Karin Loré, National Institutes of Health, Vaccine Research Center, Blg 40, Rm 3612B, Bethesda, MD 20892–3022, USA. e-mail: firstname.lastname@example.org
Dendritic cells are professional antigen-presenting cells required for generation of adaptive immunity. These cells are one of the initial target cells for HIV-1 infection or capture of virions at site of transmission in the mucosa. DCs carrying HIV-1 will migrate to the lymphoid tissue where they can contribute to the dissemination of the virus to adjacent CD4+ T cells. In addition, HIV-1–exposed DCs may have impaired antigen-presenting capacity resulting in inadequate expansion of HIV-1–specific T cell responses. Here, we review the infection of different subtypes of DCs by HIV-1 and the relevance of these cells in the transmission and establishment of HIV-1 disease. In addition, we discuss the mechanisms through which HIV-1–DC interactions could be exploited to optimise the generation and maintenance of HIV-1–specific T cell immunity.
Dendritic cells (DCs) perform an essential role in the induction and regulation of adaptive immune responses (1). There are at least two major subsets of DCs; CD11c+ myeloid DCs (MDCs) that originate from CD34+ myeloid precursors, and the more recently identified CD11c- CD123+ plasmacytoid DCs (PDCs), which are thought to share precursor cells with lymphocytes. Both DC types are professional antigen-presenting cells (APCs) but differ in some other functional aspects. MDCs are more frequently found and secrete high levels of IL-12 during antigen presentation. PDCs have the unique ability to produce high levels of IFN-alpha in response to foreign antigen structures and may, therefore, play an important role in innate antiviral immunity (2, 3). DCs, MDCs in particular, are widely distributed throughout the body and are most prevalent in the skin and mucosal tissues, where they are defined as “immature”. These immature cells express CCR5 and CCR6, which allow the DCs to migrate to sites of inflammation where the chemokines MIP-1alpha, MIP-1beta, MIP-3alpha and RANTES are produced (4). Immature DCs respond rapidly when they encounter foreign antigens and employ several antigen internalization methods such as phagocytosis, receptor-mediated endocytosis and macropinocytosis (5, 6). Antigen uptake and exposure to double-stranded RNA, bacteria, lipopolysaccharide or inflammatory cytokines induce maturation and migration of the immature DCs to regional lymph nodes (7). This maturation is characterized by upregulation of cell surface MHC class I and II molecules, costimulatory molecules such as CD80, CD86 and CD40, and the chemokine receptors CXCR4 and CCR7 (1). Expression of CCR7 directs the DCs via afferent lymphatics to T cell-rich areas in lymph nodes where its chemoattractant ligand SLC is produced (4). These phenotypic changes during maturation enable DCs to process and present antigen efficiently to both CD4+ and CD8+ T cells to prime and activate specific immune responses to invading pathogens (Fig. 1). However, these sensitive cellular interactions are also exploited by HIV-1 to facilitate infection and propagation of the virus (8).
INFECTION AND IMMUNE MODULATION OF DENDRITIC CELLS BY HIV-1
HIV-1 Infection of dendritic cells
Langerhans cells (LCs), a myeloid-derived DC located in the epidermis, were reported to be target cells for HIV-1 infection in 1987 (9–11). Both MDCs and PDCs have since been shown to be HIV-1 targets (12–19). In vitro, DCs derived from CD34+ bone marrow and CD14+ blood progenitors are susceptible to HIV-1 infection (20–22). DCs express the CD4 receptor and co-receptors such as CCR5, CXCR4, CCR2, CCR3, Bonzo/STRL33, CCR8 and CCR9, which are required for HIV-1's fusion and entry into a cell (23–27). In vitro infection of immature DCs has shown no preference for primary or laboratory adapted viruses or for any specific biological phenotype (12–14, 21, 28, 29). Maturing DCs downregulate CCR5 expression and upregulate CXCR4, thereby becoming less susceptible to the R5-HIV-1 isolates that use CCR5 as co-receptor for viral fusion (17, 30, 31). After transmission, early existing HIV-1 isolates are almost invariably of the R5-phenotype (32, 33). This selection for R5-virus strains might depend on the primary contact HIV-1 has with the CCR5-expressing DCs and CD4+ T cells residing at sites of infection (Fig. 2) (34, 35). However, the selection of R5-virus may occur without involving mucosal DCs or T cells. Mucosal epithelial cells, although resistant to virus infection, can bind these viruses and translocate them into contact with submucosal DCs and CD4+ T cells that can further spread the virus within the host (36). The efficient selection of R5-virus may therefore be due to the fact that epithelial cells, CD4+ T cells and DCs all express CCR5 and are located at the sites of initial infection.
The frequency of HIV-1-infected DCs in vivo is often 10–100 times lower than that of CD4+ T cells (37, 38). Ex vivo, the blood PDC subset contains higher levels of HIV-1 DNA than the MDC subset (18). In DCs alone, HIV-1 replication is very low. Viral replication is initially substantially enhanced when DCs are clustered with CD4+ T cells (28, 29, 39). The interactions of co-stimulatory molecules between the DCs and T cells, especially CD40-CD40 ligand ligation, may play a role by activating the T cells, rendering them more susceptible to HIV-1 infection and enhancing the replication (28, 29, 39). In vivo, the DC-T cell microenvironment in lymphoid tissue (LT) is the place where vigorous production of new HIV-1 virions takes place where (Fig. 2) (40).
Capture and uptake of HIV-1 by dendritic cells
Binding and uptake of HIV-1 by DCs can occur via cell surface interactions that are distinct from the classical infection process (14). DC-SIGN (CD209) has been found to be the main C-type lectin receptor expressed on DCs that mediates binding of gp120 from R5, X4 and R5X4 HIV-1 strains as well as HIV-2 and SIV. This suggests that this molecule may function as a universal attachment receptor for primate lentiviruses (41, 42). The biological function of DC-SIGN is to regulate adhesion, such as DC trafficking and formation of DC-T cell synapses, as well as antigen capture (42). The natural ligand for DC-SIGN appears to be ICAM-3, although functional interaction with ICAM-2 can occur as well (43). In vitro, DC-SIGN expression is found on mature and immature monocyte-derived DCs. In addition, DC-SIGN expression is found in vivo on a subset of blood MDCs, mucosal DCs and dermal DCs, whereas LCs in epidermis have no expression of DC-SIGN (44, 45). Although a small fraction of blood PDCs have been found to express DC-SIGN, most data suggest that PDCs in blood and other tissues do not express DC-SIGN.
HIV-1 that have bound to DC-SIGN can be taken up and stored in an intracellular early endosomal compartment without degradation. DC-SIGN-associated HIV virions remain infectious for several days and can be transmitted to permissive CD4+ T cells (Fig. 2) (41, 46). This may explain why DCs strongly enhance the infection of co-cultured CD4+ T cells even though the DCs themselves are either noninfected or inefficiently infected (12, 17, 41, 47, 48). The ability of rhesus macaque DCs to capture and transmit SIV to T cells is independent of DC-SIGN, although it may occur through an unidentified lectin molecule (49). Other C-type lectin receptors expressed on subtypes of DCs, e.g. mannose receptor and Langerin, have recently been shown to be involved in binding HIV-1 gp120 (50). HIV-1 engagement of Langerin, expressed on DC-SIGN-LCs, results in a similar binding of gp120 as described for DC-SIGN (50). In addition, CD14+ CD1a+ dermal DCs and PDCs can use the mannose receptor for binding of HIV-1 (50, 51). After cell surface binding, SIV and HIV-1 are internalized by immature and mature DCs via clathrin-coated pits (Fig. 3) (52). Although similar levels of SIV RNA are found in immature and mature DCs, large amounts of intact virus are only found in immature DCs (52). This suggests that different uptake and/or virus entry mechanisms exist at different stages of DC differentiation.
The role of dendritic cells in HIV-1 mucosal transmission and as a reservoir
As discussed earlier, DCs residing within epithelial surfaces (LCs) and subepithelial surfaces (dermal DCs) may be the initial targets for HIV-1 after mucosal exposure to the virus (Fig. 2). Studies of sexual mucosal transmission of SIV in rhesus macaques demonstrated that submucosal DCs rapidly become infected after virus inoculation (53–56). In contrast, studies of infection across mucosal epithelia suggest that the CD4+ T cells are the cells infected and that they are the major source of infectious virus during the acute stages of infection (57). After intravaginal inoculation of SIV in rhesus macaque monkeys, virus and infected cells are quickly disseminated throughout the LT (58). The migratory nature of DCs, along with their ability to recruit numerous T cells to the LT, identifies them as strong candidates for a central role in spreading HIV-1 in the host (39). Numerous CD1a+ CD83+ DC-SIGN+ DCs accumulate rapidly in LT during the first weeks after HIV-1 infection (Fig. 4) (59). Furthermore, an explosive increase, several hundred fold, of productively infected CD4+ T cells occurs in LT between the first and second week after transmission (57). Throughout the course of infection, HIV-1 replicates in clusters of DCs and CD4+ T cells in the LT. Many of these clustering DCs and T cells found in tonsils comprise of multinucleated syncytia with a high level of HIV-1 protein expression (60). DCs in LT may serve as depots for HIV-1 by continually infecting newly recruited T cells (37, 39, 61).
Fluctuations in levels of dendritic cells during HIV-1 infection
During the acute phase of HIV-1 infection and phases of high viremia, the frequencies of both MDCs and PDCs in blood decrease (19, 62–67). The depletion of MDCs in blood correlates closely to the increase of plasma virus load (19, 63, 67), whereas this correlation is not so clear for PDCs (19, 64, 67, 68). Frequencies of MDCs in the skin (62) and in the mucosa (69) are also reduced during HIV-1 infection. Decreased levels of blood MDCs and PDCs persist in many infected individuals even though the patients are receiving Highly Active Antiretroviral Therapy (HAART) (19, 65, 67, 70).
In contrast to the depletion of DCs in blood, skin and mucosa, an influx of DCs is seen in lymph nodes, especially during episodes of high viremia (19, 59). Taken together, these findings point to a redistribution of DCs from the peripheral tissues to the LT during active stages of the infection. Several factors, such as persistent production of HIV-1 antigens and continuous pro-inflammatory cytokine and chemokine expression in LT, may be involved in the recruitment of DCs to this site.
Failure of immune activation by dendritic cells during HIV-1 infection
As a chronic pathogen, HIV-1 has evolved many mechanisms to inhibit or delay the induction of potent HIV-1-specific immune responses. Cytotoxic T lymphocytes (CTLs) play a central role in the clearance of infected cells and the control of virus replication (71–76). The viremia in acute HIV-1 infection resolves with the appearance of HIV-1-specific CD4+ T helper cells and CTLs (73, 74, 77, 78). However, in the absence of antiretroviral drugs, the immune response eventually fails to control virus replication in all but a small population of HIV-positive patients termed longterm non-progressors. The collapse of HIV-1-specific CTLs in late progressive disease is possibly dependent on the loss of CD4+ T helper cell activity (79–81). HIV-1 destroys both activated infected and uninfected CD4+ T cells, which leads to the depletion of existing memory and naïve T cells and facilitates viral persistence (73, 82). In addition, HIV-1-specific CD4+ memory T cells are preferentially infected by HIV-1 as compared to CD4+ memory T cells specific for other antigens (82). This may be a result of the interaction between HIV-1-specific T cells and HIV-1-bearing DCs during antigen presentation. This suggests that the essential interactions for elicitation of immune responses against HIV-1 may also be used by the virus for efficient transmission between cells.
Many viruses have developed evasion mechanisms that prevent recognition and induction of immune responses. Some of these mechanisms interfere with the antigen presentation process (83). The early appearance of deficiencies in antigen presentation during HIV-1 infection may correlate with the loss of memory responses, and this could be of major importance in the pathogenesis of this infection. The DCs found in LT during acute HIV-1 infection have reduced expression of the co-stimulatory molecules CD80 and CD86 (59). Reduced expression of CD80 and CD86 may be a direct effect of HIV-1 on DCs; on the other hand, it could be the outcome of lower expression of CD40 ligand, CD28 and/or other receptors on T cells whose interactions are essential for the induction of full maturation of DCs (84–86). Mature DCs are superior to immature DCs in transmitting HIV-1 to T cells during clustering (87). Decrease and/or lack of some molecules found on either the DCs or the T cells (e.g. CD80, CD28, CD40 or CD40 ligand) may limit HIV-1 transmission from DCs to T cells (Fig. 5). Therefore, a downregulation of co-stimulatory molecules that may limit transmission of HIV-1 between DCs and T cells could negatively affect antigen presentation and generation of T cell immunity.
In vitro experiments examining the effects of HIV-1 on DC phenotype and function have generated opposing results. Peripheral blood DCs isolated from patients in different stages of HIV-1 infection have been shown to have significantly reduced capacity to stimulate allogeneic T cells compared to DCs from uninfected controls (62, 88–91). However, several reports show that monocyte-derived DCs from HIV-1-infected individuals can efficiently induce CTL responses to different antigens (92–94). In vitro, monocyte-derived DCs from HIV-1-infected patients do not express HIV-1 DNA and exhibit normal T cell activation function (94). Furthermore, monocyte-derived DCs from healthy blood donors do not change their expression of receptors that are important for APC function when they are infected with HIV-1 in vitro. However, infected DCs exhibit a downregulation of CD4 and changed cytokine production profile compared to uninfected DCs (95–97). Blood CD11c+ MDCs exposed to HIV-1 in vitro also downregulate CD4 expression (18), a change which may not affect their function but can inhibit virions from binding and fusing to the cell. HIV does not have this effect on blood PDCs even though PDCs, and not blood MDCs, differentiate into a mature phenotype and induce IFN-alpha secretion after exposure to HIV-1 (18, 51) (Fonteneau & Larsson et al., unpublished data). The HIV-1 accessory proteins nef and tat alone can induce maturation of immature DCs, e.g. upregulation of maturation markers such as CD1a, HLA DR, CD40, CD83 and CXCR4 (98, 99). The MHC class I expression on DCs is downregulated by nef, administered either as a recombinant vector or as a soluble protein (98, 100, 101). Nef has also been found to modulate the expression of DC-SIGN on DCs, which can favor DC-T cell interactions and subsequent virus spread (102).
Dendritic cells can present different forms of HIV-1 antigen to activate T cell immunity
CD8+ T cells play an important role in control of HIV-1 by direct cytolysis of infected cells and the secretion of factors that suppress viral replication. DCs are the APCs required for priming both naïve CD4+ and CD8+ T cells and turning them into functional effector cells. At the same time, the consequence of DC and T cell interactions is known to be in vigorous viral replication and cell death both in vivo and in vitro (28). How can anti-HIV-1 immunity be generated in the face of exposure to viral replication and cell death (Fig. 6)?
Induction of apoptosis in both infected and uninfected cells is one characteristic of HIV-1 infection in vivo and in vitro (103–106). Apoptotic and necrotic cells have been shown both in vivo and in vitro to be efficient sources of exogenous antigens used by DCs for activation of both CD4+ and CD8+ T cells (107). The MHC class I presentation of antigen derived from exogenous sources such as apoptotic cells is termed “cross-presentation” and requires DCs to act as APCs (107). HIV-1 DNA and protein can efficiently be transferred to DCs from apoptotic cells with integrated HIV-1 (108). DCs present HIV-1 antigens derived from apoptotic HIV-1-infected CD4+ T cells and subsequently activate both CD4+ and CD8+ HIV-1-specific memory responses (Fig. 7) (109, 110). However, HIV-1 tat protein has been shown to inhibit uptake of apoptotic bodies (111). This involves engagement of the alpha v beta 3 integrin, which is used in engulfment of apoptotic cells by macrophages but not by DCs. In addition, tat can reduce the expression of the mannose receptor, another receptor involved in antigen uptake (112).
There is an abundance of noninfectious viral particles in the plasma of HIV-1-infected individuals (113); frequent presentation of defective, nonreplicating HIV-1 particles by DCs may therefore take place in vivo (Fig. 6). Recently, an Aldrithiol-2 (AT2) inactivated HIV-1 isolate was described that binds, and fuses, in the same manner as untreated infectious HIV-1 but is unable to produce any new virions (114). DCs can take up and present antigens derived from this noninfectious virus and thereby activate both CD8+ and CD4+ HIV-1-specific memory T cells from infected individuals as well as antigen-specific CD8+ T cell clones in vitro (109, 115, 116). Furthermore, DCs presenting HIV-1 antigen from AT2-inactivated HIV-1 expand autologous CD8+ memory T cells capable of killing HIV-1-infected targets (115). These findings show that generation of effector T cells in vitro is independent of viral replication. MHC class I presentation of noninfectious virus by DCs required fusion and binding to the cell in the same fashion as infectious virus, indicating that this is classical MHC class I presentation and not cross-presentation (116). Immune complexed HIV-1 or complement-opsonized virions are also abundant sources of HIV-1 antigen occurring in vivo that DCs may utilize as sources of antigen for T cell activation.
In summary, replicating HIV-1 virions and high levels of HIV-1 proteins may have various effects on DCs, including delayed or weakened antigen capture, processing and presentation, and T cell activation. Moreover, presentation of antigens from dead cells previously infected with HIV-1, or direct presentation of antigens from noninfectious HIV-1 virus, may be key mechanisms by which immunity to HIV-1 is generated by DCs in vivo. This information may be of great importance in the rational design of an HIV-1 vaccine. AT2-inactivated HIV-1 may be a good vaccine candidate.
THE USE OF DENDRITIC CELLS IN HIV-1 VACCINE DEVELOPMENT AND IMMUNE RECONSTITUTION
Antigen-encoded vector system to target dendritic cells for activation of HIV-1-specific T cells
In vitro testing of potential vaccine candidates using DCs loaded with HIV-1 antigens has been widely studied using various delivery and expression systems. The efficacy of antigen-loaded DC to stimulate HIV-1-specific T cell activation will subsequently be evaluated. Monocyte-derived DCs or MDCs are the sources of DCs that have been tested so far in this context. DCs transfected with the HIV-1 nef gene activated existing memory CTLs specific for nef from HIV-1-infected donors and primed nef-specific CTLs from uninfected individuals (117, 118). Liposome-encapsulated HIV-1 proteins and lipopeptides given to DCs can function as efficient sources of antigen for MHC class I stimulation of HIV-1-specific CTLs. However, their soluble counterparts are not efficient in activating CTLs (119–121). Antigens from liposome-encapsulated HIV-1 proteins and lipopeptide are cross-presented by DCs and utilize the classical MHC class I pathway because they are dependent on proteasome degradation and transportation via the endoplasmic reticulum (119–121). One approach is to develop vector-based potential HIV-1 vaccines to target DCs. DCs transfected with a HIV-1 vector using vesicular stomatitis virus (VSV) protein G as envelope were found to prime and activate HIV-specific CD8+ and CD4+ T cells (122, 123). DCs infected with recombinant vaccinia encoding HIV-1 genes have also been shown to stimulate HIV-1-specific memory CTLs from HIV-1-positive individuals at different stages of HIV-1 infection (93). In addition, DCs infected with recombinant canarypox vector (ALVAC) encoding HIV-1 genes activate and expand both CD4+ T cells and CTLs (124). Activation of HIV-specific CTL responses is dependent on CD4+ T cells, which need to be antigen specific, and cannot be replaced by exogenously added cytokines (124). However, when DCs were matured and activated by CD40 ligand trimeric complex – CD40 interactions, a total or partial substitution for CD4+ T cell help was seen. In this case it was possible to expand CTLs from HIV-1-infected individuals without CD4+ T cells (125). This may indicate that targeting antigen to DCs alone is not sufficient for development of protective HIV-1 immunity. The HIV-1 antigen source/s should be administered together with factors that ensure that recruitment, antigen loading and maturation of DCs occur. It is also important to ensure that CD4+ T cells and DCs interact directly to allow optimal conditioning of DCs for the induction of CD8+ T cell responses.
Dendritic cells as vaccine adjuvants
When DCs are delivered in vivo charged with peptides, tumor lysates, or viral vectors encoding relevant antigens, they have been shown to induce protective immune responses and therapeutic immunity to tumors and viruses in animals and humans (126–129). Injection of influenza antigen-pulsed DCs leads to a several fold increase in influenza-specific CD8+ T cells circulating in the blood of healthy volunteers, as early as 7 days after injection (128). When these T cells are activated in vitro, there is an increase in the frequency of MHC I-tetramer binding cells and cytotoxic T cells following influenza-pulsed DC vaccination. These data provide the first controlled evidence of DC immunogenicity in humans and prove that a single injection of mature DCs rapidly expands T cell immunity (128). DC-based therapeutic vaccination of HIV-1-infected individuals undergoing HAART is currently being studied using either peptide-pulsed DCs or DCs infected with recombinant canarypox virus (ALVAC) encoding HIV-1 genes (Nina Bhardwaj, personal communications). The outcome of these trials will provide important information about the requirements for protection against HIV-1 infection and what is needed to control already existing infection. However, preparation of DCs for individual autologous transfusions is not feasible in practice as an ultimate HIV-1 vaccine due to limited resources, especially in developing countries. Further research focusing on how different forms of HIV-1 antigens together with potential vaccine adjuvants that target and affect DCs in vivo is clearly of importance.
Research activity focused on dendritic cells has been intensified over the last 10 years. The information emerging from this work is providing much needed insight into the underlying pathogenesis of HIV-1 infection. In addition, clues to promising avenues for immunotherapeutic intervention are being discovered. It seems clear that the biology of DCs is as central to the pathogenesis of HIV-1 infection as it is to effective adaptive immunity. A more detailed understanding of these key immunoregulators and their role in HIV-1 infection might well improve the prospects of developing an HIV-1 vaccine.