Accuracy and specificity results for genotype 1 were as follows: With the TruGene assay as reference test, accuracies of the Inno-LiPA HCV II and the ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. With the Inno-LiPA HCV II as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II and the ViennaLab HCV Strip Assay were 100% for both assays. This result was also obtained with the Inno-LiPA HCV II assay as reference test. Accuracy and specificity results for genotype 3 were as follows: With the TruGene assay as reference test, accuracies of the Inno-LiPA HCV II and the ViennaLab HCV Strip Assay were 100 and 92.9%, respectively. With the Inno-LiPA HCV II as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 100 and 92.9%, respectively. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II and the ViennaLab HCV Strip Assay were
Table 1 Concordance of HCV genotyping results (sample size, n/100) (a) TruGene assay as reference test. (b) Inno-LiPA HCV II assay as reference test. (a) Genotype 1 2 3 4 Total
TruGene assay
(b) Genotype 1 2 3 4 Total
Inno-LiPA HCV II assay
a b c
83 2 14 1 100
83b 3b 14 1 101b
Inno-LiPA HCV II assay 1 2 3 4 NIa 83 2 14 1 100
ViennaLab HCV Strip assay 1 2 3 4 NIa c 32 1 2 13 1c 1c 97
TruGene assay 1 2 3 83 1c 2 14
ViennaLab HCV Strip assay 1 2 3 4 NIa 82 1c 3 13 1c c 1 98
4
1 100
NIa
NI, non-interpretable genotype. With the Inno-LiPA HCV II assay one sample was identified with double HCV infection of two genotypes (1b, 2a/2c, 2b). Inconsistent genotype assignments.
280
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
100 and 98.8%, respectively. With the Inno-LiPA HCV II as reference test, specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. The TruGene assay detected genotype 2 in two samples; this was in concordance with both, the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay (Table 1). With Inno-LiPA HCV II assay as reference test, an additional sample was identified with genotype 2; this was in concordance with the ViennaLab HCV Strip Assay, but not with the TruGene assay. Both, the TruGene assay and the Inno-LiPA HCV II assay found one sample with genotype 4; this sample, however, was typed as genotype 3 by the ViennaLab HCV Strip Assay (Table 1). 3.2. Detection of subtypes All samples subtyped by the three assays showed 89 distinct results with the TruGene assay (11 samples could not be further subtyped), 99 results with the Inno-LiPA HCV II assay (one sample could only be genotyped as genotype 1; for evaluating concordance the sample with genotypes/subtypes 1b, 2a/2c, and 2b was counted as ‘own subtype’), and 98 distinct results with the ViennaLab HCV Strip Assay (one non-interpretable genotype and one sample could only be genotyped as genotype 1) (Table 2). With the TruGene assay as reference test, overall accuracies of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 95.5 and 92%, respectively (Table 2). With the Inno-LiPA HCV II assay as reference test, overall accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 85.9 and 87.9%, respectively. With the TruGene assay as reference test, overall specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 97.8 and 97.7%, respectively. With the Inno-LiPA HCV II assay as reference test, the overall specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 99.4 and 98%, respectively. Eleven samples could not be further subtyped with the TruGene assay because of too less agreement with the HCV sequence library (Table 2). When 10 samples with genotype 1, which could
not be further subtyped with the TruGene assay, were determined with the Inno-LiPA HCV II assay, 4 samples were found to be subtype 1b, one sample 1a, 2 samples 1a/1b, one sample 1a/ 1b, one sample 1, and one sample 1b, 2a/2c, 2b. With the ViennaLab HCV Strip Assay, 5 samples were found to be subtype 1b, 3 samples 1a, one sample 1a/1b, and one sample 2b. The one sample with genotype 4, which could not be further subtyped with the TruGene assay, was found to be subtype 4c/4d with the Inno-LiPA HCV II assay. With the ViennaLab HCV Strip Assay, this sample was found to be subtype 3a. Fifteen samples could be identified as subtype 1a with all three assays. With the TruGene assay as reference test, accuracies of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 83.3 and 88.9%, respectively (Table 2). With the Inno-LiPA HCV II assay as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 93.8% each. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 98.8 and 96.3%, respectively. With the Inno-LiPA HCV II assay as reference test, the specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 96.4 and 95.2%, respectively. Forty-five samples could be identified as subtype 1b with all three assays. With the TruGene assay as reference test, accuracies of the InnoLiPA HCV II assay and the ViennaLab HCV Strip Assay were 100 and 93.8%, respectively (Table 2). With the Inno-LiPA HCV II assay as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 87.3 and 92.6%, respectively. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 88.5% each. With the Inno-LiPA HCV II assay as reference test, the specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 100 and 97.8%, respectively. With the TruGene assay, 7 samples were found to be subtype/s 1a or/and 1b with 6 of them in concordance with both of the reverse hybridization assays.
(a) Subtype
TruGene assay
1a 1b 1a or/and 1b 2a or/and 2c 3a Total
18 48 7 2 14 89b
(b) Subtype
Inno-LiPA HCV II assay
1a 1b 1a/1b 1a/1b 2a/2c 3a 4c/4d 1b, 2a/2c, 2b Total
16 53 7 5 2 14 1 1 99c
a b c d
Inno-LiPA HCV II assay 1a 1b 1a/1b 15 3d 48 1d 6
2a/2c
3a
1
ViennaLab HCV Strip assay 4 1a 1b 1a/1b 1a/1b 16 2d 45 2d d 1 6
2
2a/2c
2b 3a NIa
1d 2
14
13
85 TruGene assay 1a 1b 1a or/and 1b 15 48 1d 6 3d
1
1d
82 2a or/and 2c
ViennaLab HCV Strip assay 3a 1b 4b 1a 1b 1a/1b 1a/1b 1d 15 1d 4d 50 2d d 1 5 2d 2d 2
2
2a/2c
2b 3a NIa
1d
2 14
13 1d
1d 1d
1d 85
NI /non-interpretable genotype/subtype. With the TruGene assay 10 samples genotyped as 1 and one sample genotyped as 4 could not be further subtyped. With the Inno-LiPA HCV II assay one sample could only be genotyped as 1 and not further subtyped. Inconsistent or incomplete subtype assignments.
87
1
1d
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
Table 2 Concordance of HCV subtyping results (sample size, n /100) (a) TruGene assay as reference test. (b) Inno-LiPA HCV II assay as reference test
281
282
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
All of the assays identified 2 samples with subtype/s 2a or/and 2c. The Inno-LiPA HCV II assay detected one sample with a double infection. In this sample genotypes 1 and 2 (1b, 2a/2c/2b) were found. Thirteen samples could be identified as subtype 3a with all of the three assays. With the TruGene assay and the Inno-LiPA HCV II assay, one additional sample was found to be subtype 3a, which was not interpretable with the ViennaLab HCV Strip Assay. With the TruGene assay as reference test, accuracies of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 100 and 92.9%, respectively (Table 2). With the Inno-LiPA HCV II assay as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 100 and 92.9%, respectively. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. With the Inno-LiPA HCV II assay as reference test, the specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. All of the evaluated genotyping assays proved to be suitable for the routine diagnostic laboratory. For the TrueGene assay, the whole procedure took 4 h including sequencing reaction, electrophoresis, and interpretation of data, which took up to 15 min per sample. When the Inno-LiPA HCV II assay was used the whole procedure could be done within 3 h. The ViennaLab HCV Strip Assay took 2 h because of shortened incubation periods and ready-to-use reagents. The hands-on time required for both of the hybridization assays was similar including that for interpretation (up to 1 min per sample). With the TruGene assay less hands-on work had to be done in comparison with the other assays.
4. Discussion In this study three different assays for HCV genotyping/subtyping based on amplification products of the 5?UTR were compared with regard to accuracy and specificity. A total of 100 consecutive clinical samples were investigated. The distribution
of HCV genotypes in this study was found to be comparable to that in a recent study on the HCV genotype distribution in Vienna and surrounding areas (Haushofer et al., 2001). This suggests that the results presented in this study are obtained from a representative selection of clinical samples that reflect the challenge of the daily routine in a clinical laboratory in Austria. The employed assays are designed to identify genotypes and subtypes from the 5?UTR of the HCV genome. Because this region is highly conserved, it has been chosen as target of choice for most nucleic acid amplification-based detection assays for detection of HCV (Germer et al., 1999). The 5?UTR, however, is still sufficiently variable to form type and subtype-specific motifs. These motifs can be detected and discriminated from each other by nucleotide sequencing or reverse hybridization using specific probes (Stuyver et al., 1995, 1996; Maertens et al., 1997). The TruGene assay offers the possibility of using a direct sequencing technique in a convenient fashion for the routine diagnostic laboratory. The TruGene approach with its semi-automated sequencing technique seems to fill a technical gap between the classical cumbersome sequencing methods. The Inno-LiPA HCV II, which is probably the most frequently used HCV genotyping assay, is well characterized towards high sensitivity and specificity and exhibits a 100% correlation with the 5?UTR nucleotide sequences (van Doorn et al., 1994; Mahaney et al., 1994; Giannini et al., 1994; Zeuzem et al., 1995). But some HCV subtypes, for example 2a and 2c, do not exhibit enough sequence polymorphism in the 5?UTR to be discriminated by the Inno-LiPA HCV II assay. The ViennaLab HCV Strip Assay is similar to the Inno-LiPA HCV II assay and has recently been introduced. The TruGene and the Inno-LiPA HCV II assays are well covered by the literature (Ansaldi et al., 2001; Ross et al., 2000; Halfon et al., 2001; van Doorn et al., 1994; Mahaney et al., 1994; Giannini et al., 1994; Zeuzem et al., 1995). Therefore, they were alternatively used as reference tests in this study. For all three assays, overall accuracies and specificities for the various genotypes were satisfying and were found in a good agreement with
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
previously published studies that used the TruGene assay and/or the Inno-LiPA HCV II assay (Ansaldi et al., 2001; Halfon et al., 2001; van Doorn et al., 1994; Mahaney et al., 1994; Giannini et al., 1994; Zeuzem et al., 1995). With the TruGene assay and the Inno-LiPA HCV II assay Halfon et al. (2001) recently obtained concordant HCV genotyping results in 100% of samples. In that study the overall accuracies of the two tests regardless of the genotype were 76% for the TruGene assay and 74% for the Inno-LiPA HCV II assay compared to NS5B sequence analysis. The accuracy and specificity for genotypes 1 and 3 obtained with the TruGene assay and the InnoLiPA HCV II assay were found to be in agreement with those published by Lee et al. (1997) for the Inno-LiPA HCV II assay and phylogenetic analysis of the NS5B region. For 11 samples only the genotype could be determined with the TruGene assay. The reason for this inconvenience may be related to the relatively short sequence subjected to the phylogenetic tree analysis. The overall accuracy for the discrimination of subtypes was found to be in agreement with the reported complete concordance in 91% of samples between the TruGene assay and the DNA enzyme immunoassay (Ross et al., 2000). With regard to problems in the discrimination of subtypes 1a and b Giannini et al. (1994) demonstrated a 100% correlation between 5?UTR sequences and Inno-LiPA results, but also showed that the polymorphism at position 99 of the 5?UTR did not always prove to be specifically linked to subtype 1a or b sequences detected in the core region. In fact, only one or two minor nucleotide differences distinguish unique subtypes from each other. For example only a single base exchange at position 99 (adenine to guanine) distinguishes between genotypes 1a and b in the 5?UTR sequence (Stuyver et al., 1995). At position 99 Smith et al. (1995), Stuyver et al. (1995) found several HCV isolates of subtype 1a with guanine instead of adenine, and subtype 1b with adenine instead of guanine. Only by using a different region such as NS5B a further investigation of this phenomenon could be done. It has been demonstrated that the reverse hybridization assay correctly interpreted genotypes 1, 2 and 3 in
283
comparison to sequence and phylogenetic analysis of the NS5B region; one of 14 samples with subtype 1a, however, was misinterpreted as subtype 1b and 3 of 36 samples with subtype 1b were misinterpreted as subtype 1a (Lee et al., 1997). Zeuzem et al. (1995) have shown a 100% correlation for genotyping and genotype 2 subtyping in a homogenous German patient population. However, 10% of samples with subtype 1b and 9% of samples with subtype 1a showed the inverted nucleotide sequence at position 99. In a recent study, Ansaldi et al. (2001) demonstrated concordant results between the TruGene assay and the Inno-LiPA HCV II assay in 90.9% of samples. Of 3 samples, which could only be determined as genotype 1 by Inno-LiPA HCV II assay, were found to be subtype 1b (one sample) and genotype 4 (two samples). Those samples, which had been determined as genotype 4, showed only a difference in 6 nucleotides to the closest genotype 1 isolate (Ansaldi et al., 2001). Furthermore, by sequence analysis one Inno-LiPA HCV II assay genotype 1a sample was typed as 1b and 2 genotype 1b samples were typed as 1a. In the present study one sample was identified as double infection with two genotypes by the InnoLiPA HCV II assay. This sample could only be determined as genotype 1 with the TruGene assay and was found to be subtype 2b with the ViennaLab HCV Strip Assay. Tuveri et al. (1997) demonstrated detection of double infections in hemophilic patients with the Inno-LiPA assay confirmed by sequencing of the core region. Furthermore, Lau et al. (1995) showed the usefulness of that hybridization test for the detection of double infection with its high sensitivity for capturing minor sequence variants. With regard to performance, the TruGene assay and the Inno-LiPA HCV II assay have shown to be equivalent. The TruGene assay, as largely automated sequencing assay for determination of HCV genotypes, proved to be useful for a highthroughput routine diagnostic laboratory with no significant differences in costs without calculation of instrument costs in comparison to the hybridization tests. The ViennaLab HCV Strip Assay offers the advantage of a markedly shortened
284
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
hybridization time and improved handling compared to the Inno-LiPA HCV II assay. In conclusion, the TruGene and the hybridization assays were found to be reliable for the determination of all common HCV types in Europe and in North America. All assays proved to be suitable for the diagnostic routine laboratory.
Acknowledgements The authors gratefully acknowledge the expert technical assistance of B. Aigner, R. Atteneder, M. Bozic, M. Deimel, L. Gonaus, A. Huber, and G. Ho¨lzl.
References Ansaldi F, Torre F, Bruzzone BM, Picciotto A, Crovari P, Icardi GJ. Evaluation of a new hepatitis C virus sequencing assay as a routine method for genotyping. Med Virol 2001;63(1):17 /21. Bendinelli M, Vatteroni ML, Maggi F, Pistello M. Hepatitis C virus. In: Spector S, editor. Viral hepatitis: diagnosis, therapy and prevention. Totowa, NJ: Humana Press Inc, 1999:65 /127. Bukh J, Miller RH, Purcell RH. Genetic heterogeneity of hepatitis C virus quasispecies and genotypes. Semin Liver Dis 1995;15:41 /63. Chemello L, Bonetti P, Cavalletto L, Talato F, Donadon V, Casarin P, et al. Randomized trial comparing three different regimens of alpha-2a-interferon in chronic hepatitis C. Hepatology 1995a;22:700 /6. Chemello L, Cavaletto L, Bernardinello E, Guido M, Pontisso P, Alberti A. The effect of interferon alfa and ribavirin combination therapy in naive patients with chronic hepatitis C. J Hepatol 1995b;23(Suppl 2):8 /12. Choo Q-L, Richman KH, Han JH, Berger K, Lee C, Dong C, et al. Genetic organization and diversity of the hepatitis C virus. Proc Natl Acad Sci USA 1991;88:2451 /5. Davidson F, Simmonds P, Ferguson JC, Jarvis LM, Dow BC, Follet EA, et al. Survey of major genotypes and subtypes of hepatitis C virus using RFLP of sequencing amplified from the 5? non-coding region. J Gen Virol 1995;76:1197 /204. Davis GL. Hepatitis C. In: Schiff ER, Sorell MF, Maddrey WC, editors. Schiff’s disease of the liver, 8th ed.. Philadelphia, USA: Lippincott-Raven, 1999:793 /836. EASL Consensus Panel. EASL International Consensus Conference on Hepatitis C. Paris, 26 /28 February 1999. J Hepatol 1999;30:956 /961.
Forns X, Bukh J. Methods for determing hepatitis C virus genotype. Viral Hep Rev 1998;4:1 /19. Germer JJ, Rys PN, Thorvilson JN, Persing DH. Determination of hepatitis C virus genotype by direct sequence analysis of products generated with the Amplicor HCV test. J Clin Microbiol 1999;37:2625 /30. Giannini C, Thiers V, Nousbaum JB, Stuyver L, Maertens G, Bre´chot C. Comparative analysis of the hepatitis C virus core PCR and LiPA genotyping assay. J Hepatol 1994;23:246 /53. Halfon P, Trimoulet P, Bourliere M, Khiri H, de Ledinghen V, Couzigou P, et al. Hepatitis C virus genotyping based on 5? noncoding sequence analysis (Trugene). J Clin Microbiol 2001;39:1771 /3. Haushofer AC, Kopty C, Hauer R, Brunner H, Halbmayer WM. HCV genotypes and age distribution in patients of Vienna and surrounding areas. J Clin Virol 2001;20:41 /7. Lau JY, Mizokami M, Kolberg JA, Davis GL, Prescott LE, Ohno T, et al. Application of six hepatitis C virus subtyping systems to sera of patients with chronic hepatitis C in the United States. J Infect Dis 1995;171:281 /9. Lee JH, Roth WK, Zeuzem S. Evaluation and comparison of different hepatitis C virus genotyping and serotyping assays. J Hepatol 1997;26:1001 /9. Leon RV, de Medina MD, Schiff ER. Hepatitis C virology antigen, antibody, and molecular testing. In: Gordon SC, editor. Management of chronic viral hepatitis. New York, USA: Marcel Dekker, Inc, 2002:33 /64. Maertens G, Stuyver L. Genotypes and genetic variation of hepatitis C virus. In: Harrison TJ, Zuckerman AJ, editors. The molecular medicine of viral hepatitis. Chichester, UK: Wiley, 1997:183 /233. Mahaney K, Tedesci V, Maertens G, Di Bisceglie A, Vergalla J, Hoofnagle JH, et al. Genotype analysis of hepatitis C virus in American patients. Hepatology 1994;20:1405 /11. McHutchison JG, Gordon SC, Schiff ER, Shiffman ML, Lee WM, Rustgi VK, et al. Interferon alfa-2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C. Hepatitis interventional therapy group. N Engl J Med 1998;339:1485 /92. Nousbaum J-B, Pol S, Nalpas B, Landais P, Berthelot P, Bre´chot C. Hepatitis C virus type 1b (II) infection in France and Italy. Ann Int Med 1995;122:161 /8. Okamoto H, Kurai K, Okada S, Yamamoto K, Lizuka H, Tanaka T, et al. Full-length of a hepatitis C virus genome having poor homology to reported isolates: comparative study of four distinct genotypes. Virology 1992;188:331 /41. Poynard T, Marcellin P, Lee SS, Niederau C, Minuk GS, Ideo G, et al. Randomised trial of interferon alpha2b plus ribavirin for 48 weeks or for 24 weeks versus interferon alpha2b plus placebo for 48 weeks for treatment of chronic infection with hepatitis C virus. International Hepatitis Interventional Therapy Group (IHIT). Lancet 1998;35:1426 /32. Poynard T, McHutchison J, Goodman Z, Ling MH, Albrecht J. Is an ‘a la carte’ combination interferon alpha-2b plus ribavirin regimen possible for the first line treatment in
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285 patients with chronic hepatitis C?. Hepatology 1995;31:211 /8. Ross RS, Viazov SO, Holtzer CD, Beyou A, Monnet A, Mazure C, et al. Genotyping of hepatitis C virus isolates using CLIP sequencing. J Clin Microbiol 2000;38:3581 /4. Simmonds P, Holmes EC, Cha TA, Chan SW, McOmish F, Irvine B, et al. Classification of hepatitis C virus into six major genotypes and a series of subtypes by phylogenetic analysis of the NS-5 region. J Gen Virol 1993;74:2391 /9. Simmonds P, Alberti A, Alter F, Bonino F, Bradley DW, Brechot C, et al. A proposed system for the nomenclature of hepatitis C viral genotypes. Hepatology 1994;19:1321 /4. Smith DB, Mellor J, Jarvis LM, Davidson F, Kolberg J, Urdea M, et al. Variation of the hepatitis C virus 5?non-coding region: implications for secondary structure, virus detection and typing. J Gen Virol 1995;76:1749 /61. Stuyver L, Wyseur A, Van Arnhem W, Lunel F, Laurent-Puig P, Pawlotsky JM, et al. Hepatitis C virus genotyping. By means of 5?-UR/core line probe assay and molecular analysis of untypeable samples. Virus Res 1995;38:137 /57.
285
Stuyver L, Wyseur A, Van Arnhem W, Hernandez F, Maertens G. Second-generation line probe assay for hepatitis C virus genotyping. J Clin Microbiol 1996;34:2259 /66. Tuveri R, Rotschild C, Pol S, Reijasse D, Persico T, Gazengel C, et al. Hepatitis C virus genotypes in French haemophiliacs: kinetics and reappraisal of mixed infections. J Med Virol 1997;51:36 /41. van Doorn LJ, Kleter B, Stuyver L, Maertens G, Brouwer J, Schalm S, et al. Analysis of hepatitis C virus genotypes by a line probe assay (LiPA) and correlation with antibody profiles. J Hepatol 1994;12:122 /9. Widell A, Shev S, Mansson S, Zhang YY, Foberg U, Norkrans G, et al. Genotyping of hepatitis C virus isolates by a modified polymerase chain reaction assay using type specific primers: epidemiological applications. J Med Virol 1994;44:272 /9. Zeuzem S, Ru¨ster B, Lee JH, Stripf T, Roth WK. Evaluation of a reverse hybridisation assay for genotyping of hepatitis C virus. J Hepatol 1995;23:654 /61.
Genotyping of hepatitis C virus* comparison of three assays /
Alexander C. Haushofer a,*, Jo¨rg Berg b, Rene´ Hauer c, Doris Trubert-Exinger a, Herbert G. Stekel b, Harald H. Kessler d a
Institute of Laboratory Medicine, General Hospital St. Po¨lten, Propst Fu¨hrer-Strasse 4, A-3100 St. Po¨lten, Austria b Institute of Laboratory Medicine, General Hospital Linz, Krankenhausstrasse 9, A-4020 Linz, Austria c Institute of Laboratory Medicine, Municipal Hospital Lainz, Wolkersbergenstrasse 1, A-1130 Vienna, Austria d Institute of Hygiene, Karl-Franzens-University, Universitaetsplatz 4, A-8010 Graz, Austria Received 29 March 2002; received in revised form 6 August 2002; accepted 7 September 2002
Abstract Background: Genotyping of hepatitis C virus (HCV) is clinically relevant to epidemiology, prognosis, and therapeutical management of HCV infection. Objectives: Accuracy and specificity of three assays for HCV genotyping/subtyping were determined. The TruGene HCV 5?NC Genotyping Kit (TruGene), which is a direct sequencing test and two assays based on reversed hybridization, Inno-LiPA HCV II assay and ViennaLab HCV Strip Assay, were compared. Amplification products generated by the Cobas AmplicorTM HCV Test were used. Study design: A total of 100 consecutive HCV RNA positive samples derived from patients with chronic hepatitis C were examined for their genotypes/subtypes by the three assays. Results: Identification of genotypes and subtypes by the TruGene assay as reference test for the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay or Inno-LiPA HCV II assay as reference test for the TruGene and the ViennaLab HCV Strip Assay showed similar results for overall accuracies (TruGene as reference test for Inno-LiPA HCV II and ViennaLab HCV Strip Assay, genotypes/subtypes: 100%/95.5% and 97%/92%; Inno-LiPA HCV II as reference test for TruGene and ViennaLab HCV Strip Assay, genotypes/subtypes: 99%/85.9% and 97%/87.9%) and specificities (TruGene as reference test for Inno-LiPA HCV II and ViennaLab HCV Strip Assay, genotypes/subtypes: 100%/97.8% and 99%/97.7%; Inno-LiPA HCV II as reference test for TruGene and ViennaLab HCV Strip Assay, genotypes/subtypes: 100%/99.4% and 99.7%/98%). Conclusions: The three assays were found to be reliable for the detection and discrimination of all HCV genotypes common in Europe and in North America and to be suitable for the routine diagnostic laboratory. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Hepatitis C virus; Genotype; Sequencing; Hybridization assay; Genotyping assay
1. Introduction * Corresponding author. Tel.: /43-2742-300-3750; fax: / 43-2742-300-2710. E-mail address: [email protected] (A.C. Haushofer).
Hepatitis C virus (HCV) is a small, enveloped human flavivirus with a positive strand ribonucleic acid (RNA) genome of approximately 9400 bases
1386-6532/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 6 5 3 2 ( 0 2 ) 0 0 1 8 3 - X
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
in length. The genome comprises one large open reading frame flanked by highly conserved untranslated regions (UTRs) at the 5? and 3?termini. Nucleotide sequences of various HCV isolates have been examined worldwide, and a substantial nucleotide sequence variation was found throughout the entire HCV genome (Choo et al., 1991). The genetic variation of HCV is thought to reflect sequence errors that are predominantly introduced during virus replication due to a non-proofreading RNA-dependent RNA polymerase. Many different but closely related HCV variants are generated within one HCV infected individual (Bendinelli et al., 1999; Davis, 1999). HCV isolates show four levels of genetic variations: types, subtypes, isolates, and quasispecies (Bukh et al., 1995). Up to date, 11 genotypes including more than 90 subtypes have been identified (Maertens et al., 1997). All HCV isolates separate into phylogenetically related clusters called subtypes. Subtypes can be classified into several major types that show sequence similarities of 65 /75% of the total genome (Leon et al., 2002). HCV genotypes may differ from each other by as much as 33% of their nucleotide sequence distributed over the entire viral genome. HCV genotypes were originally identified by a comparison of published HCV sequences (Okamoto et al., 1992). Despite a high degree of sequence conservation in the 5?UTRs, genotype-specific differences do exist in this region. Based on the sequence variations in the 5?UTR and the NS5 region Simmonds et al. proposed a classification of HCV types and subtypes, that is currently most often used (Simmonds et al., 1994). Genotyping and subtyping of HCV is relevant to the epidemiology of HCV, vaccine development, clinical management, and assessment of the risk-benefit ratio of therapeutic measures against chronic HCV infection (McHutchison et al., 1998; Poynard et al., 1998, 2000). Therapeutical management of chronic HCV infection based on the HCV genotype has been suggested in the Consensus Statement of the European Association of the Study of the Liver EASL Consensus Panel (EASL Consensus Panel, 1999). With regard to chronic hepatitis C, HCV genotyping has shown to be of predictive value,
277
which was found independent of viral load, age, ethiology, duration of disease, and histological findings (Chemello et al., 1995a; Nousbaum et al., 1995). Furthermore, differences between genotypes in their response to treatment were observed in recent studies in which a combined treatment with interferon-alpha and ribavirin was administered (Chemello et al., 1995b; McHutchison et al., 1998; Poynard et al., 1998). The HCV genotype can be determined by nucleotide sequencing of the entire genome followed by composition of a phylogenetic tree, which is presently the ‘gold standard’ for the detection and identification of the various HCV genotypes and subtypes (Lau et al., 1995). This approach, however, is cumbersome and regarded as impractical for routine clinical laboratory settings (Ross et al., 2000). More convenient methods focus on the amplification of defined regions of the HCV genome by reverse transcription (RT)-PCR followed by nucleotide sequencing, RT-PCR with genotype-specific primers, RT-PCR followed by restriction fragment length polymorphism, and RT-PCR with universal primers followed by subtype-specific hybridization (Simmonds et al., 1993; Widell et al., 1994; Davidson et al., 1995; Stuyver et al., 1996). The TruGene HCV 5?NC Genotyping Kit (Visible Genetics, Toronto, Ontario) has been introduced recently. This assay is based on a nucleotide sequencing reaction of RT-PCR products from the 5?UTR. The RT-PCR products can conveniently be retrieved from the Cobas AmplicorTM HCV Test (Roche Molecular Systems, Pleasanton, CA) or the Cobas AmplicorTM HCV Monitor Test (Roche). Those RT-PCR products are also used in genotyping assays that are based on subtype-specific hybridization, the Inno-LiPA HCV II (Innogenetics N.V., Gent, Belgium) assay and the ViennaLab HCV Strip Assay (ViennaLab Labordiagnostika GmbH, Vienna, Austria). The latter two assays have especially been designed for routine laboratory settings. The aim of the present study was to compare three different HCV genotyping/subtyping assays with regard to accuracy and specificity. Furthermore, performance of these assays in a diagnostic routine laboratory was evaluated.
278
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
2. Materials and methods 2.1. Samples A total of 100 consecutive HCV RNA positive samples derived from patients with chronic hepatitis C, analyzed at the Molecular Diagnostics Laboratory, Institute of Hygiene, Karl-FranzensUniversity Graz, were studied. All samples were negative for HIV-1 antibodies and HBsAg. Amplification products were generated by RTPCR with the qualitative Cobas AmplicorTM HCV Test (Roche Molecular Systems). Aliquots of the amplification products were stored at /70 8C until further examination. Amplification products were analyzed with all three genotyping assays. 2.2. HCV genotyping/subtyping For the TruGeneTM HCV 5?NC Genotyping Kit (Visible Genetics), amplification products were neutralized and purified with the High Pure PCR Product Purification Kit (Roche Molecular Biochemicals, Mannheim, Germany). The sequencing procedure was performed according to the manufacturer?s instructions employing the CLIPTM sequencing technique (Visible Genetics). The CLIP technique allows both directions of the amplification product to be sequenced simultaneously within the same tube using two different dye-labeled primers for each of the four sequencing reactions. Electrophoresis and data analysis were carried out with the automated OpenGeneTM DNA sequencing system (Visible Genetics) that were combined with the automated Long-Read TowerTM for the separation of the sequenced fragments. Data were acquired with the GeneLibrarianTM module of the GeneObjectsTM (Visible Genetics) software. Forward and reverse sequence data were combined and compared with a HCV sequence library. Thereafter, the sequences of the HCV 5?UTR were subjected to a phylogenetic analysis using the PHYLIP software package, version 3.5c (Forns and Bukh, 1998). Distances between pairs of sequences were estimated with the DNA-DIST program. Phylogenetic trees were constructed by the unweighted pair group method using arithmetic
averages on the previous sets of pair-wise distance according to Ross et al. (2000). For the correct identification of a certain HCV subtype, the sequence has to agree for more than 90% with a corresponding sequence of the HCV sequence library according to the guideline for the interpretation of subtypes in the ISO 9001 certified molecular diagnostics laboratory. The Inno-LiPA HCV II assay (Innogenetics N.V.) and the ViennaLab HCV Strip Assay (ViennaLab Labordiagnostika GmbH) were performed according to the manufacturer’s instructions. These assays are based on the reversehybridization principle of RT-PCR products derived from the 5?UTR of the HCV genome. Briefly, biotinylated RT-PCR products are hybridized to a selection of specific oligonucleotide probes immobilized on membrane strips. The labeled RT-PCR products will only hybridize to a probe (or line) matching the sequence of the isolate, allowing stringent discrimination at the subtype level. The high specificity is obtained by using very stringent hybridization conditions. The presence of probe/amplicon-biotin hybrids is revealed by an alkaline phosphatase-labeled streptavidin conjugate by converting a chromogen gaining a color reaction. The Inno-LiPA HCV II strip is based on 21 lines of probes, the ViennaLab HCV Strip Assay on 18 lines of probes. The genotype/subtype is found by comparing the pattern of positive (colored) lines of the hybridization assays with the special interpretations charts enclosed to the package insert by the companies. 2.3. Statistical analysis The TruGene assay or the Inno-LiPA HCV II were alternatively used as reference test for the evaluation of concordance and for the calculation of accuracy and specificity. The accuracy was defined as the number of correctly determined genotypes or subtypes from the TruGene assay or the Inno-LiPA HCV II or the ViennaLab HCV Strip Assay divided by the total number of genotypes/subtypes determined firstly by the TruGene assay and secondly by the Inno-LiPA HCV II as reference tests. Specificity was defined as the number of correctly identified negative results
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
obtained with the TruGene assay or the InnoLiPA HCV II or the ViennaLab HCV Strip Assay for a certain genotype/subtype divided by this number plus false positive results determined firstly by the TruGene assay and secondly by Inno-LiPA HCV II as reference tests.
3. Results 3.1. Detection of genotypes All samples genotyped by the three assays showed 100 distinct results with the TruGene assay, 101 results (one sample with a double infection showed two genotypes) with the InnoLiPA HCV II assay, and 99 distinct results (plus one non-interpretable genotype) with the ViennaLab HCV Strip Assay (Table 1). With the TruGene assay as reference test, overall accuracies of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 100 and 97%, respectively. With the Inno-LiPA HCV II as reference test, overall accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 99 and 97%, respectively. With the TruGene assay as reference test, overall specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were
279
100 and, 99% respectively. With the Inno-LiPA HCV II assay as reference test, the overall specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 100 and 99.7%, respectively. Accuracy and specificity results for genotype 1 were as follows: With the TruGene assay as reference test, accuracies of the Inno-LiPA HCV II and the ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. With the Inno-LiPA HCV II as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II and the ViennaLab HCV Strip Assay were 100% for both assays. This result was also obtained with the Inno-LiPA HCV II assay as reference test. Accuracy and specificity results for genotype 3 were as follows: With the TruGene assay as reference test, accuracies of the Inno-LiPA HCV II and the ViennaLab HCV Strip Assay were 100 and 92.9%, respectively. With the Inno-LiPA HCV II as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 100 and 92.9%, respectively. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II and the ViennaLab HCV Strip Assay were
Table 1 Concordance of HCV genotyping results (sample size, n/100) (a) TruGene assay as reference test. (b) Inno-LiPA HCV II assay as reference test. (a) Genotype 1 2 3 4 Total
TruGene assay
(b) Genotype 1 2 3 4 Total
Inno-LiPA HCV II assay
a b c
83 2 14 1 100
83b 3b 14 1 101b
Inno-LiPA HCV II assay 1 2 3 4 NIa 83 2 14 1 100
ViennaLab HCV Strip assay 1 2 3 4 NIa c 32 1 2 13 1c 1c 97
TruGene assay 1 2 3 83 1c 2 14
ViennaLab HCV Strip assay 1 2 3 4 NIa 82 1c 3 13 1c c 1 98
4
1 100
NIa
NI, non-interpretable genotype. With the Inno-LiPA HCV II assay one sample was identified with double HCV infection of two genotypes (1b, 2a/2c, 2b). Inconsistent genotype assignments.
280
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
100 and 98.8%, respectively. With the Inno-LiPA HCV II as reference test, specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. The TruGene assay detected genotype 2 in two samples; this was in concordance with both, the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay (Table 1). With Inno-LiPA HCV II assay as reference test, an additional sample was identified with genotype 2; this was in concordance with the ViennaLab HCV Strip Assay, but not with the TruGene assay. Both, the TruGene assay and the Inno-LiPA HCV II assay found one sample with genotype 4; this sample, however, was typed as genotype 3 by the ViennaLab HCV Strip Assay (Table 1). 3.2. Detection of subtypes All samples subtyped by the three assays showed 89 distinct results with the TruGene assay (11 samples could not be further subtyped), 99 results with the Inno-LiPA HCV II assay (one sample could only be genotyped as genotype 1; for evaluating concordance the sample with genotypes/subtypes 1b, 2a/2c, and 2b was counted as ‘own subtype’), and 98 distinct results with the ViennaLab HCV Strip Assay (one non-interpretable genotype and one sample could only be genotyped as genotype 1) (Table 2). With the TruGene assay as reference test, overall accuracies of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 95.5 and 92%, respectively (Table 2). With the Inno-LiPA HCV II assay as reference test, overall accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 85.9 and 87.9%, respectively. With the TruGene assay as reference test, overall specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 97.8 and 97.7%, respectively. With the Inno-LiPA HCV II assay as reference test, the overall specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 99.4 and 98%, respectively. Eleven samples could not be further subtyped with the TruGene assay because of too less agreement with the HCV sequence library (Table 2). When 10 samples with genotype 1, which could
not be further subtyped with the TruGene assay, were determined with the Inno-LiPA HCV II assay, 4 samples were found to be subtype 1b, one sample 1a, 2 samples 1a/1b, one sample 1a/ 1b, one sample 1, and one sample 1b, 2a/2c, 2b. With the ViennaLab HCV Strip Assay, 5 samples were found to be subtype 1b, 3 samples 1a, one sample 1a/1b, and one sample 2b. The one sample with genotype 4, which could not be further subtyped with the TruGene assay, was found to be subtype 4c/4d with the Inno-LiPA HCV II assay. With the ViennaLab HCV Strip Assay, this sample was found to be subtype 3a. Fifteen samples could be identified as subtype 1a with all three assays. With the TruGene assay as reference test, accuracies of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 83.3 and 88.9%, respectively (Table 2). With the Inno-LiPA HCV II assay as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 93.8% each. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 98.8 and 96.3%, respectively. With the Inno-LiPA HCV II assay as reference test, the specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 96.4 and 95.2%, respectively. Forty-five samples could be identified as subtype 1b with all three assays. With the TruGene assay as reference test, accuracies of the InnoLiPA HCV II assay and the ViennaLab HCV Strip Assay were 100 and 93.8%, respectively (Table 2). With the Inno-LiPA HCV II assay as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 87.3 and 92.6%, respectively. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 88.5% each. With the Inno-LiPA HCV II assay as reference test, the specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 100 and 97.8%, respectively. With the TruGene assay, 7 samples were found to be subtype/s 1a or/and 1b with 6 of them in concordance with both of the reverse hybridization assays.
(a) Subtype
TruGene assay
1a 1b 1a or/and 1b 2a or/and 2c 3a Total
18 48 7 2 14 89b
(b) Subtype
Inno-LiPA HCV II assay
1a 1b 1a/1b 1a/1b 2a/2c 3a 4c/4d 1b, 2a/2c, 2b Total
16 53 7 5 2 14 1 1 99c
a b c d
Inno-LiPA HCV II assay 1a 1b 1a/1b 15 3d 48 1d 6
2a/2c
3a
1
ViennaLab HCV Strip assay 4 1a 1b 1a/1b 1a/1b 16 2d 45 2d d 1 6
2
2a/2c
2b 3a NIa
1d 2
14
13
85 TruGene assay 1a 1b 1a or/and 1b 15 48 1d 6 3d
1
1d
82 2a or/and 2c
ViennaLab HCV Strip assay 3a 1b 4b 1a 1b 1a/1b 1a/1b 1d 15 1d 4d 50 2d d 1 5 2d 2d 2
2
2a/2c
2b 3a NIa
1d
2 14
13 1d
1d 1d
1d 85
NI /non-interpretable genotype/subtype. With the TruGene assay 10 samples genotyped as 1 and one sample genotyped as 4 could not be further subtyped. With the Inno-LiPA HCV II assay one sample could only be genotyped as 1 and not further subtyped. Inconsistent or incomplete subtype assignments.
87
1
1d
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
Table 2 Concordance of HCV subtyping results (sample size, n /100) (a) TruGene assay as reference test. (b) Inno-LiPA HCV II assay as reference test
281
282
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
All of the assays identified 2 samples with subtype/s 2a or/and 2c. The Inno-LiPA HCV II assay detected one sample with a double infection. In this sample genotypes 1 and 2 (1b, 2a/2c/2b) were found. Thirteen samples could be identified as subtype 3a with all of the three assays. With the TruGene assay and the Inno-LiPA HCV II assay, one additional sample was found to be subtype 3a, which was not interpretable with the ViennaLab HCV Strip Assay. With the TruGene assay as reference test, accuracies of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 100 and 92.9%, respectively (Table 2). With the Inno-LiPA HCV II assay as reference test, accuracies of the TruGene assay and ViennaLab HCV Strip Assay were 100 and 92.9%, respectively. With the TruGene assay as reference test, specificities of the Inno-LiPA HCV II assay and the ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. With the Inno-LiPA HCV II assay as reference test, the specificities of the TruGene assay and the ViennaLab HCV Strip Assay were 100 and 98.8%, respectively. All of the evaluated genotyping assays proved to be suitable for the routine diagnostic laboratory. For the TrueGene assay, the whole procedure took 4 h including sequencing reaction, electrophoresis, and interpretation of data, which took up to 15 min per sample. When the Inno-LiPA HCV II assay was used the whole procedure could be done within 3 h. The ViennaLab HCV Strip Assay took 2 h because of shortened incubation periods and ready-to-use reagents. The hands-on time required for both of the hybridization assays was similar including that for interpretation (up to 1 min per sample). With the TruGene assay less hands-on work had to be done in comparison with the other assays.
4. Discussion In this study three different assays for HCV genotyping/subtyping based on amplification products of the 5?UTR were compared with regard to accuracy and specificity. A total of 100 consecutive clinical samples were investigated. The distribution
of HCV genotypes in this study was found to be comparable to that in a recent study on the HCV genotype distribution in Vienna and surrounding areas (Haushofer et al., 2001). This suggests that the results presented in this study are obtained from a representative selection of clinical samples that reflect the challenge of the daily routine in a clinical laboratory in Austria. The employed assays are designed to identify genotypes and subtypes from the 5?UTR of the HCV genome. Because this region is highly conserved, it has been chosen as target of choice for most nucleic acid amplification-based detection assays for detection of HCV (Germer et al., 1999). The 5?UTR, however, is still sufficiently variable to form type and subtype-specific motifs. These motifs can be detected and discriminated from each other by nucleotide sequencing or reverse hybridization using specific probes (Stuyver et al., 1995, 1996; Maertens et al., 1997). The TruGene assay offers the possibility of using a direct sequencing technique in a convenient fashion for the routine diagnostic laboratory. The TruGene approach with its semi-automated sequencing technique seems to fill a technical gap between the classical cumbersome sequencing methods. The Inno-LiPA HCV II, which is probably the most frequently used HCV genotyping assay, is well characterized towards high sensitivity and specificity and exhibits a 100% correlation with the 5?UTR nucleotide sequences (van Doorn et al., 1994; Mahaney et al., 1994; Giannini et al., 1994; Zeuzem et al., 1995). But some HCV subtypes, for example 2a and 2c, do not exhibit enough sequence polymorphism in the 5?UTR to be discriminated by the Inno-LiPA HCV II assay. The ViennaLab HCV Strip Assay is similar to the Inno-LiPA HCV II assay and has recently been introduced. The TruGene and the Inno-LiPA HCV II assays are well covered by the literature (Ansaldi et al., 2001; Ross et al., 2000; Halfon et al., 2001; van Doorn et al., 1994; Mahaney et al., 1994; Giannini et al., 1994; Zeuzem et al., 1995). Therefore, they were alternatively used as reference tests in this study. For all three assays, overall accuracies and specificities for the various genotypes were satisfying and were found in a good agreement with
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
previously published studies that used the TruGene assay and/or the Inno-LiPA HCV II assay (Ansaldi et al., 2001; Halfon et al., 2001; van Doorn et al., 1994; Mahaney et al., 1994; Giannini et al., 1994; Zeuzem et al., 1995). With the TruGene assay and the Inno-LiPA HCV II assay Halfon et al. (2001) recently obtained concordant HCV genotyping results in 100% of samples. In that study the overall accuracies of the two tests regardless of the genotype were 76% for the TruGene assay and 74% for the Inno-LiPA HCV II assay compared to NS5B sequence analysis. The accuracy and specificity for genotypes 1 and 3 obtained with the TruGene assay and the InnoLiPA HCV II assay were found to be in agreement with those published by Lee et al. (1997) for the Inno-LiPA HCV II assay and phylogenetic analysis of the NS5B region. For 11 samples only the genotype could be determined with the TruGene assay. The reason for this inconvenience may be related to the relatively short sequence subjected to the phylogenetic tree analysis. The overall accuracy for the discrimination of subtypes was found to be in agreement with the reported complete concordance in 91% of samples between the TruGene assay and the DNA enzyme immunoassay (Ross et al., 2000). With regard to problems in the discrimination of subtypes 1a and b Giannini et al. (1994) demonstrated a 100% correlation between 5?UTR sequences and Inno-LiPA results, but also showed that the polymorphism at position 99 of the 5?UTR did not always prove to be specifically linked to subtype 1a or b sequences detected in the core region. In fact, only one or two minor nucleotide differences distinguish unique subtypes from each other. For example only a single base exchange at position 99 (adenine to guanine) distinguishes between genotypes 1a and b in the 5?UTR sequence (Stuyver et al., 1995). At position 99 Smith et al. (1995), Stuyver et al. (1995) found several HCV isolates of subtype 1a with guanine instead of adenine, and subtype 1b with adenine instead of guanine. Only by using a different region such as NS5B a further investigation of this phenomenon could be done. It has been demonstrated that the reverse hybridization assay correctly interpreted genotypes 1, 2 and 3 in
283
comparison to sequence and phylogenetic analysis of the NS5B region; one of 14 samples with subtype 1a, however, was misinterpreted as subtype 1b and 3 of 36 samples with subtype 1b were misinterpreted as subtype 1a (Lee et al., 1997). Zeuzem et al. (1995) have shown a 100% correlation for genotyping and genotype 2 subtyping in a homogenous German patient population. However, 10% of samples with subtype 1b and 9% of samples with subtype 1a showed the inverted nucleotide sequence at position 99. In a recent study, Ansaldi et al. (2001) demonstrated concordant results between the TruGene assay and the Inno-LiPA HCV II assay in 90.9% of samples. Of 3 samples, which could only be determined as genotype 1 by Inno-LiPA HCV II assay, were found to be subtype 1b (one sample) and genotype 4 (two samples). Those samples, which had been determined as genotype 4, showed only a difference in 6 nucleotides to the closest genotype 1 isolate (Ansaldi et al., 2001). Furthermore, by sequence analysis one Inno-LiPA HCV II assay genotype 1a sample was typed as 1b and 2 genotype 1b samples were typed as 1a. In the present study one sample was identified as double infection with two genotypes by the InnoLiPA HCV II assay. This sample could only be determined as genotype 1 with the TruGene assay and was found to be subtype 2b with the ViennaLab HCV Strip Assay. Tuveri et al. (1997) demonstrated detection of double infections in hemophilic patients with the Inno-LiPA assay confirmed by sequencing of the core region. Furthermore, Lau et al. (1995) showed the usefulness of that hybridization test for the detection of double infection with its high sensitivity for capturing minor sequence variants. With regard to performance, the TruGene assay and the Inno-LiPA HCV II assay have shown to be equivalent. The TruGene assay, as largely automated sequencing assay for determination of HCV genotypes, proved to be useful for a highthroughput routine diagnostic laboratory with no significant differences in costs without calculation of instrument costs in comparison to the hybridization tests. The ViennaLab HCV Strip Assay offers the advantage of a markedly shortened
284
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285
hybridization time and improved handling compared to the Inno-LiPA HCV II assay. In conclusion, the TruGene and the hybridization assays were found to be reliable for the determination of all common HCV types in Europe and in North America. All assays proved to be suitable for the diagnostic routine laboratory.
Acknowledgements The authors gratefully acknowledge the expert technical assistance of B. Aigner, R. Atteneder, M. Bozic, M. Deimel, L. Gonaus, A. Huber, and G. Ho¨lzl.
References Ansaldi F, Torre F, Bruzzone BM, Picciotto A, Crovari P, Icardi GJ. Evaluation of a new hepatitis C virus sequencing assay as a routine method for genotyping. Med Virol 2001;63(1):17 /21. Bendinelli M, Vatteroni ML, Maggi F, Pistello M. Hepatitis C virus. In: Spector S, editor. Viral hepatitis: diagnosis, therapy and prevention. Totowa, NJ: Humana Press Inc, 1999:65 /127. Bukh J, Miller RH, Purcell RH. Genetic heterogeneity of hepatitis C virus quasispecies and genotypes. Semin Liver Dis 1995;15:41 /63. Chemello L, Bonetti P, Cavalletto L, Talato F, Donadon V, Casarin P, et al. Randomized trial comparing three different regimens of alpha-2a-interferon in chronic hepatitis C. Hepatology 1995a;22:700 /6. Chemello L, Cavaletto L, Bernardinello E, Guido M, Pontisso P, Alberti A. The effect of interferon alfa and ribavirin combination therapy in naive patients with chronic hepatitis C. J Hepatol 1995b;23(Suppl 2):8 /12. Choo Q-L, Richman KH, Han JH, Berger K, Lee C, Dong C, et al. Genetic organization and diversity of the hepatitis C virus. Proc Natl Acad Sci USA 1991;88:2451 /5. Davidson F, Simmonds P, Ferguson JC, Jarvis LM, Dow BC, Follet EA, et al. Survey of major genotypes and subtypes of hepatitis C virus using RFLP of sequencing amplified from the 5? non-coding region. J Gen Virol 1995;76:1197 /204. Davis GL. Hepatitis C. In: Schiff ER, Sorell MF, Maddrey WC, editors. Schiff’s disease of the liver, 8th ed.. Philadelphia, USA: Lippincott-Raven, 1999:793 /836. EASL Consensus Panel. EASL International Consensus Conference on Hepatitis C. Paris, 26 /28 February 1999. J Hepatol 1999;30:956 /961.
Forns X, Bukh J. Methods for determing hepatitis C virus genotype. Viral Hep Rev 1998;4:1 /19. Germer JJ, Rys PN, Thorvilson JN, Persing DH. Determination of hepatitis C virus genotype by direct sequence analysis of products generated with the Amplicor HCV test. J Clin Microbiol 1999;37:2625 /30. Giannini C, Thiers V, Nousbaum JB, Stuyver L, Maertens G, Bre´chot C. Comparative analysis of the hepatitis C virus core PCR and LiPA genotyping assay. J Hepatol 1994;23:246 /53. Halfon P, Trimoulet P, Bourliere M, Khiri H, de Ledinghen V, Couzigou P, et al. Hepatitis C virus genotyping based on 5? noncoding sequence analysis (Trugene). J Clin Microbiol 2001;39:1771 /3. Haushofer AC, Kopty C, Hauer R, Brunner H, Halbmayer WM. HCV genotypes and age distribution in patients of Vienna and surrounding areas. J Clin Virol 2001;20:41 /7. Lau JY, Mizokami M, Kolberg JA, Davis GL, Prescott LE, Ohno T, et al. Application of six hepatitis C virus subtyping systems to sera of patients with chronic hepatitis C in the United States. J Infect Dis 1995;171:281 /9. Lee JH, Roth WK, Zeuzem S. Evaluation and comparison of different hepatitis C virus genotyping and serotyping assays. J Hepatol 1997;26:1001 /9. Leon RV, de Medina MD, Schiff ER. Hepatitis C virology antigen, antibody, and molecular testing. In: Gordon SC, editor. Management of chronic viral hepatitis. New York, USA: Marcel Dekker, Inc, 2002:33 /64. Maertens G, Stuyver L. Genotypes and genetic variation of hepatitis C virus. In: Harrison TJ, Zuckerman AJ, editors. The molecular medicine of viral hepatitis. Chichester, UK: Wiley, 1997:183 /233. Mahaney K, Tedesci V, Maertens G, Di Bisceglie A, Vergalla J, Hoofnagle JH, et al. Genotype analysis of hepatitis C virus in American patients. Hepatology 1994;20:1405 /11. McHutchison JG, Gordon SC, Schiff ER, Shiffman ML, Lee WM, Rustgi VK, et al. Interferon alfa-2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C. Hepatitis interventional therapy group. N Engl J Med 1998;339:1485 /92. Nousbaum J-B, Pol S, Nalpas B, Landais P, Berthelot P, Bre´chot C. Hepatitis C virus type 1b (II) infection in France and Italy. Ann Int Med 1995;122:161 /8. Okamoto H, Kurai K, Okada S, Yamamoto K, Lizuka H, Tanaka T, et al. Full-length of a hepatitis C virus genome having poor homology to reported isolates: comparative study of four distinct genotypes. Virology 1992;188:331 /41. Poynard T, Marcellin P, Lee SS, Niederau C, Minuk GS, Ideo G, et al. Randomised trial of interferon alpha2b plus ribavirin for 48 weeks or for 24 weeks versus interferon alpha2b plus placebo for 48 weeks for treatment of chronic infection with hepatitis C virus. International Hepatitis Interventional Therapy Group (IHIT). Lancet 1998;35:1426 /32. Poynard T, McHutchison J, Goodman Z, Ling MH, Albrecht J. Is an ‘a la carte’ combination interferon alpha-2b plus ribavirin regimen possible for the first line treatment in
A.C. Haushofer et al. / Journal of Clinical Virology 27 (2003) 276 /285 patients with chronic hepatitis C?. Hepatology 1995;31:211 /8. Ross RS, Viazov SO, Holtzer CD, Beyou A, Monnet A, Mazure C, et al. Genotyping of hepatitis C virus isolates using CLIP sequencing. J Clin Microbiol 2000;38:3581 /4. Simmonds P, Holmes EC, Cha TA, Chan SW, McOmish F, Irvine B, et al. Classification of hepatitis C virus into six major genotypes and a series of subtypes by phylogenetic analysis of the NS-5 region. J Gen Virol 1993;74:2391 /9. Simmonds P, Alberti A, Alter F, Bonino F, Bradley DW, Brechot C, et al. A proposed system for the nomenclature of hepatitis C viral genotypes. Hepatology 1994;19:1321 /4. Smith DB, Mellor J, Jarvis LM, Davidson F, Kolberg J, Urdea M, et al. Variation of the hepatitis C virus 5?non-coding region: implications for secondary structure, virus detection and typing. J Gen Virol 1995;76:1749 /61. Stuyver L, Wyseur A, Van Arnhem W, Lunel F, Laurent-Puig P, Pawlotsky JM, et al. Hepatitis C virus genotyping. By means of 5?-UR/core line probe assay and molecular analysis of untypeable samples. Virus Res 1995;38:137 /57.
285
Stuyver L, Wyseur A, Van Arnhem W, Hernandez F, Maertens G. Second-generation line probe assay for hepatitis C virus genotyping. J Clin Microbiol 1996;34:2259 /66. Tuveri R, Rotschild C, Pol S, Reijasse D, Persico T, Gazengel C, et al. Hepatitis C virus genotypes in French haemophiliacs: kinetics and reappraisal of mixed infections. J Med Virol 1997;51:36 /41. van Doorn LJ, Kleter B, Stuyver L, Maertens G, Brouwer J, Schalm S, et al. Analysis of hepatitis C virus genotypes by a line probe assay (LiPA) and correlation with antibody profiles. J Hepatol 1994;12:122 /9. Widell A, Shev S, Mansson S, Zhang YY, Foberg U, Norkrans G, et al. Genotyping of hepatitis C virus isolates by a modified polymerase chain reaction assay using type specific primers: epidemiological applications. J Med Virol 1994;44:272 /9. Zeuzem S, Ru¨ster B, Lee JH, Stripf T, Roth WK. Evaluation of a reverse hybridisation assay for genotyping of hepatitis C virus. J Hepatol 1995;23:654 /61.