|Home | About | Journals | Submit | Contact Us | Français|
Dendritic cells (DCs) are crucial in the generation and regulation of adaptive immunity. Given their pivotal role in marshalling immune responses, HIV has evolved ways to exploit DCs to facilitate viral dissemination and to evade antiviral immunity. Defining the mechanisms that underlie cell–cell transmission of HIV and understanding the role of DCs in this process should help us in the fight against HIV infection. This Review highlights the latest advances in our understanding of the interactions between DCs and HIV, and focuses on the mechanisms of DC-mediated viral dissemination.
In 2005, the total number of people living with HIV was estimated to be 40.3 million — the highest level so far — and resulted in 3.1 million deaths from AIDS. Moreover, approximately 4.9 million people became newly infected with HIV in 2005 1, so the fight against HIV infection and transmission has some way to go. Heterosexual transmission is the main route of dissemination of HIV infection worldwide and accounts for 80% of HIV infections 1, 2. Given that dendritic cells (DCs) are located in the mucosa (including oral and vaginal mucosal surfaces) and lymphoid tissues, they are proposed to be among the first cells that encounter HIV type 1 (HIV-1, referred to as HIV from this point) during sexual transmission. It has also been suggested that DCs mediate the spread of HIV to CD4+ T cells in lymphoid tissues in vivo, which are the main source of HIV replication and dissemination (reviewed in 2-6) (FIG. 1). So, understanding the mechanisms of HIV interactions with DCs and cellular receptors will hopefully facilitate the development of more effective interventions against HIV infection and transmission, and potentially aid in the development of novel strategies for the development of HIV vaccines.
It has been known for more than a decade that HIV-pulsed DCs facilitate viral infection of co-cultured T cells 7, 8. Recent studies of DC–HIV interactions have highlighted an important role for DCs in HIV transmission at mucosal surfaces and in viral pathogenesis. These studies have also revealed several potential mechanisms underlying DC-mediated HIV transmission. One route for viral transfer has been reported to occur through the ‘infectious synapse’ (also termed the ‘virological synapse’)that is formed between DCs and T cells 9, 10; another recently identified pathway suggests that transinfection of HIV may be mediated by DC-derived exosomes 11. Alternatively, infective HIV virions produced in HIV-infected DCs may transmit infection in cis 12-16. Further studies on the significance and relative contributions of these pathways in viral transmission in vivo may shed light on HIV pathogenesis.
This Review focuses on the interactions of HIV with different subsets of DCs, outlining the potential mechanisms of viral dissemination and the relevance to viral immunopathogenesis.
One of the enigmatic features of DC biology is the complexity of their subsets. Unlike other immune cells, such as T, cell, B cells and natural killer cells, DCs have either a myeloid or lymphoid origin17, 18. In total, DCs are a relatively rare population of cells in the blood or tissues. DC populations can be divided into several subsets based on their anatomical distribution, immunological function and expression of cell-surface markers. The main immunological functions, the susceptibility to HIV infection and their capacity to spread viral infection are summarized for each DC subset, in TABLE 1.
The main DC subsets include myeloid DCs and plasmacytoid DCs (pDCs) in the blood, and Langerhans cells in the tissues (TABLE 1). Myeloid DCs and pDCs are present at low frequencies, representing 0.5–2% of total peripheral blood mononuclear cells (PBMCs) 18, 19. In general, myeloid DCs are characterized by their ability to secrete high levels of interleukin-12 (IL-12), whereas pDCs can prime antiviral adaptive immune responses by producing high levels of type 1 interferons 20, 21. Langerhans cells are mainly found in the skin and the stratified squamous epithelia; they comprise 2%–3% of epidermal cells 17, 22. Langerhans cells express langerin (CD207), which is a Langerhans-cell-specific C-type lectin 23.
Myeloid DCs, pDCs and Langerhans cells are all susceptible to infection with HIV 12, 24-35. Each of the DC subsets express relatively low levels of the HIV receptor CD4, and the co-receptors CC-chemokine receptor 5 (CCR5), CXC-chemokine receptor 4 (CXCR4), CCR3, CXCR6 (also known as Bonzo, STRL33), CCR8 and CCR9 36-40. Although both myeloid DC and pDC subsets are susceptible to laboratory-adapted R5 and X4 HIV isolates, myeloid DCs are more susceptible to R5 HIV infection than pDCs from the same donor 35. Nevertheless, R5 HIV strains infect myeloid DCs much more efficiently than X4 HIV strains, probably because they express higher levels of CCR5 than CXCR412, 29, 30, 33, 35.
Compared with CD4+ T cells, HIV replication in DCs is generally less productive 5, 12, 30, 31, 33, 44, 45, and the frequency of HIV-infected DCs in vivo is often 10- to 100-fold lower 46. In fact, a study of cells from healthy donors indicated that on average only 1–3% of myeloid DC and pDC populations can be productively infected with HIV in vitro, as detected by intracellular staining of the HIV protein p24 35. A similar low frequency of in vitro HIV infection of Langerhans cells isolated from healthy individuals has also been reported 12. The presumed reasons for moderate HIV infection of DCs include the following: low levels of expression of HIV receptor and co-receptors; rapid and extensive degradation of internalized HIV in intracellular compartments13, 14, 47; and expression of host factor(s) that block HIV replication. An example of such a host factor is the antiretroviral protein APOBEC3G (apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like 3G), which has been shown to block post-entry HIV replication in resting CD4+ T cells and may be an important innate immune mechanism against retroviral infection 120. Nevertheless, further investigation is required to better understand how DCs restrict productive HIV infection.
FDCs are not considered to be a typical DC subset given that they are not derived from the bone marrow and are not known to process and present antigens through MHC-restricted pathways 17, 22. FDCs usually can only be found in the B-cell follicles and germinal centres of peripheral lymphoid tissues. FDCs have the unique capacity to trap pathogens, including HIV, and retain infectious viruses for long periods 48, 49. FDCs can trap and maintain large quantities of HIV early after infection, thereby establishing an insidious reservoir of infectious virus proximal to highly susceptible CD4+ T cells in lymphoid tissues 50,53. Notably, FDCs are not productively infected by HIV; rather, FDCs harbour a large and relative stable pool of virions at their surface48, 49, 51, 52. In addition, FDCs can promote the migration of resting T cells to the germinal-centre microenvironment and this may further facilitate HIV infection of T cells and thereby contribute to HIV pathogenesis 49, 52.
In addition to FDCs, HIV-infected DCs may act as viral reservoirs for persistent infection in vivo. A recent study reported that HIV-infected monocyte-derived DCs (MDDCs) could productively replicate viruses for up to 45 days 54. However, it remains to be determined whether DCs can provide a cellular reservoir for latent infection, and whether they could be a potential target for eradication of the virus.
Due to the low abundance of DC populations in vivo, to model the immunological function of DCs and study DC–HIV interactions, MDDCs are commonly used in experimental studies in vitro 55. CD14+ monocytes from human peripheral blood differentiate into immature DCs after 4–6 days in culture in the presence of IL-4 and granulocyte-macrophage colony-stimulating factor (GM-CSF). MDDCs share characteristics with myeloid DCs, immature dermal DCs and interstitial DCs, and they express high levels of the cell-surface markers MHC class II molecules, CD11c, CD25 and DC-specific intercellular adhesion molecule 3 (ICAM3)-grabbing non-integrin (DC-SIGN; also known as CD209) 56. Immature MDDCs can be converted into mature MDDCs by various stimuli, including lipopolysaccharide, interferon-γ, tumour-necrosis factor and CD40 ligand (FIG. 2) 17, 22, 57. It has been observed that different stimuli used to engender DC maturation may result in the generation of DCs with distinct HIV transmission abilities 58.
Despite their convenience, MDDCs do not fully mimic the immunological function of DC subsets in HIV infection in vivo 38, 59-61. Therefore, experimental observations made using MDDCs and the physiological implications drawn from such studies should be interpreted carefully. For example, C-type lectins expressed on the surface of MDDCs accounted for more than 80% of the binding to HIV envelope glycoprotein gp120, however, freshly isolated and cultured blood DCs only bind gp120 through CD4 and not through C-type lectins 38. So, although MDDCs have provided insights into DC biology and HIV interactions, examination of bona fide DC subsets in vivo for HIV capture, transmission and antigen-processing pathways would be beneficial for understanding of the true contribution of DCs to viral pathogenesis.
Modulation of DCs by HIV infection, in particular by modulation or interference of the antigen-presenting function of DCs, is a key aspect in viral pathogenesis and contributes to viral evasion of immunity. Compared with DCs from healthy donors, DCs derived from the peripheral blood of HIV-infected individuals at different stages of infection have significantly reduced efficiency in stimulating allogeneic T cells 62, 63. It has been reported that DC-SIGN+ DCs in acute HIV infection have reduced expression of the co-stimulatory molecules CD80 and CD86, and this may influence DC-induced T-cell responses 64. Consistent with this, HIV-infected MDDCs43 and myeloid DCs67 fail to mature in culture, and may stimulate production of the regulatory cytokine interleukin-10 (IL-10) by T cells, thereby promoting an immunosuppressive response43. These results suggest that productive HIV infection of myeloid DCs undermines the direct induction of T-cell-mediated immunity. By contrast, some studies have indicated that MDDCs from HIV-infected individuals can efficiently induce CD8+ cytotoxic T lymphocyte (CTL) responses to various antigens 65, 66, suggesting that modulation of expression of co-stimulation molecules by HIV-infected DCs might not necessarily subvert the priming of CTLs by DCs. In addition, Smed-Sorensen et al. did not observe functional defects in cytokine production following stimulation of HIV-infected myeloid DCs and pDCs 35.
HIV regulatory and accessory proteins, such as Nef and Tat, have been reported to have various effects on immature DCs, although again the results are not always consistent. Messmer and colleagues reported that expression of adenovirus-vector-encoded Nef by immature MDDCs triggers the production of various cytokines and chemokines and stimulates T-cell activation, but this occurs without upregulating the expression of DC maturation markers 68. Similarly, HIV infection or adenovirus-vector-mediated expression of Tat by immature DCs can induce interferon-responsive gene expression without inducing DC maturation71. By contrast, exposure to recombinant (Escherichia coli-expressed) Nef or Tat protein led to efficient DC maturation, including upregulation of expression of DC maturation markers 69, 70. Tat may have a role in viral pathogenesis, by inducing the production of T-cell and macrophage chemoattractants by immature DCs, and thereby may facilitate the spread of HIV infection 71. Last, Nef protein can downregulate the expression of MHC class I and CD1a molecules by immature MDDCs 72,73, which may impair antigen presentation and contribute to immune evasion of HIV.
Using the simian immunodeficiency virus (SIV) and a rhesus macaque intravaginal transmission model, several studies have suggested that SIV can rapidly penetrate the vaginal mucosa after viral exposure and then infects or associates with intraepithelial DCs, which can mediate viral transmission to CD4+ T cells 75-77. In one study of human cervical explants exposed to HIV, it was shown that migratory DC populations from the cervical tissue account for 90% of HIV dissemination 78. These results suggest that DCs may play an important role in HIV transmission from mucosal surfaces 2, 4. Langerhans cells in the mucosal epithelium have been proposed to be initial targets for HIV infection by sexual transmission 29, although this remains controversial. In addition to these Langerhans cells and immature DCs in the subepithelial mucosal tissues, CD4+ CCR5+ T cells and macrophages that reside in the subepithelium might also be early targets for HIV infection and transmission in vivo (reviewed in 2,79).
The transfer of virus from DCs to CD4+ T cells involves three discrete steps. First, DCs bind and capture HIV, the virus then traffics within the DC, before it is transferred to the CD4+ T cells in a process known as trans-infection. In the following section, we consider HIV trans-infection mediated by DCs that are not themselves infected and later we discuss the transfer of progeny virus from infected DCs to T cells, a process known as cis-infection.
DCs have a unique membrane transport pathway that facilitates the uptake pathogens to initiate adaptive immune responses. DCs constitutively take up extracellular fluid by macropinocytosis, and engulf antigens and whole pathogens by endocytosis and phagocytosis mediated by expression of Fc receptors and C-type lectins, such as DC-SIGN 17, 22, 56. C-type lectins are the main HIV attachment factors expressed by dermal and mucosal DCs. Four types of C-type lectin can bind to HIV gp120: DC-SIGN, langerin (CD207), mannose receptor (CD206) and an unidentified trypsin-resistant C-type lectin 38, 59; no single C-type lectin is fully responsible for HIV binding on all DC subsets. Binding of gp120 to DCs migrating from the tonsils seems to be CD4 dependent. By contrast, HIV capture by Langerhans cells is partly mediated by langerin 59. For DC-SIGN-dependent HIV binding, interactions between gp120 and the carbohydrate-recognition domain of DC-SIGN are required for virus capture 80, 81 (Box 1).
DC-SIGN (dendritic cell (DC)-specific intercellular adhesion molecule 3-grabbing nonintegrin) is a C-type lectin (calcium dependent) 56, 127. Human DC-SIGN expression in vivo has been reported in DCs 56, 87, 88, 93, 100, macrophages 61, 84, 93, 94, activated B cells128, the skin dermis 56, 59, placenta 99, intestinal and genital mucosae 88, 100, and lymphoid tissues 56, 61, 64, 101, 129. Interaction of the carbohydrate-recognition domain (CRD) of DC-SIGN with the HIV envelope (Env) glycoprotein gp120 is required for virus binding 80. CRD interactions with HIV depend on the high proportion of high-mannose oligosaccharides present in the Env glycoproteins 81, 130, 131. DC-SIGN functions as a tetramer. This multimerization depends on repeated sequences in the extracellular neck repeat domain, which stabilizes tetramers of DC-SIGN and increases ligand-binding avidity 130, 132.
The cytoplasmic tail of DC-SIGN contains three defined internalization motifs, a dileucine (LL)-based motif, a tri-acidic (EEE) cluster and a tyrosine (Y)-based motif. These motifs may contribute to DC-SIGN-mediated internalization of HIV 131. DC-SIGN-bound HIV can enter cells by endocytosis, some viruses may escape degradation in lysosomal compartments and usurp endosomal trafficking in transmission to CD4+ T cells, resulting in productive HIV trans-infection 11, 97. DC-SIGN facilitates the uptake of viral antigens for MHC-class-I- and MHC-class-II-restricted antigen presentation, leading to activation of antiviral-specific CD8+ and CD4+ T cells 47, 115. HIV Nef can upregulate DC-SIGN expression by DCs and enhances clustering of DCs with T cells, which promotes HIV trans-infection 133. The figure is modified with permission from REF. Pohlmann et al. Trends Immunol. 22, 643-646 (2001) Elsevier B.V. © 2001.
DC-SIGN-mediated HIV transmission seems to require steps beyond simple virus binding and sequestration 82, 83, suggesting that DC-SIGN binding and transfer functions can be dissociated. So, despite identification of endogenous expression of DC-SIGN or other C-type lectins by several cell types in vivo, these cells may not be able to support HIV trans-infection. DC-SIGN expression has been shown to increase direct HIV infection of DC-SIGN-transfected cells 84, 85, but may inhibit HIV Env-mediated cell-cell fusion 86, suggesting that competition between CD4 and DC-SIGN for binding to gp120 probably affects the access of the virus to the cytosol and the formation of syncytia.
A molecular clue to DC-mediated HIV transmission was provided by studies of DC-SIGN-expressing cells 56, 80. A subset of cells in human blood (0.01% of total PBMCs) was isolated on the basis of co-expression of DC-SIGN and the monocytic marker CD14; these DC-like cells were reported to be able to increase HIV trans-infection of T cells 87. Similarly, Gurney et al. reported that DC-SIGN-expressing cells that were isolated from the rectal mucosa and that seemed to have an immature phenotype could efficiently bind and transfer HIV to CD4+ T cells 88. These DC-SIGN+ cells comprise only 1–5% of total mucosal mononuclear cells, but they were shown to contribute more than 90% of virus binding. Binding of HIV was mainly mediated by DC-SIGN, as DC-SIGN-specific antibodies could block virus binding, when more physiological amounts of virus were inoculated 88.
Although initial observations suggested that DC-SIGN is exclusively expressed by DCs to enhance HIV trans-infection and to stimulate T-cell responses 56, 80, subsequent studies have indicated that DCs also have DC-SIGN-independent mechanisms of HIV trans-infection of CD4+ T cells 61, 85, 89-92. Blockade of DC-SIGN with specific antibodies and small interfering RNAs indicated that MDDCs and blood myeloid DCs do not require DC-SIGN to transmit HIV 61. Moreover, DC-SIGN is abundantly expressed by macrophages rather than by DCs in normal human lymph nodes 61. High levels of DC-SIGN expression by macrophages in lymph tissues or in the blood suggests that it might aid in HIV trans-infection by macrophages in vivo 61, 84, 93, 94; however, this is yet to be confirmed. Notably, similar to MDDCs, human monocyte-derived macrophages show DC-SIGN-independent HIV transmission 95, 96, possibly involving other C-type lectin molecules, such as the mannose receptor.
HIV trafficking in DCs seems to be a critical step in viral transmission and may also be DC-SIGN independent. Kwon and colleagues observed that HIV that was internalized after DC-SIGN binding retained infectivity and could be transmitted to target CD4+ T cells 97. DC-SIGN mediates rapid internalization of intact HIV into a low pH, nonlysosomal compartment 97. However, it is not clear whether internalized HIV virions continue to associate with DC-SIGN and whether DC-SIGN mediates recycling of intact viruses back to the cell surface. The precise intracellular trafficking and localization of internalized HIV within DCs remains to be elucidated.
DC-SIGN-mediated HIV transmission is cell-type dependent 83, 85, 98 and requires cell-cell contact 83. Interestingly, B-cell transfectants expressing DC-SIGN mediate HIV-trans-infection as efficiently as DCs, whereas monocytic transfectants do not 98. In related studies, we observed that co-cultures of different cell types may interfere with DC-SIGN-mediated HIV trans-infection83. These findings reinforce the notion that the cellular environment is an important consideration when examining transmission of HIV captured by DC-SIGN or other attachment factors.
Despite the detection of DC-SIGN expression by DCs in different human tissues 56, 61, 64,88, 93, 99-101, none of the DC subsets myeloid DCs, pDCs, Langerhans cells and FDCs have been clearly shown to express DC-SIGN in vivo 33, 38, 56, 59, 61, 102, suggesting that the role of DC-SIGN in DC-mediated HIV transmission may be more limited in vivo. Some submucosal DC populations at sites of initial HIV infection are DC-SIGN-positive and seem to interact with HIV quite efficiently 59, 88, 100. However, given that they reside beneath the mucosal epithelium, in the absence of breaches in the mucosal barrier, DC-SIGN-expressing cells are unlikely to be the first cells to contact HIV in vivo. Thus, the importance of HIV dissemination by DCs and the role of DC-SIGN in the transmission process remain unclear, particularly in vivo.
Unlike mature DCs, immature myeloid DCs are specialized to take up antigens at sites of infection. Following capture of antigens, immature DCs traffic to lymphoid tissues where they develop into mature DCs, which are potent stimulators of naive T cells owing upregulation of expression of MHC class II and co-stimulation molecules 17, 22 (TABLE 2).
HIV transmission efficiency can be enhanced by maturation of DCs9, 30, 58, 103, suggesting that mature DCs in lymphoid tissues may facilitate infection of lymphoid-tissue-resident CD4+ T cells. However, the mechanisms underlying this increased viral transmission have not been well defined. Downmodulation of antigen uptake and degradation in mature DCs, as well as more efficient interactions between mature DCs and T cells may contribute to the enhanced HIV transmission efficiency 9, 58. Depending on the activating signals they receive, immature DCs can develop into different subsets of mature DCs with differing capabilities in HIV transmission and T-cell activation58. Increased ICAM1 expression correlates with the increased viral transmission by mature DC subsets, possibly due to stronger DC-T-cell interactions through ICAM1 binding to T-cell-expressed leukocyte function-associated molecule 1 (LFA1) 58 (FIG. 2). Moreover, different cellular trafficking of HIV within immature DCs and mature DCs may also contribute to differences in HIV transmission potential (L.W. and V.N.K., unpublished observations). Consistent with this, mature DCs have been reported to contain high levels of intact virions in large vesicular compartments with a perinuclear localization, whereas immature DCs retain few intact virus particles in endosomes close to the plasma membrane 106.
DC maturation is associated with a diminished ability to support HIV replication, being 10- to 100-fold lower than immature DCs 30, 42 (TABLE 2). But, HIV-pulsed mature DCs have been shown to contain 15-fold more viral DNA than immature DCs, which initially suggested that virus entry was not impaired 42. However, a recent study indicated that the defect in HIV replication observed in mature DCs results at least partially from decreased viral fusion 104. By analysing the phases of viral replication, Bakri et al. reported that DC maturation does not affect HIV reverse transcription, nuclear import and integration, suggesting that the reduced viral replication in mature DCs is due to post-integration blocks at the transcriptional level 105. This might account for the high levels of viral DNA detected in mature cells but their inefficient release of viral virions.
Current evidence suggests that DC-mediated HIV infection can occur by several distinct and co-existing processes. These processes include trans-infection through infectious synapses and exosome-associated viruses, as well as cis-infection following de novo viral production in DCs (FIG. 3).
Previous studies with MDDCs indicated that cell–cell contact is required for efficient stimulation of CD4+ T-cell infection 108. More recently studies have revealed that HIV transmission may occur across the infectious or virological synapse 9, 10, 13, 109. The structure of the infectious synapse may have similarities to the immunological synapse, which is formed between antigen-presenting cells and their T-cell conjugates 110. The cell-surface molecules that contribute to the infectious synapse and potentially support the transfer of HIV from DCs to CD4+ T cells have not been completely identified. HIV itself and viral receptors are found concentrated at the infectious synapse and DC-SIGN molecules are also detected 10, 109; however, it is unclear whether these or any other molecules are essential to promote formation of the infectious synapse. Interestingly, suppression of DC-SIGN expression has been shown to impair the infectious synapse formation, and to inhibit trans-infection of X4 HIV to T cells 10, 107. We have reported that the expression of MHC class II molecules on virus donor cells are not required for efficient DC-SIGN-mediated HIV transmission 83, suggesting that virus transmission to CD4+ T cells can occur in the absence of the classically defined immunological synapse. Indeed, human fibroblasts expressing HIV receptors can serve as effective target cells for DC- or DC-SIGN-mediated HIV transmission 90, 97; and an earlier study has shown that even DCs from mice (which can not be infected with HIV) are able to transmit HIV to human T cells 7. These results suggest that the donor- and target-cell requirements for infectious synapse formation are minimal. Under conditions with large virus inoculums and in cells placed in T-cell co-cultures soon after exposure to HIV, it is likely that infectious-synapse-mediated HIV transfer is responsible for most of the rapid and efficient viral transfer between DCs and CD4+ T cells in vitro. However, in vivo, it is likely that different mechanisms of transmission also contribute or even predominate.
Wiley and Gummuluru recently showed that DC-derived exosomes can mediate HIV trans-infection 11. HIV captured by immature MDDCs11 and mature MDDCs109 is rapidly internalized into endosomal multivesicular bodies (MVBs), which are endocytic bodies that are enriched with tetraspanin proteins. Intriguingly, some of the endocytosed HIV particles are constitutively released into the extracellular milieu in endocytic vesicles (known as exosomes)111, which can fuse with target-cell membranes to deliver infectious virus. The infectious virus associated with cell-free exosomes is a fraction of the virus measured during synaptic transmission, but, importantly, they may provide a pathway by which intracellular virus reaches the infectious synapse. The remaining MVB-resident HIV in DCs may enter the lysosomal pathway and be degraded. Therefore, the exosome-release pathway may enable HIV to circumvent immune destruction after capture by DCs. Although further study is required, it is intriguing to note that the exosome-associated HIV particles released from immature MDDCs have been reported to be 10-fold more infectious than cell-free viruses on a per-particle basis 11.
In productively infected macrophages, endocytosis and exosome trafficking pathways may be used for HIV particle assembly and release112,113,114. It will be interesting to determine whether in vivo DC subsets of myeloid DCs, pDCs and Langerhans cells also use the exosomal pathway to promote HIV trans-infection. The relative importance of this pathway of trans-infection remains to be established compared with the infectious-synapse-mediated pathway.
Although direct HIV infection is less efficient in DCs than in CD4+ T cells 30, 42, increasing evidence indicates that long-term HIV transmission mediated by DCs depends on viral production by the DCs. It has been reported that immature DCs and some DC-SIGN-expressing cell lines can retain infectious virus for up to 6 days after exposure to HIV41, 80, 85. However, recent studies have indicated that most incoming viruses are degraded in MDDCs within 24 hours13, 47, 115. Therefore, after 24 hours, virus that is transmitted from DCs to T cells must be newly synthesized progeny virus. So, two phases of DCs transfer of HIV to CD4+ T cells might exist13: the first phase (within 24 hours after exposure to HIV) might involve trafficking of captured virus from the endolysosomal pathway to the DC–T-cell synapse; the second phase (24-72 hours after viral exposure) would involve de novo replication of virus in DCs. It should be pointed out that these two processes are not necessarily sequential or interdependent.
Other recent findings also support the need for HIV replication in DCs for long-term viral transmission 14, 16. Nobile and colleagues reported that transfer of incoming virions from immature MDDCs or DC-SIGN-expressing cells occurs only within a few hours after viral exposure, indicating that there is no long-term storage of original infectious HIV particles in these cells. However, a few days after viral exposure, replicative viruses can also be transmitted to CD4+ cells by infected DCs14. Similar to these studies withMDDCs, Lore and colleagues reported that soon after HIV exposure, pDCs and myeloid DCs from human blood can transfer the virus to T cells in the absence of a productive infection. However, once a productive infection is established in the DCs, newly synthesized virus is predominantly spread to T cells 15. These results indicate that HIV-infected DCs in vivo may act as viral reservoirs during the migration to the lymphoid tissues to spread viral infection.
Both phases are likely to contribute to in vivo transfer of virus from DCs to T cells; however, given the low levels of HIV replication in DCs and the low frequency of DCs in vivo, the significance of DC transmission of de novo synthesized virus in vivo remains to be examined. Our in vitro results with replication-defective, single-cycle reporter HIV suggest that viral transmission can occur efficiently in the absence of viral replication 9, 83, 89, 90, 98. So, rare HIV replication in DCs may not be necessary to facilitate viral dissemination if the immediate viral transfer has been established through infectious synapses.
In opposition of immune sentinel function of DCs to capture and present processed pathogens, HIV hijacks DCs to efficiently promote viral dissemination. Elucidating HIV interactions with DCs will be vital in uncovering their contribution to HIV pathogenesis. An improved understanding of DC biology may also facilitate studies of DC interactions with HIV and other human pathogens. It will be crucial to dissect and compare the viral antigen processing and presentation pathways in HIV-infected DCs and in DCs bearing internalized virions. In addition, the type and tissue source of DCs from humans and non-human primate models will be an important consideration in these analyses. Because CD4+ T-cell populations in the gastrointestinal tract suffer the most marked depletion in both acute and chronic stages of HIV infection 116, 117, it will be critical to characterize DC populations in the gastrointestinal tract for their susceptibility to HIV infection and ability to promote viral transmission to CD4+ T cells in gut mucosa.
Viral assembly and budding from infected DCs is a poorly understood area of HIV research that also requires further study. An early study reported that new HIV virions could bud from the surface of infected MDDCs, although this was a rare finding 28. In infected macrophages, MVB-like compartments seem to be important for viral assembly and budding112, 113, 121, but in infected DCs, the involvement of these compartments remains to be proven.
Although in this review we have focused on the role of DCs in the dissemination of virus and how HIV may hijack DC functions to their advantage, DCs are of course potent stimulators of immune responses and therefore have been used in prophylactic and therapeutic approaches for HIV infection. For example, transfer of autologous MDDCs loaded with the inactivated, whole HIV was shown to suppress viral load in individuals chronically infected with HIV125. By contrast, a pilot clinical trial using HIV-peptide-pulsed DCs did not show significant therapeutic effects in six HIV-infected patients 126. So, further investigation of the antiviral immune responses induced by the inactivated-HIV-loaded DCs will be critical to understand the correlates underlying the therapeutic effects.
We thank Dr. J. Barbieri for critical reading of the manuscript. L.W. is supported by grants from the National Institutes of Health (NIH, United States) and the Research Affairs Committee of the Medical College of Wisconsin. V.N.K is supported by the intramural research funds from the NIH. The authors apologize to all colleagues whose work has not been cited in this review due to space restrictions.