|Home | About | Journals | Submit | Contact Us | Français|
Mycobacterium tuberculosis is a leading killer of HIV-infected individuals worldwide, particularly in sub-Saharan Africa, where it is responsible for up to 50% of HIV-related deaths. Infection by HIV predisposes individuals to M. tuberculosis infection, and coinfection accelerates the progression of both diseases. In contrast to most other opportunistic infections associated with HIV, an increased risk of M. tuberculosis infection occurs during early-stage HIV disease, long before CD4 T cell counts fall below critical levels. We hypothesized that M. tuberculosis infection contributes to HIV pathogenesis by interfering with dendritic cell (DC)-mediated immune control. DCs carry pathogens like M. tuberculosis and HIV from sites of infection into lymphoid tissues, where they process and present antigenic peptides to CD4 T cells. Paradoxically, DCs can also deliver infectious HIV to T cells without first becoming infected, a process known as trans-infection. Lipopolysaccharide (LPS)-activated DCs sequester HIV in pocketlike membrane invaginations that remain open to the cell surface, and individual virions are delivered from the pocket into T cells at the site of contact during trans-infection. Here we report that M. tuberculosis exposure increases HIV trans-infection and induces viral sequestration within surface-accessible compartments identical to those seen in LPS-stimulated DCs. At the same time, M. tuberculosis dramatically decreases the degradative processing and major histocompatibility complex class II (MHC-II) presentation of HIV antigens to CD4 T cells. Our data suggest that M. tuberculosis infection promotes a shift in the dynamic balance between antigen processing and intact virion presentation, favoring DC-mediated amplification of HIV infections.
Dendritic cells (DCs) comprise a diverse family of cell types whose primary function is to initiate and drive immune responses. Myeloid DCs (myDCs) are essential antigen-presenting cells that monitor peripheral tissues for invading pathogens. myDCs bind and internalize bacteria and viruses using a variety of surface receptors. When stimulated by pathogenic or inflammatory signals, peripheral-tissue DCs migrate to lymphoid tissues and undergo maturation, degrading stored antigens into peptides that are loaded onto major histocompatibility complex class II (MHC-II) molecules and expressed on the cell surface for presentation to CD4 T cells (reviewed in reference 4). In addition to presentation of processed peptide antigens, DCs carry intact, unprocessed proteins and pathogens from peripheral tissues to lymph nodes, where they can be passed to other antigen-presenting cells to increase the breadth of the immune response (reviewed in reference 10).
HIV can exploit the natural trafficking of DCs to establish and amplify infection of CD4 T cells. DCs efficiently transfer intact, infectious HIV to T cells during immune interactions through a process known as trans-infection (14). DCs trans-infect HIV by binding and concentrating the intact virus at the cellular interface, forming an “infectious synapse” that concentrates HIV receptors on the T cell to the same site (24). Importantly, trans-infection does not require productive infection of the DCs, which are not infected efficiently by HIV in vitro or in vivo (14). Immature DCs significantly enhance infection of T cells through trans-infection, and prior activation by cytokine or bacterial stimuli markedly increases infectious synapse formation and concomitant trans-infection (2, 24, 33).
Worldwide, nearly one-third of HIV-infected people are coinfected with Mycobacterium tuberculosis, and active tuberculosis disease (TB) is the number one cause of death in HIV-infected people. Coinfected individuals are 30 times more likely to progress to active TB, which can in turn increase HIV replication and accelerate the progression to AIDS (35). The mechanisms by which coinfection with M. tuberculosis and HIV accelerates the progression of both diseases are poorly understood.
Lung macrophages are the primary target of M. tuberculosis infection, and active disease is characterized by unconstrained replication in these cells. Dendritic cells can also be infected by M. tuberculosis, but M. tuberculosis growth is restricted due to a lack of nutrient access in the DC phagolysosomal structure in which it resides (20). Importantly, M. tuberculosis-infected DCs traffic between the infected lung and draining lymph nodes, bringing bacterial antigens into lymphoid tissues to initiate CD4 T cell responses essential for disease control (39).
Others have established that M. tuberculosis binds to and is internalized by DCs via an interaction between the mycobacterial cell wall component mannosylated lipoarabinomannan (ManLAM) and the cell surface receptor DC-SIGN on dendritic cells (15). After ManLAM stimulation, DCs begin to secrete interleukin-10 (IL-10) and show defects in immunostimulatory functions (15). However, a more recent study suggests that ManLAM may not be solely responsible for these outcomes (1).
Previously, it has been shown that lipopolysaccharide (LPS) potently stimulates HIV trans-infection of CD4 T cells by DCs (24, 33). Therefore, we reasoned that M. tuberculosis and its products might similarly stimulate DC trans-infection during active M. tuberculosis infections. Further, we hypothesized that DC activation by M. tuberculosis would result in downmodulation of processing and MHC-II presentation of newly bound HIV particles, shifting the balance away from immune control in favor of viral dissemination and pathogenesis.
Here, we demonstrate that M. tuberculosis infection of DCs enhances HIV trans-infection mediated through surface-accessible, pocketlike invaginations of the plasma membrane. Increased HIV trans-infection is accompanied by decreased MHC-II processing and presentation of HIV antigens to CD4 T cells. Our results suggest one mechanism whereby M. tuberculosis infection can fuel HIV dissemination in coinfected individuals and at the same time decrease immune control of both HIV and M. tuberculosis infections.
Monocyte-derived dendritic cells (MDDCs) were generated as described previously (32, 42). Briefly, CD14-positive monocytes were isolated from peripheral blood mononuclear cells (PBMCs) from healthy donors by positive selection using anti-CD14 magnetic beads (Miltenyi Biotec). Monocytes were cultured in RPMI 1640 plus 10% fetal bovine serum (FBS) (HyClone), 100 U/ml penicillin, and 100 μg/ml streptomycin (Invitrogen) (complete medium) supplemented with 100 ng/ml IL-4 and 50 ng/ml granulocyte-macrophage colony-stimulating factor (GM-CSF) (Gentaur Biosciences) for 6 or 7 days to produce immature MDDCs. Cytokines were refreshed on day 3, and cells were resuspended in fresh medium with cytokines on day 5 of culture. Myeloid dendritic cells (myDCs) were purified from PBMCs using anti-CD1c (BDCA-1) magnetic beads (Miltenyi Biotech) and cultured in complete medium supplemented with 50 ng/ml GM-CSF. Activated, autologous CD4 T cells were prepared by culturing CD14-depleted PBMCs in complete medium with phytohemagglutinin (PHA) (10 μg/ml) and IL-2 (20 U/ml) for 7 days. The T cells were purified by negative selection using a CD4 T cell isolation kit (Miltenyi Biotec).
LuSIV cells (CEM T cell line transduced with an HIV long terminal repeat [LTR]-luciferase reporter; AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, contributed by Jason Roos) (30) and RAJI-DC-SIGN cells (human B-cell line transduced with a DC-SIGN expression vector; AIDS Reagent Program, contributed by Li Wu and Vineet N. KewalRamani) (41) were grown in RPMI 1640 complete medium. HEK293T cells were grown in complete Dulbecco's modified Eagle's medium (DMEM).
Fluorescent anti-CD80, -CD83, -CD86, -HLA-DR (all from BD Biosciences), and -HIV p24gag (KC57-FITC [fluorescein isothiocyanate]; Beckman Coulter) and matched isotype controls were used according to the manufacturer's directions and analyzed using an LSR II (Becton Dickenson) or FACSCalibur (BD Biosciences) flow cytometer. Matched isotype controls overlaid the uninfected samples and were therefore not shown in the histograms. For blocking experiments, a DC-SIGN monoclonal antibody (MAb) (clone 120507; AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH) (19) was added either at the time of M. tuberculosis addition (for MDDC experiments) or at the time of HIV addition (for RAJI-DC-SIGN experiments).
For immunofluorescence microscopy, anti-HIV gp120env MAb 2G12 (AIDS Reagent Program, contributed by Hermann Katinger) (36) and anti-CD81 MAb (BD Biosciences) were detected with fluorescent secondary antibodies (Jackson ImmunoResearch).
The Escherichia coli/M. tuberculosis shuttle plasmid pMV262, described previously (29), expresses enhanced green fluorescent protein (eGFP) under the control of the HSP60 promoter. The eGFP was replaced with either enhanced cyan fluorescent protein (eCFP) to form pNDP11 or DsRed (Clontech) to form pNDP10. The Mycobacterium bovis bacillus Calmette-Guérin (BCG) Connaught strain was transformed with pNDP11 to generate BCG-CFP. M. tuberculosis H37Ra (ATCC) was transformed with pNDP10 to generate red fluorescent protein (RFP)-labeled M. tuberculosis. Kanamycin-resistant clones were screened for stable, high-level expression of either eCFP or DsRed. BCG and M. tuberculosis strains were grown in 7H9 medium (Difco) supplemented with albumin-dextrose-catalase (ADC) (Middlebrook) and frozen at −80°C. Thawed bacteria were cultured overnight in 7H9-ADC medium before use. The titers of stocks were determined by serial dilutions plated onto 7H11 (Difco) plates supplemented with oleic acid-albumin-dextrose-catalase (OADC) (Middlebrook) and kanamycin (25 μg/ml). Heat-killed M. tuberculosis H37Ra was prepared by incubating stock cultures at 95°C for 10 min; equivalent volumes were added to approximate the equivalent live bacterial load. Purified mannosylated lipoarabinomannan (ManLAM) from M. tuberculosis strain H37Rv (NR-14848) was obtained from the NIH Biodefense and Emerging Infections Research Resources Repository, NIAID, NIH.
Single-round GFP-Vpr-labeled HIV-1 (GFP-HIV) was generated by calcium phosphate transfection of HEK293T cells with an HIV envelope-defective provirus (HIVLAIenv) along with plasmids expressing HIVHXB2 gp120 and eGFP-Vpr, as described previously (24). Virus-containing supernatants were collected 40 h posttransfection, passed through 0.45-μm filters, and stored in liquid nitrogen. Antigen presentation assays used aldrithiol-2 (AT-2)-inactivated HIVMN (31), kindly provided by Jeffrey Lifson, NIH/NCI-Frederick. HIVIIIB (Advanced Biotechnologies, Inc.) was used for live-virus antigen presentation assays. HIVBal was expanded in U87 CD4/CCR5 cells (U87.CD4.CCR5; AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, from HongKui Deng and Dan R. Littman) (5). Supernatants were harvested on days 3 to 5 postinfection, passed through 0.45-μm filters, pooled, and stored in liquid nitrogen.
Enhancement assays were performed by pulsing MDDCs or myDCs with GFP-HIV or HIVBal for 2 h at 37°C to allow for viral uptake by the DCs. Virus was used at a concentration of 5 to 20 ng HIV p24 for every 1 × 105 cells. Pulsed DCs (2 × 104) were cocultured with LuSIV cells (2.5 × 104) for 48 h, lysed, and assayed using Brite-Luc luciferase assay reagent (Promega). The plates were read using a multiwell-format luminometer (Bio-Rad). Transfer assays were performed as described above, except the pulsed DCs were washed and cultured for an additional 2 h at 37°C before coculture with LuSIV cells. RAJI-DC-SIGN cells (2 × 104) were pulsed with GFP-HIV (2 to 4 ng p24) for 2 h at 37°C in the absence or presence of anti-DC-SIGN blocking antibody (20 or 100 μg/ml). Pulsed RAJI-DC-SIGN cells were washed, cultured for an additional 2 h at 37°C, cocultured with LuSIV cells (2.5 × 104) for 48 h, lysed, and assayed for luciferase activity as described above.
Antigen presentation assays were performed as described previously (8, 21). Briefly, 2 × 104 MDDCs from HLA-DR1.01-positive donors were incubated with AT-2-inactivated HIVMN for 4 h to allow for uptake, degradation, and presentation. DCs were then fixed in 1% paraformaldehyde, washed, and cocultured with 1 × 105 HIV reverse transcriptase (RT)-specific T cell hybridomas (1ACD5). myDC experiments were performed with unfixed myDCs, and HIVMN and T cells were added at the same time. Culture supernatants were collected at 22 h, and IL-2 levels were determined using a CTLL-2 bioassay. T cell hybridomas produce IL-2 in proportion to the amount of the cognate RT peptide/MHC presentation. IL-2 levels were determined by the CTLL-2 cell bioassay and quantified spectrophotometrically by color change of Alamar Blue dye, as previously described (8).
GFP-HIV-pulsed DCs were adhered to poly-l- lysine-coated coverslips for 3 min at room temperature, fixed with 4% electron microscopy (EM)-grade formaldehyde (Polysciences) in phosphate-buffered saline (PBS) for 20 min, and rinsed in PBS. Cell membranes were extracted with 0.1% Triton X-100 in SB (PBS, 10% normal donkey serum; Jackson ImmunoResearch) for 5 min at room temperature. The actin cytoskeleton was stained with fluorescent phalloidin (Invitrogen), and nuclei were stained with Hoechst dye (0.5 μg/ml; Sigma) in SB for 15 min at room temperature. Coverslips were mounted onto glass slides using Gel Mount (Biomedia). For surface accessibility assays, GFP-HIV-pulsed DCs were incubated with 2G12 anti-HIV gp120 MAb at 4°C for 1 h. Cells were then washed with cold PBS and fixed onto poly-l-lysine-coated coverslips. Anti-gp120 antibody was detected using labeled goat anti-human antibody (Invitrogen) in SB for 15 min at room temperature.
Slides were imaged using a Deltavision RT epifluorescent microscope (Applied Precision, Inc.). Images were captured in z series and deconvolved and processed using the Softworx analysis program. Images were exported as TIFF or JPEG files, and figures were composed using Adobe Photoshop CS.
To determine whether mycobacterial infection could potentiate HIV trans-infection, we exposed monocyte-derived dendritic cells (MDDCs) to Mycobacterium bovis bacillus Calmette-Guérin engineered to express cyan fluorescent protein (BCG-CFP) (29). After overnight culture, we assessed the ability of DCs to trans-infect target T cells by using two complementary assays, termed enhancement and transfer. Enhancement refers to the increased T cell infection observed when DCs are added along with unbound HIV to the T cell cultures, whereas transfer measures the transmission of HIV bound to DCs when unbound virus is washed away prior to coculture.
MDDCs were exposed to a single-round infectious stock of GFP-Vpr-labeled HIV-1 (GFP-HIV) and then either cocultured with LuSIV T cells (enhancement) or washed and then cocultured with the reporter (transfer). LuSIV cells, a CEM T cell line, harbor an HIV Tat-sensitive luciferase reporter gene, and the level of HIV infection correlates with luciferase activity assessed after 48 h of coculture (30). Use of single-round HIV eliminates the potential for productive infection of the DCs and subsequent second-round or cis-infection of the T cells (37, 42). In this configuration, only trans-infection of T cells is measured, since direct infection of MDDCs does not result in luciferase activity and the use of replication-defective HIV precludes second-round transfer from infected DCs.
Figure Figure1A1A demonstrates that BCG exposure markedly increased DC-mediated enhancement of HIV infection. In the experiment with results shown, untreated MDDCs increased the level of HIV infection 1.7-fold over HIV infection with no DCs, and addition of BCG resulted in a dose-dependent increase in HIV transfer over a range of 0.1 to 10 CFU BCG/DC. The greatest enhancement of HIV infection was observed at 10 CFU/DC, which resulted in a 3.7-fold increase in HIV infection over that without MDDCs and a 2.2-fold increase over that with untreated MDDCs. Exposure to greater than 10 CFU BCG/DC resulted in a somewhat diminished response.
The corresponding transfer assay (Fig. (Fig.1B)1B) exhibited a similar dose-dependent increase in DC-mediated HIV infection after exposure to BCG. In this experiment, the maximal increase in HIV transfer was observed at 10 to 20 CFU/DC (3.2-fold and 3.3-fold increases, respectively, over the level with untreated MDDCs). Exposure to more than 20 CFU BCG per DC reduced HIV transfer ability.
Both the enhancement and transfer assays show defects in transfer efficiency at high doses of BCG, which is likely due to the high bacterial burden within the DCs. High titers of BCG also reduce DC viability. During overnight BCG infection, DCs exposed to >10 CFU BCG/DC typically exhibit 2- to 3-fold more cell death than those infected with ≤10 CFU BCG/DC (data not shown). These unhealthy, dying cell populations can often give inconsistent results, even between matched enhancement and transfer experiments, so subsequent assays used a maximum of 10 CFU of mycobacterium/DC to reduce variability.
LPS activation potently stimulates trans-infection and induces sequestration of intact HIV within a specialized, CD81-positive compartment within DCs (12, 18, 42). To determine whether BCG similarly affects HIV localization, MDDCs were infected overnight with CFP-BCG and then pulsed with GFP-HIV as described above. In the absence of BCG, HIV was evenly distributed within the MDDCs (Fig. (Fig.1C),1C), whereas prior exposure to BCG resulted in the concentration of HIV into a single subcellular region (Fig. (Fig.1D).1D). Interestingly, the HIV localization was distinct from that for intracellular BCG, which is known to reside in phagolysosomes within DCs (9). Furthermore, HIV concentration was observed in DCs that did not harbor BCG (Fig. (Fig.1D,1D, top), suggesting that direct infection of the DCs was not necessary to induce sequestration of HIV in BCG-infected cultures.
To confirm that M. tuberculosis infection also stimulated HIV trans-infection, MDDCs were infected with M. tuberculosis H37Ra and assessed as described above. Figure Figure2A2A demonstrates that M. tuberculosis H37Ra increased HIV transfer with a dose response similar to that seen with BCG infection. In this experiment, 10 CFU M. tuberculosis/DC optimally stimulated HIV transfer 9.7-fold over that with uninfected MDDCs. M. tuberculosis inocula greater than 10 CFU M. tuberculosis/DC negatively impacted trans-infection (not shown). Time course analysis following infection with 10 CFU M. tuberculosis/DC indicated that increased trans-infection could be observed as early as 2 h after M. tuberculosis exposure and that maximally increased trans-infection occurred after overnight culture (Fig. (Fig.2B).2B). These kinetics are similar to those of DC activation following Toll-like receptor 4 (TLR4) stimulation by bacterial LPS (6, 27).
We have demonstrated increased HIV trans-infection in LuSIV cells, a CEM T cell line transduced with an HIV LTR-luciferase reporter. We sought to verify our results using primary autologous CD4 T cell targets (Fig. (Fig.2C).2C). MDDCs were infected with M. tuberculosis overnight and then pulsed with replication-competent HIVBal. After pulsing, MDDCs were washed of unbound virus and cocultured with primary CD4 T cells. CD4 T cells were obtained from healthy donor blood and activated with IL-2 and PHA prior to coculture. After 48 h of coculture, cells were immunostained for p24 expression and assessed by flow cytometry. This experiment confirms that increased HIV trans-infection also occurs in primary T cells. The primary T cells responded more robustly to small doses of M. tuberculosis than the LuSIV cells, indicating that the primary T cells are more sensitive to M. tuberculosis in the cocultures. This suggests that, in vivo, low levels of M. tuberculosis may be sufficient to increase trans-infection.
To confirm that the observed increase in HIV trans-infection is due to enhanced DC function and not an M. tuberculosis-induced change in T cell susceptibility to infection, LuSIV cells (Fig. (Fig.2D,2D, white bars) or primary autologous T cells (Fig. (Fig.2D,2D, black bars) were exposed to M. tuberculosis and GFP-HIV or to M. tuberculosis and HIVBal, respectively, in the absence of DCs. LuSIV cells and primary T cells were infected with an HIV input equivalent to that of the enhancement assay no-DC controls. M. tuberculosis at 0, 0.1, 1, or 10 CFU/cell was added at the time of HIV infection. The highest dose, 10 CFU M. tuberculosis/cell, is much greater than the dose of M. tuberculosis in DC cocultures under our experimental conditions, as most of the M. tuberculosis added to the DCs is either internalized by the DCs or washed away during the multiple washing steps preceding coculture. Exposure to 1 CFU M. tuberculosis/LuSIV cell slightly decreased HIV infection of the cells, and 10 CFU/LuSIV cell decreased infection further. Primary T cell infectivity was reduced by all M. tuberculosis dosages. This confirms that M. tuberculosis increases trans-infection through stimulation of the DCs, as it did not cause either the LuSIV cells or the primary T cells to become more susceptible to infection. On the contrary, M. tuberculosis either caused the cells to become less susceptible to infection or, more likely, was slightly cytotoxic. This result indicates that the observed increase in trans-infection may be even greater than what is currently measured, as neither cell type is fully functional when exposed to high titers of M. tuberculosis.
M. tuberculosis is known to activate TLR2 signaling in DCs, resulting in upregulation of surface activation markers (17). Similarly, in our MDDC cultures, overnight exposure to M. tuberculosis induced the expression of HLA-DR and activation markers CD80, CD83, and CD86 (Fig. (Fig.2E).2E). Taken together, the data shown in Fig. Fig.22 indicate that DCs respond rapidly to M. tuberculosis exposure and that DC activation parallels increased HIV trans-infection.
We sought to determine if live M. tuberculosis was required to enhance trans-infection. MDDCs were untreated, infected with live M. tuberculosis H37Ra, or exposed to heat-killed M. tuberculosis overnight before pulsing with GFP-HIV. We performed a transfer assay with LuSIV cells and found that heat-killed M. tuberculosis was more potent than live H37Ra at inducing trans-infection (Fig. (Fig.3A).3A). This suggests that increased trans-infection does not require M. tuberculosis survival or replication within the DCs. Rather, increased trans-infection is due to cellular signaling from mycobacterial components, likely found in the cell wall.
Geijtenbeek et al. demonstrated that M. tuberculosis and BCG bind to DC-SIGN on dendritic cells through the bacterial cell wall component mannosylated lipoarabinomannan (ManLAM) and induce IL-10 secretion and defective DC maturation (15). More-recent findings by members of the same group (1) indicate that ManLAM is not required for mycobacterium-DC interaction and subsequent IL-10 secretion, suggesting that other DC receptors might mediate the defect in maturation. To elucidate the role of ManLAM in trans-infection, we treated MDDCs for 18 h and compared them to untreated or M. tuberculosis-infected MDDCs (Fig. 3B and C). ManLAM treatment did not induce an increase in trans-infection (Fig. (Fig.3B)3B) and in fact mildly inhibited DC function, in agreement with previous findings (15). ManLAM exposure induced a small increase in expression of the maturation markers CD83 and CD86 over that observed in untreated MDDCs, while HLA-DR expression remained unchanged (Fig. (Fig.3C),3C), consistent with the weak TLR agonist activity of ManLAM (28).
Next, DC-SIGN blocking antibody experiments were performed to more fully explore the possibility that DC-SIGN-induced signaling pathways play a role in increasing trans-infection. MDDCs were infected with M. tuberculosis in both the presence and the absence of DC-SIGN blocking antibodies and compared to untreated MDDCs in a transfer assay to assess function (Fig. (Fig.3D).3D). DC-SIGN blocking antibody used at 20 μg/ml decreased HIV transfer to the same level as that for untreated MDDCs, while 100 μg/ml of the antibody decreased HIV transfer to a greater degree. DC-SIGN MAbs are not generally effective at blocking all HIV binding to DCs, because they block only the DC-SIGN receptor. DC-SIGN expression is not required for trans-infection of T cells, as other mannose C-type lectin receptors are involved in HIV binding to DCs and can substitute for DC-SIGN when it is not available (7). Interestingly, blocking DC-SIGN reduces trans-infection to the level found with unstimulated DCs but does not completely abolish it.
To ensure that the MAb blocked DC-SIGN-mediated trans-infection, we used RAJI-DC-SIGN cells as a functional control for the DC-SIGN blocking antibody. RAJI-DC-SIGN cells are a human B-cell line transduced with a DC-SIGN expression vector. This cell line can perform efficient, DC-SIGN-mediated HIV trans-infection of target T cells. RAJI-DC-SIGN cells were pulsed with GFP-HIV in the presence or absence of the DC-SIGN blocking antibody. A transfer assay was performed using LuSIV target cells. DC-SIGN blocking antibody efficiently blocked HIV transfer from the RAJI-DC-SIGN cells to the LuSIV cells (Fig. (Fig.3E),3E), indicating that the blocking antibody is fully functional and efficiently inhibits DC-SIGN binding.
DC-SIGN is a major receptor for HIV binding on DCs and plays a role during trans-infection (14). Therefore, rather than decreasing trans-infection by blocking M. tuberculosis-induced activation, the blocking antibody may cause low transfer levels by preventing effective HIV binding to the DCs. MDDCs were untreated or exposed to M. tuberculosis alone, M. tuberculosis plus DC-SIGN blocking antibody, or DC-SIGN blocking antibody alone. Cells were stained for the activation markers CD83, CD86, and HLA-DR and assessed by flow cytometry (Fig. (Fig.3F).3F). The results show that the DC-SIGN blocking antibody had no effect on DC activation. The blocking antibody was unable to prevent DC activation when in the presence of M. tuberculosis. Additionally, the blocking antibody alone binding DC-SIGN did not induce DC activation (Fig. (Fig.3F).3F). Taken together, these data indicate that mycobacterial components are responsible for increased trans-infection, though viable M. tuberculosis is not required to induce transfer. As trans-infection is not induced by ManLAM or DC-SIGN stimulation, it is likely the result of TLR-induced activation.
To verify that the observed increase in HIV trans-infection was not an artifact of in vitro-differentiated MDDCs, we tested myDCs purified directly from peripheral blood mononuclear cells (PBMCs). myDCs are found in blood, skin, and mucosal tissues and have been associated with HIV capture and dissemination (22). Importantly, myDCs are thought to transport M. tuberculosis from sites of infection in the lung to draining lymph nodes, where they help to coordinate the immune response (40). Figure Figure4A4A and B demonstrate that myDC-mediated trans-infection is potently activated by M. tuberculosis infection. Unstimulated myDCs modestly enhanced HIV infection of T cells, approximately 1.6-fold higher than the level with HIV alone, similar to that seen with MDDCs. Overnight exposure to M. tuberculosis stimulated myDC-mediated enhancement of HIV infection 6.1- to 7-fold. Similarly, M. tuberculosis increased transfer of infectious HIV to levels 2.9- to 28-fold greater than those achieved with untreated myDCs (1 and 10 CFU/DC, respectively).
Of note, HIV trans-infection by myDCs was greater than that of MDDCs. This difference in trans-infection ability can be explained by focusing on the level of activation of these cells. Circulating myDCs are at a very low activation state in healthy individuals, while MDDCs, as a result of in vitro differentiation and culture, are at a slightly elevated activation state. Thus, larger differences are noted in the myDCs than in the MDDCs because, by starting at a lower initial activation level, myDCs have a greater overall activation potential.
Analysis of HIV subcellular localization revealed that, as in MDDCs, M. tuberculosis infection resulted in a dramatic concentration of virion particles into a single subcellular site within myDCs. As shown in Fig. Fig.4D,4D, intracellular and extracellular M. tuberculosis were both evident, and HIV sequestration occurred in myDCs harboring M. tuberculosis as well as in cells with no apparent intracellular M. tuberculosis. Taken together, these results indicate that exposure to M. tuberculosis or mycobacterial products can potently stimulate the sequestration and transmission of HIV by primary myDCs.
LPS-activated DCs sequester HIV within invaginated plasma membrane “pockets” that remain exposed to the cell surface, allowing for delivery of intact HIV particles at the infectious synapse during trans-infection of T cells (42). To determine whether HIV is concentrated into a similar structure in M. tuberculosis-stimulated DCs, we incubated HIV-pulsed DCs at 4°C with a monoclonal antibody (MAb) specific for HIV envelope glycoprotein (gp120). Incubation at 4°C blocks endocytosis, and so the membrane-impermeable MAb binds only to surface-exposed gp120. After free MAb was washed away, DCs were fixed onto coverslips and immunostained to detect bound anti-gp120 MAb. Cells were then immunostained either for actin (Fig. 5A and B) or for CD81 (Fig. (Fig.5C).5C). Figure Figure55 shows that concentrated HIV within the M. tuberculosis-exposed dendritic cells remained accessible to the surface-applied antibody. In the example shown, all of the HIV-containing structures strongly stained with the anti-gp120 MAb, indicating that HIV is contained within surface-accessible, nonendocytic compartments within M. tuberculosis-activated MDDCs (Fig. 5A and C) and myDCs (Fig. (Fig.5B).5B). The area of HIV concentration strongly colocalized with CD81 (Fig. (Fig.5C),5C), as we have also observed for LPS-stimulated DCs (42).
We reasoned that increased trans-infection in response to M. tuberculosis infection might be accompanied by a concomitant decrease in antigen processing and presentation by the DCs at least partially due to the change in trafficking and distribution of cell-associated HIV in activated MDDCs seen in Fig. Fig.1,1, ,4,4, and and5.5. We tested the ability of M. tuberculosis-infected DCs to process and present HIV by using a T cell hybridoma that recognizes a defined HIV reverse transcriptase (RT) peptide in the context of HLA-DR1.01 (21). Presentation of the HLA/peptide complex results in dose-dependent IL-2 secretion that is read out in an IL-2-dependent CTLL bioassay. Antigen recognition is peptide specific, restricted to HLA-DR1.01, and dependent on proteolytic processing of intact virions. Presentation of the exogenous synthetic peptide, by contrast, occurs by direct binding of the peptide to cell surface HLA-DR1.01 molecules and therefore serves as a cellular control for antigen presentation in the absence of uptake and processing.
Figure Figure6A6A demonstrates that overnight exposure to increasing doses of M. tuberculosis resulted in progressively decreased presentation of intact HIVMN by MDDCs, so that at 10 M. tuberculosis CFU/DC, presentation was completely inhibited. Importantly, the presentation of exogenous peptide remained intact and in fact was slightly elevated in M. tuberculosis-infected MDDCs (Fig. (Fig.6B),6B), likely reflecting the increased expression of HLA class II by DCs after M. tuberculosis activation. Similar results were obtained with live HIVIIIB and with inactivated virus (data not shown). Presentation of intact HIV was diminished as early as 4 to 8 h after M. tuberculosis exposure, and peptide presentation was again increased in that time frame (Fig. 6C and D). These results are consistent with previous reports demonstrating that activation stimuli rapidly downregulate endocytosis and induce MHC-II expression in DCs (13, 27).
Presentation of HIV by purified myDCs was similarly inhibited by M. tuberculosis infection. myDCs were sensitive to the highest doses of M. tuberculosis, but even at 5 CFU/DC, we observed complete inhibition of HIV presentation without any effect on peptide presentation (Fig. 6E and F). The time course of inhibition was similar to that in MDDCs (Fig. 6G and H). We therefore conclude that M. tuberculosis potently inhibits de novo antigen processing of HIV in MDDCs and myDCs without reducing the presentation of the exogenously loaded peptide antigen. This indicates that M. tuberculosis activates the DCs, leading to decreased endocytosis and antigen processing of intact HIV.
In this study, we have demonstrated that DC-mediated HIV trans-infection is increased significantly following exposure to M. tuberculosis. Enhancement was observed both in in vitro-cultured MDDCs and in myDCs and was independent of the mycobacterial strain used. Additionally, increased trans-infection was observed using either T cell lines or primary autologous CD4 T cells. The response was dose and time dependent and was accompanied by upregulation of DC activation markers. Heat-killed M. tuberculosis had a similar stimulatory effect, suggesting that bacterial replication is not required for increased trans-infection of T cells by DCs. The heat-killed M. tuberculosis result also indicates that increased trans-infection is caused by a mycobacterial component interacting with the DC rather than by a molecule that is newly synthesized and secreted from the mycobacterium.
M. tuberculosis can induce changes in DC function via ManLAM/DC-SIGN interactions (15); however, even in the absence of ManLAM in the bacterial cell wall, M. tuberculosis can alter DC function (1). Our results indicate that ManLAM/DC-SIGN interactions are not required to induce increases in trans-infection. Exposing the DCs to ManLAM had little to no effect on surface activation marker expression or the ability to trans-infect. DC-SIGN blocking antibodies were similarly ineffective. Blocking antibodies did not prevent M. tuberculosis-induced upregulation of surface activation markers. While the blocking antibodies appeared to decrease HIV transfer, this observation is likely due to the blocking antibodies interfering with HIV binding to DC-SIGN, one of the main receptors involved in trans-infection. Taken together, our results point toward a mycobacterial pathogen-associated molecular pattern (PAMP) interacting with a pattern recognition receptor (e.g., a Toll-like receptor [TLR]) expressed on the DCs to mediate increased trans-infection. Consistent with this, others have shown that M. tuberculosis is a potent TLR2 agonist that can also act to a lesser degree through TLR4 and -9 (3, 25, 26).
M. tuberculosis infection is chronic, with a course of disease observed in months rather than days. Granulomatous lesions that harbor persistent M. tuberculosis are thought to be highly dynamic structures, allowing constant leukocyte migration between the lung and lymphoid tissues (11). Our data show that M. tuberculosis increases the magnitude of HIV trans-infection 2.2- to 7.0-fold. In general, between 1 and 10 CFU of mycobacterium per DC was required to enhance HIV trans-infection. While these numbers appear high, the microenvironment of M. tuberculosis infection varies greatly and likely contains regions where these levels are physiologically relevant. Cavitary TB, for example, carries an extremely high burden of mycobacterium, in which the multiplicity of infection (MOI) is probably even higher than 10:1. Even the relatively modest effects observed with lower doses of M. tuberculosis could, over the long term, easily result in significant changes to the final outcome of TB. HIV-negative individuals have a 5 to 10% lifetime risk of developing TB, whereas in coinfected individuals this number jumps to a 5 to 10% risk per year (23, 34). Our data suggest that a loss of antigen presentation combined with an increased HIV trans-infection could contribute to the loss of immune control of M. tuberculosis in the setting of HIV infection.
We further showed that M. tuberculosis-exposed DCs concentrated HIV into plasma membrane-derived, pocketlike structures that remained accessible to surface-applied antibodies, indicating that the sequestered HIV was not endocytosed and instead remained intact and competent to trans-infect CD4 T cells. These data extend our previous analysis of LPS-stimulated HIV sequestration into an identical structure (42) and suggest that pocket formation is a generalized response to activating stimuli. Indeed, Candida albicans exposure leads to partial activation of DCs along with increased HIV trans-infection (38), suggesting that innate immune responses to invading copathogens can act to stimulate HIV dissemination.
Dendritic cells remain a key cell type in which to study coinfection. M. tuberculosis infects both macrophages and DCs in the lungs. While macrophages remain primarily in the lung tissues, DCs migrate to the lymph nodes after phagocytosis of M. tuberculosis in order to initiate immune responses (39, 40). Our imaging analysis showed that M. tuberculosis exposure induced HIV sequestration even in DCs that did not harbor M. tuberculosis (Fig. (Fig.1,1, ,4,4, and and5).5). These data support the hypothesis that microbial products or inflammatory cytokines released by the infected DCs can affect neighboring DCs and increase their ability to trans-infect, even though they may not harbor intact M. tuberculosis. This suggests that M. tuberculosis-exposed DCs trafficking from the lung could further exacerbate HIV disease by activating trans-infection in lymphoid tissue-resident DC populations.
Finally, we found that M. tuberculosis exposure dramatically reduced processing and presentation of HIV virions, whereas presentation of exogenously added peptides was unimpaired, indicating that the DCs retained the ability to present processed antigens but could not degrade and present newly acquired virions. Our data suggest that endocytic downregulation is accompanied by HIV sequestration off the cell surface into pocketlike membrane invaginations, leaving the virus intact and capable of infecting CD4 T cells at the infectious synapse rather than trafficking into the lysosomal degradation and antigen presentation pathway. This result is supported by the findings of Henderson et al., who demonstrated a downmodulation of endocytosis in DCs after M. tuberculosis infection (16).
We propose that one way in which M. tuberculosis infection exacerbates HIV disease is by activation of DC trans-infection and concurrent reduction of de novo antigen processing and presentation. Myeloid DCs that encounter M. tuberculosis in the lungs become activated and migrate to locally draining lymph nodes, where they can induce M. tuberculosis-specific CD4 T cell responses. In the context of HIV infection, M. tuberculosis-induced defects in DC function act both to suppress specific immune control of HIV and to enhance DC-mediated HIV trans-infection of CD4 T cells. A single DC can interact with dozens of T cells in a lymph node in a day, providing ample opportunity to acquire and disseminate HIV among the lymphoid T cells. Enhanced HIV spread, therefore, is predicted to target primarily the CD4 T cell population, further hastening the decline of this cell population and the collapse of immune control of both pathogens.
We thank Keith Olszens for technical assistance, Zahra Toosi for critical reading of the manuscript, Jonathan Karn for useful discussions, and Jeffrey Lifson and the AIDS and Cancer Virus Program, NCI-Frederick, for AT-2-inactivated HIV.
We thank the Case/UHC Center for AIDS Research (NIAID AI36219) for financial support and use of core facilities.
This work was supported by NIH training grants GM08056 (M.A.R.) and GM007250 (N.D.P.). C.V.H. was supported by NIH AI034343, AI035726, and AI069085; D.H.C. was supported by NIH AI073217 and AI080313; and D.M. was supported by NIH AI073242 and AI087511.
Published ahead of print on 30 June 2010.