Dendritic cells (DC), professional Ag-presenting cells located in mucosae and lymphoid organs, operate at the interface of innate and adaptive immunity and are likely the first cells to encounter invading HIV-1. Although the C-type lectin DC-Specific ICAM-3-grabbing non-integrin (DC-SIGN) binds to several viruses, including HIV-1, its direct involvement in viral entry remains controversial. Despite its central role in DC function, little is known about the underlying molecular mechanism(s) of DC-SIGN-mediated Ag uptake. Here, we analyzed the early stages of DC-SIGN-mediated endocytosis and demonstrate that both membrane cholesterol and dynamin are required. Confocal microscopy and clathrin RNAi showed that DC-SIGN-mediated internalization occurs via clathrin-coated pits. Electron microscopy of ultrathin sections showed the involvement of DC-SIGN in clathrin-dependent HIV-1 internalization by DC. Currently, DC-specific C-type lectins are considered potential target in anti-tumor clinical trials. Detailed information about how different Ag are internalized via these receptors will facilitate the rational design of targeted therapeutic strategies.
Dendritic cells (DC) are professional Ag-presenting cells that operate at the interface of innate and adaptive immunity 1.
Since DC are located in the mucosae and the lymphoid tissues, they are likely to be the first cells to encounter invading HIV-1 particles 2. DC express several PRR that specifically recognize distinct pathogen-associated molecular patterns displayed on microbial surfaces including C-type lectin receptors (CLR) 3 and TLR 4.
DC-specific ICAM-3-grabbing non-integrin (DC-SIGN; CD209) is a CLR initially described as an ICAM-3 binding protein 5 as well as a major HIV-1 receptor on DC 6. In addition, DC-SIGN binds several other viruses as well as fungi, bacteria, and parasites 3. Although DC-SIGN has been shown to bind to several viruses, its capacity to directly mediate viral entry has often been debated 7, 8. Recently, we demonstrated that on the plasma membrane of DC, DC-SIGN is organized in nanoclusters, some of which colocalize with lipid rafts, that specifically confer to the receptor its capacity of binding and internalizing viruses as well as Ag conjugated to fluorescent nanoparticles 9, 10. These findings were substantiated by a biophysical study by Neumann et al., who showed that these nanoclusters are enriched near the leading edge of living DC, but are preferentially endocytosed at lamellar sites posterior to the leading edge, suggesting a directed mobility of DC-SIGN from areas of concentration at the front to rearward sites of internalization 11.
Binding of HIV-1 to DC-SIGN leads to non-fusogenic uptake of virions by DC, which is required for enhancement of T-cell infection 12 and is dependent on the cytoplasmic domain of the DC-SIGN molecule 13.
Ag internalized by DC-SIGN can be presented via both MHC class I and class II 14, 15. In addition, DC-SIGN has been shown to promote both MHC class I- and class II-restricted HIV Ag presentation 16, 17.
Despite the central role of DC-SIGN-mediated Ag uptake in DC function, little is known about the underlying molecular mechanism(s) of the initial entry and subsequent sorting pathways. Here, we analyzed the early stages of DC-SIGN-mediated endocytosis. Currently, the potential role of DC-SIGN and other CLR as target in anti-tumor DC-based clinical trials is being explored 18. Detailed information about how different Ag are internalized via DC-SIGN and other CLR will be instrumental for the rationale design of CLR-based therapeutic strategies.
Results and discussion
DC-SIGN-mediated internalization requires membrane cholesterol and dynamin
We and others previously demonstrated that on DC, DC-SIGN may reside in cholesterol- and glycosphingolipid-enriched lipid rafts 9, 19. In agreement with these observations, we show that lipid raft disruption upon extraction of plasma membrane cholesterol by methyl-β-cyclodextrin treatment affects DC-SIGN-mediated binding to ligand-coated beads both in DC and in the CHO cells stably expressing DC-SIGN (CHO-DC-SIGN) (Fig. 1A). Moreover, upon cholesterol extraction, the internalization of the anti-DC-SIGN mAb AZN-D1 was almost completely abrogated (Fig. 1B). This mAb has already been used as DC-SIGN ligand mimic in previous targeting studies 14, 15.
The GTPase dynamin is involved in multiple endocytic pathways 20, 21. Therefore, we transiently co-transfected CHO cells with DC-SIGN and either dynamin WT (WT-dynamin) or the dominant-negative K44A mutant of dynamin (DN-dynamin) and subsequently determined the Ag internalization capacity of DC-SIGN. Upon endocytosis triggering at two different time points, DN-dynamin significantly inhibited the DC-SIGN internalization capacity (Fig. 1C). These results demonstrate that DC-SIGN-mediated internalization requires both membrane cholesterol and dynamin.
Ag uptake via DC-SIGN is clathrin-dependent
Cholesterol extraction is known to not only disturb raft integrity but also endocytosis via clathrin-coated pits (CCP) 22. Therefore, we sought to investigate the involvement of CCP in DC-SIGN-mediated endocytosis. In the DC-SIGN cytoplasmic domain a dileucine (LL) motif is present that may mediate the interaction with CCP 23. We exploited the DC-SIGN mutant LL-AA 14 stably expressed on K562 cells with expression levels similar to DC-SIGN WT (Fig. 1D). We show that the internalization via K-DC-SIGN (LL-AA) is almost 50% reduced when compared with DC-SIGN WT, clearly indicating that this dileucine motif plays an important role in DC-SIGN-mediated endocytosis (Fig. 1E). The presence of another internalization motif in the cytoplasmic tail of DC-SIGN, the triacidic cluster EEE, may explain the residual internalization capacity of the LL-AA mutant. In fact, the mutation of this cluster by alanine substitution markedly attenuated DC-SIGN activity for phagocytosis and endocytosis 24.
Subsequently, we established the degree of colocalization of the anti-DC-SIGN mAb with clathrin before and after endocytosis triggering. Whereas in absence of endocytosis DC-SIGN did not show any colocalization with clathrin, endocytosis triggering induced clear colocalization of DC-SIGN with CCP, as also indicated by the marked increase in the Pearson correlation coefficient, calculated over multiple images (Fig. 1F and G).
Currently, we and others are evaluating DC-SIGN as potential target in experimental DC-based anti-cancer vaccines 18. Therefore, we investigated whether CCP formation is also essential for DC-SIGN-mediated endocytosis on monocyte-derived DC. On DC, the colocalization of DC-SIGN and clathrin over time was assessed by confocal microscopy (Fig. 2A). Whereas no colocalization was detected at steady state, internalized DC-SIGN mAb strongly colocalized with CCP up to 15 min upon endocytosis triggering. This correlation was lost at longer time points, demonstrating the transient nature of this interaction, as also underscored by the Pearson correlation coefficient (Fig. 2B).
Additionally, we investigated the possible colocalization of the endocytosed ligand with markers of the early endosomal compartments. The time-dependent colocalization between DC-SIGN and EEA-1 showed a trend similar to the colocalization of DC-SIGN and CCP (Fig. 2C).
Finally, we used RNAi methods to inhibit expression of clathrin on DC and subsequently quantify DC-SIGN internalization capacity. Viability and proper phenotype of the transfected DC (data not shown) along with the residual clathrin expression (Fig. 2D) were assessed. Upon clathrin silencing, the DC-SIGN-mediated endocytosis mAb appears significantly inhibited with respect to non-targeting RNAi DC (Fig. 2E), demonstrating that on DC, DC-SIGN exploits CCP to internalize its ligands.
DC-SIGN mediates the clathrin-dependent internalization of HIV-1 on DC
Binding of HIV-1 to DC-SIGN leads to non-fusogenic uptake of virions by DC 12. Although this type of virus uptake has been considered as a dead end for productive infection, Daecke and colleagues demonstrated that a dynamin- and clathrin-dependent endocytosis could lead to a productive entry of HIV-1 into HeLa cells 25. Considering the involvement of CCP in the DC-SIGN-mediated internalization, we investigated whether DC-SIGN participates in the uptake of HIV-1 via CCP on DC. To this end, we performed transmission EM analysis of ultrathin sections of DC incubated with AT-2-inactivated HIV-1 virions, which have functional envelope glycoproteins but are not infectious 26 and have been characterized at the ultrastructural level 27, 28.
To ensure efficient virus binding and uptake, DC were first incubated with AT-2-HIV-1 at low temperature for 1 h, washed and quickly warmed up to 37°C to induce virus uptake. After fixation, DC-SIGN was labeled with a neutral mAb and 10 nm gold particles, before being processed for EM. As illustrated in Fig. 2F, structurally intact virions appeared to be bound to the DC surface as well as internalized in CCP. Gold-labeled DC-SIGN molecules were observed in close proximity to the bound virions (Fig. 2F, i and ii) and also on top of invaginating CCP (Fig. 2F, iii). To avoid morphological damage and preserve intact CCP and virions, the samples were not permeabilized prior to immunogold labeling of DC-SIGN. In this way, no gold particles labeling DC-SIGN can be expected to be found ‘inside’ the CCP (Fig. 2F, iv–vii). However, several DC-SIGN-labeling gold particles were clearly visible at the site where CCP are closing and pinching off from the plasma membrane (Fig. 2F, iv and v).
Together these data demonstrate the involvement of DC-SIGN in the clathrin-dependent internalization of HIV-1 particles by DC.
Exploiting immunological and biochemical assays as well as confocal and electron microscopy, we showed that DC-SIGN internalizes soluble Ag via CCP and is directly involved in the clathrin-mediated uptake of HIV-1 virions on DC. Many studies are directed at exploiting DC to improve anti-viral or anti-tumor vaccine efficiency, some of which focus on Ag targeting to cell surface CLR 15, 18. Therefore, a better understanding of the Ag uptake mechanism(s) of these DC-specific CLR and the subsequent Ag processing route(s) will ultimately provide relevant information for the development of new tools to improve in vivo receptor targeting. The use of DC loaded with inactivated HIV-1 particles has shown some therapeutic potential 29 in chronic HIV-1 infection. Therefore, a better understanding of the virus internalization and subsequent degradation pathway(s) in DC will provide important information for the development of effective anti-virus strategies.
Materials and methods
Immature DC were generated from monocytes as described 30. The AT-2-HIV-1 particles [HIV-1(MN)/H9 Cl.4 lot P3936 courtesy of Biological Products Section, AIDS and Cancer Virus Program, SAIC Frederick, National Cancer Institute at Frederick, Frederick, MD USA 21702] were already described 26. A list of reagents and the detailed methods are available in the Supporting Information Materials and methods document.
The authors are grateful to Mietske Wijers and Huib Croes of the Microscopic Imaging Centre for assistance with the electron microscope, and Julian Bess and Jeff Lifson, AIDS and Cancer Virus Program, SAIC Frederick, National Cancer Institute at Frederick, for providing AT-2 inactivate HIV-1 and for critical reading of the manuscript. This work was financed by a VENI grant 916.66.028 (to AC) and a TOP grant 9120.6030 (to CGF) from The Netherlands Organization of Scientific Research (NWO).
Conflict of interest: The authors declare no financial or commercial conflict of interest.