Natural History of GBV-C/Hepatitis G Virus Infection Through the Follow-Up of GBV-C/Hepatitis G Virus–Infected Blood Donors and Recipients Studied by RNA Polymerase Chain Reaction and Anti-E2 Serology

RECENTLY, CASES OF acute or chronic hepatitis were linked to the presence of genomic sequences of a newly described RNA virus termed GBV-C/hepatitis G virus (HGV) and belonging to the Flaviridae family.1,2 The prevalence of GBV-C/HGV RNA carriers has been found to behigh in the blood donor population,2-4 and the transmission of this virus by blood products, as well as by other parenteral routes of exposure such as intravenous drug use, has been clearly established.2,5-7 However, because GBV-C/HGV has only been identified recently, little is known about the clinical significance and the pathogenicity of the infection. While the presence of GBV-C/HGV RNA has been detected in patients with signs of hepatitis, GBV-C/HGV infection appears clinically benign in the majority of immunodepressed8,9 or immunocompetent infected patients.10,11 Furthermore, the responsibility of the virus in fulminant hepatitis12-16 and in aplastic anemias17-19 is debated.

In the absence of any reliable serological assay for the diagnosis of infection, GBV-C/HGV RNA detection by polymerase chain reaction (PCR) remains the only available diagnostic tool indicating an ongoing infection. Studies performed until now were often cross-sectional or prospective with a short follow-up period. Due to the limited number of long-term sequential evaluations of GBV-C/HGV viremia, the natural history of the infection over a period of several years is not fully known. GBV-C/HGV infection might be cleared with time in certain virus carriers,20 but the frequency of a self-limited infection has not been precisely documented, and the proportion of GBV-C/HGV–infected individuals who will develop a persistent viremia is not accurately known. Recently, an assay evidencing antibodies to the envelop protein E2 of GBV-C/HGV, which is considered a marker for virus clearance,21-23 has been developed, but few data are available on the prevalence and the significance of such antibodies.

As follow-up studies of blood recipients and blood donors were very accurate in defining the natural history of hepatitis C virus (HCV) infection and its potential outcome, it seemed to us interesting to use the same approach with regards to GBV-C/HGV infection. We present here our findings of two longitudinal studies, based on the follow-up of two cohorts of GBV-C/HGV–positive patients. The first cohort included 16 individuals multiply-transfused by packed red blood cells for thalassaemia major or sickle cell disease. They were retrospectively studied by analyzing frozen serum samples, which had been sequentially collected over a 9-year period. The aim of this cohort study was to determine the outcome of GBV-C/HGV infection over a long follow-up and to elaborate the natural history of infection by identifying the patients having a known date of GBV-C/HGV infection. The second cohort included 11 blood donors found positive for GBV-C/HGV RNA through a prevalence study and prospectively followed-up since the diagnosis of the infection. The aim of this cohort study was to determine whether any of these GBV-C/HGV–infected donors presented a case of acute or persistent infection, which would permit establishing the frequency of resolved infections in the months following the diagnosis in a clinically healthy population fortuitously found positive for GBV-C/HGV RNA.

The first cohort consisted of 16 individuals diagnosed as positive for GBV-C/HGV RNA through the screening of a population of patients affected with hemoglobinopathy (thalassemia major or sickle cell disease) and multitransfused with packed red blood cells. This cohort, followed over a 9-year period (1988 to 1996), has been described elsewhere.3 All 16 patients lived in the Paris area and were prospectively followed and transfused in our center. During the follow-up period of each patient, serum samples were collected at the routine visits preceding each transfusion. The serum samples were separated from whole blood within 2 hours after venipuncture, aliquoted, and stored frozen at −80°C until tested. The main characteristics (sex, age, ethnic origin, date of beginning of transfusional period, status for infection with HCV, with hepatitis B virus [HBV], and with human immuodeficiency virus [HIV], type of hemoglobinopathy) and the duration of the follow-up period of the 16 GBV-C/HGV–positive patients are detailed in Table 1.

The second cohort included 11 individuals found positive for GBV-C/HGV RNA through a prevalence study of GBV-C/HGV infection in a population of unpaid blood donors of the Paris area. Characteristics of this donor population have been previously reported in detail.3 Briefly, the donors who met the following criteria were included in this study: no transfusion in the past; negative for markers of viruses routinely screened in blood donations (HBV, HCV, HIV, human T-lymphotropic virus of type I or II); and normal alanine aminotransferase (ALT) level. The mean age of the 11 GBV-C/HGV–positive donors was 40 years (range, 32 to 50). Eight donors were women. All donors positive for GBV-C/HGV RNA were recalled several months after the time of blood donation for collection of a new blood sample. All donors gave informed consent for the study.

Determination of GBV-C/HGV RNA in serum. 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 using 500 μL of isopropanol. cDNA synthesis was primed with random hexamers or specific primers. The resulting cDNA was subjected to a reverse-transcription coupled-PCR (RT-PCR) procedure with two primer pairs located in the NS3 and the NS5 regions of the viral genome. The NS3 primers were G9 (5′ TCYTTGATGATDGAACTGTC 3′) for RT, G8 (5′ TATGGGCATGGHATHCCYCT 3′, sense primer) and G11 (5′ TCYTTACCCCTRTAATAGGC 3′, antisense primer) for PCR, with amplification of a 140-bp fragment⁹; the NS5b primers were GV57 (5′ GGACTTCCGGATAGCTGARAAGCT 3′, sense primer) and 4512MF (5′ GCRTCCACACAGATGGCGCA 3′, antisense primer),2 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, H denotes A or C or T. PCR was performed on a Perkin-Elmer 9600 thermocycler (Perkin-Elmer Corp, Norwalk, CT) for 52 cycles (consisting of denaturation at 94°C for 30 seconds, annealing at 55°C for 30 seconds, and extension at 72°C for 30 seconds), followed by an extension cycle at 72°C for 1 minute. PCR products were separated by agarose gel electrophoresis and visualized by ethidium bromide staining.

The positivity of amplified products (obtained with at least one of both primer pairs) was assessed by the GBV-C/HGV detection kit from Boehringer (Mannheim, Germany). This kit amplifies two sequences located in two independent regions (5′ noncoding and NS5a regions), with the following primer pairs: 5′-NCR primers (5′ CGGCCAAAAGGTGGTGGATG 3′, sense primer, and 5′ CGACGAGCCTGACGTCGGG 3′, antisense primer) in the 5′ noncoding region; 77F (5′ CTCTTTGTGGTAGTAGCCGAGAGAT 3′, sense primer) and 211R (5′ CGAATGAGTCAGAGGACGGGGTAT 3′, antisense primer) in the NS5a region.2 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 (5′ biotin GGTAGCCACTATAGGTGGG 3′ for the 5′ noncoding product, and 5′ biotin GTTACTGAGAGCAGCTCAGAT 3′ for the NS5a 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 antidigoxigenin peroxidase conjugate and by use of the colorimetric substrate 2,2′-Azino-di-(3-ethylbenzthiazoline sulfonate) diamonium salt cristal (ABTS). This procedure was validated by a French multicenter quality control trial.

Each serum sample underwent the entire procedure twice. Our protocol stated that only results reproducible in these two testings (with independent nucleic acid extractions in each testing) would be taken into account.

The detection of anti-E2 antibodies was performed in serum by enzyme-linked immunosorbent assay (ELISA) (Anti-HGenv Enzymun-test; Boehringer) according to the manufacturer's instructions. The principle of the test was an ELISA/two-step sandwhich assay using streptavidin technology. Each positive result was confirmed by the confirmatory test procedure recommended by the manufacturer. Briefly, if the result of the studied sample was positive or borderline, the sample was assayed analogously to the normal test procedure, but with an incubation buffer that did not contain HGV-E2 antigen. If the sample signal of the confirmatory test was in the same absorbance range as that of the normal test procedure, the sample was deemed to be negative for anti-E2 antibodies.

Clinical parameters. In patients of both cohorts, any sign of hepatitis (jaundice, abdominal pain, fever) or any unexplained clinical symptomatology potentially linked to a primary or persistent GBV-C/HGV infection was recorded in the medical files and/or interview of the patient or the parents, if a minor. If a known date of GBV-C/HGV infection (defined as a negative then positive viral RNA PCR) was evidenced in blood recipients, a symptomatology of viral primary infection (cutaneous eruption, adenopathies, fever, asthenia) was particularly watched for in the months surrounding the contamination.

The serum ALT level was determined by an automated method and expressed as IU/L. The upper limit of normal ALT level was 40 IU/L.

In the recipient cohort, the detection of GBV-C/HGV RNA and of anti-E2 antibody was performed in all sequential serum samples collected yearly between the baseline visit (in 1988 for the majority of the patients, later if the patient was first seen after this date) and the most recent visit (censor date, December 1996 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). Additional samples other than the yearly collected samples were available in all patients who were checked by blood tests during the medical follow-up of their underlying transfusional disease. Our protocol stated that all samples available between the last yearly collected sample negative for GBV-C/HGV RNA and the first yearly collected sample positive for GBV-C/HGV RNA would be systematically tested through PCR to define at best the date of infection by GBV-C/HGV.

In the donor cohort, the detection of GBV-C/HGV RNA and of anti-E2 antibody was performed on the serum sample collected several months after the time of blood donation positive for GBV-C/HGV RNA. The anti-E2 antibody assay was also performed on the baseline sample.

The results of the GBV-C/HGV RNA PCR and of anti-E2 antibody assays in the sequential samples of the 16 multitransfused patients are detailed in Table 2. Ten patients had been GBV-C/HGV–infected before the beginning of the study period; six were GBV-C/HGV–infected during the study period and thus had a well-defined date of infection. In five of these six latter individuals, the analysis of serial samples collected at each visit allowed specifying within 1 month the date of the switch “negative PCR → positive PCR.”

At the end of the study period, 11 patients were still positive for GBV-C/HGV RNA, while five patients were negative. The 16 patients were shown to carry GBV-C/HGV RNA over a mean period of 4.4 years (range, 1 to 9), for a mean observational period (defined as the follow-up period since the first sample positive for GBV-C/HGV RNA) of 5.3 years (range, 2 to 9). Among the 16 patients, five individuals remained positive for the entire follow-up period (9 years for patient 4, 5 years for patients 11 and 12, 3 years for patients 15 and 16). Within the limits of the observational period, the six patients with a well-defined date of infection were positive for GBV-C/HGV RNA during a mean period of 4.7 years (range, 2 to 8). If defined by the presence of GBV-C/HGV RNA for at least 6 months, the persistent infection rate was 100% in this recipient cohort.

Serum anti-E2 antibody was evidenced at least once in five (31.2%) of the 16 recipients (in a sixth patient, an anti-E2–positive sample, also positive for GBV-C/HGV RNA, was not confirmed by the confirmatory test procedure). In three cases (patients 3, 7, and 13), anti-E2 antibody was detectable on all the serial samples collected immediately after the last sample positive for GBV-C/HGV RNA. In one case (patient 4), a sample, collected during the last year of the study period was simultaneously positive for both GBV-C/HGV RNA and anti-E2 antibody; the three samples collected beyond the following months were negative for GBV-C/HGV RNA and positive for anti-E2 antibody (data not shown); in the other 15 patients, no serum sample was positive for both assays. In a patient in whom a clearance of GBV-C/HGV-RNA was observed after 2 years of viremia, anti-E2 antibody was evidenced only in one sample collected during the second year of the postviremic period of 7 years (patient 6). No anti-E2 antibody was detected in the six individuals with a well-defined date of GBV-C/HGV infection.

During the observational period with a positive GBV-C/HGV RNA PCR, the 16 individuals failed to show clinical signs of acute or chronic hepatitis or a persistent symptomatology not attributable to the underlying hemoglobinopathy. Five of the six individuals with a well-defined date of GBV-C/HGV infection had been symptomless during the acute phase of infection. In the fifth individual with a well-defined date of infection (patient 2), a transient and moderate fever had been observed 1 month after the first serum sample positive for GBV-C/HGV RNA. The serum ALT level was within the normal value in all patients during the study period, except in an individual (patient 9, see Table 2) in whom this increase was noted only on the first serum sample for GBV-C/HGV RNA.

Among the three recipients HCV-infected at inclusion, two (patients 12 and 15) were positive for GBV-C/HGV RNA on the first studied sample. The third one (patient 1) was superinfected by GBV-C/HGV during the study period. This patient had also been HIV-infected through a transfusion of packed red blood cells performed 4 years before the beginning of the study period (in 1984). In 1996, he still maintained a CD4⁺ T-cell count higher than 500 per μL and belonged to stage A1 of the Center for Disease Control 1993 classification. During the first years following inclusion in the present study, he presented a total seroreversion to HCV, which has been described elsewhere.24 

Figure 1 gives the results of the detection of GBV-C/HGV RNA and of anti-E2 antibody during the prospective follow-up period of the 11 donors positive for GBV-C/HGV RNA at the time of blood donation. All were still positive for GBV-C/HGV RNA on the serum sample collected after a mean follow-up period of 7.7 months (range, 6 to 12). No donor was positive for anti-E2 antibody on the baseline and the last serum samples. If defined by the presence of GBV-C/HGV RNA for at least 6 months, the persistent infection rate was 100% in this donor cohort.

All of the individuals were symptomless during the study period comprised between both studied samples. All had a normal serum ALT level on both samples.

This study, which provides a long follow-up of GBV-C/HGV-positive individuals, shows that once acquired, the infection to GBV-C/HGV generally tends to persist in immunocompetent patients, as previously observed in immunodepressed patients.8,9 This observation is consistent with the high prevalences of viremic individuals reported in the blood donor populations,3 4 which can only be explained by a frequent chronic carrying of the GBV-C/HGV in circulating blood of infected persons. Indeed, a persistent viremia was the rule in the blood recipients and donors of our study. The longest duration of infection that we observed in our recipients was 9 years, and the real duration of the viremia in this case was probably longer because the patient was already positive for GBV-C/HGV RNA when included in the study. Moreover, all of our patients having a well-defined date of GBV-C/HGV infection presented a viremia over the whole study period, suggesting that the chronic infection to GBV-C/HGV is far more frequent than acute infection with a rapid disappearance of viral RNA.

In the majority of our patients, anti-E2 antibody became detectable after the loss of GBV-C/HGV RNA in the serum. Such a phenomenon characterizes a seroconversion and, in case of GBV-C/HGV, seems to reflect a recovery from infection.21,22 However, it is not established if this recovery is due to the anti-E2 antibody itself or if a positive anti-E2 antibody is only a serologic marker of seroconversion: other mechanisms of immunity, such as other protective antibodies or cell-mediated immunity, could intervene in the recovery of GBV-C/HGV infection.25 Furthermore, it is probable, but not formally established, that the patients negative for GBV-C/HGV RNA and positive for anti-E2 antibody have totally eradicated the virus, with a definitive loss of infectivity. We failed to observe a reappearance of a viremia in the RNA-negative patients of our series, even in the patient with a transient anti-E2 detectability, but the possibility of such phenomenon may not be excluded, in particular, in case of immunodepression. Indeed, in patients undergoing orthotopic liver transplantation, a GBV-C/HGV viremia is detectable in a large proportion of individuals only after transplantation, which may be explained by a reactivation of the virus.

The simultaneous detection of viral RNA and of anti-E2 antibody seems rare and occuring only for a limited time interval. This overlapping was observed in only one sample of a patient in our study. In our series, once evidenced in serum, anti-E2 antibody persisted over the whole study period, except in one case, in which serum anti-E2 antibody titer was perhaps too low to be evidenced. Such a case suggests that future prevalence studies could underestimate the frequency of recovered GBV-C/HGV infection, if based on anti-E2 antibody assay.

Another question raised by the recovery from GBV-C/HGV infection, as evidenced by a positive anti-E2 antibody assay, is the cause of its occurrence after a variable number of years following the first infectious contact of the virus with the organism. The mechanisms allowing a recovery to take place after several years of persistent carrying of the virus are unknown. Research on this topic could identify elements about the mechanisms that allow hepatitis viruses to persist or to be eliminated from the organism by the host's immune system.

Generally, viruses that establish persistent infection in humans produce chronic diseases in their hosts. However, no clinical evidence of a liver disease potentially linked to the GBV-C/HGV infection was observed during the follow-up of our patients, who were shown to carry HGV RNA for up to 9 years. This benignity of long-term GBV-C/HGV infection on the liver was also shown in patients undergoing hemodialysis over several years9 and in other series of recipients.26 In our series, all of the patients having presented a recovery of GBV-C/HGV infection had not developed any disease potentially linked to GBV-C/HGV. However, such recoveries without clinical sequelae are also observed in the case of infection with HBV, but this latter virus may also be responsible for major pathologies such as cirrhosis or cancer after several decades of HBV carrying. A longer period of observation is necessary to establish whether certain GBV-C/HGV–infected patients may carry the virus over 3, 4, or more decades and whether a pathology may appear or not beyond such a long interval.

The authors are grateful to David Thorup for the preparation of the manuscript. The authors also thank Micheline Thauvin for her helpful assistance.