GB Virus Inhibits HIV & May Spur New Drug Research
"Inhibition of HIV-1 replication by GB virus C infection through increases in RANTES, MIP-1a, MIP-1b, and SDF-1"
Lancet June 19 2004; 363: 2040--46
Jinhua Xiang, Sarah L George, Sabina Wünschmann, Qing Chang, Donna Klinzman, Jack T Stapleton
Research Service and Department of Internal Medicine, Iowa City
VA Medical Center and University of Iowa, Iowa City, IA, USA
Why is the pace of HIV disease so varied among different HIV-infected people? Recently, several studies have found that coinfection with a common, apparently non-pathogenic human virus called GB virus C (GBV-C) is associated with delayed HIV disease-progression and mortality. The mechanism by which GBV-C might influence disease has not been characterised.
This study found that co-infection of human lymphocytes with GBV-C and HIV led to inhibition of diverse HIV strains including those that use either of the two HIV coreceptors (CCR5 and CXCR4) for viral entry. The inhibition of HIV was mediated by induction of natural ligands for CCR5 and CXCR4 and by decreased expression of CCR5 on the surface of lymphocytes.
Human infection with GBV-C is common: 1--8% of healthy blood donors worldwide have GBV-C viraemia. After infection, viraemia persists in 20--50% of exposed individuals. The modes of GBV-C transmission are the same as for HIV; sexual, parenteral, and vertical transmission have all been documented.
The authors suggest that the hepatitis G virus, also known as GBV-C, could be developed as a treatment to slow the progression to AIDS. The study authors believe that the GBV-C virus itself could be used as a treatment. Patients may only need a few exposures to see a benefit. The next step, Dr. Stapleton, the author, says, is to determine which aspect of GB virus triggers these changes in the cell. "This may allow the design of novel targets for drugs against HIV. Because these effects are at the cellular level, and are not directly targeting the virus, it will be more difficult for HIV to become resistant to the effects."
Elucidation of how GBV-C causes these changes in lymphocytes might identify novel targets for anti-HIV medications and could lead to HIV disease-modifying vaccines.
Background: People coinfected with HIV and GB virus C (GBV-C) have lower mortality than HIV-positive individuals without GBV-C infection. HIV uses either of the chemokine receptors CCR5 and CXCR4 for entry into CD4-positive cells.
Longer survival in HIV-positive individuals is associated with high serum concentrations of ligands for CCR5 (RANTES [regulated on activation, normal T-cell expressed and secreted] and macrophage inflammatory proteins [MIP] 1a and 1b) and CXCR4 (stromal-derived factor [SDF-1]), and with decreased expression of CCR5 on lymphocytes.
Methods: Peripheral-blood mononuclear cells were coinfected with GBV-C and HIV, and HIV replication was monitored by measuring infectivity and HIV p24 antigen production.
Chemokine secretion was measured by ELISA, chemokinerecept or expression by flow cytometry, and cellular chemokine mRNA expression by differential hybridisation.
Findings: GBV-C infection of peripheral-blood mononuclear cells resulted in decreased replication of both clinical and laboratory HIV strains that use either CCR5 or CXCR4 as their coreceptor. Inhibition was related to the dose and timing of the GBV-C infection. Expression of mRNA for RANTES, MIP- 1a, MIP-1b, and SDF-1 and secretion of the chemokines into culture supernatants were higher in GBV-C-infected cells than in mock-infected cells.
The inhibitory effect of GBV-C on HIV replication was blocked by incubation with neutralizing antibodies against the relevant chemokines, and surface expression of CCR5 was significantly lower in GBV-C-infected cells than in mock-infected cells.
Interpretation: GBV-C induces HIV-inhibitory chemokines and reduces expression of the HIV coreceptor CCR5 in vitro. This study provides insight into the epidemiological association between GBV-C infection and longer survival in HIV-infected individuals.
This chemokine receptor found on monocytes and lymphocytes serves as the coreceptor for HIV isolates that replicate in monocytes and primary T cells, and which do not induce syncytia in cell-culture phenotypic assays. HIV strains that use this coreceptor are referred to as R5 viruses. Polymorphisms in the coding and non-coding regions of the gene influence the natural history of HIV infection.
This chemokine receptor, found on a wide variety of cells including haemopoietic glial and dendritic cells, serves as the coreceptor for HIV isolates that replicate in transformed CD4-positive T-cell lines and that induce syncytia formation in cell-culture phenotypic assays. HIV isolates that use CXCR4 are referred to as X4 viruses.
RANTES (regulated on activation, normal T-cell-expressed andsecreted), MIP-1a (macrophage inflammatory protein 1a), and MIP-1b (macrophage inflammatory protein 1b ) are the natural ligands for CCR5. Binding of these chemokines to CCR5 mediates chemotaxis and cell activation. These chemokines inhibit R5 HIV replication by competing for binding to CCR5. They also interact with other chemokine receptors that are not HIV coreceptors.
SDF-1 (stromal-derived factor) is the only known ligand for CXCR4. Binding of SDF-1 causes mobilisation of calcium, G-protein activation, generation of inositol phosphate, phosphorylation of adhesion complexes, and activation of MAP kinases and nuclear factor KB. This chemokine inhibits X4 HIV strains by competing for CXCR4 binding.
The course of HIV-1 infection is very variable. Before the availability of highly active antiretroviral therapy, the median duration of HIV infection was estimated to be 7·8 years. Several host factors have been implicated in delayed or accelerated progression of HIV disease. However, most of these host factors are rare, and they do not explain delayed progression in most HIV-positive people. In seven separate studies, HIV-infected people who also had active infection with GB virus type C (GBV-C; also known as hepatitis G virus), defined as GBV-C RNA detected in their serum, had longer survival than HIV-infected people who were not infected with GBV-C. In three additional reports, improved clinical outcomes, though not improved survival, were associated with GBV-C coinfection: however, three other studies did not find such a difference. Persistent GBV-C infection seems to be required for this beneficial effect; HIV-infected people in whom GBV-C viraemia cleared during follow-up had a higher mortality rate than people with active GBV-C viraemia.
GBV-C is a member of the Flaviviridae. On the basis of its genome organisation and nucleotide sequence, it is the closest relative of hepatitis C virus (HCV) among human viruses. Unlike HCV, GBV-C does not replicate in hepatocytes but replicates in vitro in peripheral-blood mononuclear cells, including CD4-positive T lymphocytes, CD8-positive T lymphocytes, and B lymphocytes (SLG and JTS, unpublished). GBV-C does not seem to be associated with acute or chronic hepatitis or with any other disease entity. Human infection with GBV-C is common: 1--8% of healthy blood donors worldwide have GBV-C viraemia. After infection, viraemia persists in 20--50% of exposed individuals, and clearance of GBV-C infection is in most cases associated with the development of antibodies against the viral envelope glycoprotein E2. The modes of GBV-C transmission are the same as for HIV; sexual, parenteral, and vertical transmission have all been documented.
Thus, HIV-positive people have both a high rate of active GBV-C viraemia (15--43%) and evidence of previous exposure to GBV-C (anti-E2 antibody in 31--55%). HIV-1 attachment to cells involves interaction between HIV glycoprotein 120 and the primary receptor (CD4), whereas entry into target cells requires interaction with a secondary receptor, in almost all cases the chemokine receptor CCR5 or CXCR4. HIV-1 isolates that are transmitted in vivo generally use CCR5 (R5 viruses). By contrast, HIV-1 isolates that use CXCR4 as the coreceptor (X4 viruses) commonly emerge in HIV-infected people later in the course of infection. The natural ligands for CCR5 are the CHEMOKINES RANTES, MIP-1a, and MIP-1b,18 and the ligand for CXCR4 is SDF-1.
We and others (Reil H, University of Erlangen- Nuremberg, Germany; personal communication) have shown that coinfection of human peripheral-blood mononuclear cells with GBV-C and an R5 strain of HIV-1 results in a decrease in HIV replication. This outcome was not a result of cellular toxicity, because GBV-C did not cause any toxic effects in lymphocytes.
We now report the effect of GBV-C isolates on HIV-1 strains that use CCR5 or CXCR4 as their coreceptor and explore potential mechanisms by which GBV-C might inhibit HIV-1.
Several reports have suggested that infection with one microbe can alter disease expression of other pathogens, including HIV. For example, infection with human herpesvirus type 6 inhibits R5 strains of HIV-1 in vitro by inducing RANTES, and influenza virus infection induces type I interferons, resulting in decreased HIV-1 replication in vitro. However, neither of these viruses is associated with delays in progression of HIV-1 disease in clinical studies. By contrast, GBV-C infection has been associated with clinical benefit or longer survival in HIV-infected individuals in ten of 13 clinical studies.
Persistent GBV-C viraemia is associated with this improvement in survival and a lower plasma HIV RNA concentration over time, which suggests an inhibitory effect on HIV replication in vivo.
In our study, GBV-C infection of peripheral-blood mononuclear cells resulted in lower replication of laboratory and clinical HIV isolates representing two clades, and of HIV that uses CCR5 or CXCR4 as the coreceptor. Inhibition was mediated by chemokines as shown by higher concentrations of mRNA for RANTES, MIP-1a, MIP-1b, and SDF-1 in GBV-C-infected cells than in mock-infected cells, and by higher concentrations of these four chemokines in cell-culture supernatants than in mock-infected cell supernatants. Most importantly, the inhibitory effect of GBV-C on HIV replication was neutralised by antibodies directed against the chemokines.
The inhibitory effect of GBV-C on HIV replication was seen with R5 HIV isolates in purified CD4-positive cells, although the amount of HIV inhibition was reproducibly greater in cultures of mixed peripheral-blood mononuclear cells. Thus, accessory cells play a part in HIV inhibition by contributing to the production of chemokines. The diminished effect of GBV-C against X4 viruses in CD4-enriched peripheral-blood mononuclear cells was not surprising, because SDF-1 is generally not highly expressed by CD4-positive T cells. Finally, GBV-C infection led to downregulation of CCR5 on peripheralblood mononuclear cells in both stimulated and unstimulated cells.
Lower amounts of CCR5 on circulating CD4-positive T cells has been associated with delayed HIV disease progression in vivo and with decreased HIV replication in vitro. Although our earlier studies did not identify a change in expression of CCR5 or CXCR4 after GBV-C infection; expression of chemokine receptor was measured for only 24 h after GBV-C infection, and differences in CCR5 expression were not seen until 3 days or longer after GBV-C infection.
A limitation of our study is that GBV-C is unlikely to exert as great an effect on HIV replication in vivo as it does in our in-vitro coinfection system, because we have found that fewer than 1 in 5000 peripheral-blood cells contain GBV-C RNA (unpublished) and transcriptionally active HIV genomes are found in about 1 in 100 to 1 in 400 CD4-positive T cells in lymphoid tissue. However, if GBV-C induces chemokines in vivo, GBV-C-infected cells in lymphoid tissue would be likely to produce chemokines, resulting in downregulation of CCR5 expression and inhibition of HIV replication without the need to infect the same cell.
Our in-vitro findings, combined with the ten epidemiological studies that have shown an association between GBV-C viraemia and improved HIV disease course, suggest that GBV-C has a direct antiviral effect on HIV-1. Studies to characterise further the specific protein or RNA replication pathway by which GBV-C infection increases concentrations of chemokines and alters CCR5 density on peripheral-blood mononuclear cells are warranted. Although HIV-positive individuals who have slow progression or stable HIV disease have been studied extensively in an attempt to identify HIV and host factors related to the slow progression, GBVC viraemia or antibody status has not been assessed in most of these studies. GBV-C infection is common among HIV-positive people and has been associated with a large effect on survival in seven independent studies, so future studies of HIV disease progression should include an assessment of GBV-C status.
To investigate whether different GBV-C isolates inhibited both R5 and X4 HIV, we used GBV-C derived from an infectious clone or clinical isolates to infect peripheralblood mononuclear cells stimulated with phytohaemagglutinin and interleukin 2. 24 h after GBV-C or mock infection of the cells, they were infected with a clinical or laboratory strain of R5 or X4 HIV-1 (from subtypes A and B). HIV-1 replication, asssessed by measurement of HIV p24 antigen production in culture supernatants, was significantly and reproducibly lower in GBV-C-coinfected cells than in mock-coinfected peripheral-blood mononuclear cells.
For example, in the coinfection experiment shown in figure 1, the difference in HIV p24 antigen concentration in culture supernatants between GBV-C infected (derived from an infectious clone) and mockinfected cells super-infected with R5 HIV was 222 ng/L (95% CI 132--301; p=0·001); for a clinical GBV-C isolate the difference was 169 ng/L (153--184; p<0·001). Both clinical and laboratory strains of HIV that use either CCR5 or CXCR4 as their coreceptor were inhibited by GBV-C. Inhibition of HIV replication was observed with both the infectious-clone-derived GBV-C isolate and clinical GBV-C isolates. GBV-C infection also resulted in HIV inhibition when culture supernatants were tested for HIV infectivity in peripheral-blood mononuclear cells (R5 and X4) or for syncytia formation in MT-2 assays (X4; data not shown). Because all infections were in primary peripheral-blood mononuclear cells, the extent of HIV p24 antigen production varied between experiments.
Consequently, we report the percentage difference in HIV p24 antigen concentration between GBV-C-infected cells and mock-infected controls for subsequent experiments. In addition, because clinical GBV-C isolates replicate more efficiently than virus derived from the infectious clone, we used the clinical GBV-C isolate20 for all subsequent experiments.
To ensure that GBV-C inhibition of HIV replication was not due to toxic effects of GBV-C infection on peripheral-blood mononuclear cells, metabolic activity was assessed by measurement of incorporation of sulphur- 35-labelled methionine into the cells. Protein synthesis was greater in GBV-C-infected cells than in mockinfected cells, and the viability as shown by trypan-blue exclusion on microscopy was similar for GBV-C-infected and mock-infected cells (data not shown).
When the multiplicity of infection of GBV-C was increased, the HIV-inhibitory effect increased substantially, although when the HIV multiplicity of infection was more than 5, the inhibitory effect of GBV-C was lost (data not shown). To assess the relation between the timing of GBV-C infection and HIV inhibition, GBV-C infection was allowed to proceed for 6--72 h before HIV infection. HIV inhibition progressively increased with the delay between GBV-C infection and HIV infection; the greatest inhibition of HIV occurred when HIV infection was initiated 48--72 h after GBV-C infection.
To investigate whether chemokine expression is involved in the GBV-C-related inhibition of HIV, chemokine mRNA expression and protein secretion into culture supernatant fluids were studied in GBV-C infected and mock-infected peripheral-blood mononuclear cells. Phytohaemagglutinin and interleukin 2 were not included in the culture media in these experiments, because GBV-C infection does not require this stimulation and the background concentrations of chemokines are lower without stimulation. The concentrations of mRNA for RANTES, MIP-1a, MIP- 1b, and SDF-1 were consistently higher in GBV-C infected peripheral-blood mononuclear cells than in mock-infected cells, as measured by differential hybridisation in three separate experiments done in duplicate. GBV-C altered the expression of some other chemokines (MIP-2, SDF-2) but not monokine induced by interferon (MIG), macrophage-derived chemokine (MDC), or interleukin 8. Culture supernatants from GBV-C-infected peripheralblood mononuclear cells had significantly more RANTES (p=0·0006), MIP-1a (p=0·0008), MIP-1b (p=0·005), and SDF-1 (p=0·004) in culture supernatants 3 days after infection than mock-infected cellculture supernatants.
To confirm that the increased amounts of chemokines present in culture supernatants were the cause of the inhibition of HIV replication, GBV-Cinfected and mock-infected cells were incubated with antibodies previously shown to neutralise the HIV-inhibitory effect of RANTES, MIP-1a, MIP-1b, and SDF-1 or with isotype control antibodies before infection with HIV. When GBV-C-infected cells were incubated with isotype control antibodies before infection with HIV-1 (R5 strain), HIV p24 antigen production was 77% (95% CI 49--90) lower than in mockinfected cells maintained under the same conditions. By contrast, when a mixture of neutralising antibodies against a combination of the CCR5 ligands (RANTES, MIP-1a, and MIP-1b) was included in the cultures, the inhibitory effect of GBV-C on R5 HIV replication was completely abolished.
When antibodies to RANTES, MIP-1a, and MIP-1b were tested individually in the cells coinfected with GBV-C and HIV (R5), the inhibitory effect of GBV-C on HIV replication was significant, but less than that observed with the mixture of anti-chemokine antibodies.
Similarly, X4 virus replicated well when used to inoculate mock-infected peripheral-blood mononuclear cells incubated with either isotype control or antibody to SDF-1. When X4 virus infection followed GBV-C infection of cells maintained in media with isotype control antibody, there was 60% inhibition of p24 antigen production.
This inhibitory effect was completely abolished when the GBV-C infected cells were maintained with antibody to SDF-1 in the medium. By comparison, HIV p24 antigen production was 50% lower when exogenous RANTES (18 nmol/L), MIP-1a (82 nmol/L), MIP-1b (37 nmol/L), or SDF-1 (73 nmol/L) was added to medium of HIV-infected peripheral-blood mononuclear cells (data not shown).
GBV-C replicates in CD4-positive T cells. To find out whether GBV-C can inhibit HIV replication in purified CD4-positive T cells, we isolated cells positive for CD4 and CD3 by flow-activated cell sorting (>99% CD4-positive) or by negative selection (>96% CD4-positive) and infected the cells with GBV-C or a mock-virus preparation for 24 h before HIV inoculation. GBV-C inhibited R5 HIV isolates in these CD4-enriched cells; however, X4 viruses were not significantly inhibited. GBV-C inhibited replication of a clinical HIV R5 isolate in peripheral-blood mononuclear cells and in enriched CD4-positive cells incubated with isotype control antibodies. Neutralising antibodies against RANTES, MIP-1a, and MIP-1b abolished the GBV-C related inhibition of HIV in both peripheral-blood mononuclear cells and CD4-enriched cell cultures. Increased chemokine expression should result in decreased expression of cognate chemokine receptors.
Thus, we measured the effect of GBV-C infection on CCR5, CXCR4, and CD4 expression in peripheral-blood mononuclear cells stimulated with phytohaemagglutinin and interleukin 2 or not stimulated. CCR5 surface expression was observed on the cells before GBV-C or mock infection, and CCR5 expression was lower in the GBV-C-infected cells grown in phytohaemagglutinin-interleukin 2 after 3 or more days of infection than in the mock-infected cells maintained under identical conditions. The difference in CCR5 surface density was not detected until later in unstimulated peripheral-blood mononuclear cells; however, by 10 days after infection, CCR5 expression was much lower in the GBV-C-infected cells. Surface expression of CXCR4 and CD4 was the same in GBV-C infected and mock-infected cells (data not shown).