|Home | About | Journals | Submit | Contact Us | Français|
Human immunodeficiency virus type 1 (HIV-1)-infected and activated CD4+ T cells have short half-lives in vivo (<2 days). We have established an in vitro culture system in which infected T cells are turned over frequently to provide a model system that examines this important facet of in vivo HIV-1 replication. We observed that virus replication in T cells under rapid-turnover conditions was possible only when immature dendritic cells or DC-SIGN-expressing cells mediated HIV-1 transmission to T cells. Virus replication was initiated more rapidly in T cells infected with the cell-associated form of virus compared to infection by the cell-free route. This accelerated transfer of virus required adhesion molecule-mediated interactions between the virus-presenting cell and T cell, but surprisingly, HIV-1 transfer could occur independently of DC-SIGN (DC-specific intracellular adhesion molecule 3 [ICAM-3]-grabbing nonintegrin)in the dendritic-cell-T-cell cocultures. These results suggest that dendritic cell-mediated transmission of HIV-1 enables virus replication under conditions of rapid cell turnover in vivo.
Studies describing the viral dynamics of human immunodeficiency virus type 1 (HIV-1) infection have demonstrated high levels of viremia that are relatively stable throughout the course of viral infection (9, 26, 44, 57). Estimates have placed the average daily production of HIV-1 at approximately 1010 virions, with much of the virus replication occurring in the paracortical regions of the peripheral lymphoid organs (11, 42, 44) that are composed predominantly of dendritic cells (DC) and CD4+ T cells (23). This high level of viral replication in the lymphoid tissue also drives the infiltration of anti-HIV CD8+ cytotoxic T lymphocytes (CTLs) that target and destroy HIV-1-infected CD4+ T cells (7, 30, 33, 40). A number of studies have also found higher rates of CD4+-T-cell turnover for patients with HIV-1 infection than in uninfected people (24, 32, 35, 36). Together, these findings suggest that HIV-infected T cells in vivo have a premature loss of productive life, with some estimates placing the half-life of an infected and activated CD4+ T cell at less than 2 days (44). Thus, the virus has a short window of opportunity to infect a target cell, establish a productive infection, and produce new virus particles in the infected host.
Most in vitro models of HIV-1 infection generally fail to take into account the short half-life of infected cells in vivo. Therefore, we recently developed an in vitro model system with rapid-cell-turnover kinetics by artificially killing cells in an HIV-1-infected culture every 2 to 3 days by exposure to a cytocidal agent (22). Fresh, uninfected T cells were added back to the culture to maintain virus replication (22). Surprisingly, a sustained high level of viral replication in T cells was initially not achieved in the rapid-turnover assay; however, continued propagation of the virus under these conditions led to the emergence of a new viral variant that could replicate under conditions where an infected cell has a short half-life. This variant, which encodes a mutation in the vpu open reading frame, enhanced cell-to-cell transmission of virus (22). Thus, we speculated that replication of virus under rapid-cell-turnover conditions in vitro was dependent on cell-to-cell transfer of virus.
In this study, we sought to determine conditions under which wild-type virus could sustain viral persistence in the face of rapid host cell turnover. We hypothesized that DC could play a role in sustaining HIV-1 replication in T cells with short half-lives. Although productive infection of DC by itself has been controversial (58), virus-bearing DC may facilitate a more efficient spread of virus to surrounding permissive T cells (3, 14, 18, 19, 25, 28, 47). In vitro data have demonstrated that DC can bind HIV for a long period of time and still induce efficient trans infection via intimate interactions with activated CD4+ T cells (6, 13, 48). A number of studies have demonstrated that HIV-1 bound to DC-SIGN (DC-specific intracellular adhesion molecule 3 [ICAM-3]-grabbing nonintegrin), a DC-associated HIV attachment factor, can be transmitted to CD4+ T cells, resulting in efficient in trans virus infection (1, 2, 15, 45). Expression of DC-SIGN in cells lining the sinusoidal endothelium in the T-cell zones of the lymph nodes (2, 15, 16, 29, 46) suggests a potential for continued interaction between virus-bearing DC and T cells.
We found that immature DC or cells that express DC-SIGN in cocultures with T cells are able to sustain infection under rapid-turnover conditions by mediating cell-associated transmission of wild-type HIV-1. In contrast to a cell-free virus challenge, virus replication in activated CD4+ T cells infected with a cell-associated form of virus was initiated within 15 h of infection. Furthermore, there was a 10-fold increase in the level of virus replication observed within the first 2 days of infection in the T cells infected by the cell-associated form of virus compared to infection by cell-free virus particles. However, the presence of virus in a DC-SIGN-associated fashion was not sufficient by itself to mediate efficient virus transfer to T cells. Rather, it required cell-to-cell adhesion between DC and T cells to increase the kinetics of the early steps of viral replication. This underscores the importance of formation of an intimate contact between the virus-presenting cell and T cells and the relevance of adhesion molecules that mediate such interactions for the rapid establishment of productive virus infection. Finally, we find that immature DC can bind HIV in both a DC-SIGN dependent and a DC-SIGN independent manner and that the virus bound independently of DC-SIGN can also be efficiently transmitted to T cells.
Immature DC were generated from the blood of healthy donors by using defined protocols (51). Briefly, peripheral blood mononuclear cells (PBMC) obtained by Ficoll-Hypaque sedimentation were depleted of nonadherent cells by adherence to plastic and vigorous washing with phosphate-buffered saline. Adherent cells were maintained for 7 days in complete RPMI (RPMI, 10% fetal bovine serum,25 mM HEPES) containing 1,000 U each of granulocyte-macrophage colony-stimulating factor and interleukin-4 (IL-4) (R & D Systems, Minneapolis, Minn.) per ml. Cytokines were replenished every third day. DC were characterized by staining with monoclonal antibodies (MAbs) to HLA-DR, CD83, CD86, CD14, and CD2 (all from Coulter, Hialeah, Fla.) and DC-SIGN (MAb 507) (59) and were reproducibly >95% DC-SIGN positive and CD14low after 7 days in culture. These primary DC (with <5% contaminating T cells) were routinely used for infections 7 to 8 days after culture in the presence of granulocyte-macrophage colony-stimulating factor and IL-4. Primary activated human CD4+ T cells were generated from PBMC (107) by staining with a MAb to CD4 (Leu3A; Becton Dickinson, San Jose, Calif.) and sorted by a fluorescence-activated cell sorter (21). CD4+ T cells (2 × 105) were then cultured in T-25 flasks with 25 × 106 γ-irradiated PBMC as feeders and stimulated with anti-CD3 (OKT3, 30 ng/ml) in RPMI supplemented with 10% human serum. IL-2 was added the next day to 50 U/ml and replenished every 2 to 3 days thereafter. The purity of the CD4+-T-cell population was routinely assessed by staining with anti-CD2 and anti-CD4 MAb (>99%). These activated CD4+ T cells (HLA-DR+, CD25+, and CD69+) were used for infections at 10 to 12 days poststimulation with anti-CD3 MAb. The Jurkat T-cell line, THP1, and THP1—DC-SIGN promonocytic cell lines have been described before (2, 15) and were maintained in complete RPMI. Jurkat cells stably expressing the HIV-1 coreceptor CCR5 were constructed by infection with a retroviral vector, pBabe-puro-CCR5 (56), and selection in puromycin (0.4 μg/ml) containing complete RPMI. The molecular clone HIVLai (CXCR4-tropic) and the molecular clone NL4-3/Ba-L, expressing the CCR5-topic Ba-L env, have been described previously (10, 22). Virus stocks were generated by calcium phosphate-mediated transfections of HEK293T cells with the proviral DNA (22). Cell virus supernatants were collected 48 h posttransfection, and virus titers were determined by multinuclear activation of galactosidase indicator (MAGI) cell assays (31).
Rapid-turnover assays with DC- and DC-SIGN-expressing cells were performed with methodologies similar to those reported previously, but with the following modifications (Fig. (Fig.1A)1A) (22). Jurkat T cells (106), DC (105)-Jurkat (106), THP1 (105)-Jurkat (106), or THP1-DC-SIGN (105)-Jurkat (106) cocultures were infected with 0.1 ml of virus supernatant (500 ng of p24gag) for 2 h at 37°C. Cells were washed three times with phosphate-buffered saline and cultured at 106 cells/ml in complete RPMI in a 24-well tissue culture plates. Infected cultures on day 3 postinfection were exposed to mitomycin C (50 μg/ml; Sigma, St. Louis, Mo.) for 2 h at room temperature. The virus-containing supernatants were used for infecting new target cells, either DC, THP1, or THP1-DC-SIGN cells, for 2 h at 37°C. The cells were then mixed with Jurkat T cells at a 1:10 ratio. Alternatively, Jurkat T cells were infected directly with cell virus-containing supernatants for 2 h at 37°C. Following the 2-h virus exposure, the mitomycin C-naive cells were washed extensively and then cocultured with the mitomycin C-exposed cells, and the infection was allowed to proceed for 3 days, after which the same protocol was repeated. Under such conditions of viral replication, the life span of an infected cell was restricted to 3 days, and the half-life of an infected cell was reduced to <2 days. Cell supernatants were harvested at 3-day intervals, and viral replication was monitored by measuring p24gag levels by enzyme-linked immunosorbent assay (ELISA).
Twenty thousand immature DC harvested on day 7, THP1 cells, and THP1-DC-SIGN cells were washed and exposed to 20 μl of virus-containing supernatants (100 ng of p24gag) for 2 h at 37°C and then washed extensively to remove unabsorbed virus. Virus-exposed cells were then cocultured with primary, activated CD4+ T cells (virus-presenting cell/T-cell ratio was 1:10), and plated at 106 per ml in a round-bottom 96-well tissue culture plate. For cell-free infections of T cells, activated CD4+ T (2 × 105) cells were directly exposed to 20 μl of cell viral supernatants (100 ng of p24gag) for 2 h, washed extensively, and plated at 106 per ml. Virus replication was monitored by ELISA measurement of p24gag antigen produced in culture supernatants. For real-time PCR analysis, total cellular RNA (including viral RNA) was isolated by lysis of 2 × 105 cells by a total RNA isolation kit (SV RNA isolation kit; Promega, Madison, Wis.), per the manufacturer's instructions. Total RNA was reverse transcribed with random hexamers as primers and Superscript II reverse transcriptase (Invitrogen, Carlsbad, Calif.). Quantitative (fluorescence-monitored) PCR was performed using 2 × 103 cell equivalents to measure viral tat-rev-nef multiply spliced RNA and cellular β-actin RNA, using Taqman technology (34) on an ABI-7900 HT kinetic PCR instrument (ABI, Foster City, Calif.). The sequences of the forward and reverse tat-rev-nef primers were 5′-GCGACGAAGACCTCCTCAAG-3′ (+5992 to + 6011) and 5′-GAGGTGGGTTGCTTTGATAGAGA-3′ (+6032 to +6045 through +8379 to +8387; spans a splice junction), respectively, and the forward and reverse gag detection primers were 5′-TGACAAATAATCCACCTATCCCAGTA-3′ (+1538 to +1563) and 5′-GTCCTTGTCTTATGTCCAGAATGCT-3′ (+1630 to +1656), respectively. The sequences of the fluorescent DNA probes with conjugated minor groove binder groups for tat-rev-nef and gag amplicons are 5′-CAGTCAGACTCATCAAGT-3′ (+6013 to +6030) and 5′-TCCTGGGATTAAATAAAA-3′ (+1589 to +1606), respectively. Both the probes were labeled with the reporter fluorochrome 6-carboxyfluorescein at the 5′ termini and with a nonfluorescent quencher at the 3′ termini (ABI, Foster City, Calif.). Reactions were carried out in a 50-μl volume, with 200 nM fluorescent probe and either 60 nM forward primer and 60 nM reverse primer for the tat-rev-nef amplification or 900 nM forward primer and 900 nM reverse primer for the gag amplification. The reaction times and temperatures were 2 min at 50°C, 5 min at 95°C, and then 40 cycles of 15 s at 95°C and 1 min at 60°C. A standard curve for viral RNA was prepared using serial dilutions of synthetic gag and nef transcripts. These RNAs were in vitro transcribed by T7 polymerase (Promega) using DNA templates that contained either the 627-bp fragment of pLai gag (positions 1090 to 1717) or the tat-rev-nef cDNA fragment amplified by gag and the tat-rev-nef forward and reverse primer set, respectively (described above). To correct for variations in cell numbers and RNA recovery, a standard curve for β-actin RNA was generated using RNA isolated from 106 uninfected, activated primary CD4+ T cells. Serial dilutions of this RNA were obtained and reverse transcribed, as described above. Sequence detection software (version 2.0; ABI) was used to analyze the quantitative PCR amplification data.
Mouse MAbs of isotype immunoglobulin G1 that bind to unknown epitopes of human DC-SIGN and block virus transmission mediated by DC-SIGN (MAbs 507, 516, and 526) have been described before (59). Mouse MAbs to human CD4 (Sim.2) and CXCR4 (12G5) were obtained from the NIH AIDS Reagent Program. Function-blocking antibodies specific to ICAM-1, LFA-1, and LFA-3 (clones G43-25B, HA58, and L306.4, respectively) were obtained from Becton Dickinson (San Jose, Calif.). Immature DC, THP1, or THP1-DC-SIGN cells (2 × 104 each) were incubated with MAbs against DC-SIGN, CD4, or CXCR4 (each at 10 μg/ml), or mannan (20 μg/ml) for 30 min at 37°C, prior to virus exposure (100 ng of p24gag) for 2 h at 37°C in the presence of the inhibitory reagents. The cells were washed extensively to remove unabsorbed virus; cocultured with primary, activated CD4+ T cells (virus-presenting cell/T-cell ratio was 1:10); and plated at 106 per ml in a round-bottom 96-well tissue culture plate in the presence of each antibody or inhibitor. For blocking cell-to-cell interactions, 2 × 105 primary activated CD4+ T cells were incubated with MAbs against ICAM-1, LFA-1, or LFA-3 (each at 10 μg/ml) for 30 min at 37°C prior to coculture with 2 × 104 virus-exposed cells (DC, THP1, or THP1-DC-SIGN cells). Infected cocultures (106 per ml in a round-bottom 96-well tissue culture plate) were cultured for 48 h in the presence of the inhibitory antibodies. Virus replication was monitored by ELISA measurement of p24gag antigen produced in culture supernatants.
We had previously demonstrated that selection for a virus that could grow when the life span of HIV-1-infected cells was kept under 3 days resulted in a mutant that spread preferentially by cell-to-cell transmission (22). Therefore, we examined conditions that would allow wild-type virus to propagate in a similar manner. Since it had been previously shown that DC could transfer virus to T cells with extremely high efficiency (6, 13, 18, 19, 48), we tested whether the presence of DC would allow wild-type HIV to grow under conditions where the half-life of the infected cells was short. Thus, we adapted the rapid-turnover assay (Fig. (Fig.1A),1A), where DC-T-cell cocultures were infected with virus and subjected to the mitomycin C induced turnovers every 2 to 3 days (described further in the Fig. Fig.11 legend). HIVLai replicated to a similar extent in both the T cells and in the DC-T-cell cocultures under conditions where no restrictions were placed on the infected-cell life span (Fig. (Fig.1B).1B). However, in the rapid-turnover assay, viral replication was sustained over multiple passages only in the DC-T-cell cocultures and not when T cells alone were used (Fig. (Fig.1C1C).
To test if DC-SIGN presence on cells was sufficient to mediate cell-associated transmission of wild-type HIV-1 replication to T cells under rapid-turnover conditions, we used a promonocytic cell line, THP1, that does not express the HIV-1 receptor, CD4, and was engineered to express DC-SIGN (2, 15). Cocultures of T cells with THP1 cells with and without DC-SIGN (1:10 ratio of the cell types in both cocultures) were infected with HIVLai and either propagated under conditions where no restrictions were placed on the infected-cell life span (Fig. (Fig.1D)1D) or subjected to rapid-turnover assay (Fig. (Fig.1E).1E). The results show that while HIVLai replicated to a similar extent in all the cultures where no constraints were placed on the half-life of infected cells (Fig. (Fig.1D),1D), viral replication was sustained in the rapid-turnover assay only in the THP1-DC-SIGN-T-cell cocultures, and not when THP1 cells did not express DC-SIGN (Fig. (Fig.1E).1E). To determine the ability of CCR5-tropic viruses to replicate within the constraints of the rapid-turnover assay, we inoculated either Jurkat-CCR5 cells directly or Jurkat-CCR5 cells cocultured with immature DC, THP1, or THP1-DC-SIGN cells with a CCR5-tropic molecular clone, pNL4-3/Ba-L env. Virus replication was similar in all cultures tested when no restriction was placed on the half-life of the infected cells (Fig. (Fig.1F).1F). But, similar to the results obtained with the CXCR4-tropic virus, HIVLai, NL4-3/Ba-L env replication under rapid-turnover assay conditions was possible only in the DC/Jurkat-CCR5 or THP1-DC-SIGN/Jurkat-CCR5 cocultures (Fig. (Fig.1G).1G). These results suggest that under conditions where infected and previously activated T cells have a short half-life, infection by a cell-associated virus is required for virus persistence, and that accessory functions of DC cells (such as their ability to activate T cells) are not necessary in this system.
The fact that virus could be propagated under rapid-turnover conditions by cell-associated transmission but not by cell-free transmission implies that cell-associated transmission overcomes a rate-limiting step at the early stages of viral infection. To quantify the differences in kinetics between these two modes of virus transmission we used quantitative real time RT-PCR to measure the initial virus binding to target cells and the synthesis of the earliest RNA transcripts after productive infection.
We first measured the amount of cell-associated viral RNA (the incoming viral genome) found attached to T cells after exposure to cell-free virus stocks for 2 h or after coincubation with DC, THP1, or THP1-DC-SIGN cells that had been exposed to virus for 2 h. After extensive washing to remove unbound virions, the mean numbers of viral RNA copies were measured by real-time RT-PCR using primers and probe specific to HIVLai gag sequence. There was an ~10-fold difference in the level of viral RNA genome found associated with the DC (1,440 copies/cell) or THP1-DC-SIGN (2,110 copies/cell) cells compared to the T-cell cultures infected with cell-free virus stocks (140 copies/cell) (Fig. (Fig.2A).2A). Note that the THP1 cells do not express any HIV-specific attachment factors and hence exhibit negligible virus binding (1.6 copies/cell). These results are in good agreement with recently published findings of simian immunodeficiency virus particle binding to immature DC (12) and suggest that immature DC or THP1-DC-SIGN cells are more efficient than CD4+ T cells in their ability to capture free virus particles.
Detection of de novo viral transcription was accomplished by amplification with primers complementary to the HIV-1 multiply spliced RNA (in order not to detect incoming genomic RNA [Fig. [Fig.2B]).2B]). The level of de novo HIV RNA synthesis associated with each culture was quantified by comparison to a standard curve generated for the multiply spliced HIV RNA species. There was a more rapid establishment of productive viral infection in the DC-T-cell or THP-DC-SIGN-T-cell cocultures compared to the cell-free infection of T cells (Fig. (Fig.2B).2B). Infection of T cells by DC- or DC-SIGN bound virus led to a 20- to 25-fold increase in absolute number of viral transcripts at 15 h postinfection (307 copies/cell equivalent in the DC-T-cell coculture, and 371 copies/cell equivalent in the THP1-DC-SIGN-T-cell coculture) compared to the number of transcripts present at the five hour time point (11 copies and 21 copies/cell equivalent, respectively) (Fig. (Fig.2B).2B). Interestingly, T cells infected by the cell-free route appeared to be delayed in their ability to synthesize new viral messages, in that there was only a two- to threefold increase in viral transcript levels at 15 h postinfection compared to the 5-h time point (Fig. (Fig.2B).2B). Also, the number of multiply spliced transcripts in the DC (307 copies/cell)— and THP1-DC-SIGN (371 copies/cell)-T-cell cocultures were ~10-fold higher than in the cell-free virus-infected T-cell cultures (25 copies/cell) at 15 h postinfection. This ~10-fold difference persisted at the 24- and 38-h time points.
A majority of these viral mRNAs in the DC-T-cell cocultures were T cell derived and not synthesized in DC, since we have observed that HIV-1 infection of DC alone results in extremely low levels of viral transcripts (Fig. (Fig.2B).2B). Furthermore, the increase in HIV-specific transcription in DC-T-cell and THP-DC-SIGN-T-cell cocultures seen within 15 h postinfection represents de novo viral RNA synthesis, as cultures infected in the presence of zidovudine (50 μM) demonstrated negligible increase in HIV transcripts (data not shown). These results suggest that the observed increase in the number of viral transcripts in the DC— and THP1-DC-SIGN-T-cell cocultures at early times postinfection (~10-fold, Fig. Fig.2B)2B) correlate with the increased number of virus particles (~10-fold) found attached to the target cell surface at the zero hour time point (Fig. (Fig.2A2A).
Finally, we sought to detect the appearance of p24gag in the cell supernatants as a quantitative measure of viral burst size (Fig. (Fig.2C).2C). We were able to detect p24gag in the cell-free supernatants within 24 h postinfection in the DC-T-cell and THP-DC-SIGN-T-cell cocultures, but not in the THP1-T-cell coculture or in cultures where T cells were infected by cell-free virus particles. This early detection of p24gag protein correlated well with the early appearance of viral transcripts (15 h postinfection) in the DC- and the THP1-DC-SIGN-T-cell cocultures (Fig. (Fig.2B).2B). Furthermore, the level of virus production was reproducibly much higher at 48 h postinfection (8- to 10-fold) for the DC-T-cell and THP1-DC-SIGN-T-cell cocultures than in cultures where T cells were infected by cell-free virus stocks. In sum, these results suggest that regardless of the mode of virus transmission to T cells, the early steps of the viral replicative cycle (entry, uncoating, reverse transcription, integration, and transcription) occur at similar efficiencies. However, our results show that the accumulation of de novo-synthesized viral RNA was significantly more rapid when virus was presented to T cells in a DC (or DC-SIGN)-associated form. Thus, while both modes of infection ultimately reach the same levels of viral transcript production, the kinetics in T cells mediated by cell-associated form of virus is much quicker. Hence, the rate-limiting step in the viral life cycle in activated CD4+ T cells is the initial attachment of the viral particles to the cell surface. This explains why cell-associated virus transmission to T cells is necessary for propagation of HIV under rapid-turnover conditions.
We next evaluated whether physical interactions via adhesion molecules was necessary for the direct transfer of virus from one cell to another. Monocyte-derived DC express a variety of cell adhesion molecules, including LFA-1 (CD11a), ICAM-1 (CD54), and LFA-3 (CD58), that mediate formation of the DC-T-cell conjugate or the immunological synapse (5, 17). Therefore the roles of these adhesion molecules in transmission of DC-SIGN- or DC-bound HIV to T cells was evaluated by incubating activated CD4+ T cells with antibodies against LFA-1, LFA-3, or ICAM-1, or a combination of all three antibodies together. These T cells were then mixed with DC or THP1-DC-SIGN cells that had been previously pulsed with virus, and the cocultures were maintained in the presence of the antibodies (Fig. (Fig.33).
Virus production in both THP1-DC-SIGN- and DC-T-cell cocultures was significantly diminished in the presence of anti-ICAM-1, anti-LFA-1, or anti-LFA-3 MAbs. Furthermore, there was a 10-fold decrease in virus replication when all three antibodies were present in the cocultures (Fig. (Fig.3).3). In fact, level of virus replication achieved in the presence of antibodies against adhesion molecules was similar to that achieved by cell-free infection of CD4+ T cells. In contrast, treatment of virus-containing supernatants with anti-ICAM-1, anti-LFA-1, and anti-LFA-3 antibodies had no effect on virus binding to DC or THP1-DC-SIGN cells, or the subsequent virus transfer to T cells (data not shown), implying that putative adhesion molecules expressed in the viral membrane (4) are not required for DC- or THP1-DC-SIGN-mediated virus transfer to T cells. These results also suggest that the potential formation of DC-T-cell or THP1-DC-SIGN-T-cell conjugates through the ICAM-1/LFA-1 or LFA-3/CD2 pathways might play an important role in the initial virus transmission to T cells, and probably also for the subsequent virus spread.
We further wanted to determine if DC-SIGN-mediated binding of HIV-1 to DC could entirely account for their ability to transfer virus to T cells. To this end, immature DC and THP1-DC-SIGN cells, which express high levels of DC-SIGN, were pulsed with HIVLai or NL4-3/Ba-L env for 2 h, extensively washed, and cocultured with activated CD4+ T cells. To assess the contribution of DC-SIGN, we examined the effects of antibodies against DC-SIGN, as well as the carbohydrate mannan, which had been previously shown to competitively inhibit HIV-1 gp120 binding to DC-SIGN (2, 15, 45, 50). Similar to previously published results, preincubation of the THP1-DC-SIGN cells with anti-DC-SIGN neutralizing antibody or mannan (inhibits HIV-1 gp120 binding to all known C-type lectin receptors, including DC-SIGN and DC-SIGNR) prior to virus exposure completely eliminated virus replication in the cocultures (Fig. (Fig.4A4A and C) (2, 15). In contrast, if antibodies against DC-SIGN were added to T cells and then these T cells were then cocultured with virus-exposed THP1-DC-SIGN cells, there was a negligible inhibitory effect on virus replication. This implies that the most important function of DC-SIGN in the THP1-DC-SIGN-T-cell cocultures is to bind HIV-1, and not to its downstream interactions with ICAM-3, since it has been previously shown that the DC-SIGN MAb used here inhibited DC-SIGN-ICAM-3 interaction in vitro (59).
Since THP1-DC-SIGN cells express negligible amounts of CD4 and CXCR4 on their cell surface, antibodies against CD4 and CXCR4 had no effect on initial virus attachment, and hence on virus replication. However, addition of antibodies against CD4 and CXCR4 to activated T cells prior to coculture with virus-exposed THP1-DC-SIGN cells inhibited virus replication (Fig. 4A and C), suggesting that transmission and entry of DC-SIGN bound virus to T cells is still dependent on defined gp120-CD4 and CXCR4 interactions. Furthermore, this also demonstrates that DC-SIGN is the sole HIV-1 attachment factor present on THP1-DC-SIGN cell surface.
In marked contrast to THP1-DC-SIGN cells, however, preincubation of immature DC with neutralizing antibodies against DC-SIGN, prior to infection with either HIVLai or NL4-3/Ba-L env had negligible effects on viral replication (Fig. 4B and D). Note that fluorescence-activated cell sorter analysis demonstrated similar levels of DC-SIGN expression on DC and THP1-DC-SIGN cell surfaces (data not shown). More importantly, pretreatment of DC with saturating amounts of mannan also had negligible effects on virus replication in DC-T-cell cocultures (Fig. 4B and D), suggesting that DC are capable of binding HIV-1 particles in a mannose-receptor-independent fashion. Furthermore, pretreatment of primary DC with mannan and antibodies against CD4 or a combination of antibodies against CD4 and DC-SIGN prior to virus exposure also had no effect on NL4-3/Ba-L env replication in the DC-T-cell cocultures (Fig. (Fig.4D).4D). Addition of neutralizing antibodies against DC-SIGN to T cells prior to their coculture with virus-exposed DC also had no effect on virus replication (Fig. 4B and D). However, virus transmission to T cells was prevented by incubation of T cells with neutralizing antibodies against CD4 and CXCR4 prior to coculture with virus-exposed DC (Fig. 4B and D). Together, these results demonstrate that primary monocyte-derived DC can bind HIV-1 by DC-SIGN independent mechanisms, and that HIV bound independently of DC-SIGN and CD4 can be transmitted to T cells with high efficiency.
We found that HIV-1 replication under conditions where CD4+-T-cell life span is limited to 3 days is dependent on the cell-associated transfer of virus. In particular, dendritic cell-mediated transfer of HIV-1 to T cells was absolutely essential for wild type virus replication over multiple passages. We demonstrate here that the establishment of productive infection in CD4+ T cells mediated by the cell-associated form of HIV-1 was much more rapid compared to that of cell-free infection. Furthermore, the efficiency of virus transfer was dependent on establishment of close contacts between the virus-bound cell and the target T cell. The inability of cell-free virus to sustain replication in T cells under rapid-turnover conditions probably relates to the relative inefficiency of the initial adsorption of free virus particles to the T-cell surface, which is diffusion limited (8). We speculate that the adhesion molecule mediated juxtaposition of the two cell membranes in a cell-associated virus infection would in essence increase the net effective concentration of the virus receptor in the vicinity of the bound virus and hence enhance transfer of attached virus to target cells.
Furthermore, the intrinsic affinity between the T-cell surface CD4 and virion-associated native gp120 is low (38, 39), thus implying that the virus relies on the presence of a large number of attachment factors to bring it into close contact with the cell membrane. Our results demonstrate that DC bind HIV-1 particles with a higher affinity than that exhibited by CD4+ T cells. More importantly, we demonstrate that the kinetics of binding of free virus particles by DC or THP1-DC-SIGN cells and subsequent transfer to CD4+ T cells was faster in its establishment of productive infection than that mediated by cell-free virus stocks of HIVLai in CD4+ T cells. Finally, the viral burst size was significantly larger in the DC- and THP1-DC-SIGN-T-cell cocultures than in T-cell cultures infected by the cell-free virus stocks within 48 h postinfection. Finally, this ability of DC to mediate rapid trans infection of activated T cells is unique, since monocyte-derived macrophages, B cells, and T cells pulsed with virus in a similar manner, fail to promote trans infection to similar levels with rapid kinetics upon coculture with T cells (Fig. 4B and D and unpublished observations) (37, 41, 52, 55).
Based on these results, we propose the following model that could potentially account for the high level of virus replication in the peripheral lymphoid organs in the context of a robust virus-specific CTL response. During HIV-1 infection in vivo, large numbers of HIV-1 particles are produced daily (~1010 virions/day) (44, 49), which can be taken up by these immature DC, and transmitted to T cells in the peripheral lymphoid organs via formation of supramolecular activation clusters between the virus-bound DC and the T cell (5). These DC-T-cell conjugates hence, are the “factories” that drive virus production in vivo. The control of HIV spread in vivo is presumably due to a strong CTL response induced by the lentivirus infection in the peripheral lymphoid organs (7, 20). Under the restrictive conditions of host immune response, virus has a short window of opportunity to establish a productive infection. In fact, some estimates predict the virus generation time in vivo to be in the order of 2.6 days or less (44). Direct infection of T cells by cell-free virus particles is rate limited due to the low affinity of the oligomeric virion gp120 for CD4, the sole virus-attachment factor on CD4+ T cells (55). Furthermore, the cell-free virus half-life in plasma is less than 2 h (44, 49), thus further limiting the opportunity for HIV to directly infect CD4+ T cells. Presence of large numbers of virus attachment factors on DC (including DC-SIGN, which binds HIV-1 gp120 with a higher affinity than CD4) could allow for the efficient uptake of virus particles. Once they encounter foreign antigen, immunologically favored cellular interactions of DC with CD4+ T cells in the paracortical regions of the lymphoid organs could increase the efficiency with which virus is transmitted to these CD4+ T cells. Hence, we propose that in addition to the postulated role of DC in facilitating initial spread after mucosal exposure of virus (15, 27, 53), DC could be involved in the continued presentation of HIV-1 to activated CD4+ T cells in the lymphatic tissues, thus sustaining a high level of virus replication under rapid-turnover conditions in vivo.
Finally, we demonstrate here that initial HIV-1 attachment to DC can also occur in a DC-SIGN-independent fashion since virus binding to DC and subsequent trans infection of T cells was not inhibited in the presence of saturating amounts of mannan, DC-SIGN neutralizing antibodies, or CD4 neutralizing antibodies (Fig. 4B and D). These results are consistent with the recently published observations which demonstrate a lack of DC-SIGN expression on rhesus macaque monocyte-derived DC (59) and certain human DC subsets, namely, Langerhans cells (16, 29) and plasmacytoid DC (43). More importantly these DC subsets were competent for virus binding and subsequent transfer of virus to T cells (43, 59). It is known that peripheral blood monocyte-derived DC express at least three different C-type lectin receptors, including DC-SIGN (54). Since we observed no significant inhibition of virus attachment and subsequent transmission to T cells in the presence of mannan and CD4 neutralizing antibodies (Fig. 4B and D), we conclude that in addition to DC-SIGN, CD4, and C-type lectin (mannose) receptors, there are yet-unknown mechanisms of DC-specific virus attachment. A possible implication of this work is that methods designed to block initial attachment of HIV-1 to DC in the genital mucosa (presumably the initial cell type that encounters virus in the periphery) should include strategies beyond those that target DC-SIGN and mannose specific C-type lectin receptors.
Hence, this study has several implications for in vivo HIV-1 pathogenesis, especially in regards to the apparent ability of the virus to establish and maintain high levels of virus replication under stressful conditions of the host immune response. We predict that in the setting of an HIV-1-infected lymph node, DC-T-cell interactions would provide a favorable milieu for the multiple rounds of virus replication that contribute to the HIV-1-induced pathology and immunodeficiency. If sustaining a vigorous level of HIV-1 replication is a kinetic race fought with virus-specific CTL, then antivirals designed to prevent HIV-1 interactions with DC that do not obviate their important role in antigen presentation may tip the balance of this battle in favor of the human immune system. Finally, our in vitro culture system recapitulates ongoing virus replication in the lymphatic tissues and thus provides a unique model system to study HIV-1 replication. Studies of viral fitness in such an in vitro environment would provide novel insights into the contributions of virus accessory genes to HIV-1 replication.
We thank W. C. Goh, M. Vodicka, L. Wu, J. Overbaugh, and J. Coffin for their insightful comments; Paul Grasse, ABI, for help with real-time PCR; and members of the FHCRC flow cytometry core facility for help with cell sorting. We acknowledge R & D Systems for sharing DC-SIGN antibodies. We are grateful to the NIH AIDS Research and Reference Reagent Program for the CD4 (Sim.2) (contributed by James Hildreth) and CXCR4 (12G5) MAbs.
This work was supported by NIH grant RO1 AI30927 to M. Emerman and by the James Pendleton Charitable Trust Foundation.