DC-SIGN, a Dendritic Cell-Specific HIV-1 Binding Protein that Enhances trans-Infection of T Cells.

Teunis B.H. Geijtenbeek, Douglas S. Kwon, Ruurd Torensma, Sandra J. van Vilet, Gerald C.F. van Duijnhoven, Jeena Middel, Ine L.M. H.A. Cornelissen, Hans S.L.M. Nottet, Vineet N. KewalRamani, Dan R. Littman, Carl G. Figdor, and Yvette van Kooyk.

Cell, Vol. 100, 587-597, March 3, 2000

     The objective of this page is to review a current article from the journal Cell.  I will begin by providing a brief overview of the paper with a summary of the results.  Next, I will explain and interpret each of the figures from the paper in an effort to determine if the claims made in the paper are supported by the results.  Finally, I will discuss the implications of the paper by suggesting future research questions to be considered along with possible procedures for investigating these questions.

     In general, the purpose of this paper was to characterize the dendritic cell -specific C-type lectin, DC-SIGN, and investigate the role of this binding protein in human immunodeficiency virus type 1 (HIV-1) infection.  The researchers made several assertions regarding DC-SIGN and its role in HIV-1 infection.  First, it was found through a series of immunological approaches that DC-SIGN is preferentially expressed on mucosal dendritic cells but not on other leukocytes, such as monocytes and peripheral blood lymphocytes (PBL).  Furthermore, it was found that DC-SIGN binds to the HIV-1 envelope glycoprotein gp120.  In addition, it was found that while DC-SIGN is required for efficient HIV-1 infection of T cells, it does not mediate HIV-1 entry into T cells. Finally, it was demonstrated that DC-SIGN-bound HIV-1 retains infectivity for up to four days.  With these findings in mind, the authors proposed that DC-SIGN specifically captures HIV-1 via a high-affinity interaction.  The immature DC carrying HIV-1 then migrates to the lymphoid tissues.  Upon arrival, DC clusters with T cells and DC-SIGN enhances HIV-1 infection of T cells in trans leading to a productive and sustained infection.

     It has recently been found that DC-SIGN is a DC-specific ICAM-3 adhesion receptor that mediates DC-T cell interactions.  In order to provide evidence that DC-SIGN is expressed specifically in DC, flow cytometric analysis was performed on a panel of hematopoietic cells using anti-DC-SIGN antibodies.  The results of the analysis revealed that DC-SIGN is almost exclusively expressed on in vitro cultured DC but not on peripheral blood lymphocytes (PBL) or monocytes (Figure 1A).  The figure clearly shows a higher level of DC-SIGN expression in DC.
Identification of DC-SIGN by peptide amino acid sequencing has revealed it to be 100% identical in its amino acid sequence to the HIV-1 envelope glycoprotein gp120-binding protein previously isolated from a placental cDNA library.
Therefore, the next logical step was to investigate the role of DC-SIGN in the binding of HIV-1 to DC.

     To determine whether DC-SIGN plays a role in the binding of HIV-1 to DC, a flow cytometric adhesion assay was used to examine the ability of HIV-1 gp120-coated fluorescent beads to bind to immature DC (Figure 1B).  The gp120-coated beads bound more efficiently to DC in the presence of medium (positive control) and anti-CD4.  However, gp120 binding to DC was strongly inhibited by anti-DC-SIGN antibodies, EGTA, and mannan.  These results suggest that DC-SIGN, but not CD4, mediates binding of gp120 to DC.  Although the authors report that the gp120-coated beads bound efficiently to DC, the highest binding percentage obtained was about 34%, which seems rather low.  In addition, we cannot be certain that the conditions under which the assay was performed do not favor one interaction over another.

     The results from Figure 1C demonstrate that DC-SIGN is sufficient for the binding interaction to occur.  By using THP-1 cells (which lack expression of both CD4 and CCR5), it was shown that when transfected with a DC-SIGN expression vector, the cells bound the gp120-coated beads.  I would like to have seen a negative control condition where the THP-1 cells were transfected with CD4 and CCR5.  I would not expect these cells to bind to the gp120-coated beads.  Figure 1C also shows us that DC-SIGN is expressed at high levels in immature DC as compared to CD4 and CCR5 (which is further support that perhaps a higher level of expression of CD4 and CCR5 may be needed to induce binding).  Figure 1D shows us that the THP-DC-SIGN transfectants bind HIV-1 gp120.  Once again, we see that adhesion is blocked in the presence of anti-DC-SIGN antibodies and EGTA, but not anti-CD4 antibodies.  Interestingly, a higher percentage (~52%) of binding occurred in the transfectant cells than in normal DC.  These findings provide evidence for the authors' conclusion that DC-SIGN is a specific dendritic cell surface receptor for the HIV-1 envelope glycoprotein.

     After demonstrating that DC-SIGN is a DC-specific HIV-1 binding protein,  the authors sought to demonstrate that DC-SIGN is required for efficient HIV-1 infection in DC-T cell cocultures.  In order to investigate this inquiry, immature DC were pulsed with a strain of HIV-1 and cultured in the presence of activated T cells (Figures 2A and 2B).  The effects of antibodies against CD4, a CC5 trio, and DC-SIGN were all examined.  The results were in agreement with the earlier findings, namely that anti-DC-SIGN inhibits infection, while anti-CD4 and anti-CCR5 on their own had no effect on infection.  However, a combination of CD4 and CCR5 antibodies did block infection of DC.  The authors suggest that this effect is due to efficient inhibition of the T cell infection by the bound (or unbound) anti-CD4/chemokines.  It is worthwhile to mention that the addition of CD4 and CCR5 antibodies to the preincubation mixture with anti-DC-SIGN did not further inhibit infection (Figure 2B).

     In order to demonstrate that the DC-SIGN antibodies were not interfering with the DC-T cell interaction (the interaction between DC-SIGN and ICAM-3 on T cells), the researchers added antibodies against DC-SIGN after exposure of DC to HIV-1, but prior to the addition of activated T cells.  Under these conditions, only anti-CCR5 and anti-CD4 antibodies inhibited HIV-1 infection.  Since the addition of anti-DC-SIGN did not inhibit infection of T cells, it may reasonably be concluded that DC-SIGN interactions with ICAM-3 are not involved in the transmission of DC-bound-HIV-1 to T cells.  DC-SIGN does however serve an important function in propagation of HIV-1 in DC-T cell cocultures.

     After finding that DC-SIGN was necessary for infection, the investigators examined whether DC-SIGN acts as a receptor that permits entry of HIV-1 into T cells (similar to CD4 plus CCR5).  293T cells were transfected with either DC-SIGN or CD4 plus CCR5, and pulsed for 2 hr with HIV-1.  Cells were then cultured for 9 days and p24 levels were analyzed.  Figure 3A shows that the 293T-CD4-CCR5 cells were readily infected while the 293T-DC-SIGN cells produced an undetectable level of p24 protein and hence were unifected.  These results indicate that DC-SIGN does not mediate HIV-1 entry.  To further investigate the possibility that DC-SIGN played a role in HIV-1 entry (by complexing with other molecules), a set of experiments were conducted in which 293T cells expressing CD4, CCR5, or both were transfected with DC-SIGN and later infected with a replication-defective strain of HIV-1.  When luciferase (reporter) activity was examined (Figure 3B), entry was not detected in 293T cells that contained only DC-SIGN.  In addition, no infection was observed if DC-SIGN was expressed with either CD4 or CCR5, indicating that DC-SIGN does not form a complex with these molecules to permit viral entry.  Viral entry (as illustrated by high luciferase activity) was obtained when T cells expressed both CD4 and CCR5.  However, luciferase activity did not fluctuate with the addition of DC-SIGN, indicating that it does not contribute further to viral entry (Figure 3B).

     Since DC-SIGN did not appear to mediate virus entry, the researchers proposed that DC-SIGN might facilitate both capture of HIV-1 and subsequent transfer of the virus to CD4/CCR5 positive T cells.  In order to test this hypothesis, THP-DC-SIGN transfectants, which do not express CD4 or CCR5 were pulsed with HIV-luciferase virus pseudotyped with gp120.  Once unbound virus was removed, the cells were cocultered with CD4/CCR5-expressing T cells (which are susceptible to HIV-1 infection), or activated T lymphocytes.  The results shown in Figure 4A indicate that only THP-DC-SIGN cells were able to effectively capture the pseudotyped virus and transmit it to the target cells with the receptors necessary for viral entry.  When antibodies against DC-SIGN were present (negative control) HIV-1 infection was completely inhibited.  This control provides good support for the claim that HIV-1 capture is DC-SIGN dependent.  Figure 4B provides further support for the theory that HIV-1 capture is DC-SIGN dependent.  In this experiment capture was demonstrated using envelope proteins taken from additional strains of HIV-1.  Once again, only cells that contained DC-SIGN were able to capture and transmit HIV-1 (Figure 4B).  Futhermore, the presence of anti-DC-SIGN inhibited this activity.  In a related experiment THP-DC-SIGN cells were incubated with HIV-eGFP viruses pseudotyped with HIV-1 and then cocultured with activated T cells.  Figure 4C shows that when CD3 is not present HIV-1 infection does not occur.  On the other hand CD3-positive T cells were infected.

     Up until this point all of the experiments described have been in vitro assays.  In vitro assays provide experimenters with the benefit of knowing exactly what components are present.  The problem with in vitro approaches however, is that there may be other factors involved in the in vivo system which are not being taken into account.  In order to simulate in vivo conditions, in which HIV-1 is likely to be limiting, THP-1 transfectants were challenged by providing low titers of pseudotyped HIV-1.  These transfectants were then cocultured with HIV-1 permissive cells, without washing away unbound virus.  The results in figure 5A confirmed the expectations of the researchers.  Neither 293T-CD4-CCR5 cells nor activated T cells were noticeably infected.  When the same amount of HIV-1 was presented to THP-DC-SIGN cells, efficient HIV-1 infection was observed in trans. Similarly, enhancement of HIV-1 infection was observed with HIV-luciferase viruses pseudotyped with five other R5 envelopes (Figure 5B).  The results in figure 5B corroborate the finding that DC-SIGN first seizes HIV-1, and then enhances CD4-CCR5-mediated HIV-1 entry by presentation in trans to the HIV-1 receptor complex.

     With figure 6 the researchers continued to tie together the in vivo characteristics of DC, by examining mucosal tissues for the presence of DC.
Immunohistochemical analyses of mucosal tissues were performed at sites of first exposure during sexual transmission of HIV-1 (Figure 6A).  Staining of mucosal tissue sections such as the cervix, rectum, and uterus (6Aa, 6Ab, and 6Ac respectively) with anti-DC-SIGN mAb revealed expression of DC-SIGN.  Figure 6B shows the staining of serial sections of rectal (a-c) and uterine (d-f) tissues with antibodies against DC-SIGN, CD4, or CCR5.  The staining patterns revealed that for the most part DC-SIGN coexpressed with CD4, but not CCR5.  Therefore it can be concluded that DC present in mucosal tissues at sites of HIV-1 exposure express DC-SIGN, but are CCR5 deficient.  A control that may have provided a helpful comparison would have examined the staining of mucosal tissue from the nasal cavity or another section unlikely to be a site of HIV-1 exposure.  The staining pattern of this section would provide insight into the overall distrubution of DC in the mucosal tissues in the body.

     The final matter considered by the investigators further contributed to the elucidation of a mechanism responsible for HIV-1 infection of T cells.  The researchers reasoned that the virus would have to retain infectivity during the transport from the mucosal tissues where DC are originally located, to the T cell zones in lymphoid tissue.  A time-course experiment was employed to determine the length of time that HIV-1 remains bound to DC-SIGN expressed on transfected THP-1 cells.  The data presented in figure 7A suggest that gp120-coated beads remained bound to DC-SIGN for for more than 60 hr.  In the presence of anti-DC-SIGN the incubation period was drastically reduced (Figure 7A).  In a further investigation, the authors determined the amount of time during which HIV-1-pulsed THP-DC-SIGN cells could retain infectious virus.  DC-SIGN- expressing transfectants were pulsed with pseudotyped HIV-1 for 4 hr and then washed.  The pulsed cells were then cocultured at defined intervals with activated T cells (Figure 7B).  As figure 7B shows, HIV-1 pulsed cells were still able to infect target cells after 4 days.  In the absence of DC-SIGN-positive cells, viruses lost infectivity after 1 day (Figure 7B).  Although these findings are in accordance with the proposed mechanism (the virus must retain infectivity during transport from the mucosal tissues to the T cells), one cannot be certain that the virus-DC-SIGN complex would display the same endurance in vivo.  A more effective measure of the infectivity period may require an in vivo system.

     The final figure (7C) presents the proposed mechanism by which DC-SIGN binds HIV-1 and migrates to target T cells where infection is enhanced.  The left part of the figure shows DC containing DC-SIGN which binds HIV-1.  After migration to the lymphoid organs in-trans infection of T cells occurs by DC-SIGN.  The virus then replicates in the CD4 and CCR5 containing T cell.

     In my opinion, the authors of this paper have employed a methodical approach in order to characterize the role of DC-SIGN in HIV-1 infection.  The authors have used a series of functional immunological assays in order to discern the role of DC-SIGN in stepwise manner.  For the most part the researchers refrain from making claims that are unsupported by their data.  Instead, their claims are based on results from several related experiments.  It appears that the authors have allowed the data collected to dictate the course of their investigation, thus avoiding the pitfalls associated with presenting erroneous and unsupported conclusions.
Overall, the results of the different experiments work together to support the mechanism which is ultimately proposed.

     Although the authors have provided a set of well-supported results, there are still several questions that remain to be investigated.  Since this paper represents the first published attempt to characterize the role DC-SIGN in HIV-1 infection, the mechanisms by which HIV-1 utilizes DC cell machinery in order to achieve efficient infection of target cells remains unclear.  Although no direct approach to elucidating the mechanisms involved exists, in vitro assays in which known reactants are used to cause infection of T cells have proved useful in this paper and will continue to do so.  By keeping track of the minimum cell machinery required for infection, researchers would be able to focus on the possible interactions taking place.  However, as we have seen in this paper, it is also important to consider in vivo approaches.  Using GFP and other reporters allows the presence of particular protein interactions to be detected.  In order to determine which proteins are involved in the pathway, it may be helpful to use the cDNAs encoding DC-SIGN (which has been detected on a placental genomic library) and gp120 to probe human genomic libraries of mucosal tissue and T cells.  This approach would potentially allow for the discovery of novel proteins involved in the mechanism.

     Once the different proteins involved in the pathway have been isolated, the two-hybrid approach may prove helpful in cloning the genes for the proteins that interact with DC-SIGN.  This approach provides several advantages over coimmunoprecipitation experiments which do not yield a coding sequence.  The two-hybrid system may then allow a new series of genetic approaches to be applied.  For example, Geijtenbeek et al. raise the question of whether a transient quaternary complex is formed between DC-SIGN, HIV-1 Env, CD4, and CCR5.  In order to investigate this question, sequencing information may be used to design an approach in which different reporters are produced once a specific interaction has taken place.  This approach would theoretically allow one to keep track of the various interactions that occur during HIV-1 infection of T cells.  It may also be helpful to crystallize the DC-SIGN complex at a various stages of the infective process.  There is some evidence that this work has already begun to be considered.  To view a chime image of Hiv-1 gp120 Core Complexed with Cd4 and a neutralizing human antibody click here.

     It has been indicated that binding of gp120 to DC-SIGN may induce a conformational change that enables a more efficient interaction with CD4 and possibly CCR5.  In order to test this hypothesis it may be useful to test the differential binding affinities of these other proteins to DC-SIGN in the presence and absence of the viral envelope glycoprotein.  I would expect the binding affinitity to be different when gp120 was not present due to conformational effects on binding affinity.

     Another question raised by the researchers involves the use of mutant forms of gp120 and DC-SIGN in order to study the mechanism of enhanced infectivity in trans.  This approach would allow one to investigate which parts of the protein are necessary for DC-SIGN to efficiently enhance infectivity in trans.  Once these regions have been identified chimeric proteins could be made to determine if the regions of interest are in fact sufficient to cause infection.

     Finally, it may be interesting to study the process of HIV-1 infection in mice that have had the genes encoding for DC-SIGN deleted from DC.  Homologous recombination would theoretically allow one to knockout the gene encoding for DC-SIGN on a specific group of cells (DC in this case).  One would expect that this procedure would interrupt the normal function of the DC.  Since the DC would no longer contain DC-SIGN, one would predict that DC would be unable to bind the viral envelope protein and migrate to the lymphoid tisssues.  This approach may also allow alternative routes of infection to be elucidated.