The prevalence of GB virus C (GBV-C) infection is high in human immunodeficiency virus (HIV)-infected persons. However, the long-term consequences of coinfection are unknown. HIV-positive persons with a well-defined duration of infection were screened on the basis of their GBV-C/hepatitis G virus (HGV) RNA status and studied. GBV-C/HGV viremia was observed in 23, who carried the virus over a mean of 7.7 years. All parameters (survival, CDC stage B/C, HIV RNA load, CD4 T cell count) showed significant differences in terms of the cumulative progression rate between persons positive and negative for GBV-C/HGV RNA. When GBV-C/HGV RNA-positive and -unexposed subjects were matched by age, sex, baseline HIV RNA load, and baseline CD4 T cell count, HIV disease progression appeared worse in GBV-C/HGV RNA-negative subjects. The carriage of GBV-C/HGV RNA is associated with a slower progression of HIV disease in coinfected persons.
GB virus C (GBV-C) and so-called hepatitis G virus (HGV) are two recently identified variants of the same virus. GBV-C/HGV was shown to belong to the Flaviviridae family but to a genus different from that of hepatitis C virus (HCV) [1, 2]. Little is known about the clinical significance and pathogenicity of GBV-C/HGV infection: Cases of acute or chronic hepatitis have been linked to the presence of genomic sequences of GBV-C/HGV, but the virus is no longer viewed as a hepatitis virus [1–6]. Indeed, GBV-C/HGV infection appears clinically benign in immunocompetent [3, 4, 7] and even immunodepressed [5, 8] persons. It has been established that GBV-C/HGV carriage may be prolonged for several years . However, the majority of carriers of viral RNA clear the virus over time, with development of antibody to envelope protein E2 (anti-E2 antibody) [10–12]. The presence of such an antibody characterizes sero-conversion and reflects recovery from GBV-C/HGV infection.
Furthermore, because GBV-C/HGV is bloodborne and sexually transmitted, the prevalence of the infection is high in human immunodeficiency virus (HIV)–infected persons [13, 14]. However, the long-term consequences of coinfection with GBV-C/HGV and HIV are unknown. For this reason, we undertook a longitudinal study to determine the impact of GBV-C/HGV infection on the natural course of HIV infection. This study was based on the follow-up of a cohort of HIV-positive subjects with a well-defined duration of HIV infection who were screened on the basis of GBV-C/HGV RNA status. Survival, clinical status (according to the CDC 1993 classification), CD4 T cell count, and plasma HIV RNA load were used as progression criteria of HIV infection.
Patients and Methods
Characteristics of the Study Population
The study population consisted of 95 HIV-1-infected subjects, who were recruited over a 7-year period (1985–1991) from a larger cohort of seropositive persons (mainly diagnosed through the systematic biologic screening of blood donations), provided they had at enrollment a documented date of seroconversion determined through an HIV-negative assay within the 6 months preceding the first HIV-positive assay. Each patient's date of seroconversion was estimated as the midpoint between the most recent HIV-negative and first HIV-positive test. The diagnosis of HIV infection had been determined in all subjects included through a specific ELISA (Sanofi-Diagnostics Pasteur, Marnes-La-Coquette, France; Abbott Laboratories, Rungis, France), with confirmation of each positive ELISA by HIV-1 Western blot (Sanofi-Diagnostics Pasteur). Among the 95 subjects, 74 (78%) were male. The mean age at HIV seroconversion was 31.1 ± 6.5 years (range, 20–46). Risk factors of HIV infection were homosexual contacts (n = 57), intravenous drug addiction (n = 4), heterosexual contacts with at-risk partners (n = 24), transfusion (n = 5), and undetermined (n = 5).
Serum GBV-C/HGV RNA was detected by in-house reverse-transcription-polymerase chain reaction (RT-PCR): GBV-C/HGV RNA was extracted from 100 μL of serum by using 900 μL of RNA Plus (Bioprobe, Montreuil, France) and 200 μL of chloroform. Nucleic acids were isolated after precipitation by use of 500 μL of isopropanol. cDNA synthesis was primed with random hexamers or specific primers. The resulting cDNA was subjected to an RT-PCR procedure with two primer pairs located in the NS3 and NS5 regions of the viral genome. The NS3 primers were G9 (5′-TCYTTGATGATDGAACTGTC-3′) for RT and G8 (5′-TATGGGCATGGHATHCCYCT-3′, sense primer) and G11 (5′-TCYTTACCCCTRTAATAGGC-3′, anti-sense primer) for PCR, with amplification of a 140-bp fragment ; theNS5b primers were GV57 (5′-GGACTTCCGGATAGCTGARAAGCT-3′, sense primer) and 4512MF (5′-GCRTCCACACAGATGGCG-CA-3′, anti-sense primer) , which amplified a 165-bp fragment. In these sequences, R denotes A or G, Y denotes C or T, D denotes A or G or T, and H denotes A or C or T. PCR was done on a thermal cycler (model 9600; Perkin-Elmer, Norwalk, CT) for 52 cycles (consisting of denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 30 s), followed by an extension cycle at 72°C for 1 min. PCR products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining. The sensitivity and specificity of this procedure have been validated through Groupe Français d'Études Moleculaires des Hepatites (GEMHEP) multicenter quality control . The positivity of amplified products was assessed by a GBV-C/HGV detection kit (Boehringer Mannheim, Mannheim, Germany) according to the manufacturer's instructions. The PCR products were labeled with digoxigenin during the amplification process. The labeled PCR products were analyzed by solution hybridization to a specific capture probe that was complementary to the inner part of the amplification product. This capture probe was biotinylated to allow immobilization of the hybrid to a streptavidin-coated microplate surface. The bound hybrid was detected by an anti-digoxi-genin-peroxidase conjugate and by use of the colorimetric substrate ABTS.
Serum anti-E2 antibody was detected by an immunoassay using recombinant E2 (Anti-HGenv Enzymun-test; Boehringer Mannheim) according to the manufacturer's instructions. Plates were incubated with the diluted blood specimen, and anti-E2 antibodies were detected by use of anti-human IgG-peroxidase conjugate and ABTS as substrate. Each positive result (value higher than the cutoff level) was confirmed by the test procedure recommended by the manufacturer. The optical density of each positive result was recorded. A confirmatory test procedure was done for all positive samples.
Virologic and immunologic markers of HIV infection
The CD4 T cell count (prospectively determined at each visit) was measured with specific monoclonal antibodies by flow cytometry with commercially available monoclonal antibodies (Becton, Pont-de-Claise, France). The total number of CD4 T cells was established by multiplying the percentage of lymphocytes that were CD4 T cells by the total lymphocyte count.
Plasma HIV-1 RNA copies were quantitated through nucleic acid sequence-based amplification (NASBA; Organon Teknika, Boxtel, The Netherlands) . Briefly, quantitation of HIV-1 RNA was based on coamplification of HIV-1 RNA with an internal standard Q-RNA dilution series, thus ensuring equal efficiency of amplification . Three distinguishable Q-RNAs (Qa, Qb, and Qc) were mixed with the wild-type sample in different amounts (i.e., 104 Qa, 103 Qb, and 102 Qc molecules) and coamplified with the wild-type RNA in one tube. By use of electrochemiluminescence-labeled oligonucleotides, the wild-type, Qa, Qb, and Qc amplificates were separately detected with a semiautomated electrochemiluminescence detection instrument (NASBA QR system), and the ratio of the signals was determined. The amount of initial wild-type RNA was calculated from the ratio of wild-type signal to Qa, Qb, and Qc signals. The assay has a quantitation limit of 400 viral copies/mL. The results were expressed as log10 of the number of viral copies per milliliter of plasma. This quantitation of plasma HIV RNA load was retrospectively done on frozen longitudinal samples routinely collected over the follow-up period and stored at −80°C within 4 h of collection until the time of testing. All specimens from each patient were tested together.
Other virologic markers
Serum HCV antibodies were detected through a third-generation ELISA (EIA 3.0 HCV, Ortho Diagnostic Systems, Roissy, France; Monolisa HCV, Sanofi Diagnostics Pasteur) according to the manufacturers' instructions, with validation of each positive result by a third-generation recombinant immunoblotting assay (Riba 3.0; Ortho Diagnostic Systems). Hepatitis B virus surface antigen was detected by ELISA (Sanofi Diagnostics Pasteur).
Observation Period of the Whole Cohort and Progression Criteria of HIV Infection
From study entry to June 1998, reference date of statistical analysis, the 95 HIV seroconverters were prospectively followed up in our outpatient clinic through regular visits (at least once a year). These visits included a physical examination for determination of the clinical status (according to the CDC 1993 criteria) , and blood specimens were obtained for laboratory evaluation. The mean follow-up period of the 95 subjects was 8.2 ± 3.2 years (range, 2.0–14.5). Fifty-seven subjects (60%) were given antiretroviral therapy at some time in their follow-up (antiretroviral therapy and consensual prophylaxis treatments had been initiated according to the consensual guidelines of the time, i.e., symptomatic HIV infection or CD4 T cell count <200/mm3). In subjects positive and negative for GBV-C/HGV RNA, 39% and 68%, respectively, were treated during their follow-up period. Immunologic progression was defined as a 50% decrease in the baseline CD4 T cell count. Virologic progression was defined as an increase of 1 log in the baseline plasma HIV RNA load.
The 95 HIV-infected subjects were retrospectively screened for GBV-C/HGV infection by analyzing frozen serum samples, which had been sequentially collected over the whole follow-up period at the routine visits. GBV-C/HGV RNA and anti-E2 antibody were detected in sequential samples collected yearly between the baseline visit (in 1985 for the majority of the patients, later if the patient was first seen after this date) and the most recent visit (censoring date: June 1998 if the patient was still alive and followed in our center at that time, earlier if the patient was dead or lost to follow-up). Two groups were distinguished on the basis of presence or absence of GBV-C/HGV RNA and compared in terms of survival, CDC stage, CD4 T cell count, and plasma HIV RNA load.
In a second step, the HIV-infected subjects positive for GBV-C/HGV RNA were matched to HIV-infected subjects recruited in the cohort of the 95 HIV-positive subjects as being unexposed to GBV-C/HGV (negative for viral RNA and for anti-E2 antibody). The matching criteria were sex, age, baseline CD4 T cell count, and baseline plasma HIV RNA load (±0.5 log). The aim of this matching was to obtain 2 groups of subjects having identical baseline predictive factors of outcome of HIV disease, to better appreciate the long-term effect of GBV-C/HGV RNA carriage.
All values were expressed as the mean ±1 SD. A χ2 test or Fisher's exact test was used for the comparison of categorical data. The Mann-Whitney nonparametric test was used for quantitative data. To compare the biologic data, the results of only three visits (first, third, and fifth) were considered for analysis. Survival rates were estimated by the Kaplan-Meier method and compared between groups by the log-rank test. All analyses were done with BMDP statistical software (1D, 3D, 4F, 3S, 1L) . Differences were considered significant at P < .05.
Results of GBV-C/HGV RNA and of anti-E2 antibody screening in the whole cohort of HIV-infected subjects
The cohort of 95 HIV-infected subjects was divided into 2 groups on the basis of GBV-C/HGV RNA status: 72 subjects were negative for GBV-C/HGV RNA during the entire follow-up period, while 23 subjects were positive for GBV-C/HGV RNA (19 were positive at their first visit; 4 acquired GBV-C/HGV during the follow-up period). At the end of the study period, 20 of the 23 subjects positive for GBV-C/HGV RNA at the time of their follow-up were still positive. The 23 subjects positive for GBV-C/HGV RNA were shown to carry the virus over a mean period of 7.7 years (range, 3.0–13.0), for a mean observation period of 8.7 years (range, 3.5–14.5).
Among the 72 subjects negative for GBV-C/HGV RNA over the follow-up period, 18 (25%) had a detectable serum anti-E2 antibody. Among these 18 subjects, 15 had been positive for anti-E2 antibody since the first screening, while 3 acquired anti-E2 antibody. In the whole cohort of the study, 41 subjects (43.2%) had been exposed to GBV-C/HGV infection (23 subjects positive for GBV-C/HGV RNA and 18 subjects positive for anti-E2 antibody).
Baseline characteristics of GBV-C/HGV RNA-positive and GBV-C/HGV RNA-negative subjects
The mean age at HIV seroconversion of the 23 GBV-C/HGV RNA-positive subjects and of the 72 GBV-C/HGV RNA-negative subjects was 29.7 ± 5.9 years (range, 20–46) and 31.5 ± 6.7 years (range, 20–34), respectively (nonsignificant difference). At the first visit, all subjects of the 2 groups were classified as CDC stage A. The mean CD4 T cell count and the mean plasma HIV RNA load of the 2 groups at the first visit (table 1) were not significantly different. The groups were also similar in terms of risk factors for HIV infection, sex, HBV infection, and HCV infection (see table 1).
Overall outcome in the GBV-C/HGV RNA-positive and-negative groups
The mean follow-up period of the subjects positive and negative for GBV-C/HGV RNA was 8.7 ± 3.3 years (range, 3.5–14.5) and 8.0 ± 3.2 years (range, 2.0–14.5), respectively (nonsignificant difference).
The mean plasma HIV RNA loads at the third and fifth years of the follow-up period are given in table 1. At these two latter visits, the mean HIV RNA load was significantly lower in the GBV-C/HGV RNA-positive group than in the GBV-C/HGV RNA-negative group (see table 1). The cumulative pro gression rate of 1-log increase in the baseline plasma HIV RNA load at the end of the follow-up period was 54.4% in subjects positive for GBV-C/HGV RNA, versus 72.8% in the other group (P = .2) (figure 1).
The mean CD4 T cell counts at the third and fifth years of the follow-up period are given in table 1. The mean CD4 T cell count was significantly higher in the GBV-C/HGV RNA-positive group than in the other group at the fifth year (table 1). The cumulative progression rate of 50% decrease in the baseline CD4 T cell count at the end of the follow-up period was 43.5% in subjects positive for GBV-C/HGV RNA, versus 83.8% in the other group (P = .05) (figure 2).
At the end of the follow-up period, 5 of the 23 subjects positive for GBV-C/HGV RNA and 41 of the 72 subjects negative for GBV-C/HGV RNA had progressed to CDC stage B/C (P < .04). The cumulative progression rate to CDC stage B/C at the end of the follow-up period was 37.8% in subjects positive for GBV-C/HGV RNA, versus 80.3% in the other group (P = .005) (figure 3).
At the end of the follow-up period, none of the 23 subjects positive for GBV-C/HGV RNA and 18 of the 72 subjects negative for GBV-C/HGV RNA had died because of HIV infection (P < .004). The cumulative survival rate at the end of the follow-up period was 100% in subjects positive for GBV-C/HGV RNA, versus 46.8% in the other group (P = .01) (figure 4).
Matching of GBV-C/HGV RNA-positive subjects to GBV-C/HGV-unexposed subjects
Twenty-three subjects positive for GBV-C/HGV RNA and 33 GBV-C/HGV-unexposed controls (recruited in the HIV-positive subjects of the cohort) could be matched for sex, age, baseline CD4 T cell count, and baseline HIV RNA load. The cumulative survival rate, the cumulative progression rate to CDC stage B/C, and the cumulative progression rate of 50% decrease in the baseline CD4 T cell count were significantly different between the 2 groups: again, the subjects positive for GBV-C/HGV RNA had a significantly higher survival rate (P = .02), a significantly lower progression rate to CDC stage B/C (P = .05), and a significantly lower progression rate of 50% decrease in the baseline CD4 T cell count (P = .02).
Comparison of GBV-C/HGV RNA-positive subjects, anti-E2-positive subjects, and GBV-C/HGV-unexposed subjects
Comparison of GBV-C/HGV RNA-positive subjects, anti-E2-positive subjects, and GBV-C/HGV-unexposed subjects showed significant differences only in terms of HIV RNA load (3.4, 4.1, and 4.0 log, respectively; P = .03) and of CD4 T cell count (537, 436, and 387 CD4 T cell/mm3, respectively; P = .008) at the third visit. On the basis of pairwise comparison, this difference was mainly due to the GBV-C/HGV RNA-positive group.
The high frequency of exposure to GBV-C/HGV infection observed in our cohort of HIV-infected subjects was in agreement with that reported in similar cohorts . In addition, the prevalence of anti-E2 antibody was less than that of GBV-C/HGV RNA, as currently observed in immunodepressed patients .
The main finding of our study was that GBV-C/HGV viremia is associated with a slower progression of HIV disease in co-infected persons. Indeed, all of the clinical and biologic parameters (survival, CDC stage B/C, plasma HIV RNA load, CD4 T cell count) showed significant differences between subjects positive for GBV-C/HGV RNA and subjects negative for GBV-C/HGV RNA after several years of follow-up. Thus, GBV-C/HGV viremia could be considered as a favorable prognostic factor of clinical and biologic progression in HIV-infected persons.
A previously published study reported a higher CD4 T cell count and a lower incidence of CDC stage B/C in HIV-infected persons positive for GBV-C/HGV RNA . However, the HIV RNA load was not available in this cross-sectional study, and subjects included did not have a well-defined duration of HIV infection. In another study, the rates of progression to AIDS and to death and the HIV RNA load were lower in HIV-infected persons positive for GBV-C/HGV RNA, but not in a statistically significant manner . Our study, based on a longitudinal determination of GBV-C/HGV status in a cohort of subjects having a well-defined duration of HIV infection, was a precise approach to determining the influence over time of the coinfection, and clearly indicated a slower HIV disease progression in persons viremic for GBV-C/HGV.
The median time from HIV seroconversion until the development of AIDS is estimated at 10 years , but seems to vary widely, from a few years in rapid progressors  to far beyond 10 years in certain long-term nonprogressors [24–26]. The parameters of such individual variations are not well-known. However, age is considered to be a factor influencing the prognosis, and the HIV RNA load, measured several months after seroconversion to HIV (when the steady-state concentration is reached), has a predictive value for the clinical outcome of HIV disease . To appreciate only the influence of GBV-C/HGV viremia, we compared HIV-infected persons positive for GBV-C/HGV RNA with GBV-C/HGV-unexposed persons after matching by age and baseline HIV RNA load (as well as by sex and CD4 T cell count). Again, HIV disease progression appeared clearly worse in GBV-C/HGV RNA-negative persons.
From this observational study, we cannot exclude the hypothesis that GBV-C/HGV infection is associated with an unknown parameter inducing a slower disease progression of HIV infection. Thus, it is difficult to establish whether the presence of GBV-C/HGV viremia is only a favorable “associated factor” of HIV disease, or whether the GBV-C/HGV could interact with and decrease HIV replication and thus improve the long-term outcome of HIV disease. The significant difference observed in terms of HIV RNA loads at the third and fifth years of follow-up between the GBV-C/HGV RNA-positive and the GBV-C/HGV RNA-negative groups may be due to either of these mechanisms. It is well-established that GBV-C/HGV maintains a high and stable level of plasma viremia for many years [8, 28, 29], which indicates efficient and active replication (in still unknown sites). Such a high level of GBV-C/HGV replication could be responsible for a lower HIV RNA load by interaction with HIV replication. In addition, it has been shown that HCV, which belongs to the Flaviviridae family, as does GBV-C/HGV, may reduce HBV replication . Furthermore, a mechanism involving the immunologic systems of persons coinfected with GBV-C/HGV and HIV may not be excluded. Further studies are necessary to determine whether screening for GBV-C/HGV viremia deserves consideration in HIV-infected persons.
We are grateful to David Thorup for manuscript preparation.