Geijtenbeek, et al. (2000) DC-SIGN, a Dendritic Cell-Specific HIV-1-Binding Protein that Enhances trans-Infection of T
          cells. 100: 587-597.

David Lamar

In a separate paper, Geijtenbeek et al. identified a C-type lectin, DC-SIGN, that is expressed exclusively on dendritic cells.  Dendritic cells migrate from peripheral locations to lymphoid organs upon uptake of antigen and in turn become antigen-presenting cells which can then activate naïve T cells.  The introduction of the paper being reviewed suggests that DC-SIGN is important in transporting HIV-1 from sites of infection to lymphoid organs where high numbers of T cells can be infected with virus.  This paper attempts to characterize several properties of DC-SIGN with regards to the HIV-1 virus.  These properties include: DC-SIGN is specific to dendritic cells and that it binds to HIV-1; DC-SIGN does not substitute for the infection-mediating properties of CD4 or CCR5, which are expressed on the surface of HIV-1-permissive T cells; DC-SIGN captures HIV-1 and mediates the virus's infection of HIV-1-sensitive cells; dendritic cells in HIV-1-exposed mucosal tissue expressed DC-SIGN but not CCR5; and HIV-1 can still infect target cells after long periods of time being bound to DC-SIGN.

Figure 1 is the data showing that DC-SIGN is exclusive to dendritic cells and that it binds HIV-1.  Figure 1a is the result of a FACScan (fluorescence-activated cell sorter scan) analysis, a method used to study lymphocytes based on cell-surface proteins.  The box labeled PBL represents the number of peripheral blood lymphocytes that express DC-SIGN.  The dashed peaks in these boxes (as well as those in Figure 1c) represent background levels when DC-SIGN is absent and act as a guide to which the data can be compared.  The histogram for PBL has a similar pattern to the control and therefore does not express DC-SIGN.  The same can be said for the histogram that represents DC-SIGN expression on monocytes.  The box labeled DC represents the expression of DC-SIGN on dendritic cells.  The peak of this histogram indicates that 100 times the control amount of DC-SIGN expression exists in dendritic cells.  I believe that this figure demonstrates that dendritic cells express DC-SIGN and that monocytes and PBL do not.  This figure does not necessarily rule out the possibility that DC-SIGN is expressed on cell types that were not studied.  The claim that DC-SIGN is only expressed on dendritic cells in not supported by this data.

Figure 1b shows the results of a flow cytometric adhesion assay testing the binding of dendritic cells to gp120-coated fluorescent beads.  gp120 is believed to be the HIV-1 glycoprotein that binds to dendritic cells.  The bars represent the percent binding of gp120 in the presence of various agents in the medium that could potentially prevent binding of gp120 to the dendritic cells.  If the bars were number 1-6 from left to right, bar 1 is the positive control for dendritic cell/gp120 binding.  Bar 2 are the results from medium plus anti-CD4 antibody.  Because this bar is no different from the positive control, it can be assumed that CD4 does not affect binding.  Bar 3 and 4 four are results from medium plus AZN-D1 and AZN-D2, respectively.  These are anti-DC-SIGN antibodies and bars 3 and 4 indicated that dendritic cell/gp120 binding in these media was inhibited.  This suggests that antibody binding to DC-SIGN prevents gp120/dendritic cell binding, which is strong evidence for DC-SIGN as a protein involved in gp120/dendritic cell binding.  EGTA (bar 5) plus medium and mannan (bar 6) plus medium resulted in data similar to bars 3 and 4.  These two bars add further evidence of DC-SIGN as the gp120-binding protein because EGTA and manna are known to bind structure such as DC-SIGN.  I feel this figure strongly suggests that DC-SIGN is important in dendritic cell/gp120 binding.  It would have been useful to have another bar representing medium and dendritic cells only.  This would have provided a negative control for dendritic cell/gp120 binding, since no binding should have occurred.

Figure 1c employs the same method used in 1a.  This figure shows that dendritic cells express CD4, CCR5, and DC-SIGN while THP-DC-SIGN (a monocytic cell line transfected with the DC-SIGN gene) cells express DC-SIGN but do not express CD4 or CCR5.  This figure casts doubt that the reason anti-CD4 antibody did not prevent dendritic cell/gp120 binding in 1b was because the dendritic cells did not express CD4.  1c also shows that the cells used in Figure 1d express DC-SIGN but not CD4 and CCR5.  I believe this figure shows what the authors suggest it does.

Figure 1d is similar to Figure 1b.  However the cells used in this figure were the THP-DC-SIGN cell type used in Figure 1c as well as THP-1 cells, which express neither CD4, CCR5, or DC-SIGN.  The dark bars represent assays using THP-DC-SIGN and the white bar represents an assay using THP-1 cells.  Bar 1 is a positive control for THP-DC-SIGN/gp120 binding, similar to Figure 1b.  Bar 2 shows that anti-CD4 does not affect binding, which also acts as a positive control for THP- DC-SIGN/gp120 binding because the cells do not express CD4.  Bars 3 and 4 show that anti-DC-SIGN antibody and EDTA inhibit THP-DC-SIGN/gp120 binding, similar to figure 1b, again suggesting that DC-SIGN binds to gp120.  Because EDTA has a greater effect on binding of THP-DC-SIGN cells (which do not express CD4) to gp120, it is possible that CD4 has contributes to gp120 binding in dendritic cells.  Bar 5 acts as a negative control for cell binding to gp120 because the THP-1 cells do not express CD4, CCR5, or DC-SIGN.  I feel this figure is well controlled and supports the authors’ claims about DC-SIGN as a HIV-1 gp120 binding protein.

Figure 2 demonstrates that DC-SIGN is important for the infection of activated T cells with HIV-1 virus.  Figure 2a is a graph of p24 levels (measured by ELISA), which is a HIV-1 protein indicating the relative amount of virus present, versus time.  In this figure, dendritic cells were incubated with various substances prior to infection with HIV-1.  After infection, the dendritic cells were cultured with activated T cells and p24 was measured.  Preincubation with medium alone acts as the positive control for viral replication because nothing was added to prevent dendritic cell/virus interaction.  Medium plus anti-CD4 antibody and medium plus CCR5 binding chemokines, resulted in graphs similar to the positive control.  Thus, blocking of either of CD4 or CCR5 on the dendritic cells does not result in decreased viral replication.  However, medium plus anti-CD4+CCR5 chemokines and medium plus anti-DC-SIGN resulted in inhibition of viral replication.  Binding of anti-DC-SIGN antibodies to DC-SIGN presumably prevents the virus from binding DC-SIGN, and, because viral replication is inhibited in the scenario, DC-SIGN/virus binding must be important for the infection of T cells to occur.  The authors suggest that the anti-CD4+CCR5 chemokines act on the T cell-CD4 and the T cell-CCR5 rather than on those of the dendritic cell to prevent replication of the virus.  Binding of anti-DC-SIGN antibodies to DC-SIGN presumably prevents the virus from binding DC-SIGN, and, because viral replication is inhibited in the scenario, presumably prevent infection of T cells.  This makes sense because dendritic cells do not express very much CD4 or CCR5 in comparison to T cells, and manipulation of these cell-surface molecules would be less consequential in dendritic cells.  It seems that this figure (as could figures 2b and 2c) could use a negative control for the presence of p24.  This could be accomplished by not infecting the dendritic cells with virus and then adding T cells as done in the other instances.

Figure 2b shows a bar-graphical representation of the information from Figure 2a at day 5.  The differences among the different media/substance combinations are not as evident as they become in later days.  This is probably due to the fact that the situations in which viral replication is not inhibited have yet to yield large amount of virus.  The bottom three bars represent situations that were not reported in Figure 2 and that result in inhibition of viral replication.  Again, this figure should probably have a negative control for viral replication.

Figure 2c was included to rule out the possibility that the reason medium plus anti-DC-SIGN results in inhibition of viral replication is because the antibody prevents DC-SIGN and ICAM-3 on the T cells from interacting correctly.  In this case, the various antibody or chemokine solutions were added after the dendritic cells were infected with the virus but before activated T cells were added.  Anti-DC-SIGN antibodies only slightly inhibited viral replication whereas anti-CD4 antibodies and the CCR5 chemokines strongly inhibited viral replication.  Perhaps this strong inhibition is due to anti-CD4 antibody/CD4 and chemokine/CCR5 interactions on the T-cells that somehow prevents infection by the HIV-1 virus.  If this is the case, then it is likely that little to none of the anti-CD4 antibody and chemokines were able to bind to the CD4 and CCR5 on the dendritic cells, leaving plenty to interact with the T cells.  Thus, it is possible that the viruses are binding to CD4 and CCR5 on the dendritic cells or that viral binding to DC-SIGN inhibits anti-CD4 antibody/CD4 and chemokine/CCR5 binding on dendritic cells.  The main point of this figure is that anti-DC-SIGN antibodies did not inhibit viral replication.  Because antibody binding to DC-SIGN in Figure 2a resulted in inhibition of viral replication, than lack of inhibition in Figure 2c means that the antibodies are probably not binding to DC-SIGN.  Most likely, the viruses, that were exposed to the dendritic cells before the antibodies were, were able to bind to the DC-SIGN molecules on the dendritic cells because there was no competition from the antibodies for the binding domains.  I believe this data suggests that it is the prevention of viral binding to dendritic cells due to antibody binding to DC-SIGN, rather than disruption of the DC-SIGN/ICAM-3 interactions, that results in decreased viral replication.

Figure 3 demonstrates the DC-SIGN does promote cell entry by the HIV-1 virus.  For Figure 3a, 293 T cells were transfected either with nothing (negative control), DC-SIGN, or both CD4 and CCR5.  ELISA detected fewer p24 levels in the supernatant of the DC-SIGN tranfects than in that of the negative control, while the supernatant from CD4/CCr5 transfects yield very high p24 levels after 9 days.  These data suggest that DC-SIGN does not enhance cell-entry by HIV-1.  I feel that this figure convincing.  The reader must assume, however, that the researchers have verified that the DC-SIGN transfects do indeed express DC-SIGN.

For Figure 3b, replication-defective HIV-1 viruses that contained a luciferase reporter gene were used.  This allowed levels of infection (but not subsequent replications and reinfections) to be measured in 293 T transfects expressing various combinations of DC-SIGN, CD4, and CCR5.  Transfects expressing nothing were the negative control.  Only CD4-CCR5 and CD4-CCR5-DC-SIGN transfects results luciferase activity much higher than control levels (control: mock).  This suggests that neither DC-SIGN expressed alone, nor with CD4, nor with CCR5 allows for HIV-1 entry.  Also, luciferase activity in CD4-CCR5-DC-SIGN transfects did not appear to differ from that in CD4-CCR5 transfects which again implies that DC-SIGN does not affect HIV-1 entry.  I believe this figure supports the authors claims.  It does appear, though, that CD4-DC-SIGN transfects resulted in more-than-background amounts of luciferase activity, but the authors claim that no infection occurred.  Maybe there is a range of activity that is known to indicated no infection.

Figure 4 attempts to show that DC-SIGN can capture HIV-1 which can then infect HIV-1 sensitive cells in trans.  For Figure 4a, some THP and CD4+ T cells were exposed to mAb against DC-SIGN while others were not; then all cells were exposed to HIV-1 for 2 hours before being thoroughly washed.  They were then added to either 293T-CD4-CCR5 or activated T cells and infectivity of the T cells was measured by luciferase activity.  Only 293T-CD4-CCR5 and activated T cells in the presence of the cells transfected with DC-SIGN and preincubated in antibody-free medium resulted in luciferase activity.  THP-1 cells acted as a negative control. I believe that the result of adding anti- DC-SIGN to some cells shows that DC-SIGN is important for the  transfer of HIV-1 from pre-infected cells to HIV-1 permissive cells.

Figure 4b shows the result of an experiment similar to that in Figure 4a.  This figure shows that THP-DC-SIGN transfects can mediate infection by several different pseudotyped HIV-1 varieties in susceptible cells.  Because no luciferase activity occurred when pre-infected THP-1 cells were added to 293T cells, I feel it is safe to suggest that it is the DC-SIGN on the THP- DC-SIGN cells that was allowing for the in trans infection of the 293T cells.

In Figure 4c, HIV-1 that expressed green fluorescent protein (GFP) was used to infect THP-DC-SIGN which were then transferred to activated T cell cultures.  The cells in this coculture were sorted based on expression of CD3 (expressed by T-cells but not THP-DC-SIGN cells) and presence of GFP.  All CD3(-) cells (i.e. DC-SIGN cells) were negative for GFP, but 5.2% of CD3(+) cells (i.e. T cells) were also positive for GFP.  Because THP-DC-SIGN cells were GFP(-), they were not infected by HIV-1.  But because a percentage of the T cells were infected, the THP-DC-SIGN cells must have been carrying the viruses, which subsequently infected the T cells.  I think this is a very powerful use of GFP and cell sorting techniques which provide believable support for the idea that DC-SIGN mediate the infection by HIV-1 in T cells.

Figure 5 demonstrates that DC-SIGN improves HIV-1 infection of T cells.  Figure 5a is very similar to 4a except the THP cells were not washed prior to addition of the T cells.  So, presumably, the HIV-1 in the medium is available to infect the T cells without the help of THP-DC-SIGN cells.  But when THP-1, CD4+ T cells, or no cells were mixed with T cells very little luciferase activity resulted.  However, when THP-DC-SIGN cells were used, high luciferase activity resulted.  This is strong evidence that DC-SIGN enhances infection.

Figure 5b is similar to Figure 4b, except that T cell were exposed directly to HIV-1 in medium.  Like in Figure 5a, little activity resulted when THP-1 cells or no cells were included in the medium whereas THP-DC-SIGN cells in the medium resulted in high luciferase activity.  This is further evidence of DC-SIGN’s relevance as a mediator of HIV-1 infection.  The fact that medium alone result in zero activity acts as negative control showing that cells must be present for infection to occur.

Figure 6 uses immunohistochemical analysis to show that dendritic cells in mucosal tissues express DC-SIGN and CD4 but not CCR5.  In sexual transmission of HIV-1 between humans, the virus enters the mucosal tissue of the sex organs, thus the study of the cells in these tissues is important to understanding HIV-1 infections.  In Figure 6a, cells from the cervix, rectum, and uterus were stained with anti-DC-SIGN mAb.  The reddish orange color in each of the three images from Figure 6a represents bound anti-DC-SIGN mAb.  If the antibody is infact binding DC-SIGN, then I believe that this is a very useful figure.  However, there was no negative controls performed on tissues known not to express DC-SIGN.  I would prefer to see a control so that I could be sure that the reddish-orange that I am seeing is not just background.

Figure 6b is the results of staining of mucosal cells of the rectum and uterus with anti-DC-SIGN, anti-CD4, and anti-CCR5.  Presumably, the reddish-orange in each image represents antibody that has bound its epitope.  Again, I have a problem with what is really antibody and what is just background.  In panels c and f it appears that there is some reddish-orange yet the authors claim that those images suggest that dendritic cells in mucosal tissue do not express CCR5.  Without a negative control I cannot say that I see what they say is there.  I feel that Figure 6 is the weakest figure in this paper.

Figure 7 demonstrates HIV-1 gp120’s ability to bind to DC-SIGN on THP-DC-SIGN for long period of time.  Figure 7a is a plot of percent gp120+bead/dendritic cell binding versus time as determined by FACScan analysis.  Binding continued for more than 60 hours, which suggests that, in vivo, HIV-1 that binds to DC-SIGN on dendritic cells in mucosal tissue could remain bound as the activated dendritic cell travels to a lymphoid organ where CD4 T cells are abundant.  The data for medium+anti-DC-SIGN suggests that although a smaller percentage of gp120 bound DC-SIGN, they binding was long lasting.  I think that this data is good evidence that gp120/DC-SIGN binding can last for a long time.

Figure 7b is a plot of luciferase activity versus time.  THP-DC-SIGN cells(with and without anti-DC-SIGN), THP-1 cells, and no cells were incubated with virus.  Each day some of the pre-infected cells (which had been washed) or medium containing only virus were added to 293T-CD4-CCR5 cells and luciferase activity was recorded.  This graph shows that added THP-DC-SIGN cells to T cell resulted in high luciferase activity up to the fifth day after pre-infection.  Results from THP-DC-SIGN+anti-DC-SIGN suggest that the virus did not bind DC-SIGN. Results from medium+virus might mean that HIV-1 virus might lose its infecting ability over time if not bound to DC-SIGN.

Figure 7c is a graphical model of what happens in vivo after HIV-1 enters mucosal tissue.  Based on data presented in Figure 7, the authors present their idea about how HIV-1 binds dendritic cells and travels from mucosal tissue to lymphoid organs where T cells are infected.

This paper raises some very interesting and potentially useful ideas about the nature of HIV-1 infection in humans.  I believe that the experiments performed do indeed support the researchers’ hypothesis that DC-SIGN expressed on dendritic cells in mucosal tissue mediates HIV-1 infection of T cells in the lymphoid organs.  Good controls were used, leaving few questions in the readers mind as to the validity of the arguments presented.  The data collected in the experiment provides many potential areas for new studies that could ultimately lead to new drugs that prevent HIV-1 infections.  Some possibilities for further study are presented below.

What Domains of DC-SIGN Actually Bind to HIV-1 gp120?
To answer this question would shed light on possible ways to prevent gp120/ DC-SIGN binding and, thus, potentially prevent infection with HIV-1.  To test this question a procedure similar to that used in Figures 1b and 1d could used.  A variety of transfected THP-type cells could be used.  Different deletions could be made in DC-SIGN and the pattern of gp120/DC binding would indicate which area are the most important for binding to gp120.  Transfect expressing the unaltered DC-SIGN could be used as a positive control, while THP-1 cells could be used as a negative control.

What is DC-SIGN’s normal function in the body?
If DC-SIGN is found to perform no vital function other than binding gp120 and mediating HIV-1 infections, then perhaps HIV-1 susceptibility could be decreased by deleting the gene that codes DC-SIGN or blocking DC-SIGN’s transcription in dendritic cells.  For obvious ethical reasons, humans could not be manipulated to determine the effects of deletion of the DC-SIGN gene, but perhaps a homologue in mice could be cloned.  The human DC-SIGN gene could be used to probe a mouse cDNA library from dendritic cells.  If a homologous gene was found, it could be knocked out by homologous recombination and the effect could be observed.  Probably, the gene has a function other than mediating HIV-1 infection because it would be selectively advantageous to express a gene that does nothing but cause infections.  But if the gene’s function proves to be minor or non-vital, genetically altered humans lacking DC-SIGN could one day be created that are immune to HIV-1 infection.

What changes occur in vivo, with regards to infection by HIV-1, when the DC-SIGN gene is altered or deleted?
By testing this question, it would be possible to determine if changing or deleting DC-SIGN can decrease the infectivity of HIV-1.  Again the ethics of human experimentation prevent using human subjects.  So, the mice homologue (if one exists) to the DC-SIGN gene could be knocked out and replaced by the human DC-SIGN gene by homologous recombination.  HIV-1 could then be injected into mice and the effectiveness of infection could be measured.  The effects of deletion or alterations of the DC-SIGN gene could be tested as well.  This method could prove difficult because of a human gene being used in mice and a human virus being used to infect mice.  This problem could avoided by transfecting a mouse virus that has similar infection properties to HIV-1 with the gp120 gene.  These experiments, if successful, would probably show that alterations in or deletion of the DC-SIGN gene result in decreased infectivity of HIV-1.  This information could be used to genetically alter DC-SIGN, thereby making individuals immune to HIV-1.

How do dendritic cells interact with T cells when in trans infection of HIV-1 occurs?
This question addresses yet another step in the HIV-1 infection process where drugs could potentially be used to interfere, thus preventing infection.  A procedure similar to that used in Figure 2b (measuring p24 levels) or Figures 3b, 4a, 4b, 5a, 5b (measuring luciferase activity) could be used to test this question.  Different combination of pre-HIV-1-infected transgenic dendritic could be cocultured with different transgenic T cells.  The transgenic cells could expresses different combinations of the cell-surface adhesion molecules that are normally expressed on dendritic and T cells.  Some of these molecules could include ICAMs and LFA-1.  If infection levels decrease with certain combinations of expressed cell-surface molecules, than the absent cell-surface molecules are probably important for the T cell/dendritic cell interaction that is required for in trans infection of the T cells by HIV-1.  This information could be useful in creating drugs to prevent this interaction and consequently, HIV-1 infection.

Molecular Biology Home Page