Inhibition of HIV strains by GB virus C in cell culture can be mediated by CD4 and CD8 T-lymphocyte derived soluble factors

Objective: A number of studies concerning the pathogenesis of GB virus C (GBV-C) in HIV-infected people suggest a beneficial effect and improved survival for dually infected individuals. However there has remained controversy regarding the clinical relevance of these findings, as some studies have not confirmed these observations. To address the possibility of direct inhibitory mechanisms, we studied the impact of GBV-C on HIV-1 replication in vitro.

Methods: Peripheral blood mononuclear cells (PBMC) were infected with sera from GBV-C positive individuals or transfected with GBV-C specific RNA and superinfected with HIV. Replication kinetics of HIV were studied by quantification of HIV-p24 release. Induction of soluble antiretroviral factors were monitored with an HIV infection assay and by quantification of chemokine secretion. Changes in chemokine receptor expression were analysed by flow cytometry.

Results: We demonstrate that GBV-C infection of PBMC leads to significant replication inhibition of R5- and X4-HIV isolates representing eight HIV clades. The inhibitory effect is mediated by GBV-C infection and also by expression of GBV-C structural glycoproteins and/or of non-structural proteins NS2/NS3. Upon GBV-C infection CD4 and CD8 T lymphocytes produce soluble HIV-suppression factors. Induction of stromal cell-derived factor (SDF)-1 and subsequent internalization of CXCR4 was not observed.

Conclusions: CD4 and CD8 T lymphocytes are stimulated by GBV-C to secrete antiretroviral factors, inhibiting R5- and X4-HIV strains. As no induction of SDF-1 and no down-regulation of the respective receptor CXCR4 could be observed, it is likely that additional unidentified factors causing inhibition of X4-HIV strains are induced by GBV-C.

Introduction

GB virus C (GBV-C), a positive-stranded RNA virus of the Flaviviridae, is related to hepatitis C virus (HCV) but does not seem to be associated to hepatitis or any other disease ^[1,2]. Its genome comprises 9.4 kb, with an open reading frame encoding one single polyprotein, which is processed by host and viral proteases. Due to similar transmission routes of HIV and GBV-C (sexual, vertical, parenteral) the prevalence of GBV-C viraemia in HIV positive individuals is much higher (15-40%) than in healthy blood donors (1-2%) ^[2,3]. In 1998 two studies reported that GBV-C co-infection might be a favorable prognostic factor for HIV patients ^[4,5]. Later reports propose delayed progression to AIDS and prolonged survival of GBV-C infected HIV patients compared to GBV-C negative HIV patients, while a few studies did not identify a survival advantage with GBV-C/HIV co-infection ^[6-10]. In order to study the interference between GBV-C and HIV, the influence of GBV-C on peripheral blood mononuclear cells (PBMC) along with primary CD4 and CD8 enriched T lymphocytes has been investigated.

Material and methods

Cells

PBMC were isolated by Ficoll-Hypaque gradient centrifugation and maintained in RPMI-1640 supplemented with 10% foetal bovine serum, gentamycin-10 μg/ml phytohemagglutinin (PHA) and 10 U/ml interleukin (IL)-2. CD4 and CD8 T cells were enriched by positive selection with magnetic labelling by respective MicroBeads (Miltenyi, Cologne, Germany) and maintained in RPMI-1640 (PHA/IL-2).

Infection experiments

Sera (40-400 μl, corresponding to 2 × 10⁶ copies) from GBV-C positive individuals were used to infect PBMC from GBV-C, HCV, hepatitis B virus (HBV) and HIV-negative blood donors. PBMC (2 × 10⁶) were incubated in 500 μl RPMI with GBV-C positive or negative sera (4 h), washed three times and maintained in RPMI-1640 (PHA/IL-2) for 48 h until HIV infection. HIV infection was performed with 2 × 10⁶ cells in 500 μl (4 h), washed (15 ml phosphate-buffered saline) and maintained in RPMI (PHA/IL2). The HIV infectious dose corresponds to 20-25 ng CC chemokine receptor 5 (CCR5; R5-HIV) or 1 ng CXC chemokine receptor 4 (CXCR4; X4-HIV) isolate. p24 antigen was quantified by ELISA according to manufacturer's instructions (Abbott, Wiesbaden, Germany).

RNA transcription

Positive-strand GBV-C RNA was generated by T7 transcription (Megascript, Ambion, Texas) from a full-length GBV-C pCR2.1-construct (WT+) ^[11] and from derivatives thereof, linearized by SpeI; negative-strand RNA (WT-) was transcribed by SP6 polymerase (Promega, Mannheim, Germany) from a full-length GBV-C pGEMzf5(-) vector, linearized by EcoRV. DNA template was removed with DNAseI (15 min, 37°C) and with the Roche High Pure RNA-Isolation Kit (Roche, Penzberg, Germany). Both GBV-C full-length plasmids were kindly provided by J. Xiang and J. Stapleton, University of Iowa, Iowa City, Iowa.

RNA-transfection and constructs

PBMC (5 × 10⁶) were transfected with 2-5 μg GBV-C RNA by nucleofection (Amaxa, Cologne, Germany) according to manufacturer's protocol. Transfection efficiency was in the range of 25-35%, estimated by transfections with a GFP-expression plasmid (1 μg). Protein expression from transfected RNA was monitored by immunostaining with anti-GBV-E2 antibody (clone125, Dunn, Asbach, Germany). Based on GenBank sequence AF121950 pΔNS4-NS5 was constructed by deletion of the StuI-fragment (nucleotides 5437-8924); pΔFS was generated creating a frameshift at nt843 by deletion of the SnaBI-fragment (nucleotides 843-4034).

Virus sedimentation

Culture supernatants were centrifuged (4 h at 44 000 r.p.m. in a SW60Ti rotor at 4°C) and tested for sufficient particles sedimentation by reverse transcription (RT)-PCR (sensitivity, 15 copies/reaction) directed to the 5′ untranslated region.

Results

HIV as well as GBV-C can infect and replicate in PBMC and CD4 T lymphocytes ^[7,12,13]. To monitor a possible effect of GBV-C on HIV replication, PBMC were infected with GBV-C positive sera 48 h prior to HIV-infection. Sera from GBV-C negative individuals served as controls for mock infection. HIV replication kinetics were clearly decreased in GBV-C infected PBMC compared to mock-infected PBMC measured by HIV p24 release into the supernatant. Efficient HIV-inhibition could be seen with different HIV isolates with both CCR5 (R5) and CXCR4 (X4) co-receptor tropism. Comparing HIV p24 quantity in the supernatant of GBV-C infected with mock-infected PBMC at day 4 post-HIV infection, mean HIV antigen release was reduced by 93 ± 11% for R5- (P < 0.0001) and by 88 ± 14% for X4-tropic HIV isolates (P < 0.0001). Replication inhibition ranged from 78 ± 18% (P = 0.023) to 98 ± 2% (P < 0.0001) when classified into their respective clades and groups suggesting that GBV-C mediated HIV inhibition is neither coreceptor nor subtype specific (Fig. 1a). Cell toxicity due to GBV-C infection as a possible reason for delayed HIV replication was excluded, as there was no change in viability between GBV-C infected and control cells using the cellTiter-96 Aqueous One Solution cell proliferations assay (Promega; data not shown). Replication suppression of HIV by GBV-C was confirmed in PBMC transfected with T7-transcribed positive-strand RNA from an infectious GBV-C clone when compared to HIV replication capacity in control PBMC that were mock transfected without GBV-C specific RNA. The inhibitory effect was not dependent upon viral replication, as HIV-inhibition was also seen in PBMC transfected with a replication incompetent deletion mutant in which substantial parts of the coding region from NS4A to the polymerase NS5B had been removed. The extent of inhibition was less than that observed using the full-length clone; however, a significant inhibitory effect on HIV-replication was detectable (Fig. 1c).

Fig. 1

Although GBV-C can replicate to high titres in vivo (up to 1 × 10⁸ copies/ml serum), infection efficiency of GBV-C is low in PBMC cultures. In general only 5-10% of GBV-C infected PBMC are positive for GBV-C antigens visualized by immunostaining against GBV-C E2-glycoprotein (data not shown). Therefore it is unlikely, that in vitro viral interference takes place within a cell infected with both HIV and GBV-C. This assumption is further supported by the finding that the inhibitory effect is mediated by soluble factors. We incubated fresh isolated PBMC with supernatants harvested from GBV-C infected or transfected PBMC cultures. Residual GBV-C particles were removed from the supernatants by centrifugation. No GBV-C particles were found in supernatants following centrifugation as evidenced by the lack of GBV-C RNA detection using RT-PCR (data not shown). HIV-infection was carried out after 48 h of incubation. Up to 70-80% reduction of HIV p24 production was observed in cells incubated with virus-free supernatants from GBV-C infected or transfected PBMC compared to control cultures, indicating that inhibitory factors are released from the cells upon GBV-C stimulation (Fig. 1d). The identification of antiviral activity in supernatants from GBV-transfected cells confirms the observation that GBV-C replication was not necessary to mediate HIV-inhibition. Supernatants from PBMC transfected with positive-strand RNA derived from the full-length GBV-C construct (WT+) or from a replication incompetent deletion mutant pΔNS4-NS5 also resulted in HIV replication inhibition, whereas transfection of minus-strand RNA or RNA of a frameshift mutant that does not code for complete GBV-C proteins did not lead to HIV suppression by secretion of HIV inhibitory factors.

To identify the cell types producing the antiretroviral factors, virus-free supernatants from GBV-C infected CD4- and CD8-enriched T cells (97-99%) have been tested for HIV suppressor activity. The data (Fig. 2a) clearly indicate that both CD4 and CD8 T cells are induced upon GBV-C infection to secrete cellular factors causing replication inhibition of R5- and X4-tropic HIV strains in vitro. FACS-analysis of PBMC revealed decreased surface expression of the CCR5 up to 40% after 1 week of GBV-C infection, a receptor that is used by R5-tropic HIV strains for cell entry (Fig. 2b). Preincubation of GBV-C positive sera with antibodies directed against the GBV-C E2-glycoprotein prevented down-regulation of CCR5 specifically. However no reduction was seen in the CD4 receptor expression and in the number of X4-HIV relevant CXCR4-positive cells. Chemokine receptor down-regulation can be mediated upon ligand-receptor interaction. After GBV-C infection in supernatants of primary CD4 and CD8 cells reproducibly a moderate 2-10-fold increase of the CCR5-ligands RANTES, macrophage inflammatory proteins (MIP)-1α and MIP-1β was seen (quantified by ELISA, R&D, Minneapolis, Minnesota, USA; data not shown). However induction of the CXCR4-ligand, the stromal cell-derived factor (SDF)-1, was definitely not detected within the observation period of 1 week post GBV-C inoculation using three different blood donors and testing different anticoagulants (EDTA and citrate) with four different primary GBV-C isolates.

Fig. 2

Discussion

Even though the majority of retrospective clinical studies clearly demonstrate a benefit for HIV positive patients co-infected with GBV-C, the mode of interaction between GBV-C and HIV is incompletely understood. To determine if GBV-C directly mediates HIV-inhibitory mechanisms, we performed in vitro studies with GBV-C and HIV on PBMC. We demonstrate that GBV-C influences HIV replication in vitro. The inhibitory effect of GBV-C is neither HIV co-receptor nor subtype specific as it can be seen with all X4- and R5-HIV strains that have been tested, representing eight different HIV clades (A to H) and two HIV-groups (M and O). The interference can be induced in PBMC either by infection with GBV-C wild-type isolates or by transfection of positive-orientated RNA derived from an infectious GBV-C clone. As GBV-C isolates vary in the ability to infect lymphocytes ^[13], most co-infection experiments and quantifications of HIV replication inhibition were performed with a lymphotropic clinical isolate that replicates well in PBMC. However we cannot rule out that strain variability of GBV-C such as cell tropism may account for the capability to interfere with HIV.

The transfection experiments strongly suggest that GBV-C replication is not necessary for HIV suppression, as the mutant pΔNS4-NS5 does not encode the essential polymerase gene NS5B, yet it led to HIV-inhibition (P = 0.026). The plus-strand RNA derived from pΔNS4-NS5 contains intact 5′- and 3′-untranslated regions (UTR) including an IRES and the coding region for the structural envelope glycoproteins E1 and E2 and the non-structural proteins NS2 and NS3, suggesting that expression of at least one of these proteins is responsible for the induction of HIV replication inhibition. Protein expression from this construct was confirmed by immunostaining of E2 protein in cells transfected with pΔNS4-NS5 RNA.

We demonstrated that the inhibitory effect of GBV-C is mediated by virus-free supernatant released by CD4 and CD8 T cells, suggesting that soluble factors cause changes in PBMC cell culture, reducing replication efficiency of R5- and X4-tropic HIV strains. Lower amounts of HIV-relevant chemokine receptors can be associated with decreased HIV replication ^[14-16]. In fact we could observe a reduced surface expression of CCR5 on PBMC upon GBV-C infection, but we did not see any changes in CXCR4. Our results confirm in part the findings of Xiang et al. who showed that GBV-C mediates HIV-suppression by CCR5 down-regulation and by induction of the chemokines RANTES, MIP-1α, MIP-1β and SDF-1 ^[17]. Nattermann et al. found increased RANTES secretion and subsequent decreased presentation of CCR5 on CD4 cells exposed to recombinant GBV-C E2-protein ^[18]. But unlike Xiang et al., in supernatants from T lymphocytes we detect antiretroviral activity against X4-HIV, independent of the presence of SDF-1. Whereas CC chemokine induction and CCR5 reduction might explain R5-HIV suppression in vitro, we have no evidences of entry impairment for X4-HIV isolates. Our findings are also supported by the recent report of Giménez-Barcons et al., who do not see any elevation of HIV inhibitory chemokines in HIV patients chronically co-infected with GBV-C ^[19]. Therefore we conclude that other, unidentified antiretroviral factors can be released from T cells upon stimulation by GBV-C that are not involved in mechanism like chemokine receptor-ligand interaction.

Acknowledgements

We thank J. Stapleton (University of Iowa) for helpful discussions and for providing the full-length GBV-C constructs cloned in pCR2.1 and pGEMzf5(-) and K. Korn (University of Erlangen-Nuremberg) for providing p24 ELISA.

Sponsorship: Supported by 'Akademie der Wissenschaften und Literatur zu Mainz', Project: 2 1.223 (HR), the Deutsche Forschungsgemeinschaft (DFG), Graduiertenkolleg 1071, Teilprojekt: A5 (SJ) and 'Johannes und Frieda Marohn Stiftung', Project: Reil/2002 (ND).