Teunis B.H. Geijtenbeek, Douglas S. Kwon, Ruurd Torensma, Sandra J. van Vilet, Gerard C. F. van Duijnhoven, Jenna Middel, Ine L. M. H. A. Cornelissen, Hans S. L. M. Nottet, Vineet N. Kewal Ramani, Dan R. Littman, Carol G. Figdor, and Yvettwe van Kooyk.  2000.  Cell.  100:587-597.

Reviewed by Jennifer L. Caldwell

        Using flow cytometric analysis of hematopoietic cells with anti-DC-SIGN antibodies, the in vitro expression of DC-SIGN was examined.  Expression levels of immature DC cultured from monocytes in the presence of GM-CSF and IL-4 were compared to expression levels of resting peripheral blood lymphocytes (PBL) and monocytes that do not express DC-SIGN.  In figure 1A the expression levels of each cell type were compared to internal controls.  The monocytes provided an appropriate negative control, as no expression of DC-SIGN was detected.  The expression levels of PBL appeared to be similar to that of the monocytes, also suggesting no expression of DC-SIGN.  DC, but not other leukocytes, appear to specifically express DC-SIGN.

        A flow cytometric adhesion assay was then used to determine whether DC-SIGN has a role in the binding of HIV to DC or DC-SIGN transfectants.  HIV-1 gp120-coated fluorescent beads were allowed to interact with DC under various binding conditions, normal, added anti-CD4, added anti-DC-SIGN, added EGTA, and added mannan.  As shown in figure 1B, over 30% of the HIV-gp120 beads bound to the DC under normal conditions.  No binding affinity differences were found upon the addition of anti-CD4 antibodies, thus CD4 probably does not mediate binding of HIV-1 gp120 to DC.  Binding was blocked by anti-DC-SIGN antibodies (AZN-D1 and AZN-D2), mannan, and EGTA. About 50% of HIV-1 gp120 bound to THP cells transfected with DC-SIGN, as shown in figure 4D.  This binding level was unchanged by adding anti-CD4 antibodies to the THP-DC-SIGN cells.  Anti-DC-SIGN antibodies blocked binding of HIV-1 gp120 to the same extent as the EGTA negative control and the medium from cells lacking DC-SIGN.  The lack of binding demonstrated by mannan and EGTA provided a negative control as DC-SIGN is a calcium binding mannose-type lectin.  The reduced binding ability when anti-DC-SIGN antibodies were added but not when anti-CD4 antibodies were added suggests that binding of HIV-1 gp120 to DC is DC-SIGN rather than CD4 mediated.

        As dendritic cells express CD4, CCR5, and DC-SIGN, the researchers examined the effects of each factor on the binding ability of HIV-1 gp120 (figure 1C).  No binding of HIV-1 gp120 occurred until after the cells were transfected with a DC-SIGN expression vector.  No expression of CCR5 or CD4 was found on the monocytic THP-1 cells, providing a negative control as THP-1 cells are supposed lack CD4 and CCR5 receptors.  Although all are potential HIV-1 gp120 receptors, CCR5 and CD4 demonstrated low binding levels in DC, whereas high expression levels were found for DC-SIGN as determined by flow cytometric analysis.  Thus, the researchers concluded that immature DC expresses low levels of CD4 and CCR5 and high levels of DC-SIGN.

        To examine the contributions of CD4, CCR5, and DC-SIGN in HIV-1 gp120 infection of T cells by DC immature DC were pulsed with HIV-1, washed, and cultured with activated T cells (figures 2A and 2B).  Preincubation of DC with anti-CD4 and a trio of CCR5-specific chemokines resulted in HIV-1 replication similar to that found with normal DC.  Replication was inhibited via the administration of anti-DC-SIGN, anti-CD4 and CCR5 trio in combination, anti-DC-SIGN and anti-CD4, anti-DC-SIGN and CCR5 trio, and anti-DC-SIGN and anti-CD4 and CCR5 trio.  The inhibition by the anti-CD4/chemokine mixture might be due to inhibition of T cell infection by the unbound antibodies/chemokines, as in unshown data the researchers report that T cells challenged with the same viral load experienced weaker infection levels than that of T cells cultured with virus-pulsed DC.  Therefore, the effects of antibodies against DC-SIGN were explored to determine if the results in figures 2A and 2B were due to interference in the DC-SIGN mediated infection regulation or due to interference with the interaction between DC and T cells as DC-SIGN binds to ICAM-3 on T cells.  Anti-DC-SIGN, anti-CD4, and the CCR5 chemokine trio were then added after DC exposure to HIV-1 but prior to the addition of activated T cells (figure 2C).  The antibodies against DC-SIGN added post-infection produced results similar to normal DC, with only CCR5-specific chemokines and anti-CD4 antibodies inhibiting HIV-1 infection of T cells.  Thus, DC-SIGN appears to mediate the binding of HIV-1 in DC-T cell cocultures.  The lack of binding found in figures 2A and 2B seems to be due to the inability of DC-SIGN to bind to gp120 and not an interaction with ICAM-3.

        The permissibility of DC-SIGN in HIV-1 entry was examined in 293T cells that expressed DC-SIGN, CD4, or CCR5 alone or in various combinations by measuring expressed p24 antigen levels using ELISA (figures 3A and 3B).  High infection levels were found in 293T-CD4-CCR5 cells pulsed with HIV-1.  No detectable p24 protein was found in culture supernatants of 293T and 293T-DC-SIGN cells.  The cells were then transiently transfected with DC-SIGN and infected with pseudotyped CCR5-trophic HIV-1 virus in the presence of polybrene to explore the possibility of interactions between DC-SIGN and CCR5 and CD4 (figure 3B).  High luciferase activity, indicating high levels of infection, was found for 293T-CD4-CCR5 and 293T-CD4-CCR5-DC-SIGN cells, but not for 293T-CD4, 293T-CCR5, 293T-CD4-DC-SIGN, 293T-DC-SIGN, 293T, or 293T-CCR5-DC-SIGN cells.  Thus, DC-SIGN does not appear to mediate HIV-1 entry and cannot substitute for CD4 or CCR5.

        To test the possibility that DC-SIGN might facilitate the capture and transmission of HIV-1 to CD4/CCR5-positive T cells independent of CD4 and CCR5, THP-DC-SIGN transfectants were pulsed with HIV-luciferase virus pseudotyped with HIV-1 envelope glycoprotein, washed, and cocultured with 293T cells or activated T lymphocytes.  Figure 4A shows that THP-1 cells alone are not capable of being infected with HIV-1, providing a negative control, yet THP cells transfected with DC-SIGN infect the 293T-CD4-CCR5 and activated T cells with HIV-1.  This effect was blocked by anti-DC-SIGN antibodies in all cases.  CD4-positive T cells by themselves were not responsible for the infection results, as CD-4 positive cells demonstrated no transmission of HIV-1 to activated T cells.  The ability of DC-SIGN to transmit HIV-1 was also found with HIV-luciferase viruses pseudotyped with envelope glycoproteins from five R5 isolates (JRFL, JRCSF, 93 TH966.8, 92 US 715.6, and 92 BR020.4), whereas THP-1 cells alone demonstrated no luciferase activity in 293T-CD4-CCR5 cells (figure 4B).  DC-SIGN appears to fully transmit HIV-1 as the THP-DC-SIGN cells are not becoming infected, but are transmitting the HIV-1 to T cells (figure 4C).  After virus-pulsed THP-DC-SIGN cells and T cells were analyzed for expression of virus-encoded green fluorescent protein (GFP) only the CD3+ T cells stained positive for eGFP, while the CD-3 negative THP-DC-SIGN cells demonstrated no staining.

        In an effort to mimic in vivo effects of low virus titers infecting permissive cells thus causing replication, THP-1 transfectants were challenged with limited amounts of HIV-1 or R5 isolates of HIV-1and cocultured with permissive cells without washing.  Figure 5A shows little infection in cocultures of THP-1 cells with 293T-CD4-CCR5 or with activated T cells or in 293T-CD4-CCR5 and activated T cells alone.  Infection was observed in the presence of THP-DC-SIGN in trans, indicating the sufficiency of DC-SIGN for the infection of HIV-1 permissive cells.  Anti-DC-SIGN antibodies blocked the successful infection of the HIV-1 permissive cells in all cases.  Similar results were found for the R5 isolates of HIV-1 (figure 5B).  No infection was found in THP-1 cells by themselves, but high infection levels were demonstrated upon the transfection of THP cells with DC-SIGN.  By acting in trans, DC-SIGN appears to enhance the efficiency of HIV-1 infection of T cells at low virus titers.

        An in vivo study was performed to determine the presence of DC and DC-SIGN in tissues that are the sites of first exposure during sexual transmission of HIV-1.  Cervical, uterine, and rectal mucosal tissues were examined using immunohistochemical analyses.  All mucosal tissue in the lamina propria stained positive with anti-DC-SIGN monoclonal antibodies (figure 6A).  The DC-SIGN appears to be present on large irregular cells.  In an experiment not shown the cells were determined to be distinct from T cells, B cells, monocytes, and macrophages, thus suggesting that the cells are DC.  Figure 6B demonstrates that these DC-like cells in uterine and rectal tissues also appear to co-express CD4 and DC-SIGN but not CCR5.  When analyzed using monoclonal antibodies against DC-SIGN, CD4, and CCR5, the tissues seem to have DC that contained both CD4 and DC-SIGN, suggesting that DC-SIGN but not CD4/CCR5 capture the HIV-1 in vivo.  The lack of CCR5 in the mucosal tissue strengthens the finding that CD4/CCR5 is required for HIV-1 infection, as DC in mucosal sites are generally resistant to HIV-1 infection.

        As in vivo HIV-1 infects lymphoid organs, the long-term infectivity of HIV-1 while bound to DC-SIGN must be investigated as the virus would have to be able to retain infectivity during the transport between the primary mucosal tissues to the T cells in the lymphoid tissue.  Using flow cytometric analysis, the time course of HIV-1 gp120 binding to THP-DC-SIGN was observed (figure 7A).  DC-SIGN- gp120 binding appears to remain for up to 60 hours, with binding being blocked by anti-DC-SIGN antibodies added to the cells (negative control).  HIV-1-pulsed THP-DC-SIGN cells cocultured with activated T cells were then evaluated over several days for infectivity efficiency (figure 7B).  No viral infectivity was found in the controls, THP-1 cells alone or in THP-DC-SIGN cells with anti-DC-SIGN antibodies.  Although demonstrating infectious ability at the time of pulsing, the virus incubated without cells lost infectivity after approximately one day.  Virus bound to DC-SIGN, however, retained infectivity for more than four days, infecting T cells in trans.  Therefore, DC-SIGN bound HIV-1 appears to allow for adequate infectivity after migration to lymphoid tissue from primary mucosal tissue.

        According to the proposed model (figure 7C) DC-SIGN acts as a trans receptor on DC for HIV-1 at the initial exposure mucosal site in the lamina propria.  HIV-1 and DC-SIGN bind, and the DC migrate to the secondary lymphoid tissues.  In the lymphoid tissue DC-SIGN enhances the interaction between DC and T cells in trans.  The virus then infects the T cell and begins replication.

        Overall, the authors present a novel mechanism for the dissemination of HIV-1 from the primary site of infection in mucosal tissues to T cell zone secondary lymphoid tissue.  Dendritic cells localized in the uterine, rectal, and cervical mucosal tissue appear to express DC-SIGN and CD4, but preferentially binds HIV-1 via DC-SIGN due to its high affinity for HIV-1 gp120.  The high affinity also allows for binding in areas of limited virus titers.  The DC then transport the virus to the target T cells, maintaining the viral infectivity for up to four days.  Then via a trans mechanism DC-SIGN facilitates the infection of CD4/CCR5 positive T cells perhaps through a novel multiple cell surface receptor system involving CD4 and/or CCR5.  The virus finally infects the T cells, signaling the beginning of viral replication, without infecting the original dendritic cell.

        More research is necessary to determine the molecular mechanism for the DC-SIGN enhancement of infection of T cells and the role of the CD4/chemokine receptor complex.  The authors suggest that DC-SIGN might induce a conformational change in the complex before the binding on the envelope glycoprotein, thus providing a more stable formation and increase in infectivity.  The infectivity of the virus after the binding with DC-SIGN can be compared to the infectivity levels of the virus after binding with other anti-gp120 antibodies.  Examination of the same effect on the envelope glycoprotein is also necessary, using various antibodies to the particular envelope as a comparison for viral infectivity.

        The authors suggest two conflicting pathways for the migration and interaction of HIV-1 bound DC-SIGN.  In one design the authors suggest that the virus is maintained on the exterior surface of the cell to enhance the probability of transmission to T cells once binding occurs with the CD4/chemokine receptor complex.  Their other model suggests that the virus may be internalized as viral particles have been found within endocytic vesicles of DC.  This internalization would protect the virus during migration to the lymphoid tissue.  Further research is necessary to determine which model, if either, is correct.  The authors could pulse THP-DC-SIGN cells with HIV-1 and culture the cells for several days.  The cells could then be immunofluorescently examined using immunofluorescently-tagged anti-gp120 antibodies.  The antibodies could be added directly to the cells, only allowing for binding to externally located gp120 proteins.  Then in another group of similar cells streptolysin-O can be used to open the plasma membrane and allow the antibodies to reach any internalized proteins.  The amount of fluorescence in each condition can be compared.

         The necessity of DC-SIGN to act in a trans conformation versus a cis conformation must also be explicitly explored.  The authors purport that the transmission of HIV-1 from DC-SIGN to the target cells occurs while DC-SIGN is in the trans conformation, yet they make no comparison to the transmission ability of DC-SIGN bound HIV-1 with DC-SIGN in the cis conformation.  Manipulating DC-SIGN into a cis conformation using heat or pH without degrading the protein, the efficacy of HIV-1/DC-SIGN mediated transmission should be compared to a trans oriented DC-SIGN/HIV-1 complex.

        Further research is necessary to study the effects of DC-SIGN in vivo.  Recreation of the experiments performed in this study in a mouse or rat model of HIV-1 infection would allow researchers to better understand the pathogenesis of HIV-1/AIDS in humans.  Knockout mice lacking DC-SIGN, CD4, or CCR5 can be treated with a mouse version of HIV-1, and the infection rate as well as areas of infection can be observed via transcription of green fluorescent tagged virus.  DC-SIGN appears to work in vitro to enhance viral entry into T cells, but little is known about the role of this protein in vivo.