Mycobacteria-infected bystander macrophages trigger maturation of dendritic cells and enhance their ability to mediate HIV transinfection

Authors

  • Jolanta Mazurek,

    Corresponding author
    1. Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Stockholm, Sweden
    2. Swedish Institute for Infectious Disease Control, Solna, Sweden
    • Full correspondence: Dr. Jolanta Mazurek, Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Nobels vag 16, 17177 Stockholm, Sweden

      Fax: +46-8302566

      e-mail: Jolanta.Mazurek@ki.se

    Search for more papers by this author
  • Lech Ignatowicz,

    1. Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Stockholm, Sweden
    2. Swedish Institute for Infectious Disease Control, Solna, Sweden
    Search for more papers by this author
  • Gunilla Källenius,

    1. Department of Clinical Science and Education, Karolinska Institutet, Stockholm, Sweden
    Search for more papers by this author
  • Marianne Jansson,

    1. Department of Microbiology, Tumor and Cell Biology, Karolinska Institutet, Stockholm, Sweden
    2. Department of Laboratory Medicine, Lund University, Lund, Sweden
    Search for more papers by this author
    • These authors contributed equally to this work.

  • and Andrzej Pawlowski

    1. Department of Clinical Science and Education, Karolinska Institutet, Stockholm, Sweden
    Search for more papers by this author
    • These authors contributed equally to this work.


Abstract

Synergistic interplay between Mycobacterium tuberculosis (Mtb) and HIV in coinfected ind-ividuals leads to the acceleration of both tuberculosis and HIVdisease. Mtb, as well as HIV, may modulate the function of many immune cells, including DCs. To dissect the bystander impact of Mφs infected with Mtb on DC functionality, we here investigated changes in DC phenotype, cytokine profiles, and HIV-1 transinfecting ability. An in vitro system was used in which human monocyte-derived DCs were exposed to soluble factors released by Mφs infected with mycobacteria, including virulent clinical Mtb isolates and nonvirulent BCG. Soluble factors secreted from Mtb-infected Mφs, and to a lesser extent BCG-infected Mφs, resulted in the production of proinflammatory cytokines and partial upregulation of DC maturation markers. Interestingly, the HIV-1 transinfecting ability of DCs was enhanced upon exposure to soluble factors released by Mtb-infected Mφs. In summary, our study shows that DCs exposed to soluble factors released by mycobacteria-infected Mφs undergo maturation and display an augmented ability to transmit HIV-1 in trans. These findings highlight the important role of bystander effects during the course of Mtb–HIV coinfection and suggest that Mtb-infected Mφs may contribute to an environment that supports DC-mediated spread and amplification of HIV in coinfected individuals.

Introduction

Tuberculosis (TB) remains the leading cause of death among HIV-infected patients, accounting for about 26% of AIDS-related deaths [[1]]. Unlike other opportunistic infections affecting severely immunocompromised patients, active TB can develop throughout the entire course of HIV infection. Pulmonary TB is often diagnosed in people with relatively intact immune systems, whereas extrapulmonary TB is seen more frequently in AIDS patients (CD4 < 200/mm3) [[2]]. In principle, Mycobacterium tuberculosis (Mtb), which causes TB, can infect any organ in the body, but lymph nodes are the most common location for extrapulmonary TB [[3]].

The interplay between Mtb and HIV is two sided and syner-gistic. HIV infection contributes strongly to the reactivation of latent Mtb infection and progression to active TB. Thus, the lifetime risk of developing active TB in immunocompetent adults is estimated to be 5–10%, while this risk is increased to 5–15% annually in HIV-positive individuals [[4]]. In addition, Mtb infection severely dysregulates immune homeostasis of the host. Mycobacterial antigens enhance cytokine and chemokine production in blood [[5]] and elevated levels of proinflammatory cytokines in plasma are observed during pleural TB [[6]]. TB status is also reflected in the circulation by a changed transcriptional signature of blood cells suggesting altered gene expression profiles [[7]]. Studies in vitro suggest that growth of Mtb is augmented in HIV-1-infected Mφs [[8]]. At the same time, Mtb infection may support HIV-1 replication and dissemination in the host by dysregulation of the cytokine and chemokine balance [[9]]. Such increased replication of the virus has been demonstrated in lungs and pleural cavities of coinfected patients [[2]], along with locally and systemically augmented genetic heterogeneity of the HIV-1 quasispecies [[9, 10]]. It has also been shown that certain soluble factors released by monocytic cells may increase HIV-1 replication [[11]].

DCs are efficient antigen presenting cells that after antigen capture undergo a maturation process including upregulation of costimulatory molecules and expression of chemokine receptors, triggering the migration of DCs from the periphery to the draining lymph nodes [[12, 13]]. In the lymph nodes, DCs present antigen to naïve T cells and produce defined cytokine profiles that promote the development of an appropriate immune response. DCs are usually the first sentinels of the immune system to encounter and recognize pathogens; however, pathogens often utilize the immunosurveilling properties of DCs to facilitate host infection [[13]]. In this regard, HIV can infect DCs and may also make use of the DC-mediated immunological synapse for efficient spread in trans to CD4+ T cells [[14-17]]. The receptors involved in DC-mediated HIV transinfection are not yet fully defined; however, it has been suggested that C-type lectins, such as DC-specific intracellular adhesion molecule-3-grabbing nonintegrin (DC-SIGN) and the mannose receptor (MR), may play important roles [[18-20]].

DC-SIGN has also been implicated in Mtb–DC interactions. Owing to its high affinity for mannose-rich carbohydrates, DC-SIGN binds mannose-capped lipoarabinomannan (ManLAM), a cell-wall component of slow-growing mycobacteria. This binding has been suggested to facilitate infection by Mtb and to modify the intracellular signaling pathways of DCs exposed to Mtb [[21, 22]]. Although Mφs are the main target cells for Mtb, DCs are also readily infected. It is still unclear, however, whether DCs allow for intracellular growth of mycobacteria in vivo [[23, 24]]. Nevertheless, recently the important role of DCs in shaping the immune response during Mtb infection has been acknowledged [[25]].

Interactions between cells of innate immunity, still often neglected, have been shown to be crucial in shaping proper immune responses to pathogens [[26]]. Hence, we here analyzed the cross-talk between DCs and Mφs infected with different strains of mycobacteria using an in vitro model in which DCs were exposed to soluble factors released by infected Mφs. We found that DCs exposed to mycobacteria-infected bystander Mφs were activated with regard to proinflammatory cytokine release and the expression of DC maturation markers. We also analyzed the impact of mycobacteria-infected Mφs on the ability of DCs to mediate HIV-1 transinfection. Importantly, in the presence of soluble factors released by Mtb-infected Mφs, the ability of DCs to transmit HIV-1 in trans to T cells was augmented.

Results

Mycobacterial infection of bystander Mφs causes enhanced proinflammatory cytokine release

With the aim to mimic bystander effects mediated by an ongoing mycobacteria infection and to investigate the cross-talk between DCs and mycobacteria-infected Mφs, we devised an in vitro model system comprised of human monocyte-derived DCs and monocyte-derived Mφs (Supporting Information Fig. 1). Mφs were infected with one of three different mycobacterial strains: BCG or the clinical Mtb strains Harlingen or S96–129. Next, the infected Mφs were cultured with DCs for 48 h, separated from the DCs by a membrane insert. For comparison, experiments with cell-to-cell contact were carried out in order to assess the contribution of soluble mediators to observed effects. In the experimental set-up with the membrane insert, in which the DCs were only exposed to soluble factors released by the infected Mφs, we found increased production of TNF, IL-12p40 and IL-6 in the presence of Mφs infected with any of the mycobacterial strains used (Fig. 1A, 1B, 1C). Notably, in cultures of DCs stimulated with LPS, as positive control, the proinflammatory cytokine levels were higher than those induced by mycobacteria-infected Mφs.

Figure 1.

Presence of mycobacteria-infected bystander Mφs in DC culture triggers the production of proinflammatory cytokines. Mφs cultured in membrane inserts were infected with one of three mycobacterial strains: BCG, Harlingen (H), or S96-129 (S96). After 3 h, the inserts were transferred to wells containing imDCs. Culture supernatants were collected after 48 h, and the levels of (A) TNF, (B) IL-12p40, and (C) IL-6 produced were assessed by ELISA. Box plots show the median, 25–75% interquartile range, and whiskers denoting full range data from n = 9 donors.*p < 0.05, **p < 0.01 as determined by Wilcoxon matched pairs test in relation to noninfected Mφs.

In control experiments, when cell-to-cell contact was allowed between the DCs and Mtb-infected Mφs, we similarly observed increased production of proinflammatory cytokines, which was two- to fivefold higher than when the DCs and Mφs were separated by the semipermeable membrane (Fig. 2A, 2B, 2C). Coculturing of DCs with noninfected Mφs did not result in significant release of cytokines as compared with DCs cultured alone, regardless of whether or not there was direct contact between the Mφs and DCs (data not shown).

Figure 2.

Proinflammatory cytokine release in cocultures with DCs and mycobacteria-infected Mφs, either in direct cell-to-cell contact or separated by a semipermeable membrane. Mφs preinfected for 3 h with BCG, Harlingen (H) Mtb, or S96-129 (S96) were cocultured with DCs with (black bars) or without (white bars) separating semipermeable membrane. After 48 h, the coculture supernatants were collected, and (A) TNF, (B) IL-12p40, and (C) IL-6 concentrations were quantified by ELISA. Data are shown as mean + SD of n = 3 replicates and are representative of five experiments.

To verify that the observed effects were induced by infected Mφs and not by the presence of mycobacteria nonspecifically attached to the insert membrane, DCs were cultured for 48 h with inserts that had been preincubated with mycobacteria alone, and then subjected to extensive washing. With this protocol, no significant changes in cytokine release were observed for any of the mycobacterial strains tested (data not shown).

Furthermore, to determine which cell types contributed to production of particular cytokines in the cocultures with cell-to-cell contact, we compared the cytokine levels in cocultures containing fixed, infected Mφs with those containing live, infected Mφs, in addition to cultures of live, infected Mφs alone. Using this strategy, we found that mycobacteria-infected Mφs alone produced significant amounts of TNF and IL-6 (Supporting Information Fig. 2A), but only small quantities of IL-12p40, indicating that DCs were the main source of IL-12p40 in the cocultures. Interestingly, we observed that fixed, mycobacteria-infected Mφs induced considerable production of IL-12p40 and IL-6 in cocultures with DCs, as compared with fixed, noninfected Mφs, although at lower levels than those induced by live, mycobacteria-infected Mφs (Supporting Information Fig. 2). As noted for the set-up with the membrane insert, here also the production of cytokines tended to be augmented in a mycobacterial strain-dependent manner, regardless of whether the Mφs were live or fixed.

The source of specific cytokines was additionally identified by intracellular staining. The high frequency of Mtb-infected Mφs exhibited strong production of TNF, similar to LPS stimulated Mφs, and were also positive for IL-6 (Supporting Information Fig. 3A). DCs cultured in supernatants derived from Mtb-infected Mφs also produced TNF and IL-6, but to a lower degree (Supporting Information Fig. 3B). Staining for IL-12 was marginal for both cell types (Supporting Information Fig. 3C and D), yet, parallel experiments with cells from the same blood donor revealed that IL-12 was mainly produced by DCs, at a level similar to LPS stimulated DCs (Supporting Information Fig. 3D). Moreover, by comparing concentrations of cytokines in supernatants from Mtb-infected Mφs and in DCs cultured in conditioned medium from such Mtb-infected Mφs, we confirmed that DCs were the major source of IL-12 (Supporting Information Fig. 4).

Taken together, these observations suggest that mycobacteria-infected Mφs stimulate DCs to produce proinflammatory cytokines, at cell-to-cell contact as well as in bystander manner.

Soluble factors released by mycobacteria-infected Mφs cause partial maturation of DCs

To determine if the observed changes in cytokine profiles were accompanied by phenotypic changes in the DCs upon exposure to mycobacteria-infected Mφs, we analyzed the expression of the DC maturation markers CD86 and MHC class II/HLA-DR, in addition to the C-type lectins, DC-SIGN, and MR, on the surface of DCs. Here, culture of DCs separated from mycobacteria-infected Mφs by a membrane insert resulted in DC upregulation of CD86 and MHC class II (Fig. 3A and 3B). The upregulation of CD86 and MHC class II was observed to be mycobacteria strain-dependent and partial, as compared with the expression induced by LPS stimulation (Fig. 3A and 3B). A significant decrease in MR expression was only observed on DCs exposed to S96–129-infected Mφs. DC-SIGN expression on DCs was instead unaffected by the presence of mycobacteria-infected Mφs, irrespective of the infecting strain (Fig. 3A and 3B).

Figure 3.

Expression of maturation surface markers on DCs exposed to mycobacteria-infected bystander Mφs. DCs cultured for 48 h in the presence of Mφs infected with BCG, Harlingen (H), or S96-129 (S96) separated by a semipermeable membrane were stained for CD86, MHC class II, MR, and DC-SIGN and analyzed by flow cytometry. (A) The MFIs (mean fluorescence intensities) are plotted. CD86 and MHC class II expression levels were normalized to the levels expressed by DCs exposed to LPS (100%). MR and DC-SIGN expression levels were normalized to those observed in imDCs (100%). Box plots show the median, 25–75% interquartile range, and whiskers denoting full range data from n = 6–8 donors and *p < 0.05, **p < 0.01 as determined for nonnormalized data by Wilcoxon matched pairs test in relation to noninfected Mφs. (B) Histograms representing staining results of one representative donor of six analyzed.

In contrast, full maturation of DCs was seen when cell-to-cell contact was allowed with Mφs infected with any of the three strains of mycobacteria tested. Thus, CD86 molecules and MHC class II were upregulated to the levels observed for LPS-matured DCs (Supporting Information Fig. 3). In addition, MR and DC-SIGN were downregulated to a degree comparable with that induced by LPS. The BCG-infected Mφs induced less profound DC maturation as compared with that triggered by Mφs infected with the virulent Mtb strains (Supporting Information Fig. 5).

Coculturing of DCs with nonactivated Mφs only led to a moderate increase in CD86 and MHC class II expression and minor alteration of MR and DC-SIGN expression on the DC surface. In addition, coculturing DCs with Mφs preexposed to LPS resulted in a DC phenotype comparable with that observed in DCs directly treated with LPS, that is, upregulated levels of CD86 and MHC class II and downregulated levels of MR and DC-SIGN (data not shown).

In summary, these findings show that DCs may respond to ongoing Mtb infection in a bystander fashion, without coming into direct contact with infected Mφs. Such DCs undergo partial maturation, whereas direct contact between DCs and Mφs activated by mycobacterial infection leads to full DC maturation, similar to that observed with LPS-activated Mφs.

DCs exposed to Mtb-infected bystander Mφs display enhanced ability to mediate HIV-1 transinfection

Considering the phenotype and cytokine profile changes induced in DCs exposed to mycobacteria-infected bystander Mφs, we asked whether these changes would result in modulation of the ability of DCs to mediate HIV-1 transinfection of T cells. This ability was investigated using coculture with membrane insert, in which the DCs were preexposed to soluble factors released by mycobacteria-infected Mφs and subsequently pulsed with HIV-1, washed and added to the target cells, namely, activated PBMCs or CD4+ T cells. Viral replication was assessed by quantification of p24 antigen in the culture supernatants or by intracellular staining. The DCs partially matured in response to soluble factors, released by Mφs infected with Mtb clinical isolate S96–129 displayed enhanced ability to mediate HIV-1 transinfection of PBMCs and CD4+ T cells as compared with immature dendritic cells (imDCs) (Fig. 4A and 4B). Of note, this Mtb isolate was also the mycobacterial strain that elicited most pronounced DC maturation in regard to release of proinflammatory cytokines and upregulation of maturation markers (Fig. 1–3). The transinfection ability of DCs exposed to bystander Mφs infected with the other mycobacterial strains tended also to be augmented, but did not reach statistical significance (Fig. 4A). No significant p24 Ag levels were detected in cultures of DCs pulsed with HIV-1 but without addition of PBMCs or CD4+ T cells, neither did the proinflammatory environment of Mtb-infected Mφs alter the susceptibility of PHA-activated PBMCs to HIV-1 infection (data not shown). These experiments suggest that the observed increase in DC transinfection ability, triggered by exposure to Mtb-infected Mφs, was neither the effect of substantial HIV-1 replication in DCs nor altered replication of the virus in PBMCs.

Figure 4.

Ability of DCs to mediate HIV-1 transinfection is augmented upon exposure to Mtb-infected bystander Mφs. Mφs were infected with one of three mycobacteria strains: BCG, Harlingen (H), or S96-129 (S96) and (A and B) cocultured for 48 h with DCs where semipermeable membranes were separating the two cell types, or (C and D) cultured in conditioned medium from mycobacteria-infected Mφs for 48 h. Subsequently, DCs were (A) pulsed with HIV-1 and PHA-activated PBMCs or (B and D) CD4+ cells were added. (A) Levels of p24 antigen in PBMC cultures 7 days after transinfection by DCs exposed to mycobacteria-infected Mφs, expressed as fold change relative to the level of transinfection mediated by DCs cultured without Mφs. Box plots show the median, 25–75% interquartile range and whiskers denoting full range data from n = 8 donors, p* < 0.05 as determined by Wilcoxon matched pairs test in relation to noninfected Mφs. (B) Representative intracellular p24 staining of CD4+ T cells transinfected by DCs cocultured in the presence of Mtb-infected bystander Mφs. (C) Concentrations of p24 in supernatants collected four (left) and seven (right) days after transinfection assessed by ELISA and are shown as mean + SD of n = 3 replicates. (D) In parallel, intracellular p24 staining was performed 4 days after transinfection. (C and D) One representative donor out of four analyzed is presented. p** < 0.01 as determined by unpaired t-test.

Changes in HIV-1 transinfecting ability of DCs caused by mycobacteria-infected bystander Mφs were further studied by transfer of cell- and bacteria-free culture supernatants, collected after 48 h from Mtb-infected (S96–129 strain) Mφs, to imDCs. After another 48 h in culture, DCs were analyzed for HIV-1 trans-infection ability using activated autologous CD4+ T cells. Here, HIV-1 transinfection efficiency of DCs incubated in culture supernatant derived from Mtb-infected Mφs was significantly increased as compared with that of DCs cultured in supernatant derived from noninfected Mφs (Fig. 4C and 4D). DCs exposed to LPS also displayed increased ability to mediate HIV-1 transinfection.

Furthermore, DCs cultured in Mtb-infected Mφ supernatants displayed increased ability to capture HIV-1 (Fig. 5A). Moreover, such increased capture was inhibited by mannan, known to block interactions between HIV and C-type lectins [[27]] (Fig. 5A). Changes in HIV-1 capture by DCs after mannan preincubation were also reflected in reduced levels of transinfection (Fig. 5B). Furthermore, augmented ability of DCs to mediate HIV-1 transinfection upon bystander exposure to Mtb infection was here confirmed for another clinical Mtb isolate, BTB05–552 (Fig. 5B).

Figure 5.

Augmented HIV-1 capture and transinfection ability of Mtb-exposed DCs is caused by different virulent Mtb strains and inhibited by mannan. (A) DCs were incubated in supernatants from uninfected or S96-129 (S96) Mtb-infected Mφs for 24 h followed by a 3-h exposure to HIV-1, extensive washing and lysis of the cells. In parallel, DCs were preincubated with mannan for 1 h before virus pulsing. Cell lysates were assessed for the presence of p24 antigen. (B) DCs were cultured for 24 h in conditioned medium derived from Mφs, either uninfected or infected with one of two different Mtb clinical isolates (S96, BTB) (left). Next, DCs were pulsed with HIV-1 for 3 h, washed, and PHA-activated PBMCs were added as the target cells. In parallel, DCs were preincubated with mannan for 1 h before virus pulsing (right). Cell culture supernatants were harvested 4 days later and p24 levels were assessed by ELISA. (A and B) Data are shown as mean + SD from four donors. Statistical significance was assessed by paired t-test. Statistical differences between mannan-treated and untreated counterparts are shown above bars illustrating mannan-treated cells, and other significances are calculated for group pairs as indicated in the graphs and *p < 0.05, **p < 0.01.

Taken together, these findings suggest that DCs exposed, in bystander manner, to Mtb-infected Mφs display augmented ability to capture HIV-1 and consequently mediate infection in trans to CD4+ T cells.

Discussion

Coinfection with HIV and Mtb results in mutual acceleration of both HIV and TB disease, leading to early death for concurrently infected individuals [[2]]. The mechanisms underlying the synergy between HIV and Mtb are not known, but it has been established that both HIV and Mtb are capable of interfering with immune responses through interactions with multiple cells [[10, 22, 28]]. In the present study, we show that DCs exposed to soluble factors derived from mycobacteria-infected Mφs undergo partial maturation, including upregulation of costimulatory molecules and MHC class II, and downregulation of MR, while DC-SIGN expression is sustained. We also show that partial maturation of DCs is paralleled by enhanced release of inflammatory cytokines, the magnitude of which is related to the virulence of the mycobacterial strain. Importantly, our study also reveals that DCs partially matured by soluble factors released from mycobacteria-infected Mφs display enhanced ability to mediate efficient HIV-1 transinfection.

Under noninflammatory conditions, DC precursors are constitutively recruited from the blood to the lungs, where they migrate to the draining lymph nodes and show rapid turn-over [[29]]. With the onset of an infection, this process changes dramatically, and during the early phase of infection, DCs accumulate in the pulmonary draining lymph nodes [[30]]. In mice, it has been shown that DCs can transport live Mtb from the lungs to the mediastinal lymph nodes early in the infection [[31]]. During active pulmonary or pleural TB, however, the presence of mycobacterial compounds is not limited to the lungs and their draining lymph nodes. The mycobacterial cell wall component ManLAM can be detected in the blood [[32]] and urine [[33]] and also mycobacterial proteins can be found in sera of TB patients [[34]]. This suggests that certain mycobacterial components may induce bystander effects in distant tissues. Such effects would include induction of a strong inflammatory response. Indeed, a recent publication by Toossi et al. [[6]] revealed that during pleural TB in HIV-infected patients, the levels of proinflammatory cytokines, that is, TNF and IL-1β, were more elevated in the plasma than in the pleural fluid. In the current study, we focused on DCs exposed to the inflammatory environment generated by an ongoing mycobacterial infection. Our observations, that DCs exposed to Mφs infected with different mycobacterial strains undergo Mtb strain-dependent maturation and produce proinflammatory cytokines, are in line with recent studies [[35, 36]]. Interestingly, we show that augmented proinflammatory cytokine production could be induced without direct contact between DCs and Mtb-infected Mφs. This observation highlights the important role of bystander effects in the course of the coinfection, whereby noninfected cells are indirectly influenced by an ongoing infection. Our finding of augmented proinflammatory cytokine production by DCs either exposed to soluble factors from Mtb-infected Mφs or remaining in direct contact with Mtb-infected Mφs, confirms the development of a Th1-promoting bias that may control mycobacterial growth [[37]]. Of note, proinflammatory cytokines, particularly excess TNF, have been suggested to facilitate HIV replication [[38]], which in coinfected hosts may exacerbate HIV disease progression.

Using paraformaldehyde (PFA)-fixed mycobacteria-infected Mφs, in addition to intracellular cytokine staining, we found that TNF was mainly produced by Mφs, and that DCs were the source of IL-12p40. Similar observations were made by Giacomini et al. who found that DCs and Mφs responded to Mtb infection with distinct cytokine profiles, suggesting different and complementary functions of these two cell populations during the course of Mtb infection [[39]]. We also noted that the use of live, infected Mφs was not a precondition for proinflammatory cytokine induction when cells were in direct contact and that the cytokine levels induced by fixed infected Mφs were mycobacterial strain dependent. These observations suggest that the mechanisms of cytokine induction in direct Mφ−DC cocultures differed from those in which a semipermeable membrane was placed between the DCs and Mφs. Similarly, it was previously observed that the surface membrane interaction between fixed activated T cells and Mtb-infected monocytes enhanced TNF release [[40]].

We also found that upon exposure to mycobacteria-infected Mφs, DCs upregulated surface molecules engaged in T-cell priming, including CD86 and MHC class II. In experiments using the semipermeable membrane set-up, we found modest downregulation of MR, while DC-SIGN expression was similar or slightly elevated compared with that on imDCs. It has been shown that indirect activation of DCs by inflammatory mediators resulted in partial DC activation as compared with DCs matured by direct exposure to pathogens [[41]]. Given that DC maturation has been demonstrated to enhance the ability of DCs to transfer HIV to T cells [[42]] partially matured DCs could maintain their ability to mediate HIV transinfection. Indeed, we found that DCs, partially matured after exposure to soluble factors released by mycobacteria-infected Mφs displayed enhanced ability to mediate HIV-1 transinfection of T cells, as compared with immature DCs. Accordingly, we believe that Mtb-infected bystander Mφs may contribute to DC-mediated spread and amplification of HIV in coinfected individuals.

Many reports have postulated an important role of DCs in HIV pathogenesis (as reviewed in [[15]] and [[17]]). In particular, the ability of DCs to spread HIV infection in trans to CD4+ T cells, by making use of the immunological synapse between these cells, has been implicated [[18, 43, 44]]. Upregulation of MHC class II and costimulatory molecules on the surface of DCs, as seen in our Mtb–HIV coinfection model, enables close contact between DCs and T cells during formation of the immunological synapse and could therefore contribute to virus spread from DCs to T cells. In fact, it has been reported by others that DCs matured by LPS display an enhanced ability to transfer HIV to T cells [[42, 45]] as confirmed in the present study.

While, the role of DCs in HIV pathogenesis has mainly been discussed in relation to virus transmission, several pieces of data suggest that DCs serve as an HIV reservoir [[46-48]]. It has also been suggested that DC-mediated HIV transinfection may be a viral immune evasion mechanism as it has been shown that HIV neutralized with antibodies can be captured by DCs and later, upon release from DCs, may still be able to infect CD4+ T cells. Furthermore, we previously reported that HIV-1 variants evolving during chronic infection have sustained ability to utilize DC-SIGN for efficient spread to CD4+ T cells [[49]]. In the current study, we also confirmed an important role of C-type lectins in the transinfection process by showing that virus capture and subsequent transinfection could be inhibited by mannan, known to block interactions between HIV and C-type lectins [[27]]. Binding of mycobacterial ManLAM to DC-SIGN has also been described to activate DCs and enable productive HIV infection of these cells [[50]]. In line with these findings, and our own study, it was recently reported that Mtb-exposed DCs promoted HIV transinfection [[51]].

Although DCs are sparsely distributed in the body [[52, 53]], they can be readily affected during ongoing Mtb infection since signaling molecules and biologically active mycobacterial products released from infected Mφs are found systemically [[6, 32-34, 54]]. In addition, the studies by Berry et al. revealed that active TB altered gene expression patterns and cellular composition of peripheral blood, which also correlated with TB disease progression [[7]]. Thus, we believe that results from our in vitro model on bystander effects mediated by Mtb-infected Mφs may well reflect an in vivo context. Conceivably, our findings suggest that mycobacterial infection alters the immunological homeostasis and affects cells in distant tissues. However, further studies are necessary to explore the complex interactions between DCs and other cell types that take place during Mtb–HIV coinfection in order to understand the mechanisms that lead to deleterious disease progression in coinfected hosts.

Materials and methods

Generation of DCs and Mφs in vitro

Monocyte-derived dendritic cells

PBMCs were isolated from the buffy coats of blood samples from healthy donors by density-gradient centrifugation (Lympho-prep, Axis-Shield, Oslo, Norway). Monocytes were separated from PBMCs using CD14+ magnetic beads (Miltenyi, Bergisch Gladbach, Germany) and were more than 98% pure. Monocytes were cultured in R10 medium, that is, RPMI 1640 medium supplemented with 10% FBS, 2 mM L-glutamine, 50 U/mL penicillin, 50 μg/mL streptomycin (all from Gibco, Paisley, UK), and the following cytokines: 100 ng/mL rhIL-4 and 75 ng/mL rhGM-CSF (both from Peprotech, Rocky Hill, NJ, USA). On day 2 in culture, half of the cell culture medium was replaced, and new cytokines were added. On day 5, imDCs were obtained and used in further experiments. In some experiments, DCs were exposed to 100 ng/mL LPS (Sigma, Saint Louis, MO, USA) for 24 h or 48 h as a positive control for DC maturation.

Monocyte-derived Mφs

Monocytes prepared as described above were cultured in DMEM supplemented with 10% FBS, 1 mM sodium pyruvate, 50 U/mL penicillin, 50 μg/mL streptomycin and 40 ng/mL rhM-CSF. On day 2 in culture, half of the medium was replaced, and the cells were cultured for another 3 days. Culture in DMEM was found to be optimal for differentiation of monocytes into Mφs; at the later stage of culture during and post-Mtb infections Mφs were maintained in R10 medium. Preliminary experiments showed no difference in the outcome of the infection experiments, when DMEM/R10 and R10 only cultures were compared (data not shown). Antibiotics were present in the cell culture medium during differentiation process and were withdrawn 24 h before infection of cells with mycobacteria.

Infections

Mycobacterial infection

Mφs were cultured in inserts with a semipermeable membrane (0.2-μm pore size) (Nunc, Roskilde, Denmark) or in 96-well, U-bottom plates (105/insert/well) and infected with mycobacteria at a multiplicity of infection of one bacterium per cell. Four different mycobacterial strains were used: BCG and three clinical Mtb strains, the Harlingen isolate [[55]], and two isolates from Swedish outbrakes: S96–129 and BTB05–552 [[56, 57]]. After 3 h, the cells were extensively washed. DCs were then added to the infected Mφs and the inserts with infected Mφs were placed in the wells containing DCs for 48 h or cell culture supernatants were harvested after 24 h, filtered, and added to DC cultures. Two days after infection, more than 95% of the cells were alive as assessed by a trypan blue exclusion assay. To exclude the possibility that nonspecifically adhering bacteria affected the DCs in culture, mycobacteria alone were added to the inserts for 3 h, and the membranes were washed extensively and placed in the wells with imDCs for 48 h. In experiments using fixed Mφs, the cells were infected for 3 h as described above and then fixed with 4% PFA for 45 min. Fixed Mφs were extensively washed with saline and then added to imDCs for 48 h.

HIV-1 infection

After 48 h of imDC (5 × 105) and Mφ (104) culturing inserts containing Mφs were removed and DCs were pulsed with a primary HIV-1 R5 (CCR5-restricted) isolate, J1874 [[58, 59]], at an inoculum dose of 75 pg reverse transcriptase per well as determined by the CAVIDI HS Lenti kit (Cavidi Tech AB, Uppsala, Sweden). After 3 h of incubation with HIV-1, the cells were extensively washed, and PHA-activated PBMCs or CD4+ T-cells pooled from two donors were added at 105/well. The PBMCs were prepared in advance by culturing for 3 days in R10 supplemented with 2.5 μg/mL PHA. CD4+ T-cells were isolated from CD14-depleted PBMC fractions by positive selection using CD4+ MACS and precultured for 3 days in R10 supplemented with 2.5 μg/mL PHA. The mixture of target cells and HIV-pulsed DCs was incubated for another 4 or 7 days in the presence of R10 medium supplemented with 10 U/mL rhIL-2. Alternatively, in some experiments PHA-activated autologous CD4+ T cells were used in the transinfection assay. After 4- or 7-day coculture, the supernatants were harvested, and Triton X-100 was added to 1%. The samples were stored at –80°C until assayed. For p24 intracellular staining, PBMCs were harvested 4 days after transinfection. In the C-type lectin-blocking experiments, DCs were preincubated with 0.5 mg/mL mannan (Sigma-Aldrich) for 1 h followed by addition of HIV-1. Next, HIV-1 pulsed DCs were either lysed with 1% Triton X-100 and p24 determined by ELISA or PHA-activated PBMCs were added for another 4 days to assess the transinfection efficiency. The presence of HIV-1 p24 Gag antigen in the cell culture supernatants or cell lysates was quantified using an ELISA according to the manufacturer's protocol (Vironostika HIV Uni-Form II Ag/Ab, Biomerieux, Boxtel, the Netherlands). Alternatively, transinfection was assessed by p24 intracellular staining (see description next). In control experiments, PHA-activated PBMCs were directly infected with HIV-1 for 24 h and then washed and cultured for additional 4 days in medium containing IL-2.

Cytokine assays

Supernatants from 48-h cocultures of mycobacteria-infected Mφs and DCs were harvested and stored at –80°C until assayed. TNF, IL-6 and IL-12p40 levels in the cell culture supernatants were quantified using an ELISA according to the manufacturer's protocol (BD OptEIA Sets, BD Biosciences, San Diego, CA, USA). For intracellular staining of cytokines, Mφs were infected with Mtb and incubated with brefeldin (eBioscience, Hatfield, UK) for 12 h, thereafter cells were fixed with 4% PFA for 1 h and stained with mAbs specific for IL-6, TNF, and IL-12 or corresponding isotype control mAbs in the presence of permeabilization buffer. Samples were analyzed with FACSCalibur (BD Pharmingen) and data analyzed with FlowJo (Tree Star). The same procedure was used for analysis of cytokines produced by DCs cultured in supernatant derived from Mtb-infected Mφs.

Analysis of cell surface marker expression and intracellular p24 staining

After 48 h of coculturing, DCs and mycobacteria-infected Mφs were washed and detached from the plates with 2 mM EDTA for 10 min at 37°C. In some experiments, DCs were cultured in the presence of 100 ng/mL LPS; alternatively, Mφs were pretreated with LPS for 3 h before adding them to the DCs for the following 48 h as a positive control for activation. The harvested cells were then stained with fluorochrome-conjugated mAbs to assess the surface expression of CD86, DC-SIGN, CD14, MHC class II/HLA-DR, and MR. Corresponding isotype mAbs were used as negative controls. The expression of CD14 was used to distinguish between DCs and Mφs in cocultures: DCs were CD14dim, and Mφs were CD14bright. Furthermore, CD14 expression did not change upon Mtb infection (data not shown). For intracellular p24, staining cells (PBMCs or CD4+ cells) were harvested, stained with anti-CD14 and anti-CD3 mAbs, and fixed with 4% PFA for 1 h. Next, cells were washed and incubated in permeabilization buffer for 30 min at 20°C followed by staining with anti-HIV-p24 FITC-conjugated antibody (Beckman Coulter, San Diego, CA, USA). Samples were analyzed with a FACSCalibur or Canto II instrument and processed with FlowJo software. The data are presented as the normalized MFI (see below for normalization procedure) or as distribution histograms.

Statistical methods

The data were analyzed using unpaired t-test, paired t-test, and nonparametric statistical methods, Wilcoxon matched pairs test, or Friedman test with Dunn's posttest, as described in figure captions. p values < 0.05 were considered to be significant. Surface marker expression (MFI) and p24 concentration in cell supernatants were normalized for each donor to minimize donor-to-donor variability. Normalization of the data was performed using either the baseline level (appropriate negative control) or the positive control (LPS treatment) of the same donor, and the data are expressed as the fold change or % relative to control. For calculation of statistical significances, nonnormalized data were used. All experiments were repeated several times as indicated in the figure caption. Statistics were analyzed using GraphPad Prism 4 software.

Acknowledgements

We would like to thank Dr. Markus Sköld for critical reading and comments on the manuscript. This work was supported by European Commission Grant MEST-CT-2005-020872 (to J.M., L.I. and A.P.) and by the Swedish Research Council grants No. K2007-58X-13027 (to G.K.), K2010-56X-14588-08-3 (to M.J.), and No. K2010-79X-21374-01-3 (to A.P.)

Conflict of interest

The authors declare no financial or commercial conflict of interest.

Abbreviations
DC-SIGN

dendritic cell-specific intracellular adhesion molecule-3-grabbing nonintegrin

imDC

immature dendritic cell

ManLAM

mannose-capped lipoarabinomannan

MR

mannose receptor

Mtb

Mycobacterium tuberculosis

PFA

paraformaldehyde

TB

tuberculosis

Ancillary