SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Exocytosis of lysosomes from macrophages has been described as a response to microbial cytotoxins and haemolysins, as well as for releasing pro-inflammatory cytokines interleukin (IL)-1β and IL-18 during inflammasome activation. The mycobacterial ESX-1 secretion system, encoded in part by the Region of Difference-1, is a virulence factor necessary for phagosome escape and host cell lysis by a contact-dependent haemolysin in Mycobacterium marinum. Here we show that ESX-1 from M. marinum and M. tuberculosis is required for Ca2+-dependent induction of lysosome secretion from macrophages. Mycobacteria-induced lysosome secretion was concurrent to release of IL-1β and IL-18, dependent on phagocytosis of bacteria containing ESX-1. Synthesis but not release of IL-1β and IL-18 occurred in response to dead bacilli and bacteria lacking ESX-1, indicating that only cytokine release was regulated by ESX-1. Release of these cytokines and exocytosis of lysosomes were independent of intracellular mycobacterial growth, yet correlated with mycobacteria-encoded haemolytic activity, demonstrating a parallel pathway for the two responses. We further identified inflammasome components caspase-1, ASC and NALP3, but not Ipaf, required for release of IL-1β and IL-18. Collectively, these results reveal a role for ESX-1 in triggering secretion of lysosomes, as well as release of IL-1β and IL-18 during mycobacteria infection.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Microbial pathogens have evolved numerous strategies to manipulate the host and target immune responses to ensure survival. One mechanism of virulence involves direct secretion of microbial products into the cytosol of infected cells, either through specialized secretion systems or by cytolysin-mediated escape from phagosomes that initially segregate the invading pathogen from host cytosol. Many pathogenic mycobacteria, including Mycobacterium tuberculosis (Mtb), express cytolytic activity thought to be important for host cell lysis and cell-to-cell spread (Hsu et al., 2003; Gao et al., 2004). Tuberculosis (TB) is the world's most prevalent infectious disease, with about one-third of the world's population infected, leading to approximately 1.7 million deaths each year (WHO, 2008). Its virulence factors and mechanisms of pathogenesis have been widely studied, yet much of the molecular basis for the extraordinary success of this pathogen remains elusive. Dissemination is thought to result from migration of infected inflammatory macrophages, followed by the development of a generalized immune response. Through production of chemokines and pro-inflammatory cytokines, such as TNFα, interleukin (IL)-1β and IL-18, infected macrophages in all tissues attract an abundance of monocytes, lymphocytes and neutrophils. The resulting granulomas, which physically constrain but fail to eliminate the infection, are a hallmark of TB disease (Russell, 2007). Production of these pro-inflammatory cytokines may also be important for attracting cells that can disseminate infection.

Mycobacterium marinum (Mm) is genetically closely related to Mtb. It is the aetiological agent of a TB-like disease in fish and amphibians and has an in vitro growth rate that is three to five times faster than that of Mtb, making it an ideal model for studying pathogenesis of mycobacterial diseases (Stamm and Brown, 2004). Mm possesses virtually all of the currently known virulence factors associated with Mtb, including the ESX-1 secretion system, encoded in part by the Region of Difference 1 (RD1) (DiGiuseppe Champion and Cox, 2007). In addition to its role in secreting T cell antigens ESAT-6 and CFP-10, ESX-1 is required for mycobacterial cytolytic activity, although effector molecules for its functions have not been identified. In vitro studies have shown that mutations in individual genes within ESX-1 result in attenuation of cytolytic activity, host cell lysis and cell-to-cell spread of bacteria (Gao et al., 2004). In vivo, ESX-1 mutants of Mm are deficient in the formation of granulomas in zebrafish, demonstrating its role in manipulation of immune signalling (Volkman et al., 2004; Swaim et al., 2006).

Recognition of conserved microbial products by pattern recognition receptors triggers a wide range of responses for limiting proliferation of the microbe to minimize host damage. The ‘inflammasome’ is a cytosolic protein complex that is activated upon recognition of intracytoplasmic products of intracellular pathogens by nucleotide-binding and oligomerization domain (NOD)-like receptors (NLRs) (Mariathasan and Monack, 2007). Activation of the inflammasome results in activation of caspase-1 and subsequent processing and release of pro-inflammatory cytokines IL-1β and IL-18. It has been shown recently that the immature forms of IL-1β/IL-18, along with caspase-1, are transported to a subset of lysosomes where they are cleaved into their active forms and released into the macrophage milieu upon Ca2+ influx and K+ efflux (Andrei et al., 1999; 2004). Once released, these cytokines have a variety of effects on immune cells, including chemoattraction of additional phagocytes, and a likely role in granuloma formation (Russell, 2007). While inflammasome activation and subsequent IL-1β and IL-18 synthesis are strongly upregulated in Mtb-infected macrophages (Giacomini et al., 2001; Montero et al., 2004), the mechanism by which mycobacteria activate the inflammasome for processing and release of these cytokines has not been examined closely.

Previously, we reported an observation that Mm infection of macrophages induces secretion of lysosomes (Koo et al., 2008). In addition Mtb phagosomes were recently shown to express lysosomal markers briefly in very early stages of infection (van der Wel et al., 2007), suggesting that lysosomes may fuse with forming phagosomes. To understand the role of lysosomes in early events during mycobacterial infection, we have characterized the mechanism of lysosome secretion and have found an essential role for ESX-1 in inducing secretion of lysosomes and in release of IL-1β/IL-18. ESX-1 was not necessary for mycobacteria-induced synthesis of IL-1β/IL-18, but was required for release of these cytokines from infected macrophages. Secretion of these potent pro-inflammatory cytokines may contribute to pathogenesis by recruiting new mononuclear phagocytes to sites of infection, increasing the likelihood of granuloma formation and perhaps the spread of mycobacteria to uninfected cells.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Lysosome secretion during initial Mm infection of macrophages

We previously showed that Mm induces secretion of lysosomal contents into the supernatant of bone marrow-derived macrophages (BMDM) (Koo et al., 2008). To determine whether the secretion was targeted towards invading bacilli or a general cell stress response, confocal time-lapse microscopy was used to observe the movement of lysosomes in BMDM during early stages of infection by Mm. BMDM expressing lysosome-associated membrane protein 1 (LAMP1) tagged with red fluorescent protein (RFP) were infected with green fluorescent protein-expressing Mm (MmGFP), and observed for 1 h after addition of the inoculum. When MmGFP came in contact with the macrophage plasma membrane, there was rapid lysosome recruitment towards the periphery of the cell where mycobacteria were localized (Fig. 1A, Movie S1), demonstrating specific targeting of lysosomes towards invading bacteria. To confirm the localized secretion of lysosomes, BMDM were fixed after 5 min of exposure to MmGFP and stained for externalization of LAMP1. Immunofluorescent staining of unpermeabilized cells showed LAMP1 colocalization with GFP-labelled Mm (MmGFP) at the plasma membrane (Fig. 1B), demonstrating that exocytosis of lysosomes occurred in the vicinity of mycobacterial contact. These results demonstrate that lysosomes are recruited to the site of mycobacteria invasion at the periphery of the cell.

image

Figure 1. Mm triggers exocytosis of lysosomes at sites of invasion. A. Confocal time-lapse microscopy of RFP-LAMP1 and MmGFP. Cells were visualized 5 min after inoculum was centrifuged on to monolayer. Minutes after the beginning of recording are indicated. Last frame shows outline of cell at the end of recording. Arrows indicate new invading MmGFP at periphery of cell. B. Staining of externalized LAMP1 after addition of MmGFP. Cells were fixed 5 min after addition of inoculum, and cells were immunostained for lumenal LAMP1 in the absence of cell permeabilization. Arrows indicate invading MmGFP in close proximity to exocytosed lysosomes.

Download figure to PowerPoint

Mycobacterial cytolytic activity is a trigger for secretion of lysosomes

Lysosome secretion can occur after mechanical injury to the plasma membrane, providing supplemental membrane for sealing discontinuities in the membrane (Reddy et al., 2001). Likewise, lysosome secretion also occurs in response to pore formation by a number of bacterial pathogens, including Salmonella typhimurium, Yersinia pestis and Streptococcus pyogenes (Reddy et al., 2001; Hakansson et al., 2005). Because mycobacteria are cytolytic towards host macrophages (Hsu et al., 2003; Gao et al., 2004), we hypothesized that mycobacteria-induced cytolysis could trigger targeting of lysosomes towards sites of invasion. Lysosome secretion was measured by release of the lysosomal enzyme β-hexosaminidase in response to infection by wild-type (WT)Mm, a non-cytolytic ΔRD1 mutant (Volkman et al., 2004), and heat-killed mycobacteria (also non-cytolytic). Both ΔRD1 and heat-killed Mm failed to induce lysosome secretion, even at very high multiplicities of infection (moi) (Fig. 2A), demonstrating that both bacterial viability and ESX-1, whose function depends on genes within RD1, are essential for inducing secretion of lysosomes. Infection with a ΔRD1 mutant transformed with a RD1-encoded plasmid (ΔRD1::RD1-2F9) (Pym et al., 2002) resulted in secretion of lysosomes at levels higher than that induced by WT Mm (Fig. 2A), perhaps because this strain is diploid for several genes involved in ESX-1 function. Secretion of lysosomes was next compared in macrophages infected by the Erdman strain of Mtb, an ESX-1 mutant 11D6 (Rv3849::tn) and heat-killed Mtb. Wild-type Mtb induced secretion of lysosomes, whereas 11D6 and heat-killed Mtb did not (Fig. 2B), demonstrating that Mtb and Mm share properties required for such secretion.

image

Figure 2. Mm-induced lysosome secretion depends on ESX-1-encoded haemolytic activity. A. Secretion of lysosomes in response to Mm strains (moi = 50): no bacteria (NB), wild type (WT), ESX-1 mutant (ΔRD1), heat-killed Mm (HK) or the complemented ESX-1 mutant (ΔRD1::RD1-2F9). B. Secretion of lysosomes in response to Mtb strains (moi = 50): no bacteria (NB), WT (Mtb), Rv3849::tn (11D6) or heat-killed Mtb (HK). C. Flow cytometric analysis of LAMP1 externalization 10 min and 2 h after infection by MmGFP, ΔRD1-GFP and HK-GFP. Histogram shows GFP-positive cells stained for LAMP1 externalization. D. Contact-dependent sheep erythrocyte haemolysis after incubation with Mm mutants. E. Secretion of lysosomes in response to Mm mutants. Secreted β-hexosaminidase was measured from supernatants collected 2 h after infection. All experiments were performed in triplicate.

Download figure to PowerPoint

To quantify ESX-1-dependent exocytosis of lysosomes as a function of bacterial load, macrophages infected with GFP-expressing WT M. marinum (MmGFP), a GFP-expressing ΔRD1 mutant (ΔRD1-GFP), or heat-killed MmGFP (HK-GFP) were stained for externalized LAMP1 and analysed by flow cytometry. By 10 min post infection, 7.4% of macrophages exposed to MmGFP were both GFP- and LAMP1-positive (Fig. 2C). Of those cells that were positive for ingested MmGFP, 16.2% showed increased externalized LAMP1. At 2 h post infection, externalization of LAMP1 increased to 22.2% of all cells exposed to MmGFP, and 33.6% of macrophages that had ingested MmGFP were LAMP1-positive. Cells infected with ΔRD1 or HK organisms showed minimal differences in LAMP-1 staining from control macrophages (Fig. 2C), despite equivalent phagocytosis to MmGFP. These results demonstrate the necessity of ESX-1 for macrophage exocytosis of lysosomes at the plasma membrane.

Because it was possible that failure to induce lysosome secretion represented a general property of mycobacterial mutants with decreased virulence, we examined Mm mutants, kasB::tn and iipA::tn, with transposon insertions outside RD1. To assess cytolytic activity, Mm-induced contact-dependent haemolysis of sheep erythrocytes was measured. Although kasB::tn and iipA::tn are both severely attenuated in macrophages (Gao et al., 2003a; 2006), both mutants retained haemolytic activity (Fig. 2D) and induced lysosome secretion (Fig. 2E) similar to WT, suggesting that this secretion correlates with membrane lysis, rather than intracellular mycobacterial growth. The Mh3866::tn mutant, whose growth in macrophages is similar to that of WT (Gao et al., 2004), had reduced haemolytic activity and induced less secretion of lysosomes (Fig. 2D and E). Therefore, the ability of a mycobacterial strain to induce lysosome secretion correlates with its cytolytic potency rather than with its ability to replicate in macrophages.

Mycobacteria-induced lysosome secretion is Ca2+-dependent and necessary for host cell survival

Secretion of lysosomes in response to membrane damage is triggered by an increase in intracellular Ca2+, either through the release of intracellular stores or via an influx from the extracellular milieu (Reddy et al., 2001). If WT Mm were damaging the plasma membrane, then the influx of extracellular Ca2+ could serve as the trigger for lysosome secretion. To determine the importance of Ca2+ in regulating mycobacteria-induced lysosome secretion, the effect of removal of extracellular Ca2+ and of buffering intracytoplasmic Ca2+ was assessed. Macrophages infected with WT Mm in the absence of extracellular Ca2+ or treated with a membrane-permeable Ca2+ chelator prior to infection failed to secrete lysosomes, with depletion of intracellular Ca2+ having a slightly greater effect (Fig. 3A).

image

Figure 3. Mm-induced lysosome secretion is regulated by intracellular Ca2+ and is necessary for host cell survival. A. Effect of depletion of extracellular Ca2+ (no [Ca2+]e) or intracellular Ca2+ (BAPTA/AM) on secretion of β-hexosaminidase. B. Measurement of cell death induced by Mm in the absence of Ca2+. Cell death was measured by the CellTiter Blue assay (Promega). Asterisks (*) indicate levels beyond detection. Assays were performed on samples collected 2 h after infection. All experiments were performed in triplicate.

Download figure to PowerPoint

If the influx of Ca2+ required for eliciting lysosome secretion were necessary for repair of membrane damage induced by mycobacteria-encoded cytolysin, then inhibition of secretion should be detrimental to the host cell. To test this prediction, macrophage viability was measured in the absence of Ca2+. In the absence of extracellular Ca2+, infected macrophages showed a 40% reduction in viability, whereas in the absence of both extra- and intracellular Ca2+, infected macrophages showed an 80% reduction in viability (Fig. 3B). These results demonstrate that secretion of lysosomes is triggered by increases in cytosolic Ca2+ arising from both extra- and intracellular sources and is necessary for host cell survival.

ESX-1 is necessary for mycobacteria-induced release of IL-1β and IL-18

Membrane damage and cellular stress can serve as signals for activation and release of pro-inflammatory cytokines IL-1β and IL-18 (Ogura et al., 2006), recently demonstrated to be secreted via a subset of lysosomes (Andrei et al., 2004). We hypothesized that these cytokines could be released via the same or parallel mechanism during infection by mycobacteria, ultimately resulting in modulation of the host immune system. Release of IL-1β in response to WT, ΔRD1 and heat-killed Mm was measured. After 24 h of infection, WT Mm induced release of IL-1β at all moi tested, in contrast to the ΔRD1 mutant and heat-killed mycobacteria (Fig. 4A). Synthesis and release of IL-1β require two distinct signals: the first signal (e.g. TLR activation) triggers the synthesis of pro-IL-1β, and the second signal activates caspase-1 processing and release of the cytokine (Mariathasan and Monack, 2007). Importantly, although IL-1β was not released in response to ΔRD1 or heat-killed mycobacteria, the cytokine was detected in cell lysates, particularly in macrophages infected with higher moi, at which macrophages infected by ΔRD1 and heat-killed Mm remain viable (Fig. 4B). The decrease in IL-1β release observed at the highest moi in response to WT Mm is likely a consequence of early host cell cytotoxicity induced by these bacteria (Gao et al., 2004). IL-18 was also released in an ESX-1-dependent manner from BMDM (Fig. 4C). No differences were observed in TNFα release, which occurs via cleavage from the plasma membrane and is thus unrelated to lysosome secretion (Fig. 4D). Release of IL-1β was next measured from BMDM infected with WT Mtb, the ESX-1 mutant 11D6, and heat-killed Mtb. Wild-type Mtb induced high levels of IL-1β release, whereas infection by 11D6 and heat-killed Mtb resulted in a much more attenuated response (Fig. 4E), demonstrating that ESX-1 and bacterial viability are shared features of Mtb and Mm, important in signalling release of IL-1β during infection.

image

Figure 4. ESX-1 is necessary for secretion, but not synthesis of IL-1β and IL-18. A. Release of IL-1β from BMDM in response to Mm strains: wild-type (WT), ESX-1 mutant (ΔRD1) and heat-killed Mm (HK). Multiplicities of infection (moi) are indicated as bacteria : macrophage ratio. B. IL-1β in lysates of infected BMDM. Asterisks (*) indicate levels below detection. C. Release of IL-18 in response to Mm infection. D. Secretion of TNF-α in response to Mm infection. E. Release of IL-1β from BMDM in response to Mtb strains: WT (Erdman), Rv3849::tn and heat-killed Mtb (HK). All assays were performed in triplicate, using supernatants and lysates collected 24 h after infection.

Download figure to PowerPoint

To further isolate the role for ESX-1 in IL-1β release, macrophages were primed with LPS for 4 h to activate IL-1β synthesis. Primed macrophages were infected with WT, ΔRD1 or heat-killed Mm for 2 h, during which time live Mm are not replicating, and thus eliminating replication differences among the strains as a potential variable in IL-1β release. At 2 h post infection of LPS-primed macrophages, WT Mm induced IL-1β release, whereas ΔRD1 and heat-killed Mm did not (Fig. 5A). The complemented ΔRD1::RD1-2F9 strain induced release of WT levels of IL-1β (Fig. 5A), proving the requirement for ESX-1 to activate the signal for IL-1β release.

image

Figure 5. ESX-1 is necessary for triggering the release of IL-1β and IL-18 from primed BMDM. BMDM were primed with LPS, wild-type (WT), ESX-1 mutant (ΔRD1) or heat-killed (HK) Mm overnight. Cytokine release was measured 2 h after infection (A and C) or re-infection (B, D–F) using indicated Mm strains. A. Release of IL-1β from LPS-primed BMDM in response to infection by Mm. B. Release of IL-1β from LPS-primed BMDM in response to Mm mutants. C. Release of IL-1β from LPS-primed BMDM in the absence of phagocytosis. Cytochalasin D was used to treat cells prior to infection by Mm. D. Stimulation of IL-1β release from Mm-primed BMDM in response to re-infection by Mm. E. Release of IL-18 from Mm-primed BMDM in response to re-infection by Mm. F. Secretion of TNF-α from Mm-primed BMDM in response to re-infection by Mm. All assays were performed in triplicate.

Download figure to PowerPoint

To determine whether the requirement for ESX-1 reflected its effects on virulence or its role in cytolysis, kasB::tn, iipA::tn and Mh3866::tn were added to LPS-primed macrophages, and release of IL-1β was measured after 2 h. Both kasB::tn and iipA::tn, which are very haemolytic (Fig. 2D) but severely attenuated in intracellular growth in BMDM (Gao et al., 2003a; 2006), induced release of IL-1β (Fig. 5B). Mh3866::tn, which has low haemolytic activity (Fig. 2D) but grows similarly to WT in macrophages (Gao et al., 2004), failed to induce release of IL-1β. These results suggest that the ESX-1-dependent cytolysin is involved in inducing release of IL-1β. Furthermore, no IL-1β was released in response to Mm in the presence of cytochalasin D (Fig. 5C), suggesting that phagocytosis of WT Mm is required to trigger pro-IL-1β processing and release.

To examine the possibility that failure to release IL-1β or IL-18 in response to ΔRD1 or heat-killed mycobacteria despite adequate synthesis resulted from inhibition of the process in the absence of functional ESX-1, BMDM were infected with WT, ΔRD1 or heat-killed mycobacteria to activate synthesis of pro-IL-1β and pro-IL-18. After 24 h, addition of WT Mm to each sample of infected cells induced release of IL-1β (Fig. 5D) and IL-18 (Fig. 5E), whereas ΔRD1 and heat-killed mycobacteria did not. TNFα was secreted in response to all three stimuli (Fig. 5F). Thus pre-incubation with ΔRD1 or heat-killed bacteria did not inhibit the signal from WT Mm that leads to cytokine processing and release.

NALP3 is necessary for inflammasome activation by mycobacteria

The inflammasome is comprised of caspase-1, the adaptor ASC and a NLR protein that recognizes bacterial products and triggers its activation. Caspase-1 is the central effector protein, necessary for processing and release of IL-1β and IL-18. Macrophages were infected with WT Mm and ΔRD1, and levels of IL-1β in the supernatant and lysate were measured in the presence of the caspase-1 inhibitor YVAD. Cells pre-treated with YVAD showed a significant decrease in the amount of IL-1β released in response to WT Mm infection (Fig. 6A), demonstrating a role for caspase-1 in mycobacteria-induced IL-1β release.

image

Figure 6. Mm activates the NALP3/ASC inflammasome. A. Effect of caspase-1 inhibitor YVAD (50 μm) on secretion of IL-1β from BMDM. B. Secretion of IL-1β from macrophages of ASC−/− and NALP3 knockout (NALP3−/−) mice. C. Secretion of IL-18 from macrophages of ASC−/− and NALP3 knockout (NALP3−/−) mice. D. Secretion of IL-1β from LPS-primed BMDM in response to ATP and nigericin. E. Secretion of β-hexosaminidase from untreated or LPS-primed BMDM in response to ATP and nigericin. Assays were conducted using samples collected 2 h after infection. Asterisks (*) indicate levels below detection.

Download figure to PowerPoint

Most bacterial pathogens activate the inflammasome through one of two NLRs: NALP3/Cryopyrin (Listeria monocytogenes, Staphylococcus aureus) or IPAF (S. typhimurium, Shigella flexneri, Legionella pneumophila) (Mariathasan and Monack, 2007). To determine which of these is responsive to ESX-1, macrophages isolated from NALP3−/− and IPAF−/− mice were infected with WT and ΔRD1 Mm, and levels of released IL-1β and IL-18 were determined. In contrast to WT macrophages, NALP3−/− macrophages failed to release IL-1β or IL-18 in response to WT Mm, whereas IPAF−/− macrophages released both cytokines normally (Fig. 6B and C). ASC−/− macrophages, which fail to assemble either NALP3 or IPAF containing inflammasomes, also failed to release IL-1β and IL-18 in response to infection with WT Mm. These results indicate that release of the cytokines is regulated by the NALP3 inflammasome. Lysosome secretion from the knockout macrophages did not differ significantly from WT macrophages, suggesting that it was independent of inflammasome activation (data not shown).

Activation of the inflammasome and subsequent release of IL-1β appeared to correlate with secretion of lysosomes. To test whether the correlation was specific to mycobacteria, β-hexosaminidase release was examined in response to ATP or nigericin, known activators of inflammasome-mediated release of IL-1β. ATP and nigericin induced release of both IL-1β and β-hexosaminidase (Fig. 6D and E), further supporting a parallel pathway for lysosome and cytokine release.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We here show that interaction of cytolytic mycobacteria at the macrophage plasma membrane leads to lysosome secretion. Induction of macrophage lysosome secretion requires the bacterial virulence-related machine ESX-1 that is involved in expression of a contact-dependent haemolysin on the surface of mycobacteria. Our data suggest that this cytotoxin signals localized Ca2+-dependent lysosome secretion. Localized secretion of lysosomes was not observed in response to infection with the ΔRD1 mutant (data not shown). ESX-1 also was required for mycobacteria activation of the NALP3 inflammasome and subsequent release of IL-1β and IL-18, pro-inflammatory cytokines that are likely involved in granuloma formation and dissemination of infection. The release of these cytokines followed a parallel pathway to secretion of lysosomes, as both occurred in response to similar stimuli, including the ‘classical’ inflammasome activators nigericin and ATP.

We have shown that ESX-1-expressing mycobacteria activate NALP3-, but not IPAF-containing inflammasomes for post-translational processing and release of IL-1β and IL-18, but ESX-1 is not necessary for synthesis of these cytokines. Recent evidence demonstrates that the NALP3 inflammasome can be activated by a peptidoglycan fragment, muramyl dipeptide (MDP) in the host cell cytoplasm (Martinon et al., 2004). As the ESX-1-dependent cytolytic activity is required for phagosome escape by both Mm and Mtb (Gao et al., 2004; van der Wel et al., 2007), it seems logical that this cytolysis could be a major mechanism by which mycobacterial MDP or other bacterial products enter the macrophage cytoplasm. All other bacteria known to activate the NALP3 inflammasome also express cytolysins or haemolysins, suggesting parallel mechanisms of action (Mariathasan and Monack, 2007).

ESX-1 is a well-described virulence determinant in pathogenic mycobacteria. As it likely has evolved for optimal dissemination of the pathogen, we were interested in determining how induction of lysosome secretion might be beneficial for Mm. The role of ESX-1 in pathogenesis is yet to be fully defined, but is clearly involved in altering cytokine signalling, vesicular trafficking, phagosome escape and host cell lysis (Stanley et al., 2003; Gao et al., 2004; MacGurn and Cox, 2007). Because lysosome secretion occurs at the site of mycobacterial invasion, one possible outcome of this response is a remodelling of the phagosome, resulting in the characteristic non-fusigenic mycobacteria-containing phagosome. MacGurn and Cox recently identified several ESX-1-related genes in a screen for Mtb mutants that fail to inhibit phagosome–lysosome fusion, thus defining a role for ESX-1 in vesicle fusion with the mycobacteria-containing phagosome (MacGurn and Cox, 2007). In addition, Van der Wel et al. recently showed that Mtb briefly colocalizes with lysosomes within minutes of phagocytosis, prior to their escape into the cytoplasm (van der Wel et al., 2007). In conjunction with the findings reported here, it is reasonable to consider that ESX-1 is involved in each of these events and that mycobacteria-induced lysosome secretion is related to phagosome remodelling very early in infection, even during the process of bacterial uptake.

An alternative but not mutually exclusive role for mycobacteria-induced lysosome secretion in pathogenesis is that ESX-1-dependent secretion of IL-1β and IL-18 promotes successful infection of the host. IL-1β, which lacks a classical secretion signal, is processed and released by a subset of lysosomes in response to Ca2+ influx and K+ efflux (Andrei et al., 2004), and the structurally related IL-18 is expected to follow the same pathway of secretion. There is evidence that IL-1β and IL-18 are both necessary for the formation of the mycobacterial granuloma, in which bacteria can establish long-lived latent infection (Pechkovsky et al., 2006). Thus, IL-1β and IL-18 appear to be involved in the creation of a niche for the bacteria protected from a sterilizing immune response. As a result, mycobacteria can establish a prolonged phase of infection in which they persist, but do not kill their host. Another possible role for IL-1β and IL-18 in pathogenesis of TB may involve recruitment of new mononuclear phagocytic cells for dissemination of infection.

Andrews and colleagues have defined a role for lysosome secretion in membrane repair, demonstrating that an influx of Ca2+ from the extracellular milieu is responsible for triggering secretion (Reddy et al., 2001). We have demonstrated that Ca2+-dependent secretion of lysosomes is associated with host cell survival, consistent with the membrane repair hypothesis, as chelation of either extra- or intracellular Ca2+ results in host cell cytotoxicity. Under all experimental conditions that included Ca2+, no significant host cell death occurred (data not shown). Our findings also suggest a close correlation between exocytosis of lysosomes and signalling for release of IL-1β and IL-18. Clearly, secretion of lysosomes occurs even in the absence of TLR-activated IL-1β and IL-18 synthesis. However, in LPS-primed macrophages, a very strong correlation was found between stimuli of lysosome secretion and those of IL-1β and IL-18 release. It is possible that these cytokines are released via the same subset of lysosomes that exocytose for a membrane repair response to mycobacterial stimulation, or that the secretion pathways occur in parallel, responding to the same stimuli and signalling molecules.

ESX-1 is clearly required for virulence, but the precise mechanisms of its involvement in disease are unknown. The ESX-1-dependent cytolytic activity, necessary for phagosome escape, host cell lysis, and, as described here, secretion of lysosomes and release of IL-1β and IL-18, may have a key role in TB pathogenesis. It is our hypothesis that this contact-dependent cytolytic activity underlies the role of ESX-1 in pathogenesis. More work is currently underway to identify the molecular nature of the cytolysin responsible for these events, as well as to characterize this specialized secretion machinery. There is evidence that ESAT-6, one of the secreted effectors of ESX-1, displays membranolytic activity (Hsu et al., 2003; de Jonge et al., 2007), but whether this molecule represents the key toxin that signals lysosome secretion remains to be determined. Alternatively, it is also possible that indirect release of ATP during mycobacterial infection may induce secretion via stimulation of P2X7 receptors (Mehta et al., 2001).

In conclusion, we have described here a previously unrecognized role for the ESX-1-dependent cytolytic activity in eliciting secretion of lysosomes and release of IL-1β and IL-18 from mycobacteria-infected macrophages, adding to the numerous virulence effects currently identified for this novel secretion system. Understanding the mechanism of ESX-1-mediated secretion as well as its effectors responsible for the multiple virulence traits associated with this novel pathway should lead to valuable insights into the pathogenesis of TB and other mycobacterial diseases, as well as to potentially new avenues for therapy.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Bacterial strains and cell culture

Wild-type Mm (strain M), MmGFP, ΔRD1, ΔRD1-GFP and ΔRD1::RD1-2F9 were cultured in Middlebrook 7H9 (Difco) supplemented with 0.2% glycerol/0.05% Tween 80/10% ADC enrichment (Fisher). ΔRD1 was transformed with the pRD1-2F9 plasmid (Pym et al., 2002), as previously described (Gao et al., 2003b), to generate ΔRD1::RD1-2F9. Heat-killed bacteria were prepared by incubation for 30 min at 65°C, and loss of viability was confirmed by plating on 7H10 plates. Mtb (Erdman strain) and the 11D6 mutant were cultivated as previously described (Cox et al., 1999). Macrophages were derived from the bone marrow of C57BL/6, ASC−/−, NALP3−/− and IPAF−/− mice and harvested after 7–21 days, as described (Sango et al., 1995; Mariathasan et al., 2004; 2006).

Microscopy

Infected BMDM were fixed and stained with rat anti-mouse LAMP1 antibody (BD Pharmingen) and Alexa 594 anti-rat antibody (Molecular Probes). Coverslips were mounted with Prolong antifade reagent (Molecular Probes). Images were acquired on a Nikon Eclipse TE300 microscope with an Evolution QEi cooled charge-coupled device (Media Cybernetics, Silver Spring, MD) with IPLAB acquisition software (Scanalytics).

BMDM were transduced using retroviral vectors encoding Lamp1-RFP (pLZRS-Lamp1-RFP), as described (Czibener et al., 2006). Time-lapse microscopy was performed using the 100× NA 1.4 objective of an epifluorescent microscope (TE2000; Nikon). RFP and GFP channels were imaged every 10 s, using acquisition software (Micro-Manager).

Macrophage infections

Mycobacteria were washed twice in Hanks' Buffered Salt Solution (HBSS) without Ca2+ and Mg2+ and disrupted into single bacilli by passage through a 26-gauge needle, added to BMDM, centrifuged at 500 g for 5 min, and incubated at 32°C in 5% CO2. Opsonized mycobacteria were prepared by incubation of Mm with human serum for 45 min at 37°C.

For co-infection studies, BMDM were plated at 5 × 104 cells well−1 in a 96-well plate. Bacteria were added to monolayers for 2 h at an moi of 10, washed twice with HBSS and incubated in DME-3 at 30°C in 5% CO2. The following day new preparations of bacteria were made and spun on to infected monolayers in DME-3. After 2 h incubation at 30°C, supernatants and lysates were collected.

Cytokine assays

Cytokine levels were measured by enzyme-linked immunosorbent assay (ELISA) according to manufacturers' instructions (IL-1β, BD Biosciences; IL-18, Invitrogen; TNF-α, eBioscience).

Lysosome secretion assays

BMDM from C57BL/6 (WT) were plated at 5 × 104 cells well−1 in a 96-well plate. Mycobacteria were prepared as described above and spun on to BMDM monolayers in HBSS or cells pre-incubated in HBSS without Ca2+ and Mg2+, 10 μM BAPTA/AM or 1 μM cytochalasin D, as indicated. After 2 h incubation at 30°C, supernatants were collected, and cells were lysed in 0.5% Triton X-100 in 0.1 M citrate buffer, pH 5. To measure β-hexosaminidase activity 25 μl of 2 mM 4-Methylumbelliferyl N-acetyl-β-d-glucosaminide (Sigma) in 0.2 M sodium acetate buffer, pH 4.8, was incubated with an equal volume of supernatant or lysate and incubated at 37°C for 1 h. The reaction was stopped with 100 μl of 0.1 M sodium carbonate buffer, pH 10.5, and released 4-methylumbelliferone was quantified in a fluorometer with excitation at 355 and emission at 460 nm. The percentage of cellular β-hexosaminidase that had been secreted was calculated from enzyme activity of supernatants and lysates. Viability of infected cells was measured using the CellTiter-Blue (Promega) cell viability assay according to manufacturer's instructions (data not shown). All experiments were performed in triplicate.

Flow cytometry

Flow cytometric analysis of LAMP1 externalization was performed as described (Hakansson et al., 2005). Briefly, BMDM were infected at an moi of 50 for 2 h. Cells were washed and detached by addition of EDTA, then stained with anti-LAMP1 antibody for 30 min. Cells were washed and fixed in 2% paraformaldehyde for 1 h at 4°C, followed by secondary labelling with Alexa 647 anti-rat antibody (Molecular Probes) for 30 min. FACS data were acquired with FACSCalibur (Becton Dickinson) and analysed by FlowJo (Treestar) according to standard procedures.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank S. Mariathasan for macrophages from knockout mice, F. Carlsson for technical assistance with experiments, L. Ramakrishnan for the ΔRD1 mutant strain, N. Andrews for the LAMP1-RFP retroviral construct and F. Carlsson for critical reading of the manuscript. This work was supported by the National Institutes of Health Grants AI55614 (to E.J.B.) and AI63302 and AI51667 (to J.S.C.).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Movie S1. Time-lapse video of a BMDM transduced with Lamp1-RFP prior to exposure to MmGFP. MmGFP, green; Lamp1-RFP, red. At the beginning of the video, several bacilli have already invaded in the bottom left and right quadrants, and lysosomes are predominantly perinuclear. New bacilli invade the cell in the upper left quadrant, followed by rapid recruitment of lysosomes towards the periphery of the cell. Supporting information for Fig. 1C.

This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/j.1462-5822.2008.01177.x

Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
CMI_1177_sm_Movie_S1.avi8763KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.