|Home | About | Journals | Submit | Contact Us | Français|
Human immunodeficiency virus type 1 (HIV-1) infection of dendritic cells (DCs) has been documented in vivo and may be an important contributor to HIV-1 transmission and pathogenesis. HIV-1-specific CD4+ T cells respond to HIV antigens presented by HIV-1-infected DCs and in this process become infected, thereby providing a mechanism through which HIV-1-specific CD4+ T cells could become preferentially infected in vivo. HIV-2 disease is attenuated with respect to HIV-1 disease, and host immune responses are thought to be contributory. Here we investigated the susceptibility of primary myeloid DCs (mDCs) and plasmacytoid DCs (pDCs) to infection by HIV-2. We found that neither CCR5-tropic primary HIV-2 isolates nor a lab-adapted CXCR4-tropic HIV-2 strain could efficiently infect mDCs or pDCs, though these viruses could infect primary CD4+ T cells in vitro. HIV-2-exposed mDCs were also incapable of transferring virus to autologous CD4+ T cells. Despite this, we found that HIV-2-specific CD4+ T cells contained more viral DNA than memory CD4+ T cells of other specificities in vivo. These data suggest that either infection of DCs is not an important contributor to infection of HIV-2-specific CD4+ T cells in vivo or that infection of DCs by HIV-2 occurs at a level that is undetectable in vitro. The frequent carriage of HIV-2 DNA within HIV-2-specific CD4+ T cells, however, does not appear to be incompatible with preserved numbers and functionality of HIV-2-specific CD4+ T cells in vivo, suggesting that additional mechanisms contribute to maintenance of HIV-2-specific CD4+ T-cell help in vivo.
Dendritic cells (DCs) constitute a highly sophisticated component of innate immunity, as they are specialized to process foreign antigens collected in the periphery and then transport them to T-cell-rich areas of lymphoid tissue. DCs initiate adaptive immune responses to their cargoes of antigens; in particular, they stimulate antigen-specific memory CD4+ T cells from the resting pool (2). Early studies showed a physical association between DCs and human immunodeficiency virus type 1 (HIV-1) in cutaneous and mucosal tissues of infected individuals (24, 25, 32), and simian immunodeficiency virus (SIV)-infected mucosal DCs have been documented very early after infection of rhesus macaques (33, 61). It is now apparent that DCs support HIV replication by at least two distinct mechanisms, namely, by becoming productively infected with HIV-1 or by capturing virions and transferring them to CD4+ T cells in the absence of overt infection (12, 50-53). Although HIV-1 can enter immature DCs because they express the necessary surface CD4 molecule and the coreceptors CCR5 (R5) and CXCR4 (X4) (21, 48), it is only once DCs come into contact with T cells in lymphoid tissue that extensive viral replication takes place (28). Since HIV-1 replicates preferentially in activated T cells, these DC-T-cell conjugates present the virus to a pool of highly susceptible target cells and allow the infection to become established in T cells within lymphoid tissues.
It is well documented that CD11c+ myeloid DCs (mDCs) and CD123+ plasmacytoid DCs (pDCs) are susceptible to productive infection with both CCR5- and CXCR4-tropic HIV-1 isolates (16, 21, 38, 48, 59). We have previously shown that these infected DCs preferentially transfer HIV-1 to antigen-specific CD4+ T cells responding to antigens presented by infected DCs (38). This suggests that the infection of DCs by HIV-1 and subsequent processing and presentation of HIV-1 antigens on the surfaces of DCs are a double-edged sword. In order to induce an HIV-specific immune response, it is necessary for DCs to deliver appropriate signals to memory CD4+ T cells. However, in doing so, the intimate contact of DCs and T cells leaves HIV-1-specific CD4+ T cells susceptible to becoming infected by these HIV-1-carrying DCs. This could help to explain mechanistically how HIV-1-specific CD4+ T cells become preferentially infected in vivo.
In this study, we sought to explore the interaction of primary DCs with HIV-2. We hypothesized that HIV-2 isolates may be less efficient at infecting DCs than HIV-1 isolates or that HIV-2-infected DCs may be incapable of transferring virus to antigen-specific CD4+ T cells, helping to explain mechanistically the preservation of CD4+ T cells specific for HIV-2 in vivo. Additionally, we investigated whether HIV-2-exposed mDCs could transfer virus to antigen-specific CD4+ T cells. We exposed mDCs and pDCs to four HIV-2 isolates, namely, the lab-adapted strain CBL-20 (56) and primary isolates from long-term asymptomatic HIV-2-positive (HIV-2+) individuals (RH2-3, RH2-13, and RH2-14) (6), and monitored infection by intracellular staining for HIV-2 capsid protein or by quantitative PCR (qPCR) for HIV-2 gag. Our results indicate that primary mDCs and pDCs are not susceptible or are poorly susceptible to infection by R5- or X4-tropic HIV-2 strains and that DCs exposed to HIV-2 do not transfer the virus to autologous CD4+ T cells. However, despite this lack of transfer of HIV-2 from DCs to antigen-specific CD4+ T cells in vitro, we show here that HIV-2-specific CD4+ T cells in vivo contain more HIV-2 viral DNA than do memory CD4+ T cells of other specificities.
DC subsets were isolated directly from peripheral blood as described previously (37). Briefly, peripheral blood mononuclear cells (PBMC) from healthy volunteers were collected by automated leukapheresis. Lymphocytes and monocytes were enriched by counterflow centrifugal elutriation. mDCs and pDCs were isolated from elutriated monocytes by using magnetic bead isolation kits (Miltenyi) and sequential separation with an AutoMacs instrument (Miltenyi). CD1c and BDCA-4 isolation kits were used to isolated mDCs and pDCs, respectively. mDCs and pDCs were cultured in complete medium (RMPI 1640 supplemented with 10% heat-inactivated fetal calf serum [FCS], 100 U/ml penicillin G, 100 μg/ml streptomycin sulfate, and 1.7 mM sodium glutamate) in the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF) (2 ng/ml; PeproTech Inc., Rocky Hill, NJ) and recombinant interleukin-3 (IL-3) (1 ng/ml; R&D Systems, Minneapolis, MN), respectively.
HIV-1 BaL and IIIB (NIH AIDS Research and Reference Reagent Program) and HIV-2 CBL-20 (NIH AIDS Research and Reference Reagent Program), RH2-3, RH2-13, and RH2-14 were grown on phytohemagglutinin (PHA; Sigma)-stimulated PBMC in complete medium supplemented with IL-2 (Chiron). RH2-3, RH2-13, and RH2-14 are primary strains of HIV-2 isolated from long-term aviremic individuals, as described previously (6). p24 and p27 Gag in cell culture supernatants was monitored by enzyme-linked immunosorbent assay (ELISA) for the HIV-1 and HIV-2 strains, respectively (Coulter). Viruses were harvested at the peak of infection, on either day 4 or day 7. Viruses were concentrated by ultracentrifugation at 30,000 rpm for 70 min at 4°C, and virus pellets were resuspended in fresh RMPI at 10× to 50× to obtain clean, concentrated virus stocks. Control concentrated conditioned supernatant from uninfected PHA-stimulated T cells was collected as a control. The viral titers were determined by sensitive 14-day end-point titration assays using PHA- and IL-2-stimulated PBMC, as described previously (43). Virus stock 50% tissue culture infective doses (TCID50) were as follows: for HIV-1 BaL, 2.6 × 106 TCID50/ml; for HIV-1 IIIB, 1.75 × 106 TCID50/ml; for HIV-2 CBL-20, 5 × 106 TCID50/ml; for HIV-2 RH2-3, 2.9 × 105 TCID50/ml; for HIV-2 RH2-13, 7.2 × 104 TCID50/ml; and for HIV-2 RH2-14, 2.0 × 105 TCID50/ml.
mDCs and pDCs were cultured at 1 × 106/ml (with 0.3 × 106 to 0.4 × 106 cells/tube) for 12 h in the presence of GM-CSF or recombinant IL-3 for 12 h at 37°C in round-bottomed polystyrene tubes (Becton Dickinson). HIV-1 or HIV-2 isolates were then added to the DCs and cultured for 6 to 72 h, as described previously (38, 59, 60). Viruses were normalized by MOI (MOIs of 0.05 and 0.25 were used in various experiments, as noted), and volumes were equalized in each culture with concentrated conditioned supernatant. In some experiments, a Toll-like receptor 7/8 (TLR7/8) agonist (1 μg/ml; 3M Pharmaceuticals, St. Paul, MN) was added to induce maturation, as previously described (37). TLR7/8 was added either in the 12 h prior to infection or in the final 24 h of infection to DCs exposed to virus for a total of 72 h. In some experiments, an anti-CD4 monoclonal antibody (MAb) (clone RPA-T4; Pharmingen) was added to the cells at 0.5, 1.0, or 2.0 μg/ml 30 min prior to the addition of the virus.
HIV-exposed DCs were harvested, washed, and surface stained with different combinations of anti-CD11c-fluorescein isothiocyanate (FITC) (Caltag Laboratories), anti-CD14-peridinin chlorophyll protein (Pharmingen), anti-CD86-allophycocyanin (APC) (Pharmingen), anti-CD123-APC (Miltenyi), and anti-CD86-FITC (Pharmingen) at pretitrated concentrations at 4°C for 20 min. Cells were washed and fixed/permeabilized for 10 min at room temperature, using a 2× fixation-permeabilization solution (Becton Dickinson). Cells were then washed twice and stained intracellularly for HIV Gag, using anti-p24 RD1 (KC57 clone; Beckman Coulter). There is no commercially available MAb against the HIV-2 capsid for use in flow cytometry. We attempted to conjugate several MAbs specific for the HIV-2 capsid and to optimize them for use in flow cytometry, but none gave acceptable staining of HIV-2-infected cells. The anti-p24 MAb KC57 cross-reacts with the HIV-2 capsid, and the median fluorescence intensities of p24 staining of infected cells in CBL-20- versus BaL-infected cultures of CD4+ T cells are not different (see Fig. Fig.1I),1I), and as such, we used this MAb for all flow cytometry analyses of HIV-2-exposed DCs. Antibody-labeled cells were then analyzed by flow cytometry (FACSCalibur; Becton Dickinson). Data were analyzed using FlowJo software, version 8.1.1 (TreeStar Inc., Ashland, OR).
Viral DNA in HIV-exposed DCs was assessed by qPCR with an ABI7900HT machine (Applied Biosystems, Foster City, CA). HIV-1 Gag primers and probe were designed against the Los Alamos HIV database as described previously (17). Sequences were as follows: HIV-1GagFwd, GGTGCGAGAGCGTCAGTATTAAG; HIV-1GagRev, AGCTCCCTGCTTGCCCATA; and HIV-1GagProbe, 6-carboxyfluorescein (FAM)-AAAATTCGGTTAAGGCCAGGGGGAAAGAA-QSY7. HIV-2 Gag primers and probe were designed against 15 subtype A HIV-2 sequences, and their sequences were as follows: HIV-2GagFwd, AGGCTGCACGCCCTATGA; HIV-2GagRev, TACTGTGCTYGTTGTYCCTGCTAT; and HIV-2GagProbe, FAM-CTTAAYTGTGTGGGCGAYCAYCAAGC-BHQ1. Sites where base sequences differed among HIV-2 isolates were made degenerate and are noted as follows: Y = C or T. To quantify the cell number in each reaction mix, the human albumin gene copy number was also assessed by qPCR. Albumin primer and probe sequences were as follows: hAlbFwd, TGCATGAGAAAACGCCAGTAA; hAlbRev, ATGGTCGCCTGTTCACCAA; and hAlbProbe, FAM-TGACAGAGTCACCAAATGCTGCACAGAA-QSY7. Additionally, we designed primers and a probe to detect HIV-2 minus-strand strong-stop DNA (−sssDNA). −sssDNA primer and probe sequences were as follows: HIV-2sssDNAFwd, AARGARGARGGGRTAMTWSCAGATTG; HIV-2sssDNARev, CYAGYTTCCAYAGCCAYCCRWA; and HIV-2sssDNAProbe, FAM-CAGAAYTAYACTCATGGGCYAGGRRYRAGRTAYC-BHQ1, where degenerate bases are denoted as R (A or G), M (A or C), W (A or T), S (G or C), and Y (C or T). Virus-exposed DCs were lysed in 0.1 mg/ml proteinase K buffered with Tris-Cl at 10,000 cells/μl buffer for 1 h at 56°C and then for 10 min at 95°C to inactivate the enzyme. Five microliters of lysate was used in a total reaction volume of 25 μl containing 12.5 μM primers, 5 μM probe, a 10 mM concentration of each deoxynucleoside triphosphate, 50 mM MgCl2, 1.25 mM Blue 636 reference dye, and 0.625 U Platinum Taq in the supplied buffer (18). Reaction conditions included a 5-minute activation at 95°C followed by 15 s at 95°C and 1 min at 60°C for 45 cycles. Quantification was generated using standard curves for HIV-1 Gag, HIV-2 Gag, HIV-2 sssDNA, and albumin.
Supernatants from HIV-1- or HIV-2-exposed pDCs were collected and frozen at −80°C until use. Supernatants were analyzed for alpha interferon (IFN-α) by ELISA (PBL Biomedical Laboratories) according to the manufacturer's instructions (37).
Autologous elutriated lymphocytes were obtained for each donor. CD4+ T cells were isolated by negative selection; lymphocytes were depleted of CD14+, CD19+, CD56+, and CD8+ cells by magnetic bead separation on an AutoMacs instrument (Miltenyi). Sorted CD4+ T cells were either used immediately or frozen in FCS with 10% dimethyl sulfoxide at −80°C until use. CD4+ T cells were then thawed for coculture, washed thoroughly, and labeled with 0.25 μM carboxyfluorescein diacetate succinimidyl ester (CFSE) fluorescent dye (Molecular Probes) as described previously (8). CFSE-labeled cells were washed thoroughly and resuspended in complete medium for coculture.
mDCs exposed to HIV-1 or HIV-2 for 6 or 72 h were washed thoroughly three times in complete medium to eliminate any residual virus not associated with mDCs. DCs were pulsed with cytomegalovirus (CMV) lysate (10 μl) derived from the CMV isolate AD-169 grown on MRC-5 cells and inactivated by gamma irradiation. In some experiments, the superantigen staphylococcal enterotoxin B (SEB; 1 μg/ml) was used instead of CMV lysate. mDCs were then cocultured with sorted autologous CFSE-labeled CD4+ T cells at a ratio of 1:10 (DCs:T cells) in 1 ml complete medium in a round-bottomed polystyrene tube (38). DC-T-cell cocultures were incubated for 3.5 days, with brefeldin A (1 μg/ml; Sigma-Aldrich) present during the final 12 h. Cells were harvested after 3.5 days of coculture, surface stained with anti-CD11c-Cy5PE, and then permeabilized using 2× fixation-permeabilization solution. Cells were then stained intracellularly with anti-p24-RD1, anti-IFN-γ-APC, anti-tumor necrosis factor alpha (anti-TNF-α)-APC (BD Biosciences), and anti-IL-2-APC (Caltag). Cells were analyzed immediately by flow cytometry on a FACSCalibur flow cytometer.
PBMC from 18 HIV-2-infected individuals were isolated and cryopreserved at −80°C in FCS with 10% dimethyl sulfoxide until use. PBMC were thawed and rested overnight at 37°C in complete medium supplemented with 10 U/ml DNase I (Roche Diagnostics, Indianapolis, IN). Costimulatory antibodies (anti-CD28 and anti-CD49d) (1 μg/ml each; BD Biosciences), monensin (Golgistop) (0.7 ml/ml; BD Biosciences), brefeldin A (10 μg/ml; Sigma-Aldrich), and anti-CD107a-Alexa 680 were then added to all tubes. Overlapping HIV-2 Gag peptide pools (final concentration, 2 μg/ml/peptide) were used to stimulate cells. Cells were incubated for 5.5 h at 37°C. Following incubation, cells were washed and stained with pretitrated surface antibodies. Cells were then washed and fixed/permeabilized with Cytofix/Cytoperm per the manufacturer's instructions (BD Pharmingen). Following permeabilization, cells were stained intracellularly with pretitrated antibodies specific for CD3, cytokines, and chemokines. Cells were then washed, fixed, and analyzed immediately via flow cytometry. Directly conjugated MAbs were obtained from the following: BD Pharmingen, San Diego, CA, IFN-γ-FITC, macrophage inflammatory protein 1β (MIP-1β)-phycoerythrin (PE), IL-2-APC, TNF-α-Cy7PE, and CD3-Cy7APC; Caltag, Burlingame, CA, CD4-Cy5.5PE; and Beckman Coulter, Miami, FL, CD45RO-ECD. The following antibodies were conjugated in our laboratory according to standard protocols (http://drmr.com/abcon/index.html): CD107a-Alexa 680, CD27-Cy5PE, CD14-Cascade blue, and CD19-Cascade blue. The following antibodies were conjugated in our laboratory as described previously (13): CD8-QD705 and CD57-QD545. Unconjugated antibodies were obtained from BD Biosciences. Quantum dots were obtained from Quantum Dot Corporation (Hayward, CA), Alexa 680 and Cascade blue were obtained from Molecular Probes (Eugene, OR), and Cy5 was obtained from Amersham Biosciences (Pittsburgh, PA). A violet fluorescent reactive dye (Molecular Probes) was used as a viability marker to exclude dead cells from analysis, as described previously (49). HIV-2-specific CD4+ T cells were sorted with a modified FACSDiva flow cytometer by gating tightly on live CD3+ CD4+ T cells producing any of five functions (IFN-γ, IL-2, TNF-α, MIP-1β, or CD107a) above background. Two hundred to 9,000 HIV-2-specific CD4+ T cells were sorted, depending on the number of PBMC available and the response level for each individual. CD45RO+ CD57− memory and CD45RO+ CD57+ memory CD4+ T cells not specific for HIV-2 were also sorted for each subject.
Statistical tests were performed using the Mann-Whitney U test, using GraphPad Prism, version 4.0.
We previously described a procedure for directly isolating highly pure human CD11c+ mDCs and CD123+ pDCs from elutriated monocytes (37, 38, 59, 60). These cells are isolated based on unique expression of BDCA-1 (mDCs) and BDCA-4 (pDCs) and the lack of lineage markers (CD3, CD19, CD14, and CD56). Overnight culture in GM-CSF and IL-3 yields mDCs and pDCs, respectively, with distinctive DC morphology and an immature phenotype: these DCs have low expression of major histocompatibility complex class II, CD40, CD80, CD83, and CD86. Freshly isolated mDCs and pDCs express CD4, CCR5 (R5), and CXCR4 (X4) on their surfaces (21, 48) and are susceptible to both R5- and X4-tropic HIV-1 isolates (16, 21, 38, 48, 59). There have been no studies to date investigating the susceptibility of primary DCs to infection with HIV-2.
We exposed directly isolated immature mDCs and pDCs to primary R5-tropic HIV-2 isolates, a lab-adapted X4-tropic HIV-2 strain (CBL-20), or HIV-1 BaL (R5-tropic) or IIIB (X4-tropic) for 72 h at matched MOIs. The frequency of infection was determined by intracellular p24 staining. Exposure of mDCs (Fig. (Fig.1A)1A) or pDCs (Fig. (Fig.1B)1B) to HIV-2 isolates at doses as high as an MOI of 0.25 did not result in productive infection. Exposure of DCs to neither the lab-adapted CBL-20 strain nor any of the three primary HIV-2 isolates resulted in detectable levels of p24 staining after 72 h of exposure. In contrast, HIV-1 BaL efficiently infected both mDCs and pDCs, even though the amount of virus used was substantially less than that reported previously to induce productive infection and detectable p24 staining at 72 h (38, 59). This lack of productive infection by HIV-2 was observed in 5 to 10 individual donors, where levels of p24 staining after 72 h of exposure were consistently negligible in mDCs exposed to any of the four HIV-2 isolates (Fig. (Fig.1C)1C) (0.05% + 0.009% [mean + standard error of the mean [SEM]). Similarly, pDC exposure to HIV-2 isolates resulted in no substantial productive infection, as measured by p24 expression (Fig. (Fig.1D)1D) (0.02% + 0.004%). In contrast, HIV-1 BaL induced high levels of productive infection of both mDCs (0.74% + 0.18%) and pDCs (1.8% + 0.83%). Consistent with previous observations, mDCs were more susceptible to the R5-tropic HIV-1 BaL strain than to X4-using IIIB (38, 59). This preferential susceptibility of mDCs to R5 strains clearly did not extend to HIV-2 isolates, as all the primary HIV-2 strains use CCR5 exclusively as the entry coreceptor. Since we reported previously that there is wide donor variability in the susceptibility of mDCs and pDCs to IIIB infection (38, 59), in this study we exposed DCs to a much lower viral load than those in previous studies (up to 10-fold less), and in most donors this resulted in very little infection of mDCs or pDCs by IIIB (Fig. 1A to D).
We speculated that very low levels of HIV-2 Gag may be present in exposed DCs, but at levels undetectable by intracellular p24 staining. We used a very sensitive qPCR to assess whether there was any HIV-2 gag DNA present in HIV-2-exposed mDCs and pDCs. In contrast to HIV-1 BaL, with a mean of 7 × 104 gag copies per 1 × 105 mDCs at 72 h postinfection, little to no gag was detected in mDCs exposed to either CBL-20 or any of the primary HIV-2 isolates (Fig. (Fig.1E).1E). A similar lack of productive infection, as shown by qPCR for HIV-2 gag, was seen in pDCs exposed to primary HIV-2 isolates for 72 h, although a small proportion of CBL-20-exposed pDCs had detectable levels of HIV-2 gag (mean, 4,600 gag copies/105 pDCs; range, 770 to 16,500 copies) (Fig. (Fig.1F).1F). This low-level infection of pDCs by CBL-20 was CD4 dependent, as it could be blocked with an anti-CD4 MAb (Fig. (Fig.1G1G).
RH2-3, RH2-13, and RH2-14 are primary subtype A HIV-2 isolates from long-term aviremic HIV-2-infected individuals (6); they use CCR5 as the coreceptor for entry (5). These were the first viruses isolated from plasma virus-negative HIV-2-infected individuals, signifying that replication-competent virus could in fact be isolated from PBMC of individuals with no apparent plasma viremia. Attempts to isolate infectious virus from PBMC of HIV-2-infected individuals with high CD4+ T-cell counts have largely been unsuccessful in the past, suggesting that HIV-2 variants from individuals with nonprogressive disease may have impaired replication capacity. Thus, these primary HIV-2 isolates are quite rare and offer the unique opportunity to study the replication kinetics of these viruses in T cells and other potential targets.
Consistent with previous observations, we found that these primary HIV-2 isolates clearly infected primary CD4+ T cells in vitro (7), but the levels of p24 positivity were lower than those observed for lab-adapted HIV strains (Fig. (Fig.1H).1H). CBL-20 is a lab-adapted X4-tropic HIV-2 strain isolated from a Gambian AIDS patient (56), and it also replicated well in primary CD4+ T cells and yielded levels of p24 not substantially lower than those of HIV-1 BaL (Fig. (Fig.1H).1H). Thus, each of these HIV-2 isolates is capable of productively infecting susceptible CD4+ T cells, and as such, they are clearly replication-competent viruses. We considered that the lack of p24+ DCs after HIV-2 exposure could occur if the anti-p24 MAb binds the HIV-2 capsid protein less tightly or less efficiently than it binds the HIV-1 capsid and may therefore miss HIV-2-infected DCs in culture. We found this to be unlikely, as CBL-20-infected T cells showed a median fluorescence intensity of p24 staining that was not different from that of p24+ T cells infected with HIV-1 BaL (Fig. (Fig.1I1I).
We examined whether the time course of infection of DCs with HIV-2 isolates was different from that observed for HIV-1. We exposed directly isolated mDCs to HIV-1 or HIV-2 isolates for 0, 24, 48, 72, and 96 h and monitored levels of infection by intracellular p24 staining (Fig. (Fig.2A).2A). mDCs productively infected with HIV-1 BaL could be detected by flow cytometry as early as 48 h postexposure, and the proportion of BaL-infected mDCs steadily increased over the first 96 h of exposure (Fig. 2A and B). In contrast, neither CBL-20 nor any of the three primary HIV-2 isolates showed any productive infection of mDCs over 4 days by flow cytometry (Fig. 2A and B). This lack of productive HIV-2 infection in mDCs exposed to HIV-2 isolates for up to 96 h was also observed by qPCR for gag (Fig. (Fig.2C).2C). Extending the culture time beyond 96 h was not feasible, as primary DCs are extremely prone to death in long-term culture. Increased cell death was evident in virus-exposed DC cultures at time points longer than 96 h.
Because the production of p24 Gag is a relatively late stage of the virus replication cycle, we hypothesized that some block to reverse transcription of these HIV-2 isolates may occur in DCs. We questioned whether the very earliest minus-strand DNA intermediate (−sssDNA) was being generated in HIV-2-exposed DCs. We found very low copy numbers of −sssDNA in RH2-3-, RH2-13-, and RH2-14-exposed mDCs and pDCs by qPCR at up to 24 h postexposure. Substantially greater quantities of −sssDNA were detected in CBL-20-exposed DCs. These levels of −sssDNA were consistent with the amounts we found in our viral stocks and were resistant to DNase treatment. We cannot eliminate the possibility that the −sssDNA we detected in virus-exposed DCs was residual virus-associated −sssDNA from our original inoculum (unpublished data).
Viral replication is observed in HIV-1-exposed mDCs and pDCs in the absence of prior stimulation or activation of the DCs (38, 59). However, infection of primary CD4+ T cells is much more efficient when the cells are activated. Nonproliferating, quiescent T cells can be infected by HIV-1 (20, 77, 78), but the virus often does not undergo complete reverse transcription and remains in a labile state until the infected cell is stimulated (76). We hypothesized that HIV-2 may remain in an incompletely reverse-transcribed labile state in DCs and that complete reverse transcription and productive infection may be induced by stimulation.
We exposed immature mDCs and pDCs to HIV-1 or HIV-2 isolates for 72 h and stimulated some cultures with a TLR7/8 agonist in the final 24 h of exposure. We found that stimulation of mDCs (Fig. (Fig.3A)3A) and pDCs (Fig. (Fig.3B)3B) did not induce an increase in viral replication or the proportion of DCs productively infected with HIV-1 BaL (38, 59), as measured by qPCR for gag. Similarly, stimulation did not lead to productive infection of mDCs with any of the four HIV-2 isolates (Fig. (Fig.3A).3A). In CBL-20-exposed pDCs, where small numbers of gag copies could be detected by qPCR, there was a diminution in the proportion of gag-positive cells in cultures where pDCs were stimulated in the final 24 h with the TLR7/8 ligand (Fig. (Fig.3B).3B). Presumably, this decrease in productive infection is related to downregulation of entry molecules CD4 and CXCR4 on pDCs upon TLR7/8 stimulation. In mDCs and pDCs stimulated with TLR7/8 for 12 h prior to exposure to virus, there was also no induction of productive infection by HIV-2 isolates (Fig. 3C and D).
In vitro, primary mDCs and pDCs exposed to HIV-1 BaL and IIIB have been shown to partially upregulate the costimulatory molecules CD40, CD83, and CD86 (38, 59). This maturation was partial; it was always exceeded by the full maturation induced by stimulation through TLR7/8 and was present even in the absence of productive HIV-1 infection. We observed a similar phenomenon in DCs exposed to HIV-2. We documented here that mDCs and pDCs exposed to any of the four HIV-2 isolates did not become productively infected (Fig. (Fig.11 and and2);2); however, HIV-2 exposure did result in partial maturation of mDCs (Fig. (Fig.4A).4A). Expression of the costimulatory molecule CD86 increased on a proportion of mDCs exposed to HIV-2 isolates for 72 h, to comparable levels to those on mDCs exposed to HIV-1 BaL or IIIB. However, this level of maturation was only partial, as only a proportion of exposed cells upregulated CD86, in contrast to complete upregulation of CD86 on mDCs exposed to a TLR7/8 agonist in the final 24 h of exposure. These data were consistent for 10 individual donors. Additionally, mDCs stimulated through TLR7/8 produced IL-12p70, but neither HIV-1- nor HIV-2-exposed mDCs secreted detectable IL-12 in the absence of TLR stimulation (unpublished data).
pDCs exposed to HIV-1 or HIV-2 showed little to no upregulation of CD86 expression (Fig. (Fig.4B).4B). These data were reproduced in a total of six donors. However, in the presence of the TLR7/8 agonist, there was a substantial upregulation of this costimulatory marker, indicating that these pDCs were capable of responding to TLR7/8 stimulation and undergoing maturation. As reported previously, HIV-1-exposed pDCs produced substantial amounts of IFN-α (22, 38, 73), with up to 3,000 pg/ml in pDCs exposed to HIV-1 BaL or IIIB at an MOI of 0.25 (Fig. (Fig.4C).4C). Similarly, although HIV-2 exposure did not result in detectable productive infection of pDCs, IFN-α was induced at concentrations not substantially lower than those induced by HIV-1 exposure. In both HIV-1 and HIV-2 cultures stimulated with TLR7/8 ligand, pDCs produced maximal amounts of IFN-α upwards of 10,000 pg/ml. These data indicate that although HIV-2 exposure does not result in productive infection or maturation of pDCs, these cells respond functionally to exposure to this virus.
Physical associations between DCs and adjacent CD4+ T cells, so-called “infectious synapses,” are prime sites for HIV-1 replication in vitro and in vivo (12, 23, 25, 50, 53). We have previously shown that HIV-1 is transferred to autologous CD4+ T cells by productively infected mDCs and pDCs; in particular, activated antigen-specific CD4+ T cells were the primary recipients of this transferred HIV-1, and conjugates of DCs and autologous CD4+ T cells were the primary source of the transferred virus (38).
We examined whether HIV-2-exposed mDCs transferred HIV-2 to autologous CD4+ T cells during antigen presentation (Fig. (Fig.5).5). Primary mDCs and pDCs were isolated from healthy CMV-seropositive donors and exposed to HIV-1 or HIV-2 isolates for 72 h. Exposed mDCs were then washed thoroughly and cocultured with autologous CFSE-labeled CD4+ T cells in the presence of inactivated CMV lysate or SEB for 3.5 days. This protocol ensured that autologous CD4+ T cells were exposed exclusively to mDCs infected with HIV-2 or associated with HIV-2 virions; hence, any virus transferred would come from mDCs, not from residual input virus inoculum. Transfer of virus to autologous T cells was then analyzed by measuring intracellular p24 expression. To quantify p24 expression in CD4+ T cells exclusively, only CFSE+ cells negative for CD11c expression were analyzed. This excluded any T cells that were in conjugation with CD11c+ mDCs and any CD11c+ mDCs expressing p24.
We have previously shown that HIV-1 preferentially infects HIV-1-specific CD4+ T cells (17). In our model in vitro system, primary DCs isolated from CMV-seropositive donors were exposed to virus for 72 h and then cocultured with autologous CFSE-labeled CD4+ T cells in the presence of CMV lysate for 3.5 days. Since CMV-specific CD4+ T cells are preserved in healthy donors and are stimulated by CMV lysate, we created a scenario very similar to what we hypothesize happens when HIV-infected DCs come into contact with HIV-specific CD4+ T cells in vivo. We specified four distinct populations of autologous CD4+ T cells, namely, those that are specific for CMV (and do or do not divide or produce IFN-γ, IL-2, or TNF-α in response to CMV antigen) and those that are not CMV specific (and do or do not divide or produce cytokine in response to CMV antigen), as shown in Fig. Fig.5A.5A. As previously reported, 3.5 days after coculture, HIV-1 BaL-exposed mDCs transferred virus to autologous CD4+ T cells (38). CMV-specific CD4+ T cells, those cells that were responding to CMV antigens presented by BaL-infected mDCs, were preferentially infected (Fig. 5B and C). Significant proportions of cytokine-positive divided cells and cytokine-positive undivided cells were p24 positive at 3.5 days postcoculture. A substantial but smaller proportion of cytokine-negative divided cells were infected. In contrast, virtually no transfer of virus to the nonresponding (cytokine-negative, undivided, and presumably non-CMV-specific) CD4+ T cells was detected.
In contrast, little to no HIV-2 was transferred to autologous CD4+ T cells (Fig. 5B and C). There was virtually no transfer of the primary isolate RH2-3 to autologous CD4+ T cells in two individual donors. The same lack of transfer was seen for RH2-13 and RH2-14 in the one donor we tested using these two viruses (unpublished data). This lack of transfer of primary HIV-2 isolates from mDCs to autologous CD4+ T cells was observed in two donors by using CMV as the model antigen and also when SEB was used to stimulate T cells polyclonally (unpublished data). Although we detected virtually no infection of CBL-20-exposed mDCs at 72 h by intracellular p24 staining or qPCR (Fig. (Fig.11 and and2),2), we found that CBL-20 was transferred to a small proportion of autologous CD4+ T cells (Fig. 5B and C) in an antigen-specific manner.
We have previously shown that mDCs and pDCs can capture HIV-1 in the first few hours of exposure and can transfer virus to autologous CD4+ T cells in the absence of productive infection of the DCs (38). We questioned whether mDCs could transfer HIV-2 captured soon after virus exposure to autologous CD4+ T cells, without the need for initiating a productive infection of these DCs. We exposed mDCs to HIV-1 or HIV-2 isolates for 6 h, washed them, and added autologous CFSE-labeled CD4+ T cells for 3.5 days. Transfer of virus to T cells was measured by intracellular p24 staining. HIV-1 BaL was transferred to T cells, exclusively to CMV-specific CD4+ T cells, in this shorter virus exposure period, but there was little to no passing of HIV-2 to T cells (Fig. (Fig.5D).5D). As seen with the 72-hour exposure period, CBL-20 was transferred to a small proportion of T cells, but the primary isolate RH2-3 was not. In total, these data suggest that the lack of productive infection of DCs by HIV-2 and the absence of transfer of virus to antigen-specific CD4+ T cells may be a mechanism for preserved HIV-2-specific CD4+ T-cell help in vivo. It should be noted, however, that our experimental conditions cannot rule out the transfer of HIV-2 by mature DCs to T cells in trans.
We hypothesized that the lack of transfer of HIV-2 from DCs to autologous CD4+ T cells we observed in vitro may manifest itself in a lack of or reduced transfer of HIV-2 to HIV-2-specific CD4+ T cells in vivo. PBMC were isolated from 18 HIV-2-infected individuals, and we sorted HIV-2-specific memory CD4+ T cells producing any of five functional parameters (IFN-γ, IL-2, TNF-α, MIP-1β, or CD107a) in response to an HIV-2 Gag peptide stimulus. Additionally, we sorted bulk CD57− memory and CD57+ memory CD4+ T cells which were not specific for HIV-2. We used qPCR to measure total HIV-2 gag DNA in each of these three memory CD4+ T-cell populations as a quantitative record of infection history. We were able to sort sufficient HIV-2-specific CD4+ T cells to complete this analysis for seven individuals. These HIV-2-positive individuals had high magnitudes of HIV-2-specific CD4+ T-cell help and preserved CD4 percentages, and six of the seven had undetectable viral loads. Clinical data are shown in Table Table1.1. In a striking contradiction to our hypothesis, we found significantly more HIV-2 viral DNA in HIV-2-specific memory CD4+ T cells than in CD45RO+ CD57− memory or CD45RO+ CD57+ memory CD4+ T-cell populations (Fig. (Fig.6).6). HIV-2-specific CD4+ T cells contained 1.95- to 55.64-fold higher copy numbers of HIV-2 DNA (mean, 13.6-fold) than did CD57− memory cells. Assuming one DNA copy per cell, 0.34% to 6.54% (mean, 1.6%) of HIV-2-specific CD4+ T cells contained HIV-2 viral DNA compared with 0.009% to 0.56% (mean, 0.23%) of CD57− memory CD4+ T cells. CD57+ memory CD4+ T cells are less frequent in peripheral blood, and as such, we sorted sufficient cells for this analysis from only four individuals. This population contained the least amount of HIV-2 viral DNA. CD57+ memory cells contained 1.9- to 34.9-fold (mean, 10.5-fold) less HIV-2 viral DNA than did CD57− memory CD4+ T cells. As reported for HIV-1-infected individuals (9), within individual HIV-2-infected individuals we found that CD57− memory CD4+ T cells were approximately 10 times more likely to harbor HIV-2 than were CD57+ memory CD4+ T cells. In summary, these data suggest that preferential infection of HIV-specific CD4+ T cells may not be restricted to HIV-1-infected individuals, but rather may be a hallmark of HIV infection in general. Like HIV-1-specific CD4+ T cells, HIV-2-specific CD4+ T cells were also more likely to be infected in vivo than memory CD4+ T cells of other specificities.
DCs are primarily located in mucosal and lymphoid tissues, including both oral and vaginal mucosal surfaces, and have been shown to be associated with HIV-1 virions in mucosal tissues of infected humans (24, 25) and SIV-infected rhesus macaques (33, 61) very early in infection. It has been suggested that DCs may in fact be the first cell type encountered by HIV-1 after sexual transmission (15, 32, 79). Early in vitro studies documented that DCs pulsed with HIV-1 could facilitate infection of cocultured T cells (12, 50-53), and recent work shows that DCs may be a key mediator in the spread of HIV-1 to CD4+ T cells in vivo. DCs appear to support HIV-1 replication by several distinct pathways, i.e., by becoming productively infected and passing virus to T cells (11, 35, 38, 47, 66), by transmitting virus through an infectious synapse of DCs and T cells (1, 45), or by capturing virions in the absence of productive infection and transferring the virus to CD4+ T cells in trans (3, 26, 27, 29, 30, 65, 68, 70, 71). Thus, it appears that the role of DCs as antigen-presenting cells may be sabotaged to promote and enhance HIV-1 dissemination. DCs infected with or carrying HIV-1 virions may inadvertently present the virus to a pool of highly susceptible target cells during antigen presentation and allow the infection to become established in lymphoid tissues, contributing to HIV-1 pathology and enhancing viral dissemination.
In contrast to HIV-1, HIV-2 is associated with a considerably slower progression to disease (41, 42), an overall mortality rate only twice that of uninfected individuals (54), a much lower plasma viral load (4), and a decreased frequency of transmission (34, 69). HIV-1 infection of DCs appears to be intimately linked with transfer of virus to CD4+ T cells, which may be an important contributor to HIV-1 pathogenesis. There have been no studies of the association of DCs with HIV-2 in vivo or in vitro. In this study, we sought to compare the susceptibilities of DCs to infection with HIV-1 and HIV-2 in an attempt to understand more fully the differential pathogenesis of these two infections.
We exposed mDCs and pDCs to four HIV-2 isolates, namely, the lab-adapted CBL-20 strain and three primary isolates from long-term asymptomatic HIV-2-positive individuals, i.e., RH2-3, RH2-13, and RH2-14, and monitored the cells for infection by flow cytometry and qPCR. We were unable to infect either mDCs or pDCs with any of the HIV-2 isolates at levels detectable by intracellular p24 staining, even after 4 days of exposure (Fig. (Fig.11 and and2).2). We were also unable to detect infection by a more sensitive DNA PCR assay in any of the HIV-2-exposed DCs, except for a small PCR signal indicative of infection of pDCs by CBL-20 (Fig. (Fig.1F).1F). In contrast, HIV-1 BaL efficiently infected both mDCs and pDCs. The inability of the HIV-2 isolates to infect DCs did not extend to T cells, as all of the HIV-2 isolates were clearly capable of infecting primary CD4+ T cells (Fig. (Fig.1H).1H). The in vitro replication capacity of these primary HIV-2 isolates in primary CD4+ T cells is lower than that for HIV-2 strains isolated from progressing patients and slightly lower than that for HIV-1 variants from long-term nonprogressors (7). However, levels of infectivity of susceptible T-cell lines do not differ between these primary HIV-2 isolates and HIV-1 strains from long-term nonprogressors (7). The lab-adapted HIV-2 strain CBL-20 infects primary CD4+ T cells as well as HIV-1 BaL does (Fig. (Fig.1H1H).
It was shown recently that primary isolates of HIV-2 and HIV-1 have distinct patterns of replication kinetics in primary monocyte-derived macrophages (MDMs) (40). HIV-2 can enter, reverse transcribe, and produce infectious virions in MDMs, but unlike HIV-1, where infectious virions are produced at a steady rate, HIV-2 showed an initial burst of viral replication followed quickly by a state of latency and an absence of infectious virion production after 3 days. In contrast, we found that HIV-2 isolates infected primary DCs very poorly, although these cells express the necessary CD4, CCR5, and CXCR4 entry receptors for these viruses. This suggests that either the HIV-2 virions were not entering DCs or the viral RNA was not being fully reverse transcribed. Notably, there were clear differences between isolates. For the X4 isolate, CBL-20, we were able to detect low-level infection of pDCs by qPCR using gag primers, indicating that near-to-full reverse transcription was occurring. Furthermore, we showed that this infection was inhibitable by anti-CD4, indicating that the infection was CD4 dependent. In contrast, we found no evidence, by p24 staining or gag qPCR, of DC infection with the R5 isolates of HIV-2. Investigating this block to infection further, we found very low copy numbers of −sssDNA in RH2-3-, RH2-13-, and RH2-14-exposed mDCs and pDCs by qPCR at up to 24 h postexposure. However, these levels of −sssDNA were consistent with the amounts of −sssDNA we found in our viral stocks. We detected −sssDNA which was resistant to DNase treatment (i.e., it was contained within the virion) in our cell-free HIV-2 viral preparations, suggesting that there may have been some amount of endogenous reverse transcription of our HIV-2 strains (80). We cannot eliminate the possibility that the small quantity of −sssDNA we detected in virus-exposed DCs was residual virus-associated −sssDNA from our original inoculum and therefore cannot document that even partial reverse transcription occurs in DCs exposed to primary isolates of HIV-2.
Thus, since the levels of −sssDNA were not sufficient to indicate that early steps of reverse transcription were occurring and since no gag DNA was detected by 4 days postinfection, it appears that HIV-2 may be deficient at completing reverse transcription in DCs. It is possible that the lack of productive infection of primary DCs exposed to HIV-2 could be attributed to inhibition of the viral life cycle at a very early postentry step. Tripartite motif protein 5α (TRIM5α) mediates a dominant block to reverse transcription by interfering with capsid uncoating (63, 64, 72). The human version of TRIM5α is not effective against HIV-1, but it is plausible that HIV-2 may be more sensitive to the effects of this cytoplasmic body protein. If HIV-2 isolates were more susceptible to this dominant block to infection by TRIM5α, this could help to explain the lack of productive infection of primary DCs by HIV-2 that we observed and potentially help to explain the lower pathogenicity of HIV-2 in vivo. However, we have also shown that primary CD4+ T cells accommodate both entry and reverse transcription of HIV-2. Thus, for TRIM5α to be implicated in this specific inability of HIV-2 to infect DCs, one would have to hypothesize a greater level of activity of TRIM5α in DCs than in T cells.
APOBEC (apolipoprotein B mRNA-editing enzyme, catalytic polypeptide-like) (57) cytidine deaminase proteins can assemble into HIV particles and mediate hypermutation of early reverse transcripts, leading to a loss of infectivity (31, 36, 39, 81). The HIV Vif protein has evolved to protect wild-type HIV strains from the effects of APOBEC (57) by recruiting the protein to the proteasome (46, 58, 62, 74, 75). Although HIV-2 Vif has been shown to function much like HIV-1 Vif in blocking the incorporation of APOBEC proteins into the virion (67), it is possible that APOBEC may still be an active contributor to the lack of productive infection we have seen in HIV-2-exposed DCs. It was recently shown that cellular APOBEC3G strongly protects unstimulated peripheral blood CD4+ T cells from HIV-1 infection (14). It is possible that APOBEC proteins affect the ability of HIV-2 isolates to infect nondividing, nonactivated cells (such as DCs) preferentially and may be relatively ineffective at blocking HIV-2 infection of T cells. More studies are needed to assess the effects of APOBEC proteins on HIV-2 infection of both T cells and cells of other lineages.
Productive infection of primary CD4+ T cells often requires proliferation of the infected cell. Infection of nonproliferating, quiescent T cells can lead to a state where the virus does not undergo complete reverse transcription and remains in a labile state until the infected cell is stimulated (20, 76-78). Similarly, Marchant et al. have recently shown that HIV-2 can infect MDMs but that the infection resolves into a latent state soon after infection (40). HIV-2 production could be induced from latently infected MDMs by stimulation with lipopolysaccharide. We hypothesized that HIV-2 may remain in an incompletely reverse-transcribed labile state in DCs and that reverse transcription and productive infection may be induced by stimulation. However, activation of DCs through TLR7/8 stimulation did not induce productive infection of DCs with any of the HIV-2 isolates, suggesting that either HIV-2 does not reside in a partially reverse-transcribed labile state in DCs (as suggested by the lack of −sssDNA) or TLR stimulation was not sufficient to induce the completion of reverse transcription and subsequent productive infection.
We showed previously that HIV-1 is transferred to autologous CD4+ T cells by productively infected mDCs and pDCs in vitro; in particular, activated antigen-specific CD4+ T cells were the primary recipients of this transferred HIV-1, and conjugates of DCs and autologous CD4+ T cells were the primary source of the transferred virus (38). In this study, we found that in addition to a lack of productive infection of mDCs and pDCs by HIV-2, there was virtually no transfer of HIV-2 from exposed DCs to autologous CD4+ T cells. This suggested that HIV-2-specific CD4+ T cells may not be infected through DC interactions in vivo, which could be a mechanistic explanation for our finding of preserved HIV-2-specific CD4+ T-cell help (19). Interestingly, this lack of transfer of HIV-2 from exposed DCs to autologous CD4+ T cells in vitro was not reflected in a lack of preferential infection of HIV-2-specific CD4+ T cells in vivo. We found significantly more HIV-2 viral DNA in sorted HIV-2-specific memory CD4+ T cells than in CD45RO+ CD57− or CD45RO+ CD57+ memory CD4+ T cells of other specificities, suggesting that HIV-2-specific CD4+ T cells are infected more frequently than other peripheral blood memory CD4+ T cells.
Increased levels of HIV-2 viral DNA in HIV-2-specific CD4+ T cells are not incompatible with our finding of preserved polyfunctional CD4+ T cells in HIV-2-infected individuals (19; M. G. Duvall, M. L. Precopio, D. A. Ambrozak, A. Jaye, A. J. McMichael, H. C. Whittle, M. Roederer, S. L. Rowland-Jones, and R. A. Koup, submitted for publication). Indeed, vaccinated animals challenged with SIVmac251 have vaccine-induced SIV-specific CD4+ T cells that harbor increased levels of SIV DNA compared to other memory CD4+ T cells (44) yet are still protective after challenge in vaccinated animals. Thus, it appears that preferential infection of de novo or vaccine-generated CD4+ T cells and well-controlled infection are not mutually exclusive. Additionally, it has been known for some time that defective HIV-1 proviruses accumulate in CD4+ T cells in vivo (55). Our assay does not discriminate integrated from nonintegrated and replication-competent from incompetent (graveyard) viral DNA. It is possible that some of the DNA we detected in HIV-2-specific CD4+ T cells could be defective “graveyard sequences.” This would suggest that even if HIV-2-specific CD4+ T cells contained more HIV-2 DNA than other memory CD4+ T cells, graveyard sequences would not contribute to the production of infectious virions. This could be reflected in the low to undetectable viral loads in most HIV-2-positive individuals and would not hinder the development of highly polyfunctional HIV-2-specific CD4+ T cells.
Another possible explanation for our finding of increased quantities of viral DNA in HIV-2-specific CD4+ T cells is that these cells could have an increased life span compared to other memory CD4+ T cells. HIV-1-specific CD4+ T cells are preferentially infected in vivo (17) and are difficult to detect at appreciable levels in most HIV-1-infected individuals. It was recently shown that preferential infection shortens the life span of HIV-1-specific CD4+ T cells (10), helping to explain the virtual absence of HIV-1-specific CD4+ T cells in vivo. It is possible that the relative absence of hyperimmune activation and antigenic stimulation in HIV-2-infected individuals may select for the preferential survival of HIV-2-specific CD4+ T cells. Finally, despite our inability to measure infection of DCs by HIV-2 in vitro, there may be adequate infection in vivo to mediate preferential infection of HIV-2-specific CD4+ T cells. Alternatively, despite our previous work, DC infection may not be the mechanism through which preferential infection of HIV-specific CD4+ T cells occurs. Despite these unknowns, our data clearly indicate that HIV-2 interacts with DCs in a fundamentally different manner from that for HIV-1, and this difference may be crucial to the differences in pathogenicity of these two viruses.
This work was funded by the National Institutes of Health.
Published ahead of print on 3 October 2007.