Low levels of co-receptor CCR5 are sufficient to permit HIV envelope-mediated fusion with resting CD4 T cells

The susceptibility of phenotypically CCR5-negative resting CD4 T cells for membrane fusion with a CCR5-specific HIV-1 envelope was analysed using a novel sensitive fusion assay. A very low overall density of CCR5 on T cells expressing high levels of CD4 was shown to be sufficient for HIV envelope-mediated membrane fusion. These findings are relevant to the understanding of how HIV-1 R5 strains enter and replicate in resting CD4 T cellsin vivo. HIV-1 entry into target cells is a well-ordered process during which the envelope glycoprotein subunit gp120 interacts with cell surface receptors CD4 and chemokine receptors CCR5 or CXCR4. These interactions ultimately lead to fusion of the viral and target cell membrane and virion entry. Depending on the chemokine receptors used for entry, HIV-1 strains are classified as R5, X4 or R5X4 isolates [1]. In the infected patient, HIV genomes can be detected in activated and resting CD4 T cells in lymphoid tissues and in the peripheral blood [2–6]. Their consistent presence even during prolonged highly active antiretroviral therapy (HAART) prompted the hypothesis that resting CD4 T cells might constitute a viral reservoir that can produce infectious progeny virus upon cellular stimulation [4–7]. Analysis of the subtype of infected CD4 T cells revealed that naive as well as memory CD4 T cells are infected by both R5 and X4 viruses [8–11]. This appears to be paradoxical as the CCR5 co-receptor cannot be detected on naive cells and is present only on a small number of memory cells, contrasting with the more abundant expression of CXCR4 [12,13]. Earlier in-vitro studies [14,15] showed that unstimulated peripheral blood monocuclear cells (PBMC) allowed the entry of R5 strains but not replication, whereas more recent data obtained with purified resting cells [10,16] suggested that entry occurs only if high multiplicities of R5 virus are used for infection. However, recent in-vitro data evidence suggest that in the microenvironment of lymphoid tissue, resting cells might be infected and actively replicate virus [17,18]. We analysed the role of CCR5 expression levels on resting CD4 T cells for their susceptibility to Env-mediated membrane fusion. A methodology was developed to analyse membrane fusion independently of subsequent steps in the viral life cycle, e.g. reverse transcription. The assay, named ‘fluorescence-based assay of syncytia induced by HIV envelope’ is based on cytoplasmic dye transfer subsequent to viral envelope-mediated fusion of cells (Fig. 1a) [19,20]. Small syncytia resulting from the fusion of a single Env-expressing cells with single target cells can thus be detected by this method.Fig. 1.: (a) Effector BHK cells are infected with defective recombinant SFV vectors expressing the HIV envelope glycoprotein [19]. Target cells are labelled with the fluorescent dye CMFDA (Molecular Probes) and effector cells are labelled with SNARF-1 and then co-cultured for 3–9 h. Fused cells are visualized by fluorescence microscopy as yellow syncytia and quantified by FACScan using CellQuest software. Cells are scored as syncytia if they are positive for both fluorescent dyes (SNARF-1+/CMFDA+). (b) Control staining of MT4-R5 cells with mAb 2D7 directed against CCR5 (continuous line). The dashed line represents the isotype control labelling of MT4-R5 and the dotted line mAb 2D7 labelling of MT4 cells. CD4 T cells were purified using magnetic beads (Myltenyi) and purity controlled by CD3+/CD4+ labelling. The expression of HIV (co)-receptors CXCR4 and CCR5 and activation marker CD25 were analysed on purified CD4 T cells (three upper panels) and purified CD4 T cells depleted of CD25- and CD69-expressing cells (three lower panels). (c) Fusogenicity of CD4 T cells (open bars) and CD4 T cells depleted of CD25+/CD69+ cells (closed bars). Co-culture was performed with BHK-Env-X4 (NL4-3) or BHK-Env-R5 (BX08) in the presence of 5 μg/ml mAb Q4l2O (anti-CD4) [20], 200 ng/ml AMD3100 or 200 nM TAK779 to control for specificity. The results shown are representative of three independent experiments. Similar results were obtained with HLA-DR-depleted CD4 T cells (data not shown). Fusion is expressed as the ratio of double-positive syncytia obtained with BHK-Env compared with BHK-β-gal, ± SD.The specificity of the fluorescence-based assay of syncytia induced by HIV envelope method was tested using phytohaemogglutinin/IL-2 stimulated human PBMC as target cells and known HIV entry inhibitors. Two envelopes, which had previously been shown to use CD4 cells for entry were used in this study: NL4-3, which uses CXCR4 (designated X4-Env) [21], and BX08, which uses CCR5 (designated R5-Env) as the co-receptor [22,23]. The fusion of both envelopes required binding to CD4 cells as it was completely blocked by the monoclonal antibody to CD4 cells (data not shown). The chemokines SDF1α and MIP1β, which are ligands of CXCR4 and CCR5 selectively inhibited fusion in a co-receptor-specific manner, as did the CXCR4- and CCR5-specific antagonists AMD3100 [24] and TAK779 [25]. The R5-Env exclusively uses CCR5 as a receptor for fusion with PBMC as cells from a donor carrying a 32 base pair deletion in CCR5 (CCR5Δ32/Δ32) readily fused with X4-Env but not with R5-Env-expressing cells (data not shown). Before fusion analysis with resting cells, receptor expression was analysed on bead purified CD4 T cell populations depleted of activated CD4 T cells using anti-CD25, anti-CD69 or anti-HLA-DR antibodies (Fig. 1b). Although CD4 and CXCR4 could be detected on more than 97% of purified resting CD4 T cells, CCR5 was detected on only 1% of cells using monoclonal antibody (mAb) 2D7. 2D7 staining was as low on unstimulated CD4 T cells which still contained activated cells. Similar results were obtained with CD25-depleted and HLA-DR-depleted CD4 T cells. To rule out that the conformation of CCR5 was responsible for the lack of recognition by mAb 2D7, which binds to the extracellular loop 2a of CCR5, three additional antibodies were tested. MAb CTC5 is directed against the N-terminus of CCR5, mAb 181 is a multiepitope antibody and mAb 182 recognizes the extracellular loop 2b. None of these antibodies allowed the detection of CCR5 on more than 1% of purified CD4 T cells (data not shown). Furthermore, no increased CCR5 expression was observed after 12 h of co-culture with Env-expressing cells (data not shown). Both unstimulated and resting CD4 T cells were tested for their capacity to fuse with R5- or X4-Env-expressing cells (Fig. 1c). X4-Env and R5-Env clearly induced fusion with resting cells. R5-Env fusion could be inhibited by mAb Q4120 and by TAK779, providing evidence that the fusion step was CD4 and CCR5-dependent in resting CD4 cells. Given the undetectable levels of CCR5, it was surprising to note that the R5-Env BX08 efficiently fused with CD4 T cells in a CD4 and CCR5-dependent manner. This finding illustrates the fact that minute levels of co-receptor in primary resting CD4 T cells are sufficient for R5-Env mediated fusion in cells expressing high levels of CD4. Our results extend ealier data obtained in transformed cell lines expressing defined levels of CCR5, in which infection by an R5 isolate of HIV-1 was observed despite undetectable levels of co-receptor [26,27] when CD4 was expressed at high levels. The results presented in this paper might help to clarify how HIV enters resting CD4 T cells in vivo. Resting CD4 T cells are permissive for direct fusion of R5 strains in the presence of very low levels of CCR5. Integration and viral gene expression might occur once these infected resting cells encounter cytokine signals at sites of inflammation or in secondary lymphoid organs. Experimental evidence for this hypothesis comes from studies with pseudotyped HIV, which can enter resting CD4 T cells in a CD4 and co-receptor-independent fashion. Integration and viral gene expression can be obtained by stimulation with cytokines such as IL-2, IL-4, IL-7, or IL-15 without the induction of cell activation markers [28]. These mechanisms might operate in vivo when viral replication and gene expression have been observed in resting CD4 T cells [18,29]. Acknowledgements The authors would like to thank Marc Girard, Françoise Bachelerie, Yann Percherancier, Eric Cabannes and Olivier Schwartz for helpful discussions and critical reading of the manuscript. They are grateful to José Alcami for providing blood samples from CCR5Δ32/Δ32 donors.