A complete molecular biology assay for hepatitis C virus detection , quantifi cation and genotyping

Introduction: Molecular biology procedures to detect, genotype and quantify hepatitis C virus (HCV) RNA in clinical samples have been extensively described. Routine commercial methods for each specifi c purpose (detection, quantifi cation and genotyping) are also available, all of which are typically based on polymerase chain reaction (PCR) targeting the HCV 5’ untranslated region (5’UTR). This study was performed to develop and validate a complete serial laboratory assay that combines real-time nested reverse transcription-polymerase chain reaction (RT-PCR) and restriction fragment length polymorphism (RFLP) techniques for the complete molecular analysis of HCV (detection, genotyping and viral load) in clinical samples. Methods: Published HCV sequences were compared to select specifi c primers, probe and restriction enzyme sites. An original real-time nested RT-PCRRFLP assay was then developed and validated to detect, genotype and quantify HCV in plasma samples. Results: The real-time nested RT-PCR data were linear and reproducible for HCV analysis in clinical samples. High correlations (> 0.97) were observed between samples with different viral loads and the corresponding read cycle (Ct Cycle threshold), and this part of the assay had a wide dynamic range of analysis. Additionally, HCV genotypes 1, 2 and 3 were successfully distinguished using the RFLP method. Conclusions: A complete serial molecular assay was developed and validated for HCV detection, quantifi cation and genotyping.

Hepatitis C virus (HCV) is recognized as one of the main causes of chronic liver disease worldwide 1 .Hepatitis C virus infection is usually asymptomatic during the acute phase, but more than 80% of patients progress to chronic hepatitis C (CHC).Approximately 130-170 million people worldwide are chronically infected with HCV, and these people are at risk of developing hepatic complications, such as cirrhosis and/or hepatocarcinoma 2 .
Hepatitis C virus is an enveloped, single-stranded, positivesense ribonucleic acid (RNA) virus with a 50nm diameter viral particle and is a member of the Hepacivirus genus of the Flaviviridae family 3 .Its RNA genome encodes a unique polyprotein of approximately 3,000 amino acids 4,5 .The HCV genome is extremely heterogeneous.Published sequence data indicate that the 5' untranslated region (5´UTR) is highly conserved among various HCV isolates and is the target of most HCV molecular biology assays.This region, however, also contains genotypically variable sequences that allow the virus to be classifi ed into six classical genotypes, which were recently updated to seven main genotypes that differ by more than 30% in their nucleotide sequence [6][7][8] .Hepatitis C virus genotyping is clinically relevant because improved treatment response rates have been observed with genotypes 2 and 3, compared to genotypes 1 and 4 2 .Consequently, patients infected with genotypes 1 or 4 are treated for longer times and with more potent combinations of antivirals than patients with genotypes 2 and 3 9,10 .Chronic hepatitis C is usually associated with mild symptoms.Elevation of aminotransferases, particularly alanine aminotransferase (ALT), is commonly observed, although up to one-third of asymptomatic patients have persistently normal enzymes 11 .Liver histology in CHC generally consists of infl ammatory infi ltrates and some degree of fi brosis, ranging from minimal expansion of portal tracts to cirrhosis.The fi brotic process, which is driven by liver infl ammation, can progress over time, although patients differ greatly in this process depending on several viral and host factors 12 .Cirrhosis occurs in at least 20% of patients with CHC within 20 years 13 .In contrast to other viral chronic infections, hepatitis B virus (HBV) and human immunodefi ciency virus (HIV), HCV viral load is not a pivotal factor in monitoring disease progression.However,

METHODS
it is an important prognostic factor for predicting the treatment response 14 .Additionally, monitoring viral load is essential during treatment because this information is necessary for achieving an early virological response (defi ned as negativity or a more than 2-log decrease in viral load by week 12) and allows clinicians to continuously monitor patient treatment 15 .
Several molecular biology assays have been described for hepatitis C virus-ribonucleic acid (HCV-RNA) detection and quantifi cation.Most commercial methods are polymerase chain reaction (PCR)-based and target the 5'UTR, which is the most conserved region among various genotypes/subtypes 16 .The introduction of real-time PCR-based assays improved both detection and viral load analysis 17,18 .These techniques have a broad dynamic range of quantifi cation, which is well suited to the clinical needs (upper range of quantifi cation: 7-8 log10IU/mL).Additionally, real-time PCR is more sensitive than classical PCR, with limits of detection of 10-15IU/mL.More recently, commercial real-time platforms have become available for the detection and quantifi cation of HCV-RNA: the Cobas TaqMan platform, which can be used together with automated sample preparation with the Cobas AmpliPrep_system (CAP-CTM; Roche Molecular System, Pleasanton, CA), and the Abbott platform (Abbott Diagnostic, Chicago, IL), which uses the m2000RT amplifi cation platform together with the m2000SP device for sample preparation 19 .
Molecular methods for genotyping HCV that target various HCV genomic regions have also been described 20,21 .Widely used laboratory procedures include the line probe assay (LiPA) and 5'UTR sequencing 22,23 .Restriction fragment length polymorphism (RFLP) analysis of the 5´UTR of the HCV genome was one of the fi rst assays used in large genotyping studies.In this procedure, a PCR-amplifi ed hepatitis C virusdeoxyribonucleic acid (HCV-DNA) fragment is digested into fragments with restriction enzymes that recognize cleavage sites specifi c for each genotype 24 .Genotyping based on the amplifi cation of this region has the advantage that it can be performed on PCR amplification products obtained from HCV-RNA detection tests 25 .However, this procedure is not widely used because it involves the use of 5 to 6 different restriction enzymes and is hampered by partial digestions and indeterminate results, making it laborious and time-consuming 26 .This study aimed to develop and validate a complete analytical assay based on nested reverse transcription polymerase chain reaction (RT-PCR) and simpler RFLP methodology for the detection quantifi cation and genotyping of HCV.

Sequences and comparative analysis
A total of 1,080 HCV 5´UTR nucleotide sequences (accession numbers EF558854-EF558890, EF564603-EF564609, EF571224-EF571247, AY306229-AY306686, AY309974-AY310119 and AY310921-AY311334) were retrieved using Entrez from the National Center for Biotechnology Information (NCBI, Bethesda, MD).These sequences were obtained from previous HCV genotype prevalence studies with 1,080 different HCV infected patients from different states of the 5 geographic regions from Brazil 27 .Additional 26 reference sequences from different genotypes (according to consensus proposals) were retrieved for comparative analysis 6 .
Sequences were edited and aligned with EditSeq and MegAlign (using the Clustal method) programs from the DNAstar package (LaserGene Inc., Madison, WI, USA).Primers and one probe were selected directly from the aligned sequences.The presence of the restriction sites for Hae III, Hinf I, BstN I, Rsa I, BfuC I and BstU I was determined using Mapdraw program within the DNAstar package (LaserGene Inc., Madison, WI, USA).

Reference and clinical samples
A panel of 57 HCV-positive and HCV-negative plasma samples was obtained from a Brazilian company (ControlLab, Rio de Janeiro -RJ, Brazil).Fifty anti-HCV-negative blood donor samples were obtained from a hospital blood bank (Hospital das Clínicas de Porto Alegre -HCPA).A panel of 267 HCV-RNApositive clinical samples (HCV genotypes 1, 2 and 3) was provided by a molecular diagnostic laboratory (Simbios Biotecnologia, Cachoeirinha, RS, Brazil).All of the plasma samples were collected with ethylenediaminetetraacetic acid (EDTA) in different clinical laboratories and were conserved at -20ºC until analysis.

RNA extraction
RNA was purifi ed from 100µL of plasma using a silica RNA extraction method that has been previously described 28 .Briefl y, 100μL of plasma was added to 900μL of lysis buffer in a microtube and incubated at room temperature for 10 min.Then, 20μL of the silica particle suspension was added to each micro-tube and centrifuged at 10,000rpm for 30 sec.The pellet was washed twice with washing solution, twice with 75% ethanol and once with acetone.After the last centrifugation step, supernatant was removed, and the pellet was dried at 56-60°C for 15 min.RNA was eluted with 50μl of elution buffer, and the tube was incubated at 65°C for 5 min.

RT-PCR and nested real-time PCR assay
Reverse transcription and the fi rst round of PCR amplifi cation was performed in a 20µL reaction volume using 14.5µL of a mastermix solution with a fi nal concentration of 75mM KCl, 50mM Tris-HCl, pH 8.3, 3mM MgCl 2 , 2.5mM DTT, 1mM dNTPs, 2.0µM of the primers, 24U of MMLV-RT (Life Technologies, Carlsbad, CA, USA), 4U RnaseOut (Life Technologies, Carlsbad, CA, USA), 1U Taq DNA polymerase (Cenbiot Enzimas, Porto Alegre, RS, Brazil) and 5µL of extracted RNA.Reverse transcription polymerase chain reaction amplifi cation was performed for 30 min at 37°C followed by 15 cycles of the following temperatures and times: 94°C for 30 sec, 60°C for 30 sec and 72°C for 60 sec.A nested realtime PCR (second amplifi cation stage) was performed in a 30µL reaction volume using 28.6µL of a mastermix solution with a fi nal concentration of 50mM KCl, 10mM Tris-HCl, pH 8.3, 1.5mM MgCl 2 , 1mM dNTPs, 0.25µM of the primers, and 0.125µM of the probe.The PCR amplification was performed for 35 cycles of 94°C for 15 sec and 60°C for 60 sec.Positive and negative samples were added as controls in all detection steps.Quantitative analysis was conducted using standard samples with pre-defi ned viral loads.Hepatitis C virus concentrations were log 10 -transformed for analysis.The linear range was examined by plotting the data and comparing them to a line of equality.Correlation coeffi cient calculations and linear regression analyses were performed on scatter plots of log-transformed HCV RNA levels using Microsoft Excel (Microsoft Corp., Redmond, WA, USA).

5'UTR genotyping
Genotyping of all HCV-positive RNA samples was performed using an RFLP procedure adapted from the original method 24 .Briefl y, three restriction enzymes were used in two separate digests to cleave each nested PCR product.First, the HCV-amplifi ed fragment was digested with Hae III alone in the appropriate buffer and conditions (New England BioLabs, Ipswich, MA, USA).The fragment was also digested with Hinf I and BstN I simultaneously in appropriate buffer and conditions (New England BioLabs, Ipswich, MA, USA).Some samples were digested with three other enzymes: Rsa I (Promega, Fitchburg, WI, USA), BfuC I and BstU I (New England BioLabs, Ipswich, MA, USA).Single digests were prepared with the respective enzyme and the appropriate buffer and conditions: Rsa I (Promega, Fitchburg, WI, USA), BfuC I and BstU I (New England BioLabs, Ipswich, MA, USA).Digested products were separated by electrophoresis on a 12.5% polyacrylamide gel and visualized after rapid silver staining.Banding patterns of the various HCV genotypes were deduced from those previously established by analysis of the 5'UTR sequences obtained from gene databases.Genotypes were then identifi ed according to previous fi ndings 6 .

Confi rmation of RFLP analysis by sequencing
PCR products were sequenced to confirm the results.Forward and reverse sequencing reactions were performed using template DNA, inner primers and BigDye Terminator v3.1 Cycle Sequencing reagent (Applied Biosystems Inc., Norwalk, CT, USA).Sequencing was performed using the thermocycler Veriti 96 (Applied Biosystems Inc., Norwalk, CT, USA) with an initial denaturation step at 95ºC for 3 min followed by 40 cycles of 95ºC for 10 sec and 60ºC for 240 sec.Samples were purifi ed using an ethanol/EDTA/sodium acetate protocol, and DNA products were injected in the automated DNA sequencing ABI 3130 XL Genetic Analyzer (Applied Biosystems Inc., Norwalk, CT, USA).Sequence data were collected and quality analysis was performed using the Sequencing Analysis v.5.3.1 software by evaluating the main technical parameters as raw data, electropherograms and quality values of sequenced bases (Applied Biosystems Inc., Norwalk, CT, USA).The nucleotide sequences from the same amplicon (performed with sense and antisense primers) were edited and assembled using SeqMan software (DNAStar, Madison, WI, USA).Nucleotide sequences were aligned using the MegAlign program (DNAStar, Madison, WI, USA), and genotypes were deduced using the phylogenetic analysis protocols in this software.Additionally, data obtained were compared with sequences available in GenBank database (www.ncbi.nlm.nih.gov).

Ethical considerations
This project was approved by the Ethical Committee of the Universidade Luterana do Brasil (Canoas, RS, Brazil).

Real-time nested RT-PCR primers and probe
All 5´UTR HCV sequences obtained from GenBank were aligned using PrimerSelect software (Lasergene Inc., Madison, WI, USA).Four primers and one probe were designed to accommodate all HCV types (Table 1).These oligonucleotides were used to develop and standardize a molecular method to detect, quantify and genotype HCV.The complete procedure consists of reverse transcription and a nested PCR.Reverse transcription and the fi rst amplifi cation were performed in the same reaction tube.The second amplifi cation was performed using TaqMan targeting all HCV genotypes.

RFLP patterns and HCV genotyping
A total of 1,080 NCBI sequences were evaluated for genotypespecifi c RFLP patterns.The differentiation of the three main genotypes (1, 2, 3) present in Brazil could theoretically be accomplished using a double digest with restriction enzymes BstN I and Hinf I.The banding patterns of Genotype 3 samples consist of a 165bp fragment, while almost all genotype 1 samples (98.1%) contain a 117bp fragment and the majority of the genotype 2 samples (94.5%) have a 221bp fragment (Table 2).To further differentiate HCV genotypes 4 and 5 (which present the same patterns as genotypes 3 and 2, respectively) and to differentiate a signifi cant proportion (5.5%) of the genotype 2 samples that have a 117bp-fragment, the restriction enzyme Hae III was selected because it was informative for all of these discriminations.Table 1 shows the predicted patterns using BstN I and Hinf I double digestion and Hae III single digestion for the HCV Brazilian sequences.Based on these results, the following workfl ow for genotyping HCV was defi ned: BstN I + Hinf I double digestion and Hae III single digestion in a fi rst round for all samples.Samples with indeterminate patterns in the fi rst round would be secondarily digested with Rsa I, BfuC I and BstU I.

HCV detection and quantifi cation validation tests
First, an HCV-positive (genotype 1) standard sample at a concentration of 10,000,000 IU/mL was diluted ten-fold to 1IU/mL, and the complete procedure (RNA extraction, RT-PCR and nested real-time PCR assay) was performed on all dilutions.Real-time PCR results obtained from the dilutions were linear and reproducible.High correlations (> 0.97) were observed between samples with different viral loads, and the corresponding cycle threshold (Ct) had a wide dynamic range of analysis across the entire spectrum of clinically relevant viral loads (100 to 10,000,000IU/mL).Additionally, reproducibility was tested using four different HCV RNA genotype 1 samples (A = 6,000,000IU/mL, B = 30,000IU/mL, C = 3,000IU/mL, D = 1,500IU/mL) in 20 runs on different days.The total coeffi cients of variation of the cycle threshold (Ct) were 8.77% for sample A (mean Ct = 10.32 ± 0.90), 3.72% for sample B (mean Ct 18.36 ± 0.68), 3.96% for sample C (mean Ct = 21.72 ± 0.86) and 3.67% (mean Ct = 22.81 ±0.84).
The limit of detection was estimated using the Probit test by evaluating nine replicates of seven diluted HCV RNA samples ranging from 51 to 3,300IU/mL.The assay sensitivities were 1,500IU/mL for 95% positive repetitions and 500IU/mL for 50% positive repetitions.Specifi city was tested using hepatitis C-negative blood donors, and no false-positive results were observed in a total of 50 anti-HCV-negative samples.
The procedure was then used to analyze the 57 plasma samples obtained from an inter-laboratorial program (ControlLab).In total, 36 samples tested positive and 21 negative, with 100% of agreement with the results reported by other laboratories.In the viral load analysis of the 36 HCV-positive samples, values ranging from 1,350 (log 3.13) to 5,250,000 (log 6.72) IU/mL were obtained.These data correlated well with the values reported by the ControlLab program.Linear regression of this comparison returned an intercept of -0.09 and a slope of 0.94.The overall correlation coeffi cient was 0.91 (Figure 1).

HCV genotyping validation test
To validate the genotyping procedure, we performed a blind study comparing RFLP analysis and nucleotide sequencing for 20 random clinical samples.RFLP results with the Hinf I + BstN I   double digest and the Hae III digest revealed nine samples with the pattern predicted for genotype 1 (A), nine samples with the predicted pattern for genotype 3 (B), one sample with the primary predicted pattern for genotype 2 (C) and one sample with an unpredicted pattern (D) (Figure 2).All 19 samples from genotypes 1, 2 and 3 had exacting the same genotyping results with 5´UTR sequencing.The banding pattern of the remaining sample was not previously observed.Digests with the three other restriction enzymes (Rsa I, BfuC I and BstU I) were performed, and the banding patterns were consistent with genotype 2. The sequencing result and the subsequent phylogenetic analysis revealed similarity to genotype 2c (Figure 3).

Analysis of clinical samples
A total of 267 consecutive HCV-positive samples were tested using the nested real-time RT-PCR protocol in a routine laboratory.The mean viral load was 6,316,241 (log 6.32)IU/mL (standard deviation of log 0.80), ranging from 514 (log 2.71) to 93,850,240 (log 7.97)IU/mL.All samples were genotyped using the RFLP method and were classifi ed according to the banding pattern (Table 2).Patterns A, B, C, E, F, G, H and I were identifi ed in 155 (58.1%), 96 (36.0%), 9 (3.4%), 1 (0.4%), 1 (0.4%), 1 (0.4%), 2 (0.7%) and 2 (0.7%) samples, respectively (Figure 2).Banding pattern D was not observed in any sample.A total of four randomly selected samples representing the three main banding patterns (A, B and C) and all samples presenting patterns E, F, G, H and I were sequenced and confi rmed to be of the expected genotypes (A and I genotype 1, C, E, F and G genotype 2, B and H genotype 3).In this group of samples, 157 (58.8%) were from genotype 1, 12 (4.5%)from genotype 2 and 98 (36.7%) from genotype 3. Genotypes 4, 5 and 6 were not found.

DISCUSSION
Hepatitis C virus detection, quantifi cation and genotyping are the crucial tests required to defi ne and monitor the treatment of hepatitis C. Consequently, the introduction of assays with good analytical performance, high-throughput capacity and low cost for these three analyses will benefi t both laboratories and patients.Studies have described the development and validation of in-house and commercial real-time PCR assays for HCV detection and quantifi cation 17,18 .We developed a new procedure based on a nested PCR approach.Although this method is based on two rounds of PCR amplifi cation, the results obtained by realtime PCR were linear and reproducible, similar to other in-house or commercial real-time techniques previously described.The use of RT-PCR amplifi cation with a limited number of cycles (in a fi rst round) followed by a real-time PCR with a higher number of cycles (in a second round) allowed the precise quantifi cation of clinical samples.Further, this real-time nested RT-PCR assay demonstrated similar sensitivity and specifi city values compared to other methods 29,30 .According to the analyses of 57 samples from an inter-laboratorial program, the correlation with other techniques was also very good.This assay could be useful in detecting and quantifying HCV in plasma samples.
Several methods have been used for assessing HCV genotypes in the clinical laboratories.The gold standard technique for genotyping HCV involves sequencing one or more genes in the HCV genome (mainly the 5´UTR, core, E1, NS3 and NS5) and comparing these sequences to the established genotypes by computer analysis.This approach is considered too expensive and time-consuming for large-scale diagnostics and has not been widely used in the majority of clinical laboratories 26 .Commercial genotyping tests, such as the Invader assay, Trugene 5´NC and the INNO-LiPA HCV II test are available; however, they all require an initial PCR amplifi cation step 26 .
Reverse transcription polymerase chain reaction analysis of the 5´UTR of the HCV genome was the fi rst genotyping method used for large-scale epidemiological studies 20,24 and became the preferred method of routine HCV genotyping in clinical laboratories for some years.However, its use declined because of the diffi culty involved in performing the necessary restriction analyses.The methodology was considered laborious, time-consuming and diffi cult to evaluate (primarily when the banding patterns are visualized in ethidium bromide gels) 26 .
In the present study, we propose a simplifi ed procedure for HCV genotyping using only three restriction enzymes in two digests.The analysis of the 1,080 available HCV sequences from different Brazilian regions demonstrated that only four RFLP banding patterns (A, B, C and E), characteristic of the three main genotypes (1, 2 and 3), would be obtained in the great majority of the HCVs present in Brazil.Samples with different banding patterns could be either confi rmed using a more complete set of restriction enzymes or submitted to sequencing.This procedure could also identify genotypes 4 and 5 (patterns H and G, respectively), which are rarely found in Brazil 27 .Concordance between this RFLP procedure and direct nucleotide sequencing was observed in the analysis of all HCV-RNA positive samples.Further analysis of clinical samples demonstrated a total of eight possible banding patterns (A, B, C, D, E, F, G H and I), which are easily visualized in polyacrylamide gels.The most frequent banding patterns are characteristic of specifi c genotypes (A -genotype 1; B -Genotype 3; C, D, E and F -genotype 2).Only a few exceptions (banding patterns G, H and I presented in less than 3% of the positive samples) must be analyzed using a more informative technique (e.g., sequencing).This RFLP procedure is certainly easier and less laborious than the original method 20,24 .The banding patterns are also clear, and the identifi cation is easy to perform in a routine service (Figure 2).
Finally, the complete molecular biology assay was successfully applied for the analysis of 267 HCV positive plasma samples in a clinical laboratory setting.The whole in-house procedure could be performed completely in less 48 hours, including the determination of viral load and genotype in HCVpositive plasma samples.Moreover, this approach eliminates the need for sequencing or hybridization with specifi c probes after PCR amplifi cation, eliminating the costs of expensive automated sequencers and/or hybridization platforms.

FIGURE 2 -
FIGURE 2 -Silver-stained polyacrylamide gels showing the different banding patterns obtained in the validation of the methodology (A) and in the analysis of the 804 clinical samples (B).The numbers indicate the digestion system (1 = Hinf I + BstN I and 2 = Hae III), and the letters indicate the banding patterns (A, B, C, D, E, F, G, H and I).The 50-bp molecular weight is also shown.MW: molecular weight.

FIGURE 1 -
FIGURE 1 -The relationship between HCV RNA viral load levels given by ControlLab and obtained by nested real-time RT-PCR.HCV RNA: hepatitis C virus ribonucleic acid; RT-PCR: Reverse transcription polymerase chain reaction.

TABLE 2 -Restriction fragment sizes and frequency of HCV restriction patterns found in the samples of the different genotypes, originated from the NCBI sequences.
HCV: hepatitis C virus; NCBI: National Center for Biotechnology Information; BstN I + Hinf I and Hae III: restriction enzymes.