Review: Early dendritic cell-driven events governing the mucosal transmission of HIV: targets for vaccines and microbicides

By Melissa Pope, Ph.D.*

HIV breaches mucosal barriers as either cell-free or cell-associated virus, subsequently interacting with local leukocytes within the tissues to establish infection. The exact role of cell-free vs. cell-associated virus (and which cells within the inoculum are infected) in mucosal transmission remains unresolved. Most research has focused on transmission of cell-free virus, but more innovative investigations are needed to verify and characterize the nature of incoming infectious inoculum. However borne, in penetrating the epithelial surface HIV might either infect epithelial cells, pass between these cells or directly through breaks in the tissue, or transcytose epithelial cells (reviewed in 1). Subsequently, multiple white blood cell types within the tissues can be targeted by HIV, initiating the sequences of events facilitating the onset and dissemination of infection. The earliest events of virus-cell interplay, with an emphasis on the particular role of dendritic cells (DCs), will be discussed here, highlighting what needs to be considered to improve vaccine and microbicide strategies to prevent mucosal transmission and dissemination of HIV.

Initial leukocyte targets within the epithelial tissues

DCs, macrophages, and CD4+ T cells can all serve as targets for incoming HIV. Increasing evidence supports the notion that DCs positioned within and just underneath the epithelia are one of the first leukocytes to interact with the virus and that they are critical to the onset of infection (reviewed in 1-4) (Figure 1). Like macrophages and CD4+ T cells, immature DCs can be productively infected with CCR5-using HIV (R5 HIV). DCs have the additional attribute that they can also efficiently capture virions (independent of infection) and this, as well as virus newly produced by infected immature DCs, can be rapidly transmitted to CD4+ T cells (5, 6). Macaque studies documented that virus-positive Langerhans cells (LCs, the DCs in the outer stratified squamous epithelia of the vagina, anus, and oral mucosa) can be detected within the first 1-2d after vaginal exposure to SIV and virus-positive T cells appear 2-3d after mucosal (vaginal or oral) challenge (reviewed in 2, 3). This suggests that DCs entrap virus during the first moments after exposure, possibly replicating R5 HIV themselves (at least at low levels), before passing it to the more permissive CD4+ T cells (and macrophages) that then amplify infection both locally and in distal tissues (Figure 1).

Figure 1. Early events during sexual transmission of HIV. R5 and X4 viruses can cross the epithelia as (i) cell-free virus passing through the barrier and/or upon interaction with cells in the epithelium (DCs, T cells, epithelial cells) or (ii) cell-associated virus (not shown). The exact mechanisms determining why R5 infection dominates is not completely understood. DCs, T cells, and macrophages can be productively infected via a CD4/CCR5-dependent mechanism and DCs capture viruses (potentially R5 and X4) via CLRs. Cell-to-cell spread of virus likely occurs within the local epithelial tissues (not shown). Virus-carrying and -infected cells move via the afferent lymphatics to the draining lymph nodes, resulting in virus dissemination to and amplification in resident CD4 T cells. This is exacerbated by the migrated DCs being able to transmit virus extremely efficiently to T cells (captured viruses and those newly produced by the DCs) while activating poor anti-viral immune responses. Amplified virus (cell-free or as infected CD4 T cells) then move via the efferent lymphatics facilitating systemic infection. [View Larger Image]


DCs are in an immature state in healthy epithelial tissues, but an influx of DCs and DC maturation occurs in inflamed tissues. Although mature DCs are more resilient to productive infection by HIV, mature DCs readily capture HIV and transfer it to CD4+ T cells driving robust virus replication (reviewed in 2, 3). So inflammation caused by other sexually transmitted infections potentially results in mixtures of immature and mature DCs as well as the influx of CD4 T cells, thereby providing additional cellular targets to increase the chances of HIV transmission. Furthermore, DC activation can occur during migration from the mucosal tissues to the draining lymph nodes. If the migrating DCs are carrying HIV, the activation process could influence the subsequent fate of the virus: infectious virus might be retained for subsequent transmission, de novo synthesis of infectious virus might be shut down, or the captured virions might be degraded. Understanding how HIV interacts with different DC subsets (the activation state of which might have been modulated following exposure to other pathogens or factors) is therefore vital to identifying what needs to be tackled to restrict mucosal transmission.

An additional point to keep in mind is that R5 HIV isolates predominate in the earliest stages of infection, despite the likelihood of exposure to both R5 and X4 viruses. Although the exact mechanisms controlling this are not understood, the presence of numerous cellular targets capable of replicating R5 viruses might contribute at least in part to this phenomenon (DCs and macrophages are poorly infected by X4 viruses). However, the fact that DCs can capture both X4 and R5 HIV and transmit them to T cells to promote virus growth in the DC-T cell milieu (reviewed in 2, 3) suggests that X4 viruses should be able to establish infections in vivo. Indeed a recent report showed that upon co-exposure of macaques to a mixture of X4 and R5 SHIVs the animals became infected with both isolates, but that the R5 SHIV quickly dominated and the X4 SHIV receded coincident with the appearance of virus-specific CD8+ T cells (7). X4 SHIV re-emerged upon depletion of the CD8+ T cells. This suggests that at least one of the governing features includes the differential immune control of X4 and R5 viruses, enabling the outgrowth of R5 virus. The mechanism underlying this differential immune control needs to be elucidated to advance vaccine design—perhaps it is dictated by how the DC “presents” an X4 compared to an R5 virus?

Virus-cell interactions in the early moments

Virus interactions with most leukocyte receptors are largely dependent on virus envelope. While DCs (especially immature DCs) have terrific capacities to non-specifically ingest particles, they also appear to capture virus predominantly by envelope-dependent, receptor-mediated mechanisms (8). Productive infection of immature DCs with R5 HIV requires CCR5, probably involving the classical CD4/CCR5-dependent binding and virus-cell fusion events (reviewed in 2, 3). Down-regulation of CCR5 expression upon DC maturation at least partially explains the more limited infectibility of mature DCs. More recent evidence underscores how immature and mature DCs can also capture viruses via mannose-dependent C-type lectin receptors (CLRs) such as mannose receptor (CD206), DC-SIGN (CD209), and Langerin (CD207) (reviewed in 2, 3, 9-11). CLR-dependent capture of virus is very efficient and may augment CCR5-dependent infection of DCs in cis (12). However, CLR-entrapped virus is mostly internalized by the cells and subsequently transmitted to CD4+ T cells (in trans) in the absence of DC infection (6, 13). CLRs are differentially expressed on distinct DC subsets and as a result unique virus-CLR interactions occur with each subset (reviewed in 2, 3, 11). For instance, LCs lack CD209 but express CD207, while submucosal DCs lack CD207 yet express CD206 and CD209. So as well as being susceptible to infection with R5 HIV, immature DCs express a variety of CLRs that enable proficient entrapment of virus that can then be disseminated.

These two dominant modes of DC-virus interplay (CCR5-dependent and CLR-dependent) are manifest as two phases of transmission of virus from DCs to T cells (Figure 1). Using model monocyte-derived DCs (moDCs) to closely follow the kinetics of retention of infectious virus, we recently demonstrated the transfer of entrapped virus to T cells independent of DC infection (both immature and mature DCs) as well as the transfer of newly-synthesized virions by productively infected immature DCs (6). Comparable biology has been described using a cervical tissue explant model, where the tissue is exposed to HIV in vitro (in the presence or absence of specific blockers) (14). Actual infection of cells within the tissue can be monitored as well as the ability of the cells that migrate from the tissue—mimicking the migration to the lymphoid tissues—to transmit infection to permissive cells. Cells within the mucosal tissue explants are preferentially infected by R5 HIVs and blocking CD4/CCR5-dependent interactions between the virus and the mucosal cells prevents subsequent infection. In contrast, the ability of the migrated cells to transmit infection is not affected by CD4/CCR5 blockers, but is impaired when CLR-virus interactions within the tissues are inhibited (e.g., with mannan). Of note, the migrated CD3-HLA-DR+ fraction comprising numerous DCs is ultimately responsible for virus transmission.

Despite the considerable involvement of CD4, CCRs and CLRs in the various virus-DC interactions, virus capture by DCs is rarely blocked 100% by blocking strategies targeting these molecules (6, 14, 15). This may simply reflect only partial efficacy for these in vitro analyses and that we need to identify more effective CD4, CCR, and CLR blocking agents. However, determinants other than CD4, CCRs, and CLRs on DCs also likely contribute to virus-DC interplay and need to be considered when preventing DC-driven HIV spread.

Once the virus is trapped by DCs (by whatever mechanism) it can be very rapidly transmitted to neighboring T cells, exacerbating virus dissemination. Earlier work highlighted the ability of DC-T cell conjugates to amplify virus infection (16) and more recent studies revealed that virus transmission from DCs occurs preferentially to proliferating T cell subsets (17, 18). Strikingly, virus transfer from DCs to T cells has now been visualized (5, 6). Virus moves quickly to the synapse between the DC and T cell, where CD4 and CCR5 molecules congregate (5). Once virus has moved from the DCs to the T cells, there is T cell-T cell spread of virus (6) through CD4/CCR-dependent mechanisms (19). While some virus moving from the donor cell might be newly produced budding virions, whole internalized virus particles are also released at the contact points to fuse with the recipient T cell membrane (unpublished observations).

Virus exploitation of the immune system drives early spread

The natural function of DCs in the immune system is to capture pathogens, present them to the immune system and stimulate potent pathogen-specific immunity. But HIV manages to subvert the antigen-presenting cell (APC) system to favor infection instead of robust protective anti-viral immunity. Both immature and mature moDCs that have captured virus are able to stimulate virus-specific T cell responses in vitro (20-24). Notably, immature DCs preferentially stimulate CD4+ T cells while mature DCs induce both CD4+ and CD8+ responses (24). Therefore, when an immature DC entraps incoming HIV, virus-specific CD4+ T cells may get activated, but not CD8+ T cells—a response insufficient to eradicate infection. Additionally, virus-specific CD4+ T cells are more susceptible to infection (25) and so the activation of virus-specific CD4+ T cells might further augment virus dissemination (Figure 1).

Unlike mature DCs, immature DCs typically induce poorer Th1 effector responses, and in fact stimulate regulatory T cell (Treg) responses (26) that may dampen any virus-specific innate or adaptive responses elicited during primary infection. Recent work suggests that Tregs control the immune responses to HIV infection (27) and that natural Tregs are especially susceptible to HIV infection (28). Additionally, a recent report indicates that HIV-infected immature DCs favor the induction of IL-10 responses that would dampen Th1 immunity (29). Hence, by targeting immature DCs within the epithelial tissues, HIV avoids the activation of strong effector responses and favors the activation of Tregs, further limiting effective clearance of infection and, if anything, creating an even more permissive milieu for virus replication.

Adding to this is increasing evidence that determinants within the virus can modulate APC functions to drive virus infection while avoiding potent immune activation. Unlike other pathogens, HIV does not stimulate DC maturation and as such limits the likelihood that a virus-bearing DC will elicit strong effector immunity unless an exogenous DC stimulus is provided. In fact, HIV seems to hijack selective attributes of the APC machinery to favor its own replication. HIV Nef triggers DCs and macrophages to secrete chemokines and cytokines to attract additional T cells to the initial focus of infection (reviewed in 3), thereby providing more targets for virus amplification. Moreover, Nef-signaled macrophages activate B cells that in turn signal resting T cells to become permissive for HIV infection (30). Nef-expressing immature DCs also signal resting T cells and drive virus growth (31). It has been suggested that Nef modulates CD209 expression to promote DC-T cell contact needed to drive infection, although this is not seen in all Nef-bearing DCs (reviewed in 3). Despite this, Nef-bearing DCs do not up-regulate costimulatory molecules (an event typical for mature DCs and essential for effective immune stimulation) and therefore remain poor stimulators of anti-viral effector responses. Similar modulation of DC biology in the absence of classical phenotypic activation is also induced by Tat (32). As a result, HIV factors selectively exploit specific aspects of APC biology to encourage APC-T cell communication and drive virus spread while sidestepping the activation of effector immune responses.

Considerations for preventing HIV transmission

As we learn more about the early events of virus crossing the mucosal barriers, it is clear that we must bear in mind (i) the variety of cells with which HIV interacts, (ii) the multitude of receptors that are utilized by HIV to enable infection and/or entrapment, and (iii) the complexities of the subsequent efficient spread of virus between cells. As reviewed recently (1, 33), the immune system is faced with an enormous challenge within a relatively short window of time to control the initial stages of virus amplification. These challenges exist for the development of both vaccines and microbicides against HIV.

Whether it is a vaccine-elicited immune response or a topically applied microbicide, a broad acting strategy is needed to impede the wide array of different HIV envelopes from interacting with all potential cellular targets (and the various molecules on their surfaces). Anti-envelope approaches should limit most envelope-mediated interactions and act fairly broadly to this end. Passive transfer of neutralizing antibodies (NAbs) protects against intravenous SHIV challenge (34, 35), indicating the importance of NAbs in controlling infection. But NAb responses will most probably need to be elicited by vaccines at the mucosal surfaces to have significant impact in preventing transmission (36). NAbs probably have greater impact by preventing the infection of new targets, while cellular responses will be required to eradicate already infected cells (as in therapeutic strategies for infected individuals). It is vital that vaccines are presented appropriately to the immune system to ensure that potent effector T (and B) cell responses, not regulatory responses, are induced, such as can be achieved through targeting mature DCs (37). This might also require coordination of boosting innate responses by DCs (e.g., IFN-alpha, defensins) to assist directly in virus control and also enhance the activation of adaptive responses (37, 38). Thus, vaccines face the challenge of having to induce antigen-specific effector responses with wide specificities in order to clear infected cells and prevent new infections, as well as dealing with the ever-mutating virus.

While still a daunting task, microbicide strategies may be designed to target more generalized features of the virus or even host molecules and thereby be less restricted by the continuously evolving virus. For instance, anti-envelope NAbs protected at least 70% of monkeys against vaginal infection with SHIV (34, 39). These data provide proof of principle that mucosal transmission can be impeded by blocking envelope-host interactions (as well as emphasizing the importance of inducing mucosal Ab responses through vaccination). In agreement with the need for broad-acting modalities, the negatively-charged sulfated polysaccharides like Carraguard (a carrageenan-based formulation) represent a promising approach to potentially interfere with all virus-cell interactions as well as cell-to-cell spread through their (charge-based) non-specific actions. In fact, Carraguard significantly impaired virus capture by immature and mature moDCs (unpublished observations) and protected approximately 70% of the monkeys vaginally challenged with infectious SIV (David Phillips and Louis Martin, personal communication). Not surprisingly, just blocking CCR5 (with a single CCR5 inhibitor) had a less dramatic effect, preventing vaginal SHIV infection in only 2 of 11 macaques (40) (although the viral replication in all animals was reduced compared to the control group). Therefore, broad-acting and/or combinatorial approaches will probably exert the most effective preventive microbicide strategies.

In summary

The primary events of cell-virus interactions following the immediate penetration of the epithelial barrier are multifaceted, involving multiple cell types that express a variety of molecules to bind virus (CD4, CCRs, CLRs, others). Moreover, cell-to-cell spread of HIV is especially efficient and the tight junctions between the cells may afford “protection” for the virus being transmitted, making it difficult to block critical interactions. Adding to this, the immune system is exploited by HIV to foster the stimulation of suboptimal immunity, further exacerbating the intricacies of the onset of infection. Defining the complexities of these events will help develop vaccine and microbicide modalities with sufficient strength and breadth that are needed to limit HIV transmission and dissemination.


1. M. Pope, A. T. Haase, Nat. Med. 9, 847 (2003). PubMed  
2. I. Frank, M. Pope, Curr. Mol. Med. 2, 229 (2002). PubMed  
3. N. Teleshova, I. Frank, M. Pope, J. Leukoc. Biol. 74, 683 (2003). PubMed  
4. C. J. Miller, R. J. Shattock, Microbes Infect. 5, 59 (2003). PubMed  
5. D. McDonald, L. Wu, et al., Science 300, 1295 (2003). PubMed  
6. S. G. Turville, J. J. Santos, et al., Blood 103, 2170 (2004). PubMed  
7. J. M. Harouse, C. Buckner, et al., Proc. Natl. Acad. Sci. USA 100, 10977 (2003). PubMed  
8. I. Frank, M. J. Piatak, et al., J. Virol. 76, 2936 (2002). PubMed  
9. S. G. Turville, Arthos, et al., Blood 98, 2482 (2001). PubMed  
10. S. G. Turville, J. Arthos, et al., J. Clin. Virol. 22, 229 (2001). PubMed  
11. S. Turville, J. Wilkinson, et al., J. Leukoc. Biol. 74, 710 (2003). PubMed  
12. B. Lee, G. Leslie, et al., J. Virol. 75, 12028 (2001). PubMed  
13. D. S. Kwon, G. Gregorio, et al., Immunity 16, 135 (2002). PubMed  
14. Q. Hu, I. Frank, et al., J. Exp. M. 199, 1065 (2004). PubMed  
15. S. Gummuluru, M. Rogel, et al., J. Virol. 77, 12865 (2003). PubMed  
16. M. Pope, M. G. H. Betjes, et al., Cell 78, 389 (1994). PubMed  
17. S. Gummuluru, V. N. KewalRamani, et al., J. Virol. 76, 10692 (2002). PubMed  
18. M. Sugaya, K. Lore, et al., J. Immunol. 172, 2219 (2004). PubMed  
19. C. Jolly, K. Kashefie, et al., J. Exp. Med. 199, 283 (2004). PubMed  
20. F. Buseyne, S. L. Gall, et al., Nat. Med. 7, 344 (2001). PubMed  
21. S. M. Santini, C. Lapenta, et al., J. Exp. Med. 191, 1777 (2000). PubMed  
22. M. Larsson, J. F. Fonteneau, et al., AIDS 16, 1319 (2002). PubMed  
23. M. Larsson, D. T. Wilkens, et al., AIDS 16, 171 (2002). PubMed  
24. I. Frank, J. J. Santos, et al., J. Acquir. Immune Defic. Syndr. 34, 7 (2003). PubMed  
25. D. C. Douek, J. M. Brenchley, et al., Nature 417, 95 (2002). PubMed  
26. R. M. Steinman, M. C. Nussenzweig, Proc. Natl. Acad. Sci. USA 99, 351 (2002). PubMed  
27. E. M. Aandahl, J. Michaelsson, et al., J. Virol. 78, 2454 (2004). PubMed  
28. K. Oswald-Richter, S. M. Grill, et al., PLoS Biol. 2, E198 (2004). PubMed  
29. A. Granelli-Piperno, A. Golebiowska, et al., Proc. Natl. Acad. Sci. USA 101, 7669 (2004). PubMed  
30. S. Swingler, B. Brichacek, et al., Nature 424, 213 (2003). PubMed  
31. D. Messmer, J.-M. Jacqué, et al., J. Immunol. 169, 4172 (2002). PubMed  
32. E. Izmailova, F. M. Bertley, et al., Nat. Med. 9, 191 (2003). PubMed  
33. C. D. Pilcher, J. J. Eron, et al., J. Clin. Invest. 113, 937 (2004). PubMed  
34. J. R. Mascola, Vaccine 20, 1922 (2002). PubMed  
35. W. Xu, R. Hofmann-Lehmann, et al., Vaccine 20, 1956 (2002). PubMed  
36. J. R. Mascola, Curr. Mol. Med. 3, 209 (2003). PubMed  
37. M. Pope, Curr. Mol. Med. 3, 229 (2003). PubMed  
38. C. A. Biron, Immunity 14, 661 (2001). PubMed  
39. R. S. Veazey, R. J. Shattock, et al., Nat. Med. 9, 343 (2003). PubMed  
40. R. S. Veazey, P. J. Klasse, et al., J. Exp. Med. 198, 1551 (2003). PubMed 

*Melissa Pope is a scientist at the Center for Biomedical Research at the Population Council in New York researching dendritic cell biology in the context of HIV transmission to improve microbicide and vaccine strategies for prevention of HIV.