Macrophage Autophagy and Bacterial Infections

1Macroautophagy, hereafter referred to as autophagy, is a lysosomal degradative process that participates in cellular homeostasis by enabling the removal of defective organelles, protein aggregates, or intracellular microorganisms. The process is highly regulated by multiple signaling pathways and orchestrated by more than 30 autophagy-related (Atgs) proteins organized in several functional units. Upon autophagy activation, Atgs, serine/threonine kinase ULK1, and Beclin-1, in association with Atg14 and type III phosphatidylinositol 3-kinase Vps34, promote the formation of a cup-shaped isolation membrane to engulf the cargo. Through concomitant activity of two ubiquitin-like conjugation systems, the covalent linkage of Atg12 with Atg5/Atg16L1 and LC3 lipidation with phosphatidylethanolamime, the isolation membrane elongates into a double-membrane vesicle, called autophagosome. The autophagosome then fuses with lysosomes to form an autolysosome in which the engulfed cargo is degraded. This latter step is mediated by a second Beclin-1 complex, lysosomal-associated membrane protein 1 (LAMP1), and a fusion machinery including SNARE syntaxin-17.

2In addition to its role in cellular homeostasis, autophagy is essential to immunity. The autophagy machinery targets intracellular pathogens for degradation, modulates inflammation, and participates in adaptive immune responses. Here, we review the many functions of autophagy in bacterial infections with a focus on macrophages, the first line of host defenses, and the replicative niche of numerous pathogens.

3Bacteria induce autophagy mainly via their pathogen-associated molecular patterns (PAMPs) and pathogen-induced damage-associated molecular patterns (DAMPs). Cell surface recognition and cytosolic sensing of these molecules result in signaling cascades that promote rapid and localized autophagy machinery recruitment. Autophagy can further be regulated by several transcriptional factors such as NFkappaB and TFEB to promote expression of different autophagy genes and thus prolong autophagy activation. Depending on PAMP/DAMP nature and localization, autophagy can selectively capture bacteria, such event is called xenophagy, damaged organelles, and other signaling platforms activated during the infection. Furthermore, Atgs proteins have non-autophagic functions essential for innate immunity against bacteria (Figure 1).

4In macrophages, xenophagy has mainly been characterized during Mycobacterium tuberculosis infection but the mechanism is quite similar in epithelial cells infected with Salmonella typhimurium. M. tuberculosis through its ESX-1 secretion system damages the phagosomal membrane and access the cytosol. Cytosolic sensor c-GAS recognizes bacterial DNA, which results in ubiquitination of the bacterium or its phagosome by ubiquitin ligases Parkin and Smurf1. Subsequently, ubiquitin chains bind to autophagy adaptors, such as p62 and NDP52, which recruit LC3 to deliver M. tuberculosis into an autophagosome. Damaged phagosome can also be targeted by autophagy via the recognition of host glycan present on the phagosomal lumen by cytosolic lectins of the galectin family. Additionally, in human macrophages, immunity-related GTPase family M protein participates in xenophagy by promoting mitochondrial reactive oxygen species (ROS) production and recruiting autophagy machinery after PAMP exposure.

5Ultimately, autophagosome sends bacteria to lysosome for degradation. In parallel, autophagy can also generate and deliver antimicrobial peptides to bacterial compartment to enhance killing. Several, in vitro, studies have shown that xenophagy reduces intracellular survival of M. tuberculosis, however, its role, in vivo, is unclear. Mice with monocyte-derived cells and neutrophils lacking Atg5 are more susceptible to M. tuberculosis, but not those lacking other Atgs such as Beclin-1 or Atg14, suggesting that autophagy is not involved in controlling the infection. Nonetheless, separate studies suggest that autophagy may play a role in a latter chronic phase of infection and/or that M. tuberculosis may inhibit the process in vivo. As we will discuss below, M. tuberculosis and other pathogens have developed multiple strategies to block autophagy and some of them have been relevant in vivo.

6Aside from xenophagy, a non-canonical autophagy process named LAP is known to play an important role in antibacterial defenses. Upon phagocytosis, particles or pathogens that engage surface receptors, such as toll-like receptors, Fcgamma receptors or Dectin-1, trigger LC3 conjugation directly onto the phagosomal membrane and independently of autophagosome formation. In contrast to canonical autophagy, this process does not require ULK1 or Beclin-1/Atg14 complexes but instead relies on Beclin-1/Rubicon complex and NADPH oxidase-2 (NOX2) activation. Rubicon activates phosphatidylinositol 3-phosphate (PI3P) synthesis, which in turn, stimulates ROS production. PI3P and ROS then promote recruitment of the two ubiquitin-like conjugation systems to trigger LC3 conjugation. Another protein MORN2 has recently been implicated in that pathway, however, its action mechanism is unknown.

7In several instances, LC3 conjugation onto the phagosome enhances the fusion between phagosome and lysosomes. Consequently, macrophages with defects in LAP pathway are less efficient in controlling intracellular growth of various bacteria such as Legionella pneumophila, Staphylococcus aureus, M. tuberculosis, M. bovis BCG, and Listeria monocytogenes. Importantly, in vivo studies have shown that Rubicon-deficient mice have greater bacterial load and are more susceptible to L. monocytogenes infection. Another important function of LAP is to assist antigen presentation via MHC class II molecules, thus to bridge innate to adaptive immunity. Of note, in some specific contexts, LC3 conjugation delays or does not affect at all phagosome maturation suggesting that LC3 alone is not sufficient to boost fusion with lysosomes.

8Autophagy machinery can also regulate phagocytosis indirectly by altering surface expression of phagocytic receptors. Lack of Atg protein Atg7 in macrophages results in upregulation of two class A scavenger receptors, MARCO and MSR1 that facilitate phagocytosis of M. bovis BCG and M. tuberculosis. Autophagy-deficient cells accumulate p62, which promotes dissociation of transcription factor nuclear erythroid-related factor 2 from Keap1 and then its translocation into the nucleus to mediate expression of these receptors. Notably, in another instance, autophagy deficiency can lead to reduced phagocytosis depending on the nature of the bacteria.

9Although inflammation is central to control bacterial infection, excessive inflammatory responses can lead to host tissue injury, and disease progression. Numerous studies have demonstrated the importance of autophagy in inflammation regulation in infectious and non-infectious settings, the interested readers can refer to recent reviews for a comprehensive view on this subject. Here, we will only discuss on the beneficial role of autophagy machinery in modulating inflammation in a context of bacterial infections and macrophages. One key role of autophagy is the down-regulation of inflammasome activation through multiple mechanisms. Pseudomonas aeruginosa infection results in mitochondrial damage that leads to NLRC4 inflammasome. Elimination of damaged mitochondria via autophagy, i.e., mitophagy, limits inflammasome activation both in vitro and in vivo. Similarly, in a P. aeruginosa septic model, atg7fl/fl mice have an enhanced susceptibility to infection with important neutrophil infiltration and severe lung damage. Loss of Atg7 in alveolar macrophages results in upregulation of IL-1beta and pyroptosis. Besides mitophagy, autophagy can also restrain inflammasome activation by capturing inflammasome subunits. Lastly, autophagy can prevent non-canonical caspase-11 inflammasome activation by targeting S. typhimurium and thus limiting LPS release into the cytosol.

10Finally, mechanisms and levels of autophagy dependent greatly on macrophage microenvironment. While IL-4, IL-13, and IL-10 appear to inhibit autophagy, proinflammatory cytokines such as IFNgamma, IL-1beta, and TNFalpha, activate this process. Specifically, IFNgamma-mediated xenophagy requires additionally ubiquilin-1 and guanylate-binding proteins 1 and 7 to promote recruitment of autophagy proteins p62 and Atg4B. On another hand, the presence of vitamin D in serum enhances significantly macrophage autophagy via expression of cathelicidin antimicrobial peptide. Specific T cells can also stimulate autophagic activity in M. tuberculosis-infected human macrophages. Finally, microbiota may influence autophagy response too, as recently, probiotic Bacillus amyloliquefaciens has been shown to upregulate autophagy genes in macrophages, which lead to enhanced Escherichia coli killing.

11Intracellular bacterial pathogens have developed a wide array of tactics to counterbalance macrophage antibacterial defenses and autophagy is no exception. Most of the uncovered strategies are directed against xenophagy and target different steps of the process (Figure 2).

12Cytosolic L. monocytogenes avoids recognition by autophagy machinery by secreting two virulence factors ActA and InlK. These factors promote recruitment of host proteins, actin, and major vault protein, respectively, to form a protective coat that masks the bacteria. Francisella tularensis disguises itself directly by producing a surface polysaccharide, the O-antigen, which prevents cytosolic sensing of the pathogen and thus xenophagy. Importantly, cytosolic O-antigen mutants are killed by Atg5-dependent autophagy inside murine macrophages.

13In macrophages, S. typhimurium prevents autophagy by activating mTOR, a master repressor of autophagy. Two mechanisms seem to be at play, first, the degradation of the energy sensor, Sirt1/AMPK complex, which negatively regulates mTOR resulting in autophagy activation and, second, the recruitment of non-receptor tyrosine kinase focal adhesion kinase (FAK), an activator of Akt/mTOR pathway. SsrB, a response regulator of a two-component system involved in the regulation of SPI2 encoded virulence factors, promotes AMPK down-regulation. Notably, in vivo, macrophages deficient in FAK are more efficient in clearing S. typhimurium infection than their wild-type counterpart.

14Conversely, some pathogens can inhibit signaling pathways that promote autophagy. M. tuberculosis secretes an N-acetyltransferase, Eis, to target JNK-dependent autophagy; however, it does not seem to be sufficient to affect intracellular growth. In addition, M. tuberculosis limits LAP by inhibiting NADPH oxidase recruitment onto its phagosome, which favors pathogen intracellular growth in vitro and in vivo. Interestingly, antioxidant enzymes SodB and SodC of F. tularensis prevent ROS-induced xenophagy and possibly LAP. Further, vacuolar L. pneumophila translocates a sphingosine-1 phosphate lyase into the cytosol to alter sphingosine metabolism implicated in autophagy activation. Lastly, some pathogens can take advantage of negative feedback regulatory circuits present in macrophages to target autophagy. P. aeruginosa infection leads to NLRC4-dependent caspase-1 activation which results in cleavage of TRIF, an important mediator of TLR4-induced autophagy. Importantly, in vivo, preventing TRIF cleavage restores autophagy and bacterial clearance.

15L. pneumophila can also block autophagy directly by secreting two specific proteases. First, Lpg1137 targets ER-mitochondria contact sites to cleave syntaxin-17, a key SNARE protein implicated in autophagosome formation. On another hand, RavZ, a cysteine protease, cleaves phosphatidylethanolamine-conjugated LC3 to produce a permanent unlipidated form of LC3 that is inadequate for the formation of autophagosomes. Another example is L. monocytogenes, which secretes phospholipase PlcA to degrade PI3P, a key lipid produced by Beclin-1 complex, and involved in LC3 lipidation.

16Besides post-translational modification, autophagy can also be modulated at a post-transcriptional level via expression of various small non-coding RNAs (miR). M. tuberculosis thwarts autophagy by inducing expression of miR-33 which down-regulates expression of several autophagy proteins along the pathway (Atg5, Atg12, LC3, and LAMP1). Remarkably, mice with hematopoietic miR-33 deficiency have a superior capacity to control M. tuberculosis infection. Another miR induced by M. tuberculosis, is miR-125a, which targets UVRAG in complex with Beclin-1 to inhibit autophagy and thus promote pathogen intracellular survival inside macrophages. Interestingly, miR-33 and miR-125a can be induced by PRR ligands isolated from M. tuberculosis. Expression of these miR may be a normal negative feedback loop present in host cells to prevent excessive autophagy, which M. tuberculosis exploits to its advantage. Additionally, M. tuberculosis upregulates miR-144* to reduce expression of DRAM2, a recently discovered autophagy protein, that promotes activation of a second Beclin-1 complex involved in autophagosome maturation. Finally, M. tuberculosis can down-regulate miR-17, which targets Mcl-1, an inhibitor of Beclin-1 complex involved in autophagy initiation.

17As mentioned above, M. tuberculosis can block autophagosome maturation via miR-144*, although, the bacterial factor(s) involved in that process remain to be identified. Separate studies have shown that Esat-6, an important M. tuberculosis virulence factor, plays also a role in blocking autophagosome maturation in macrophages as well as in dendritic cells. The action mechanism of Esat-6 is unknown but might be mediated, in part, by miR modulation. Several other bacterial pathogens can impair autophagosome maturation for survival or growth in macrophages such as Chlamydia trachomatis, Yersinia pestis, Y. pseudotuberculosis, but like for M. tuberculosis the underlying molecular mechanisms are unknown. Furthermore, it is unclear whether these mechanisms applied to both xenophagy and LAP.

18While some pathogens neutralize autophagy to replicate intracellularly, others exploit autophagy or some autophagy proteins to acquire nutrients and remodel their vacuoles. Y. pseudotuberculosis inhibits autophagosome maturation; however, it relies on autophagosome formation for its growth in macrophages. An early study has shown that exogenous autophagy activation promotes the development of Coxiella-replicative vacuole. In macrophages and epithelial cells, C. burnetti resides in an large acidified compartment, which fuses with mature autophagosomes via Cig2, a type IV secretion system effector. Interestingly, in vivo, Cig2 reduces host tolerance to C. burnetti infection without affecting bacterial load.

19Brucella abortus subverts ULK1 and Beclin-1 complexes to remodel its ER-derived vacuole into a compartment with autophagic properties. Nonetheless, this conversion is independent of the two ubiquitin-like conjugation systems, Atg5-12 and LC3. This selective manipulation of autophagy proteins is important for Brucella lifecycle and cell-to-cell spreading. Similarly, M. marinum, a species closely related to M. tuberculosis, uses autophagy machinery to promote bacterial egress from its natural host, Dictyostelium, a macrophage-like unicellular organism. How Brucella and M. marinum manipulate autophagy proteins to favor transmission is still unknown.

20Some pathogens have a more complex relationship with autophagy in a sense that they can evade, as outlined above, or exploit autophagy depending on the infection stages or the type of autophagy. At early stage, vacuolar Legionella interacts with autophagy pathway and seems to rely on autophagosome for survival in permissive macrophages. Similarly, cytosolic F. tularensis triggers an Atg5-independent autophagy pathway to acquire nutrients and replicate while avoiding xenophagy. The molecular differences between these two autophagy-related pathways, one that favors bacterial growth and the other promotes killing, definitely call for further investigation. Finally, immediately after entering macrophages, L. monocytogenes manipulates LAP to form spacious non-acidified Listeria-containing phagosomes (SLAPs) which are believed to participate in persistent infection. Low expression of virulent factor, Listeriolysin O, appears to play a role in SLAP formation whereas high-expression triggers bacterial escape into the cytosol.

21One example is Vibrio parahaemolyticus, which limits NLRC4 inflammasome-mediated IL-1beta production by activating macrophage autophagy via a type III secretion effector, VopQ. Other bacteria seem to benefit from the autophagy machinery in a more passive way. For example, Atg16L1-deficient mice clear more efficiently uropathogenic E. coli (UPEC) and Citrobacter rodentium than wild-type controls. The improved protection against C. rodentium is associated with an enhanced immune response dependent on monocytes; however, it does not rely on a cell-intrinsic role of Atg16L1 in myeloid cells. In contrast, loss of Atg16L1 in macrophages is responsible for the phenotype observed during UPEC infection. Atg16L1-deficient macrophages, infected with UPEC, activate more NLRP3/Caspase-1 inflammasome, and IL-1beta production. Importantly, in vivo, IL-1beta neutralization reduces the capacity of Atg16L1-deficient mice to control UPEC infection. Testing other autophagy proteins will be essential to confirm the role of autophagy entire pathway in that process.

22In the past few years, a number of studies have highlighted the potential of targeting autophagy for the control of bacterial infections. FDA-approved drugs with proautophagy activity, such as statin, gefitinib, carbamazepine, and metformin, have been shown to limit M. tuberculosis growth in mice model of infection. Similarly, rapamycin enhances clearance of P. aeruginosa and B. cepacia, in vivo, in addition to reducing lung inflammation. However, these drugs can module several other cellular functions, so it is unclear whether their action, in vivo, is mediated by autophagy. Another difficulty is the seemingly opposing effects of a same drug depending on the infection settings. Indeed, while rapamycin induces M. tuberculosis killing in macrophages, it promotes bacterial survival in endothelial cells and in HIV-coinfected macrophages. The only compound tested, in vivo that specifically activates autophagy is a small cell permeant peptide, TAT-Beclin-1. This designed peptide has been shown to improve outcome of chikungunya and West Nile virus infections in mice, thus it would be a good candidate to test in the context of bacterial infection. However, one has to bear in mind that some pathogens use autophagy to thrive in host cells and other can either evade or exploit this pathway according to the infection stage. Therefore, a detailed knowledge of the role(s) of autophagy and its molecular mechanisms for each bacterial pathogen is mandatory to develop more specific autophagy modulators, as well as, the availability of relevant preclinical models to evaluate the efficacy of these compounds.

23Manipulation of autophagy in antigen-presenting cells including macrophages holds also great promise in vaccine development. Autophagy plays an important role in antigen processing and MHC presentation. Studies already indicate that boosting autophagy can enhance antigen presentation and vaccine efficacy. For instance, mice immunized with BCG-infected dendritic cells treated with mTOR inhibitor, rapamycin, have greater Th1-mediated protection after being challenged with M. tuberculosis. Recently, BCGΔureC::hly, a live vaccine with improved immunogenicity and in clinical trial phase II, has been shown to enhance xenophagy. However, the importance, in vivo, of phagocyte autophagy in vaccine enhanced-efficacy is unclear as both cited strategies trigger other cellular pathways.

24Macrophage autophagy is central to host defenses against bacterial infections, sending intracellular pathogens to lysosomes for degradation while controlling inflammation to limit host damages. Since novel functions for autophagy proteins have emerged in various physiological and pathological situations, it is likely that further contributions of autophagy machinery to macrophage biology will be unveiled in the context of bacterial infection. On the other hand, pathogens have found multiple subterfuges to manipulate this machinery in order to persist or proliferate, still, their action mechanisms and significance in vivo are far from being thoroughly understood. A more comprehensive and integrated view of bacteria-autophagy interplay will definitely help in designing more specific HDT and vaccine based on autophagy modulation.

25IV and AB reviewed the relevant literature and wrote the manuscript. IV prepared and AB revised the figures. IV and AB have read and approved the final manuscript.

26The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.