The central finding of this study was that PLC-expressing Mycobacterium tuberculosis is more virulent than Mtb lacking these enzymes, through inducing necrosis of alveolar macrophages, which is associated to subversion of PGE2 production. This is the first study to demonstrate such a role for mycobacterial PLCs using clinical isolates, which actually cause tuberculosis, instead of models of recombinant expression of these enzymes in non-pathogenic mycobacteria.
We showed that PLC-expressing Mtb (isolate 97-1505) induced high rates of alveolar macrophage death, especially through necrosis, whereas the PLC-deficient Mtb (isolate 97-1200), despite its ability to cause cell death, did not induce necrosis as efficiently. Control of macrophage cell-death pathways by Mtb has been extensively described as a strategy to avoid innate and adaptive immune responses [12, 26, 27]. Manipulation of cell-death modality has been successfully used by other intracellular pathogens such as Chlamydia, Legionella pneumophila, Listeria monocytogenes, Shigella flexineri, and Salmonella enterica subsp. enterica serovar Typhimurium [28–30]. It has been demonstrated that host-cell apoptosis confers protection to the host, once the uptake of apoptotic bodies derived from macrophages by dendritic cells allows an effective activation of the immune response . In contrast, host-cell necrosis can benefit the pathogen because disruption of the cell membrane releases the bacteria to efficiently spread and infect adjacent cells . Recently, descriptions of the manipulation of cell-death fate by Mtb have shown that a virulent bacillus, the H37Rv strain, caused macrophage necrosis whereas the attenuated strain H37Ra was related to apoptotic death . Likewise, a Ndk- (nucleoside diphosphate kinase) knockout Mtb showed reduced virulence, which was demonstrated by the susceptibility to macrophage microbicidal activity and increased ability to induce host-cell apoptosis .
Pulmonary macrophages are the primary niches for Mtb replication, thus host resistance is critically dependent on innate immune functions played by these cells. In this scenario, proinflammatory cytokines and nitric oxide (NO) are essential for host control of Mtb. Macrophage recognition and phagocytosis of Mtb stimulates mostly the production of TNF-α, IL-1α and β, and IL-6, which are fundamental for the resolution of Mtb infection in mice . Our results highlighted the proinflammatory response triggered by 97-1505 Mtb isolate, which induced a higher production of those cytokines by alveolar macrophages than the isolate 97-1200. Surprisingly, the higher production of proinflammatory cytokines did not result in better outcome for the host cell, as shown by the decreased macrophage survival. Stimulation of NO generation can cause oxidative stress leading to dysfunction in mitochondrial respiration and also block caspase-3 activity by nitrosylation, which may inhibit apoptosis and thereby promote necrosis . Beyond the effects on the immune response, TNF-α has been associated with necrosis in a caspase-independent mechanism through activation of receptor TNFR1 and engagement of RIP1 kinase . Recently, it was suggested that alveolar macrophages infected by an attenuated BCG (Bacillus Calmette–Guérin) show high expression of the TNF-α-receptor TNFR1 associated with increased cell apoptosis . However, in that particular study, only apoptosis rate was analysed and necrosis was not shown. In addition, host-cell necrosis induced by the T3SS pore-forming protein, YopB, from pathogenic Yersinia has been associated with increased production of proinflammatory cytokines, such as IL-1β and TNF-α . These findings support our data showing that proinflammatory cytokines are involved in cell death induced by intracellular bacteria.
Activation of the MAPK pathway has been directly linked to cytokines production in proinflammatory cell responses to bacterial stimulus , including Mtb . In addition, MAP kinases have an essential role in production of lipid mediators, such as LTB4, since activation of 5-LO is dependent on phosphorylation mediated by ERK1/2 and p38 . In this study, higher phosphorylation of MAPK p38, ERK1/2, and JNK1/2 was observed in cells infected with 97-1505. Although phosphorylation of ERK1/2 and p38 can also be triggered by mammalian PLCs, as demonstrated by LPS activation of the PLC–PKC pathway , we observed no differences in PLC-γ phosphorylation induced by the Mtb isolates 97-1200 or 97-1505 when compared to uninfected cells. Moreover, different mycobacterial PLC isoforms can trigger MAPK signalling by directly activating PKC through DAG production from cell membrane phospholipids [7, 39]. Based on these findings, we hypothesise that the differential activation of the MAPK pathway in 97-1505-Mtb-infected alveolar macrophages may be due to mycobacterial PLC actions.
Macrophages infected by mycobacteria increase the production of LTB4 itself , which mediates host immunopathology by enhancing Th1 responses and by exacerbating inflammation [16, 40]. LTB4 production induced by both isolates in this study was considerably amplified by PLCs; however, no significant differences were observed at the early stages of infection, which suggests that, besides PLCs, other mechanisms such as the overproduction of proinflammatory cytokines can contribute to immunopathology of Mtb infection. The emergent knowledge that the balance in LTB4 production is fundamental for the outcome of Mtb infection points out that the excessive production of this lipid mediator, associated to dysregulated production of TNF-α, increases Mtb susceptibility in the zebrafish model, demonstrated by necrosis of infected macrophages . We also found a lower production of PGE2 to be associated with decreased mRNA expression of COX-2 and EP-2/4 receptors in Mtb 97-1505-infected alveolar macrophages. Our group previously demonstrated that pharmacological inhibition of COX-2 results in increase of LTB4 synthesis, during Mtb infection in mice . In the present study, we show that addition of exogenous LTB4 to the culture impairs PGE2 production by infected cells. These data are in accordance with the concept of a shift in lipid mediator production toward one eicosanoid subpathway , which may explain the higher LTB4 and lower PGE2 production observed here. Moreover, the finding that down-regulation of PGE2 and higher necrosis were both impaired after incubation of the isolate 97-1505 with PLC inhibitors, supports the hypothesis that virulent mycobacterium subverts eicosanoid synthesis to manipulate host-cell death to promote proliferation and dissemination . Here, when exogenous PGE2 was added to 97-1505-infected alveolar macrophages, the necrosis rate decreased. On the other hand inhibition of PGE2 by celecoxib enhanced necrosis in cells infected by both isolates. It has been reported that PGE2-preventing necrosis is due to PGE2 involvement in the synthesis of the lysossomal Ca2+ sensor SYT7, which is essential for prevention of mitochondrial damage, enabling repair of plasma membrane disruption . Although virulent mycobacteria sabotage of PGE2 to induce necrosis has been associated with increased production of LXA4[12, 13, 41], we did not detect LXA4 in the supernatant of Mtb-infected alveolar macrophages (data not shown). Nevertheless, the potential relationship between mycobacterial PLCs and host-cell necrosis through down-regulation of PGE2 production shown in this study is new evidence of the relevance of this virulence factor.
Indeed, despite the described plc gene polymorphism , there is no genome or proteome characterised for either Mtb isolate, and further studies are necessary to better understand the differences between 97-1505 and 97-1200, and the role of PLC in Mtb virulence. However, our data make a valuable demonstration of subversion of lipid mediator synthesis and its association with cell necrosis. Furthermore, our data are consistent with the recent finding of Bakala N’Goma and colleagues , who showed for the first time the cytotoxic effect of mycobacterial PLCs on macrophages. Finally, the relevance of PLCs as determinants of virulence in Mtb expands our understanding of how these virulence factors can act to the detriment of the host, and highlights eicosanoids, such as PGE2 and LTB4, as mediators with functions that extend beyond innate immune mechanisms.