Hepatitis C virus (HCV) is a major etiological agent of liver disease in most regions of the world, with approximately 170 million infected individuals.1,2 Coinfection with HIV is found in approximately one third of HCV-infected persons.3 This is important because HIV coinfection increases HCV-related liver diseases.4
HCV is a single-stranded RNA virus, which encodes a polyprotein of approximately 3000 amino acids, which is cleaved into 3 structural (Core, E1, and E2) and several nonstructural (p7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B) proteins.5 HCV, like HIV, exists in each infected person as a quasispecies or “swarm” of closely related but divergent genetic sequences.6,7 These divergent sequences are generated by an error-prone polymerase (the NS5B protein), driven by the highly dynamic nature of HCV replication.8 HCV quasispecies arise by nucleotide substitutions that either preserve protein structure (synonymous substitutions) or that change the amino acid composition (nonsynonymous substitutions). Although both types of substitutions arise from error-prone viral replication, their relative proportions may be informative due to different selective forces. Synonymous changes are generally well tolerated by the virus except in regions of RNA secondary structure9 and thus are thought to be nearly neutral in the envelope gene region. In contrast, nonsynonymous substitutions may be deleterious due to effects on protein structure and function or may be advantageous when they mediate immune escape. Thus, a reduced ratio of nonsynonymous to synonymous changes (dN/dS) is an indication of decreased immunologic pressure.10,11
There is conflicting information on the effect of HIV on the course of HCV evolution. It has been shown that HCV RNA levels rise approximately 0.5 log10 in HIV-coinfected subjects, suggestive of increased viral replication.12-15 If this increased replication is due to reduced immunologic pressure, then a decrease in dN/dS would be anticipated. Therefore, we hypothesized that HIV infection would alter the host-HCV interaction resulting in lower dN/dS. To test this hypothesis, we examined HCV envelope sequences over time in persons with chronic HCV infection who were either HIV negative or HIV positive, with or without severe immunodeficiency.
MATERIALS AND METHODS
We studied subjects participating in a longitudinal study of the natural history of HCV infection among drug users who were enrolled between November 2, 1999, and December 13, 2000, in the Bronx, NY.14,16 Subjects underwent semiannual standardized interviews and phlebotomy for several tests including HIV-1 viral loads, HIV antibody, T-lymphocyte subset assays, and HCV RNA assays as previously described.14,16 To assess the possible relationship of HIV infection and immunodeficiency to changes in HCV sequences over time, we chose a nested sample of 115 individuals having chronic HCV viremia who could be classified at the time of their last study phlebotomy into 1 of the following 3 groups: group 1 consisted of HIV-negative, HCV RNA-positive subjects; group 2 consisted of subjects with a persistently high CD4+ T-cell count of >350 cells per cubic millimeter; and group 3 consisted of subjects with a persistently low CD4+ lymphocyte count of <200 cells per cubic millimeter.
HCV RNA Analyses
To characterize the HCV RNA sequence, we amplified a 1026-nt region of the HCV genome, which encodes the envelope glycoproteins E1 and E2, including hypervariable region-1 (HVR-1), as previously described.17,18 Prior work has indicated that the consensus sequence is an excellent approximation of the master sequence or most commonly observed sequence.19 Therefore, reverse transcriptase-polymerase chain reaction (RT-PCR) products were purified and directly sequenced using the reverse primer 1868a21 (positions 1868-1848 relative to HCV-H77, GenBank accession number AF009606) 5′-GAAGCAATAYACYGGRCCACA-3′.17 Sequences generated in this study are available in GenBank, with accession numbers EU925401 through EU925556.
Sequences were aligned from raw tracings using Aligner version 2.0.4 (CodonCode Corporation) and trimmed to 756 nt (positions 1041-1796 relative to HCV-H77) encompassing approximately most of E1 and one third of E2 including HVR-1, using BioEdit.20 Sites containing a gap in any sequence were stripped during analysis. Genetic distance (divergence) between paired sequences was calculated using the distance matrix calculation scripts provided with HyPhy version 0.99,21 with model selection and parameter estimation by maximum likelihood (inferred model parameters available on request from the authors). Synonymous and nonsynonymous genetic distances were determined using the MG94custom model as implemented in HyPhy.22 To determine the specific regions of the HCV genome where these changes were occurring, we used the VarPlot program (available from S.C.R. at http://sray.med.som.jhmi.edu/SCRoftware),18 which uses the Nei and Gojobori method.23 Using this software, we calculated values for dS, dN, and the dN/dS ratio in a sliding window of 20 nt.18 This process was then repeated for an overlapping segment of 20 base pair (bp), which was shifted by 1 bp (the step size), and continued across the alignment. At each step, all pairwise comparisons for a subject were performed and values averaged. The mean values for all subjects were then averaged, ensuring that each subject was given equal weight. The Jukes-Cantor correction was used to correct for underestimation of distance due to multiple substitutions at the same site.24
The χ2 test was used to analyze categorical data, and a Wilcoxon signed-rank test was used to compare HIV viral loads and CD4+ T-cell counts nonparametrically. The Kruskal-Wallis rank test was used for multiple comparisons among the 3 groups of subjects.
Initially, 115 subjects were chosen for analysis at 2 time points. Twenty-two subjects were excluded because HCV-specific RT-PCR failed at both time points, and 8 were excluded because only 1 of 2 specimens was successfully amplified. Two subjects were excluded because sequencing of RT-PCR products yielded sequence product length less than 80% of the mean length of other subjects, and 4 subjects were excluded because of possible HCV reinfection. Among the excluded subjects, 20 were from group 1, 14 from group 2, and 2 from group 3. The primers used in this study were designed to amplify genotype 1, which is the most common genotype in the United States, and as expected, a majority of the pairs that we failed to amplify were nongenotype 1. Therefore, a total of 79 subjects (74.7%) at 2 time points were evaluable (Table 1). The median time between visits was 933 days (range 360-1303). Thirty-eight subjects were HIV negative, 21 had CD4+ T-cell counts >350 cells per cubic millimeter at both time points, and 20 had CD4+ counts <200 cells per cubic millimeter at both time points. No significant differences were observed in the sex or age between the 3 groups. The subjects were primarily Hispanic, and a significant difference in race composition was detected between the 3 groups. By design, the median CD4+ T-cell count was significantly higher in group 2 compared with group 3; in addition, a significantly higher HIV viral load was observed for group 3 compared with group 2. No significant differences were detected in the HCV viral load or the estimated duration of HCV infection among the 3 groups at either visit.
Effective antiretroviral therapy (ART) has been shown to alter the complexity and diversity of HCV quasispecies. Among the 41 HIV-infected individuals in our study, 20 had reported ART. Seven of the 21 subjects in group 2 reported ART, 3 at both time points and 4 at 1 time point. Thirteen of the 20 subjects in group 3 reported ART, 2 at both time points and 11 at a single time point. Over the time course of this study, only 4 of 20 subjects on ART maintained HIV RNA levels less than 2.6 log10 copies per milliliter, which precluded an analysis of the effect of ART on HCV sequence divergence.
Overall, divergence between paired sequences in the 3 groups was similar (Table 2). In addition, no significant differences in divergence were detected between groups when the 4 excluded subjects (possible reinfection) were included in the analysis. Because a significant difference in racial composition was detected between groups, we analyzed divergence controlling for race. However, because our study only had 7 white subjects and group 2 did not contain white subjects, they were not included in this analysis. No significant differences in divergence stratified by race were detected (P > 0.50, Kruskal-Wallis test).
The median number of days between specimens was significantly different among the 3 groups (P = 0.036, Kruskal-Wallis test), with group 1 having a significantly longer interval between the specimens compared with group 3 (Dunn pairwise multiple comparison). The rate of divergence per year was also calculated and was not significantly different among the groups (Table 2). Divergence rate was also stratified by race, but no statistical differences were detected (P > 0.50, Kruskal-Wallis test).
To specifically address our hypothesis that increased immunosuppression would alter the host-HCV interaction by decreasing amino acid-changing mutations, we determined synonymous and nonsynonymous substitution frequencies. The median dS, dN substitution frequencies and rates, and dN/dS ratios are listed in Table 2. Analysis of these measures for the 3 groups showed no significant differences (Kruskal-Wallis test).
Because we found no global differences between the 3 groups in selective pressure, a high-resolution analysis was performed to determine if specific regions within the genome segment we analyzed were under greater selective pressures (Fig. 1). Using the program VarPlot, we calculated values for dS, dN, and the dN/dS ratio in a sliding window. Comparison of the 3 groups using this method revealed little difference within dS or dN (Figs. 1A, B). However, in general, synonymous distances were greater than nonsynonymous distance with median values for the 3 groups being 0.436 versus 0.081, respectively. Interestingly, synonymous changes were suppressed in the region centered on codon 362. The highest rate of nonsynonymous mutations was observed for the HVR-1 region centered on codon 398 in the E2 protein, as has been described previously.18
Based on our initial assumptions, we would expect that increasing immunosuppression as a result of HIV infection would tend to decrease immune selection (dN) and thus decrease the dN/dS ratio particularly in the group with the lowest CD4 counts. The HVR-1 dN/dS ratios in this study reached a maximum of approximately 0.9. However, the 3 groups essentially had overlapping dN/dS curves with the highest ratio centered on HVR-1.
In this study, we examined the hypothesis that immunosuppression resulting from HIV infection would modify the host-HCV interaction, resulting in decreased HCV sequence evolution over time. By measuring genetic divergence and synonymous and nonsynonymous nucleotide substitutions in HCV sequences over time, we found no significant differences in the master sequence between HCV-monoinfected control subjects and HIV/HCV-coinfected subjects with various levels of immunodeficiency as measured by CD4+ T-cell counts.
Most previous studies have primarily analyzed complexity and diversity of the quasispecies. These analyses are efficient in determining the temporal composition of the HCV population and have sensitivity to allow determination of both minor and major variants. We decided to examine evolution (divergence) by directly sequencing RT-PCR products at 2 time points. This approach gives a more global or average weight to the quasispecies but is insensitive for detection of minor variants of the population. Based on our approach, our data suggest that the shift in the quasispecies master sequence was similar among the 3 groups. However, a shift in minor variants could be occurring but would not be detectable by our approach.
In our study, we analyzed HCV sequence variability for more than a 2- to 3-year interval in essentially 2 groups, HIV-positive and HIV-negative subjects. One possible explanation for our inability to detect differences among these groups is that our assay was insensitive. However, the median dS value observed in our subjects was similar to synonymous distances found when analyzing full-length genomes in this region and the HVR-1 dN/dS ratios, which reached a maximum of approximately 0.9, were also similar to another study.25 The synonymous and nonsynonymous substitution rates that were observed in our study were also consistent with previous reports of HCV sequence evolution.26 Furthermore, the highest rate of nonsynonymous mutations in our subjects was observed for the HVR-1 region centered on codon 398 in the E2 protein as expected. In addition, synonymous changes in our subjects were suppressed in the region centered on codon 362, which has been previously described and may represent RNA secondary structure.25 Therefore, our results are consistent with previous reports and validate that our methodology was sufficient to detect earlier recognized HCV sequence variability.
In our analysis, we looked at other factors that could influence HCV sequence evolution such as race, age, duration of HCV infection, sampling interval, and HCV viral load. Keenan et al27 examining HCV-infected subjects found significantly higher frequencies of dN and dN/dS ratios in the HVR-1 region for white versus African American subjects. However, no significant differences were detected in complexity and diversity among these groups. A limitation of our study was that the group of HIV-seropositive persons with CD4+ counts >350 cells per cubic millimeter had no white subjects, and we therefore could not examine our data for differences between this group and African Americans. However, our HIV-negative group contained 6 African Americans and 5 whites, but no significant differences were detected in divergence, dS, dN, or the dN/dS ratio (data not shown). We therefore stratified our results based on race and omitted the data from white subjects in all groups based on the small sample size and found no significant differences between Hispanics and African Americans.
Previous HCV quasispecies analyses of HIV/HCV-coinfected subjects have produced conflicting results. In one report examining HCV HVR-1 diversity in HIV-negative versus HIV/HCV-positive subjects, no differences were detected until the CD4+ T-cell count decreased below 50 per microliter.28 Still others have detected either an increase29,30 or decrease in HCV sequence variability.17,31,32 The role of highly active antiretroviral therapy in this setting has also been examined with some groups finding increased HCV quasispecies diversity after immune restoration32,33 and some finding no difference.34 It is difficult to reconcile the differences in these studies because they used different study designs (cross sectional versus longitudinal), and variable duration of HCV infection existed among study subjects. In addition to the above-mentioned factors, the HCV sequence analysis methodology also differed. A majority of these studies utilized gel shift patterns to determine the diversity and complexity of the quasispecies, whereas some utilized direct sequencing of polymerase chain reaction products. The gel shift analyses are used to examine viral heterogeneity (eg, complexity and diversity among clones), which has been shown to be influenced by the factors listed above. Because we did not examine viral heterogeneity, we could have missed changes in minor quasispecies variants on which evolution could be acting. Although we agree that these changes in minor variants are likely occurring, we found in our study the major or most fit variant seems to be stable over the time period studied. Because the interaction of HIV and concomitant immunosuppression on HCV is likely a multifactorial process, small differences in experimental designs with small and diverse study populations make it difficult to precisely model this interaction.
There were limitations to our study. Antiretroviral treatment has been shown to influence complexity and diversity of the quasispecies.32,33 Because only 4 subjects in our study had effective ART, we could not assess its role in HCV sequence divergence. Similarly, we did observe higher median rates of divergence in subjects with HIV and CD4 depletion compared with those without HIV infection that were not statistically significant, suggesting that future studies using similar methods may need to be larger than the one described here to determine whether there may be small significantly different divergence rates. Another potential limitation is that we only examined approximately 600 nt in the E1 and E2 regions of the HCV genome and could have missed changes in other B- and T-cell epitopes outside this region. Data on liver histology and aminotransferase levels were not available, precluding an examination of the effect of liver disease activity on HCV sequence divergence.
A possible explanation for our results could be that amino acid substitutions in the envelope glycoproteins, especially HVR-1, have saturated. By studying HCV sequence evolution in a cohort of Irish women infected by contaminated anti-D immunoglobulin, McAllister et al35 found that amino acid substitutions become saturated over short durations of divergence. Therefore, it is plausible with the greater than 20 years of estimated HCV infection that the subjects in the current study have reached their divergence or evolutionary plateau.
1. World Health Organization. Hepatitis C: global prevalence. Wkly Epidemiol Rec
2. Alter MJ, Kruszon-Moran D, Nainan OV, et al. The prevalence of hepatitis C virus infection in the United States, 1988 through 1994. N Engl J Med
3. Sherman KE, Rouster SD, Chung RT, et al. Hepatitis C virus prevalence among patients infected with human immunodeficiency virus: a cross-sectional analysis of the US Adult AIDS Clinical Trials Group. Clin Infect Dis
4. Graham CS, Baden LR, Yu E, et al. Influence of human immunodeficiency virus infection on the course of hepatitis C virus infection: a meta-analysis. Clin Infect Dis
5. Choo QL, Richman KH, Han JH, et al. Genetic organization and diversity of the hepatitis C virus. Proc Natl Acad Sci USA
6. Martell M, Esteban JI, Quer J, et al. Hepatitis C virus (HCV) circulates as a population of different but closely related genomes: quasispecies nature of HCV genome distribution. J Virol
7. Kato N, Ootsuyama Y, Tanaka T, et al. Marked sequence diversity in the putative envelope proteins of hepatitis C viruses. Virus Res
8. Neumann AU, Lam NP, Dahari H, et al. Hepatitis C viral dynamics in vivo and the antiviral efficacy of interferon-alpha therapy. Science
9. Tuplin A, Wood J, Evans DJ, et al. Thermodynamic and phylogenetic prediction of RNA secondary structures in the coding region of hepatitis C virus. RNA
10. Evans DT, O'Connor DH, Jing P, et al. Virus-specific cytotoxic T-lymphocyte responses select for amino-acid variation in simian immunodeficiency virus Env and Nef. Nat Med
11. Ray SC, Mao Q, Lanford RE, et al. Hypervariable region 1 sequence stability during hepatitis C virus replication in chimpanzees. J Virol
12. Sherman KE, O'Brien J, Gutierrez AG, et al. Quantitative evaluation of hepatitis C virus RNA in patients with concurrent human immunodeficiency virus infections. J Clin Microbiol
13. Thomas DL, Shih JW, Alter HJ, et al. Effect of human immunodeficiency virus on hepatitis C virus infection among injecting drug users. J Infect Dis
14. Fishbein DA, Lo Y, Netski D, et al. Predictors of hepatitis C virus RNA levels in a prospective cohort study of drug users. J Acquir Immune Defic Syndr
15. Thomas DL, Astemborski J, Vlahov D, et al. Determinants of the quantity of hepatitis C virus RNA. J Infect Dis
16. Strasfeld L, Lo Y, Netski D, et al. The association of hepatitis C prevalence, activity, and genotype with HIV infection in a cohort of New York City drug users. J Acquir Immune Defic Syndr
17. Mao Q, Ray SC, Laeyendecker O, et al. Human immunodeficiency virus seroconversion and evolution of the hepatitis C virus quasispecies. J Virol
18. Ray SC, Wang YM, Laeyendecker O, et al. Acute hepatitis C virus structural gene sequences as predictors of persistent viremia: hypervariable region 1 as decoy. J Virol
19. Cabot B, Esteban JI, Martell M, et al. Structure of replicating hepatitis C virus (HCV) quasispecies in the liver may not be reflected by analysis of circulating HCV virions. J Virol
21. Pond SL, Frost SD, Muse SV. HyPhy: hypothesis testing using phylogenies. Bioinformatics
22. Muse SV, Gaut BS. A likelihood approach for comparing synonymous and nonsynonymous nucleotide substitution rates, with application to the chloroplast genome. Mol Biol Evol
23. Nei M, Gojobori T. Simple methods for estimating the numbers of synonymous and nonsynonymous nucleotide substitutions. Mol Biol Evol
24. Jukes TH, Cantor TR. Evolution of protein molecules. In: Munro HN, ed. Mammalian Protein Metabolism
. New York: Academic Press; 1969:21-132.
25. Smith DB, Simmonds P. Characteristics of nucleotide substitution in the hepatitis C virus genome: constraints on sequence change in coding regions at both ends of the genome. J Mol Evol
26. Ina Y, Mizokami M, Ohba K, et al. Reduction of synonymous substitutions in the core protein gene of hepatitis C virus. J Mol Evol
27. Keenan ED, Rouster SD, Shire NJ, et al. Complexity and diversity of hepatitis C virus RNA in African Americans and whites: analysis of the envelope-coding domain. J Infect Dis
28. Toyoda H, Fukuda Y, Koyama Y, et al. Effect of immunosuppression on composition of quasispecies population of hepatitis C virus in patients with chronic hepatitis C coinfected with human immunodeficiency virus. J Hepatol
29. Sherman KE, Andreatta C, O'Brien J, et al. Hepatitis C in human immunodeficiency virus-coinfected patients: increased variability in the hypervariable envelope coding domain. Hepatology
30. Tanaka Y, Hanada K, Hanabusa H, et al. Increasing genetic diversity of hepatitis C virus in haemophiliacs with human immunodeficiency virus coinfection. J Gen Virol
31. Roque-Afonso AM, Robain M, Simoneau D, et al. Influence of CD4 cell counts on the genetic heterogeneity of hepatitis C virus in patients coinfected with human immunodeficiency virus. J Infect Dis
32. Shuhart MC, Sullivan DG, Bekele K, et al. HIV infection and antiretroviral therapy: effect on hepatitis C virus quasispecies variability. J Infect Dis
33. Blackard JT, Yang Y, Bordoni P, et al. Hepatitis C virus (HCV) diversity in HIV-HCV-coinfected subjects initiating highly active antiretroviral therapy. J Infect Dis
34. Babik JM, Holodniy M. Impact of highly active antiretroviral therapy and immunologic status on hepatitis C virus quasispecies diversity in human immunodeficiency virus/hepatitis C virus-coinfected patients. J Virol
35. McAllister J, Casino C, Davidson F, et al. Long-term evolution of the hypervariable region of hepatitis C virus in a common-source-infected cohort. J Virol