| |
Significance of liver negative-strand HCV RNA quantitation in chronic hepatitis C
|
| |
| |
Journal of Hepatology
Volume 44, Issue 2, Pages 302-309 (February 2006)
Nobukazu Yukia, Shinji Matsumotob, Kenichi Tadokorob, Kiyoshi Mochizukic, Michio Katoa, Toshikazu Yamaguchib
Background/Aims: Liver negative-strand hepatitis C virus (HCV) RNA is the most direct indicator of active viral replication but has only been examined in a few semiquantitative studies.
Methods: Positive- and negative-strand HCV RNA in the right (R) and left (L) liver lobes was quantified by rTth-based strand-specific real-time polymerase chain reaction for 48 chronic hepatitis C patients.
Results: Close correlations between lobes were seen for positive- and negative-strand amounts (r=0.950; P<0.001 and r=0.920; P<0.001, respectively). The ratio of negative to positive strands (median, 0.14 for R and 0.13 for L) varied by 2log directly in relation to HCV replication assessed by liver negative strands but had no relation to liver positive strands and circulating HCV. Only negative-strand quantitation was inversely correlated with age (r=-0.322; P=0.026 for R and r=-0.340; P=0.018 for L), while liver tissues with hepatitis B virus DNA contained larger amounts of each strand. In 27 patients treated with enhanced interferon monotherapy, the amounts of liver negative strands (<4log copies/100ng RNA) were the only independent predictor of a sustained virologic response.
Conclusions: Negative-strand quantitation is uniform in the liver and bears distinct relevance to the disease.
1. Introduction
Hepatitis C virus (HCV) replication, like that of other single-strand, positive-sense RNA viruses, is presumably preceded by the synthesis of negative-strand RNA. Thus, the amounts of negative-strand RNA-replicative intermediates in liver tissues should serve as a more reliable marker of active viral replication than positive-strand HCV RNA in the liver or in circulation. Serum HCV loads are affected by replication within the liver and extrahepatic sites and by immunologic clearance of the virus. The detection of liver positive-strand (genomic) HCV RNA can simply imply contamination by such circulating virions. Thus far, only a few semiquantitative studies have been done on the clinical relevance of liver negative-strand HCV [1-4], and controversy remains. Patients with chronic hepatitis C can show uneven distribution of liver injury, but intrahepatic variation of HCV replication also remains to be clarified. To further address these issues, we quantitatively analyzed positive- and negative-strand HCV RNA in each liver lobe by strand-specific real-time polymerase chain reaction (PCR) using rTth.
4. Discussion
Little has been known about the clinical significance of quantifying negative-strand RNA-replicative intermediates in the liver. The present study analyzed the ratio of liver negative- to positive-strand RNA. This ratio is the most reliable parameter since it does not depend on genotypes or normalization to the cellular GAPDH mRNA quantitation. For each liver lobe, the median ratio of 0.1 was similar to that found with cell-based HCV replicon systems [8]. Importantly, it was disclosed that the ratio was not constant but varied by 2log values in relation to the intrahepatic HCV-replicative status. These observations suggest that the negative-strand quantitation is not merely a reflection of liver positive strands but should serve as a distinct HCV replicative marker.
Chronic hepatitis C is known as a disease with uneven distribution of lesions in the whole liver [9]. Previous studies have shown a correlation between positive-strand HCV RNA levels of the right and left liver lobes [9,10]. The present study demonstrated a close correlation between lobes not only for positive strands but also for negative strands. Thus, HCV replication within the liver was shown to be uniform, and a single biopsy seemed generally representative of the whole liver. Although the between-lobe variation of HCV RNA loads should be interpreted with caution when the difference is small, it was only found in women, raising a possibility that sex hormone(s) and sex-linked genetic factor(s) are involved in the heterogeneity of HCV replication. In the present study, the amounts of positive- and negative-strand HCV and the ratio of negative to positive strands showed no correlation with the necroinflammatory grade and the fibrosis stage. However, we must stress the possibility that the HCV replication level, especially that assessed by negative strands, may have some relevance to histologic features such as steatosis [4].
Factors affecting HCV replication within the liver have been the subject of controversial discussions from the standpoint of the liver and circulating positive strands. Based on the negative-strand level, HCV replication in each liver lobe was shown to be inversely correlated with age. The efficiency of negative-strand RNA synthesis can be influenced by various host factors at multiple levels [11]. The data obtained raise the possibility that some age-related factor(s) may be involved in the regulation of HCV replication within the liver. The present study further showed that liver tissues with concomitant occult HBV contained larger amounts of negative- and positive-strand HCV RNA. Among HCV patients, those carrying occult HBV can manifest severer liver disease and display a poor response to IFN [12]. Occult HBV may also have relevance for hepatocarcinogenesis [13], although the mechanism remains to be clarified. Although further studies are necessary, the data obtained raise the possibility that occult HBV exerts virulence partly by enhancing HCV replication.
As for IFN-based therapy, only limited data are available on the significance of the liver negative-strand HCV RNA quantitation. In a previous semiquantitative study, the negative-strand levels were not related to the outcomes of short-term IFN-a therapy (3 MU thrice weekly for 10 weeks) [2]. Our patients were treated with 6-month enhanced IFN monotherapy [14]. A sustained virologic response was only associated with small amounts of liver negative-strand HCV RNA (<4log copies/100ng liver RNA). Based on these preliminary data, further studies are warranted in populations treated with the currently standard regimen of peginterferon and ribavirin.
In conclusion, our findings combined indicate that liver negative-strand HCV RNA quantitation offers clinically relevant information distinct from that available from positive strands within the liver and in the circulation.
3. Results
3.1. Amounts of positive- and negative-strand HCV RNA in right (R) and left (L) liver lobes
Normalized positive-strand HCV loads in the right liver lobe (median, 5.9; range, 2.5-8.5log copies/100ng liver RNA) showed a correlation with those in the left liver lobe (median, 6.0; range, negative to 6.8log copies/100ng liver RNA) (r=0.950; P<0.001) (Fig. 2A). Six (13%) of the 48 patients had a between-lobe discrepancy of 0.3-2.2log. The discrepancy was related to gender (6 [26%] of 23 women vs. none of 25 men) (odds ratio 10.9 [95% CI 1.3-90.9], P=0.027). Fig. 2B shows a correlation between normalized negative-strand HCV loads in the right lobe (median, 4.9; range, negative to 7.2log copies/100ng liver RNA) and the left lobe (median, 5.0; range, negative to 6.3log copies/100ng liver RNA) (r=0.920; P<0.001). A discrepancy of 2.0log was seen in one (2%) patient (Table 2).
In 38 patients with detectable levels of positive and negative strands in each liver lobe, the ratio of negative- to positive-strand HCV (median, 0.14; range, 0.01-0.81 for R and median, 0.13; range, 0.03-0.45 for L) increased according to negative-strand liver HCV (r=0.282; P=0.086 for R and r=0.441; P=0.006 for L) (Fig. 3). The ratio showed no correlation with positive-strand liver HCV (r=-0.192; P=0.248 for R and r=-0.097; P=0.564 for L) and circulating HCV (r=0.154; P=0.355 for R and r=0.106; P=0.527 for L). Serum HCV RNA loads ranged between 3.1 and 7.6log copies/mL (median, 6.1), and displayed a relation to the positive-strand liver HCV quantitation (r=0.604; P<0.001 for R and r=0.634; P<0.001 for L) and the negative-strand liver HCV quantitation (r=0.632; P<0.001 for R and r=0.609, P<0.001 for L).
3.2. Determinants of positive- and negative-strand HCV RNA amounts in the liver
The amounts of positive- and negative-strand HCV in each liver lobe were correlated with patient characteristics including age, gender, mode of infection, duration of infection estimated from years after blood transfusion, serum alanine aminotransferase (ALT) levels, histologic grade and stage, HCV genotypes and detection of HBV DNA in the corresponding liver lobe. An inverse correlation was found between the negative-strand liver HCV quantitation and age (r=-0.322; P=0.026 for R and r=-0.340; P=0.018 for L). The positive-strand liver HCV quantitation, however, had no relation to age (r=-0.237; P=0.104 for R and r=-0.216; P=0.140 for L) (Fig. 4). The amounts of positive- and negative-strand liver HCV did not differ between 38 patients with HCV genotype 1b and 10 patients with genotype 2 (seven with genotype 2a and three with genotype 2b), but were affected by concomitant liver HBV. By using X primers, HBV DNA was detected in both liver lobes in three patients and only in the left lobe in another patient. None of the patients tested positive for liver HBV DNA using surface primers. The four HBV DNA-positive liver tissue samples from the left lobe contained larger amounts of positive and negative strands than the 44 HBV DNA-negative tissues (6.6±0.2 vs. 5.4±1.4; P=0.007 and 5.8±0.1 vs. 3.9±2.2log copies/100ng liver RNA; P=0.006, respectively). For the right liver lobe, the positive- and negative-strand liver HCV quantitation also tended to be high in the three HBV DNA-positive liver tissues (6.5±0.2 vs. 5.6±1.2; P=0.081 and 5.7±0.2 vs. 4.0±2.3log copies/100ng liver RNA; P=0.049, respectively). None of the patient characteristics examined showed a relationship to the ratio of negative- to positive-strand HCV and serum HCV RNA load.
3.3. Histologic variation between right and left liver lobes
The total necroinflammatory grade ranged between 2 and 10 (median 7) in each liver lobe (P=0.295 by signed rank test). The fibrosis stage ranged from 1 to 6 (median 4) in the right lobe and from 2 to 6 (median 3) in the left lobe (P=0.614). Fig. 5 shows the histologic between-lobe variation among the 47 patients studied. Eleven (23%) patients showed differences of the necroinflammatory grade defined as a difference of ≥2 points, and 19 (40%) patients of the fibrosis stage defined as difference of ≥1 point. The between-lobe variation in the HCV quantitation had no impact on the histologic variation. The mean grading score of the right and left liver lobes was <7 in 10 (91%) out of the 11 patients with a grade difference compared with 16 (44%) out of the 36 patients without it (odds ratio 6.5 [95% CI 1.3-33.3], P=0.025). The difference in the fibrosis stage, however, had no relation to any of the patient characteristics examined.
3.4. Factors influencing the efficacy of IFN treatment
Eighteen (67%) out of the 27 patients were negative for serum HCV RNA at the end of treatment, and eight (30%) patients displayed sustained HCV clearance over 6 months posttreatment. The end-of-treatment virologic response was independently associated with an absence of between-lobe discrepancy of the necroinflammatory grade (odds ratio 0.2 [95% CI 0-0.9], P=0.042). However, the amounts of negative-strand HCV RNA in the liver were identified as the only independent predictor of a sustained virologic response. The mean negative-strand quantitation of the right and left liver lobes was <4log copies/100ng liver RNA in all sustained virologic responders (SVRs) compared with 1 (5%) of the 19 non-SVRs (odds ratio 85.4 [95% CI 5.4-999], P=0.002).
2. Patients and methods
2.1. Patients
Forty-eight patients with chronic hepatitis C underwent laparoscopic liver biopsies. All patients were positive for serum HCV RNA (Amplicor HCV Test, Roche Diagnostics K.K., Tokyo, Japan). No confounding etiology of liver disease was found in any patient. They were negative for hepatitis B surface antigen in the serum. The group was comprised of 25 men and 23 women ranging in age from 33 to 70 years (median, 57 years). Sixteen (33%) patients had a history of blood transfusion 8-52 years (median, 36 years) earlier. Biopsies were performed using 13-gauge Tru-Cut needles (Hakko Medical Co., Ltd, Nagano, Japan), and liver tissues sufficient for histologic and virologic evaluation were obtained from the anterior segment of the right lobe and the lateral segment of the left lobe. Specimens 15mm long and 2mm wide were embedded in paraffin for histopathological study. The remaining portions were immediately frozen and then stored at -80 C until PCR testing. With one patient, the specimen from the left lobe was subjected to only virologic evaluation due to its limited size. Paired serum samples were obtained from all patients at laparoscopy and stored at -80 C without thawing until virologic tests. Of the 48 patients, 27 (Table 1) were treated with enhanced interferon (IFN) monotherapy. After laparoscopy, 3 MU of IFN-B(Feron, Toray Co., Tokyo, Japan) was administered twice a day for 2 weeks followed by 9 MU of IFN-a (Sumiferon, Sumitomo Pharm. Co., Osaka, Japan) daily for 2 weeks and thrice weekly for 20 weeks. The study was approved by the local research ethics committee in accordance with the 1975 Declaration of Helsinki, and all patients provided written informed consent.
2.2. Virologic testing
Circulating HCV genomic RNA was quantified by a PCR assay (Amplicor HCV Monitor Test version 2.0, Roche Diagnostics K.K.). HCV RNA of ≥6.4log copies/mL was measured after serum dilution. HCV genotypes were determined by a PCR genotyping system [5].
2.3. Positive- and negative-strand HCV RNA quantitation by rTth-based strand-specific real-time reverse-transcription polymerase chain reaction (RT-PCR)
Strand-specific TaqMan RT-PCR was designed to quantify the 5' untranslated region of the HCV genome using a thermostable enzyme, rTth (Applied Biosystems, Foster City, CA). Total hepatic RNA, 100ng, was added to an RT reaction mixture containing 2μL of 10X RT buffer (Applied Biosystems), 20nmol of MnCl2, 5U of rTth, 24U of RNasin (Promega, Madison, WI), 4nmol of each dNTP, and 10pmol of sense primer HCV-20F (5'-CGACACTCCACCATGAATCACT-3') for the negative-strand assay or antisense primer HCV-114R (5'-GAGGCTGCACGACACTCATACT-3') for the positive-strand assay. The RT reaction was performed in a final volume of 20μL at 70 C for 60min. The reaction was then treated with 5μL of 10X chelating buffer (Applied Biosystems), 75nmol of MgCl2, 10nmol of each dNTP, 10pmol of antisense primer HCV-114R for the negative-strand assay or sense primer HCV-20F for the positive-strand assay, and 5pmol of TaqMan probe HCV-P43 (5'FAM-CCCTGTGAGGAACTACTGTCTTCAC-GCAGATAMRA3'). The final volume was adjusted to 50μL. The samples were promptly set in an ABI PRISM 7700 Sequence Detection System (Applied Biosystems) and incubated at 70 C for 2min and then at 94 C for 2min. Real-time PCR amplification and data analysis were subsequently performed for 45 cycles (94 C for 20s and 62 C for 1min). Copy numbers of the 95-base target sequence were determined using the standard curve based on measurements of serial 10-fold dilutions of synthetic positive- and negative-strand HCV RNA. The sensitivity was 2log copies/reaction for the positive-strand assay and 3log copies/reaction for the negative-strand assay. The dynamic ranges were 2-7log copies/reaction and 3-7log copies/reaction, respectively. In each assay, false detection of an incorrect strand occurred when the amount of incorrect strand added reached 7log copies. The positive- and negative-strand quantitation before normalization was ≦6.4 and ≦5.9log copies/100ng liver RNA, respectively, in this study. Thus, the strand-specificity was unlikely to be affected by an excess of incorrect strands. Self-priming or endogenous priming was ruled out by the lack of amplification product following RT-PCR of total hepatic RNA without primers in the RT mixture. All assays were done in duplicate, and the mean values were obtained. Hepatic RNA samples from the same liver were always measured in the same run.
The HCV-specific primers and probe used are conserved among genotypes. To verify that HCV genotypes 1b, 2a and 2b could be quantified with similar efficiency, high-concentration serum samples of each genotype were obtained from eight patients and diluted to 4.4log copies/reaction by Amplicor HCV Monitor version 2.0, which is known to equally amplify all genotypes. The positive-strand HCV quantitation by the TaqMan RT-PCR was the same for genotypes 1b (5.3±0.7), 2a (5.5±0.4) and 2b (4.9±0.5log copies/reaction) (P=0.141 by one-way analysis of variance).
2.4. Normalization of hepatic HCV RNA amounts and criteria for between-lobe discrepancies
GAPDH mRNA in total hepatic RNA, 100ng, and control total RNA (Raji cell line), 100ng, was also quantified by real-time RT-PCR, and copy numbers were determined using the standard curve (Human GAPD Endogenous Control, Applied Biosystems). Hepatic HCV RNA and GAPDH mRNA quantitation, which were performed in separate tubes, showed a linear relationship with the amounts of target RNA (Fig. 1). The HCV RNA copy number was divided by the ratio of the sample GAPDH mRNA amounts to the TaqMan control value. Thus, normalized hepatic HCV RNA amounts were obtained and used for data analysis. In preliminary experiments, assay variance for the log10 transformed HCV RNA quantitation before normalization was evaluated based on five measurements of 10 liver samples (intra-assay coefficients of variation (CVs)=0.88-2.85% and inter-assay CVs=1.19-6.91% for the positive-strand assay; intra-assay CVs=2.27-9.72% and inter-assay CVs=1.52-18.11% for the negative-strand assay). Assay variance was greater for the negative-strand assay, which may be attributable to interfering factor(s) such as a large amount of positive strands in the RT reaction. The mean SDs of intra-assay variance were 0.106 and 0.081 for <5 and ≥5log copies, respectively, in the positive-strand assay, whereas they were 0.374, 0.256 and 0.158 for <4, 4-5 and ≥5log copies, respectively, in the negative-strand assay. The HCV RNA quantitation was assumed to vary within twice these SDs. Between-lobe HCV RNA differences were considered significant when the normalized HCV RNA amounts differed by more than the estimated variance for normalized values. All discrepancies were confirmed by repeating the assays.
2.5. Detection of liver hepatitis B virus (HBV) DNA by nested PCR
Total hepatic DNA, 100ng, was subjected to nested PCR to amplify HBV DNA. The primers were set in the surface region
(outer sense 5'-
TCGTGTTACAGGCGGGGTTT-3'; outer antisense 5'-
CGAACCACTGAACAAATGGC-3'; inner sense 5'-
CAAGGTATGTTGCCCGTTTG-3'; inner antisense 5'-
GGCACTAGTAAACTGAGCCA-3') and the X region (outer sense
5'-GCATGGAGACCACCGTGAA-3'; outer antisense 5'-
CAGACCAATTTATGCCTACAG-3'; inner sense 5'-
TACATAAGAGGACTCTTGGACT-3'; inner antisense 5'-
CAGACCAATTTATGCCTACAG-3'). PCR products (233 and 151bp, respectively) were visualized by 3% agarose electrophoresis and ethidium bromide staining. All assays were done in duplicate. The sensitivity was 1copy/100ng liver DNA for each primer set. To avoid contamination in all PCR assays, the contamination avoidance measures [6] were strictly applied throughout the study, and positive and negative controls were used.
2.6. Histologic evaluation
After routine staining with hematoxylin-eosin, all liver biopsy specimens were examined by the same experienced pathologist without knowledge of their source. Biopsy specimens were semiquantitatively evaluated using the modified histologic activity index described by Ishak et al. [7].
2.7. Statistical analysis
Viral load was log10 transformed to obtain a more symmetrical distribution without outliers. An arbitrary value of 0log copy/100ng liver RNA was attributed to the liver tissues negative by PCR. Data on continuous variables were presented as mean±SD unless otherwise stated. Statistical analysis for group comparisons was performed using the Wilcoxon nonparametric test. Correlations between the variables were calculated using Spearman rank order correlations. To assess variables potentially related to virologic and histologic between-lobe discrepancies and responses to IFN, stepwise multivariate logistic regression models were used. All analyses were done with SAS (version 8.02) (SAS Institute, Inc., Cary, NC). A P value of less than 0.05 (2-tailed) was considered to indicate significance.
|
|
| |
| |
|
|
|
