BST-2/tetherin was demonstrated to be a potent restriction factor for enveloped virus release [24–27, 29, 30]. However, its role in cell-to-cell viral transfer is still unclear. In vivo viral restriction activity of BST-2/tetherin was demonstrated in a report showing that tetherin-deficient mice could not inhibit Moloney murine leukemia virus (Mo-MLV) replication and disease progression . While this restriction factor is constitutively expressed at low levels at the surface of some hematopoietic cells, its antiviral action is only unveiled after IFN-α induction and was, therefore, proposed to be critical for the antiretroviral activity of IFN-α in vivo. This agrees with a recent report showing that induction of BST-2/tetherin expression after pegylated IFN-α/ribavirin treatment strongly correlated with reduction in HIV-1 load and with the antiretroviral capacity of IFN-α in vivo. As an inflammatory environment is a pre-requisite to increase both BST-2/tetherin expression and restriction activity in specific cell types, we aimed to investigate the potential restriction activity of BST-2/tetherin in different DC subsets while modulating their maturational state with IFN-α or LPS. One of the limiting steps to studying HIV-1 infection in DC from myeloid origin is the fact that these cells are quite refractory to HIV-1 infection, more particularly when cells are matured, due to potent post-entry restriction blocks [6, 7, 10]. Taking advantage of the recent characterization of SAMHD1-mediated restriction in myeloid DC [6, 7], we used Vpx-expressing lentiviral vectors in order to remove this restriction block and to allow higher levels of HIV-1 infection to be reached in DC from myeloid origin. This provided a more accurate way to measure the effects of BST-2/tetherin in this cell type. This setting did not compromise either DC maturation or the response to IFN-α or LPS treatment but allowed higher levels of HIV-1 infection, as previously shown  and as evidenced by BST-2/tetherin and RIG-I up-regulation. Interestingly, BST-2/tetherin up-regulation was recently proposed to occur in most immune cells from patients in the acute phase of HIV-1 infection  as well as in simian CD4+ T cells upon SIV infection . In fact, and in agreement with a recent report from Coleman et al. , we found that, although HIV-1 infection led to BST-2/tetherin up-regulation, viral spread from DC was not efficiently restricted in the presence of BST-2/tetherin, even after removal of the SAMHD1-mediated post-entry restriction block with Vpx. Our evidence also showed that while constitutive BST-2/tetherin expression could be detected at the surface of both monocyte-derived DC and myDC, its expression was greatly up-regulated after IFN-α or LPS treatment of both DC subsets. When primary DC subsets were challenged with HIV-WT and HIV-ΔVpu, we observed that BST-2/tetherin surface expression could also be decreased in a Vpu-dependent manner as previously reported for other cells [24, 25, 28]. Surprisingly, however, Vpu-mediated down-regulation of BST-2/tetherin from the cell surface did not correlate with significant enhancement of viral release suggesting that the lack of tetherin-mediated HIV-1 restriction might be Vpu-independent. Hence, BST-2/tetherin down-regulation at the cell surface of both DC subsets was not always followed by its degradation, as previously observed in other cells [37, 38, 48]. Overall, these results demonstrated that the absence of Vpu did not result in significant alteration of HIV-1 replication in these cells and thus precluded BST-2/tetherin-mediated restriction activity against HIV-1 in primary DC.
Interestingly, some recent reports have shown that BST-2/tetherin expression is up-regulated after stimulation with TLR agonists [42, 57, 59, 60], which reinforces the idea that this cellular restriction factor could be a critical component of the intrinsic innate immune response. We have demonstrated here that, while IFN-α treatment was unable to efficiently restrict DC-to-T cell HIV-1 transfer, LPS-mediated maturation of DC induced a much more potent blockade of viral infection and transfer in-cis toward autologous CD4+ T cells. Strikingly, when LPS-matured DC were challenged with HIV-ΔVpu, viral replication and transfer were decreased by more than 4 fold and 2 fold (respectively) compared to HIV-WT.
As tetherin surface levels were modulated in IFN-α or LPS-treated cells and given that previously described results highlighted a more potent restriction of HIV-ΔVpu in LPS treated DC compared to IFN-α treated DC, we hypothesized that a potential mechanism of BST-2/tetherin restriction in DC could rely on its cellular localization rather than overall expression. Additionally, treatment with IFN-α or a TLR agonist was shown to differentially modulate DC maturation and immune function. In comparison to LPS stimulation, IFN-α was proposed to only partially affect DC maturation [61, 62] while also inhibiting the production of IL-12 and thus potentially skewing the T helper response [63, 64]. However, IFN-α is considered as a major regulator of the antiviral response and has been shown to be a potent inducer of CD8+T cell cross-priming and therefore critically required to induce DC full maturation upon engagement of a TLR [65–67]. One of the marked differences between both of these biological compounds in the process of DC maturation is that, while LPS-driven maturation is followed by a decreased rate of endocytosis, antigen-processing and a stabilization of MHC-II/peptide complexes, IFN-α does not alter MHC-II synthesis, antigen processing or endocytosis rates [68, 69]. Thus, vesicular trafficking pathways could behave differently depending on the applied stimulus to mature DC. In fact, we observed that BST-2/tetherin was relocalized to tetraspanin-enriched compartments only after LPS stimulation. This localization was similar to the HIV-containing tetraspanin-rich compartment previously described [46, 70, 71]. When LPS-matured DC were infected with HIV-1, we observed that BST-2/tetherin was co-localized with tetraspanin CD81. In contrast, BST-2/tetherin did not accumulate significantly with CD81 in IFN-α matured DC.
The intracellular relocalization of BST-2/tetherin observed in DC after LPS treatment did not influence DC-mediated viral transfer in trans toward CD4+ T cells. This was demonstrated by the comparable data obtained with DC-mediated HIV-WT and HIV-ΔVpu transfer of infection to CD4+ T cells. This was also confirmed when BST-2/tetherin expression was knock-downed in DC loaded with HIV-1. Of note, LPS pre-treatment of DC even increased DC-mediated HIV-1 transfer in trans, independently of BST-2/tetherin expression as reported by Coleman et al. . While in agreement with the previously described model of increased HIV-1 transmission by mature DC through trans-infection [19, 42, 72–74], our data exclude a BST-2/tetherin effects during early DC-mediated HIV-1 capture and transfer to CD4+ T cells.
In contrast, our data with LPS pre-treated DC suggested that tetherin potently restricts DC-mediated cis-infection of CD4+ T cells. This restriction correlated with the polarization of BST-2/tetherin into the HIV-1-containing tetraspanin-enriched compartments which is part of the trafficking pathway required for DC-T cell transfer during HIV-1 infection . A plausible hypothesis would be that this virus-containing compartment, previously thought to be an escape route for HIV-1, could in fact become an innate antiviral cell host defense mechanism when fueled with BST-2/tetherin molecules after LPS-mediated maturation of DC.
In agreement with its restrictive action in LPS-matured DC-mediated viral transfer in cis, we found BST-2/tetherin relocalized to the virological synapse along with HIV-1 and TEM markers. Hence, it appeared that BST-2/tetherin localization at the virological synapse was further enhanced with HIV-Δvpu infected LPS-matured DC in contact with CD4+ T cells, which correlated with the increased accumulation of BST-2/tetherin observed in HIV-1 containing TEM within LPS-matured DC infected with HIV-Δvpu. These results also suggest that, while the lack of BST-2/tetherin-mediated HIV-1 restriction observed in DC-mediated trans-infection appeared to be Vpu-independent, Vpu might indirectly contribute to overcome BST-2/tetherin antiviral activity on DC-mediated HIV cis-infection toward CD4+ T cells. Indeed, in the absence of Vpu, more BST-2/tetherin molecules from the cell surface are available when polarised into the HIV-containing compartment upon LPS treatment, reinforcing the retention of de novo budding virions and the restriction of vpu-deficient HIV-1 transmission compared to wild-type viruses. However, as evidenced by our data, the presence of Vpu in infected DC does not fully overcome BST-2/tetherin antiviral activity. Interestingly, a recent publication  reported that localisation of BST-2/tetherin to the virus-containing compartment (VCC) of infected macrophages correlated with a decreased HIV release and cell-to-cell transmission. Importantly, this report also suggested that BST-2/tetherin was essential for VCC formation. If BST-2/tetherin molecules at the cell surface are a major source for the virus-containing compartment, this could explain the apparent increased size of the HIV-containing compartment observed in HIV-ΔVpu infected DC.
As shown here, and as previously reported [42, 76], BST-2/tetherin is expressed at low levels on DC from myeloid origin but its expression is increased after HIV-1 infection . Indeed, conditions where cytokines and/or pathogen associated molecular patterns (PAMPs) could up-regulate BST-2/tetherin expression would fit with the concept of the dysregulated cytokine storm observed during HIV-1 acute infection (for review see ). Furthermore, this might even be exacerbated during co-infection events engaging specific TLR pathways. Surprisingly, however, we found that IFN-α and TLR agonist treatments were both differentially regulating BST-2/tetherin intracellular trafficking.
The lack of BST-2/tetherin restriction in IFN-α treated DC could be due to the complex trafficking pathway used by HIV-1 in DC for viral assembly and release. Indeed, in contrast to T cells and some epithelial cell line models in which viral release occurs mainly at the plasma membrane, HIV-1 budding and release from DC and macrophages was shown to take place in complex compartments connected, at least in part, to the cell surface and enriched in tetraspanins [46, 70, 78–80]. Therefore, in conclusion, we demonstrated that the innate antiviral BST-2/tetherin activity could be modulated depending on the context of DC maturational state and inflammatory environment.