Autophagy balances inflammation in innate immunity

ABSTRACT

Macroautophagy/autophagy is a homeostatic process with multiple effects on immunity. One of the pivotal contributions of autophagy in immunity is the cell autonomous control of inflammation. This property leads to systemic consequences and thereby influences the development of innate and adaptive immunity, which promotes or suppresses pathology in various disease contexts. In this review we focus on the intersections between autophagy and inflammasome activation, autophagy and interferons, and autophagy and inflammation in association with infection.

Abbreviations

CD

Crohn disease

FA

Fanconi anemia

GWAS

genome-wide association studies

Mtb

Mycobacterium tuberculosis

PAMPs

pathogen-associated molecular patterns

ROS

reactive oxygen species

SLR

SQSTM1-like receptor

Introduction

As a metabolic, cytoplasmic quality control and general homeostatic process, autophagy is primarily cytoprotective, tissue protective and anti-inflammatory. Its manifestations in adaptive immunity and in sterile and infection-associated inflammation, cytokine- and innate immune cell-processes intended to help clear cell-damaging sterile irritants or invading pathogens, are numerous.^1,2 The general theme of the present review is how autophagy balances and modulates immune activation to avoid excessive inflammation, with a focus on single-cell level intracellular processes that may serve as initial stimuli to trigger or skew broader tissue responses. An emerging concept underlying this theme is the important intricate relationship between endomembrane and organelle homeostasis and inflammatory outputs (Fig. 1) as a key determinant of innate immunity.

Figure 1.

Common themes in autophagic homeostasis of endomembrane damage and proinflammatory signaling. Organelles such as mitochondria (A), lysosomes (B), and phagosomes (C), can be depolarized (ΔΨ_m), perforated, or otherwise damaged. E3 ligases, such as PRKN/PARK2/parkin, SMURF1, etc., can ubiquitinate targets (K48, K63) on damaged organelles. The AAA-ATPase VCP/p97 can either turn over ubiquitin (K48) by recruiting deubiquitinases to permit further steps in lysosomal homeostasis, or unfold targets such as MFNs (mitofusins) on mitochondria to present them to the proteasome for degradation. K48 ubiquitin also presents phagosomal membrane to proteasomes thus contributing to membrane processing or further damage. Galectins can recognize membrane tears and bind to exposed lumenal glycans while in turn binding to receptors such as SLRs (e.g., LGALS8-CALCOCO2/NDP52), which deliver organelles to phagophores or receptor-regulators (e.g., LGALS3-TRIM16 or LGALS8-TRIM16) that function as receptors and in addition assemble and promote ubiqitination and activation of regulatory ATG factors. SLRs, or other types of receptor such as PHB2 (exposed on the inner membrane of ruptured mitochondria) or integral outer membrane receptors on mitochondria bind Atg8-family proteins (e.g., LC3) via their LC3-interacting region motifs. Damaged organelles are either removed (note crescents representing phagophores, i.e. autophagosome precursors) or otherwise repaired, and if not, they activate inflammasomes (and potentially type I IFNs) as described in the text. If bacteria are not removed along with the remnants of their vacuoles/phagosomes (e.g., via LGALS8), linear ubiquitin chains are formed to activate IKBKG/NEMO and NFKB, leading to inflammation. These processes share common principles and contribute to either suppression of pro-inflammatory responses (by repair or removal of offending organelles and bacteria) or, when they are overwhelmed or otherwise fail, elicit inflammatory cytokines as a second line of defense but at a cost due to associated tissue damage. mATG8s, mammalian Atg8-family proteins.

Genetic links between autophagy and inflammatory diseases

The strong connections between inflammatory diseases—which encompass a vast variety of diverse disorders with dysregulated inflammatory responses causing tissue pathology in a wide spectrum of human organs—and alterations in autophagy were initially gleaned from genome-wide association studies (GWAS) examining associations between genetic polymorphisms and predisposition for a range of human diseases.³ Most notably, early studies connected Crohn disease (CD) susceptibility with polymorphisms in genes encoding ATG16L1 and IRGM, whose products interact to control autophagy in coordination with NOD2, a familial risk factor in CD.⁴ Links between CD (or another form of inflammatory bowel disease, ulcerative colitis) and other autophagy-related genetic polymorphisms span a full range of genes that function in different steps of autophagy and which may, in some cases, also have additional functions.¹ These include, for example, ULK1, which acts at an early step in the autophagy regulatory cascade⁵ and selective autophagy factors, such as SMURF1,^6–9 and CALCOCO2/NDP52.¹⁰

Other connections have been reported between several autophagy loci and genetic predispositions for chronic inflammatory disorders and autoimmune diseases. These include systemic lupus erythematosus (IRGM,¹¹ ATG5,¹² PRDM1-ATG5¹³; and DRAM1¹⁴), asthma (ATG5¹⁵), rheumatoid arthritis (albeit somewhat marginally significant ATG5 rs6568431 association¹⁶), Vici syndrome (EPG5¹⁷), celiac disease¹⁸ and other conditions (CLEC16A, ULK3, and MIR's that target autophagy genes),^18–21 multiple sclerosis (CLEC16A²¹), and other neurological disorders, including amyotrophic lateral sclerosis and frontotemporal dementia, especially in patients with stronger neuroinflammatory components (where connections to autophagy have been made through mutations in TBK1²²). CLEC16A is also genetically linked to type 1 diabetes mellitus, a condition associated with immune infiltration, and plays a role in mitophagy and autolysosome function.²³ The latter function of CLEC16A is interesting as it points to stages beyond autophagosome-lysosome fusion. Finally, a recent study has connected the autosomal dominant mutations in the gene (MEFV/PYRIN/TRIM20) responsible for familial Mediterranean fever to the control of selective autophagy.²⁴ Thus, autophagy and inflammation show genetic links in a broad spectrum of human disorders that have inflammatory and/or autoimmune components.

Autophagy and the inflammasome

Inflammasomes, which come in several variants, are cytosolic responders to microbial products and sterile endogenous agonists. Once activated, inflammasomes proteolytically process pivotal pro-inflammatory cytokines including IL1B/IL-1β.²⁵ A canonical inflammasome consists of pro-CASP1 (caspase 1), PYCARD/ASC adaptor, and one of the proteins sensing diverse endogenous agonists (often associated with mitochondrial reactive oxygen species [ROS] or lysosomal damage) or bacterial products such as NLRP1, NLRP3, NLRC4, or cytosolic DNA sensors such as AIM2 and IFI16.²⁵ Once inflammasome components assemble, activated CASP1 processes cytosolic pro-IL1B into mature IL1B ready to be secreted from the cells.²⁵ Noncanonical inflammasomes do not depend on ASC and NLRPs, and instead the agonists, such as cytosolic LPS, directly activate CASP4/CASP11 (mouse) or CASP4 and CASP5 (human) resulting in proteolytic processing of GSDMD (gasdermin D) causing a type of cell death termed pyroptosis. There are crossovers between canonical and noncanonical inflammasomes as CASP1 can also proteolytically activate GSDMD.²⁶

The intersections between autophagy and the inflammasome are numerous and represent some of the earliest and best examples of the anti-inflammatory role of autophagy. The initial proof-of-principle for a genetic linkage between autophagy gene mutation and inflammation, including the role of IL1B, processed by activated inflamamsomes as described above, came from studies modeling Crohn disease in mice lacking functional ATG16L1 in hematopoietic cells,²⁷ and in mice expressing a whole body hypomorphic allele of Atg16l1.²⁸ Mice lacking ATG16L1 in hematopoietic cells show elevated levels of IL1B and IL18, the key pro-inflammatory cytokines triggered by inflammasome activation, and pathology in these mice can be countered by blocking antibodies against IL1B and IL18. Thus, this study,²⁷ provided the first in vivo evidence that an autophagy gene can function to control inflammasome activation. More recently, macrophage-specific deletion of autophagy genes in mice has been shown to lead to inflammasome-mediated IL1B release and uveitis, an inflammation-mediated eye disease often observed in patients with CD.²⁹

Several early reports^30,31 describe modes of indirect suppression of inflammasome activation by autophagy. Because autophagy removes damaged or irreversibly depolarized mitochondria, intact autophagy is necessary to prevent the accumulation of depolarized mitochondria (a process called mitophagy) that release endogenous inflammasome agonists such as ROS and oxidized mitochondrial DNA.^30,31 These earlier findings have been strengthened by recent new lines of evidence linking intact mitophagy to prevention of inflammasome activation. For example, macrophage-specific deletion of the gene encoding the autophagy adaptor SQSTM1/p62 results in accumulation of damaged mitochondria, excessive inflammasome activation-IL1B-dependent inflammation, and macrophage death.³² Similarly, in response to treatment with inflammasome activators, bone marrow-derived macrophages from mice lacking the Fanconi anemia (FA) gene, Fancc (a newly identified mediator of selective autophagy) accumulate damaged mitochondria and have increased mitochondrial ROS-dependent inflammasome activation.³³ Intriguingly, a common naturally-occurring mutation in FANCC associated with a milder clinical phenotype preserves the mitophagy function but not the DNA damage repair function of FANCC. This observation suggests that defects in mitophagy—and consequent aberrant inflammasome activation—may underlie some of the pathology in patients with FA (a congenital disease) or oncogenesis in patients with tumors due to mutations in the FA gene family.

In addition to preventing inflammasome activation by removing damaged mitochondria, autophagy has also recently been shown to play a more direct role in regulation of the inflammasome. Notably, individual inflammasome components are substrates for autophagic degradation, thus representing another modality by which autophagy prevents excessive inflammasome activation. AIM2 is removed by autophagy in a process involving SQSTM1 recruitment to K63-ubiquitinated PYCARD/ASC.³⁴

Several newly described autophagy receptors³⁵ from the family of TRIM proteins play a role in autophagic degradation of a number of individual inflammasome components. For example, MEFV/TRIM20, targets pro-CASP1, NLRP1, and NLRP3 for autophagic degradation.²⁴ MEFV/TRIM20 recognizes these targets via its SPRY domain, has 3 LC3-interacting region motifs allowing it to interact with various members of the mammalian homologs of yeast Atg8 (LC3s and GABARAPs), and assembles locally the core autophagy regulators ULK1, BECN1/Beclin 1 and ATG16L1.²⁴ Moreover, mutations in MEFV/TRIM20 associated with familial Mediterranean fever, diminish interactions between MEFV/TRIM20 and the core regulators of autophagy.²⁴ In addition, TRIM11 acts as a receptor for AIM2; AIM2 binds to the PRY-SPRY domain of TRIM11; and TRIM11 auto-polyubiquitinates and recruits SQSTM1 to eliminate AIM2 via SQSTM1-dependent selective autophagy.³⁶

However, autophagy does not exclusively play a negative role in the regulation of inflammasomes. Despite its general suppressive effects on the inflammasome itself, autophagy components play a positive role in the unconventional secretion of IL1B^37-39 from the cytosol into the extracellular milieu where this cytokine exerts its biological activity. Thus, autophagy, in its engagement with inflammasomes and their substrates, appears to play a balancing act in supporting productive inflammatory responses while simultaneously preventing excessive inflammatory responses and tissue damage.

Autophagy intersects with type I and type II IFN

IFNs (interferons) are immunomodulators secreted by immune and other cells in response to pathogens or tumors. Type I IFN is often associated with induction of protective antiviral states but may be counterproductive in certain bacterial infections at least in part due to inhibition of IL1 production.⁴⁰ Autophagy factors can directly suppress activation of protein complexes that stimulate type I IFN production. Type I IFN activation can occur through several signaling pathways. These come from toll-like receptor (TLR) and other signaling molecules that activate NFKB/NF-κB, including intracellular nucleic acid sensors such as DDX58/RIG-I that interact with mitochondrial-localized MAVS and activate TBK1 kinase, and a second messenger cGAMP (generated by CGAS/cGAMP synthase in response to cytosolic DNA) which binds to the ER adaptor protein TMEM173/STING, which in turn promotes dimerization of IRF3 and its phosphorylation by TBK1. IRF3 (and IRF7) then activates transcription of type I IFN genes.

Early studies⁴¹ indicated that the ATG12–ATG5 complex inhibits DDX58/RIG-I signaling responsible for type I IFN induction. This was due to direct binding of ATG12–ATG5 to the CARD domains of DDX58/RIG-I and MAVS. Additionally, ATG9A negatively controls trafficking of the ER-associated TMEM173/STING and inhibits activation of TBK1. The absence of autophagy may also amplify DDX58/RIG-I like receptor signaling through increased MAVS levels on accumulating mitochondria, and, as is the case with the inflammasome, increases in depolarized mitochondria pools in the absence of mitophagy lead to increased ROS and enhanced DDX58/RIG-I like receptor outputs.⁴² A very recent study indicates that MAVS itself is a direct target for autophagic removal coordinated by the BST2/tetherin-recruited E3 ligase MARCH8 and CALCOCO2/NDP52.⁴³

Other recent studies have shown that individual components of type I IFN activation pathways are targets for degradation by selective autophagy. TRIM21 targets both IKBKB/IKKβ (of the NFKB activation pathway)⁴⁴ and IRF3 dimers (but not its inactive monomers)²⁴ for selective degradation through autophagy. TRIM21, similarly to MEFV/TRIM20, binds the core autophagy machinery components including mammalian Atg8-family members, ULK1, BECN1, ATG16L1, and SQSTM1.²⁴ Finally, CGAS, a cytosolic DNA sensor⁴⁵ that activates type I IFN via the production of cGAMP, which activates TMEM173 and TBK1, itself is a substrate for selective SQSTM1-dependent autophagy.⁴⁶ Interestingly, a TRIM (TRIM14) negatively controls this process. Unlike TRIM11, MEFV/TRIM20 and TRIM21, which promote degradation of the components of the inflammasome or type I IFN activation pathways, TRIM14 recruits a deubiquitinase to CGAS thus protecting it from autophagic degradation. TRIM14 itself is a member of type I interferon-stimulated genes, and thus its increased expression in response to viral infections amplifies the cellular capacity to produce type I IFN when needed.⁴⁶

Type II interferon, IFNG/IFN-γ, a key cytokine associated with protective Th1 responses against intracellular bacteria, also intersects with autophagy. IFNG acts upstream of autophagy by activating systems that modulate autophagy. This regulation can occur through an IFNG stimulated kinase, DAPK1, which in turn phosphorylates and activates BECN1.⁴⁷ However, this can also proceed through immunity-related GTPases. For example, IRGM1, a mouse paralog of human IRGM, is IFNG inducible.⁴⁸ IRGM organizes core autophagy factors including ULK1, BECN1 and ATG16L1, and links them with receptors of innate immunity.⁴⁹ IFNG may also act through increased expression of TRIM subsets. Among the TRIMs potently activated by IFNG is MEFV/TRIM20, which targets inflammasome components for selective autophagy.²⁴ The subsequent degradation can contribute at least in part to the tapering/balancing effects of IFNG on inflammation.⁵⁰ UBQLN (ubiquilin) also seems to work in selective autophagy upon IFNG activation by promoting ubiquitination of targets.⁵¹

Autophagy may reduce triggers of inflammation during bacterial infection

Several organelles in eukaryotic cells, such as mitochondria, may originate from bacterial endosymbionts. Thus, it is not surprising to observe parallels between autophagic control of bacteria and autophagic removal of a range of endomembranous organelles during sterile homeostasis.^9,52–61 These phenomena include cooperation between several quality control systems and autophagy, and the examples thus far include mitochondria, lysosomes and phagosomes. These organelles, when damaged by crystals such as silica or monosodium uric acid,⁶² cholesterol crystals,⁶³ and possibly protein fibrils/amyloid,^64–66 and bacteria or viruses, become dysfunctional and all are capable of triggering inflammasome activation.⁶² This property is not limited to mitochondria alone, because many classical agonists of the inflammasome such as silica, alum, and monosodium urate crystals, are primarily lysosome damaging agents.⁶² Thus, we speculate that by limiting consequences of damage to lysosomes, endosomes, and phagosomes^58,59,61,67 (potentially including pathogen-containing phagosomes) as well as mitochondria, autophagy may reduce triggers of cell-autonomous inflammation.

A parallel between mitophagy and autophagic elimination of bacteria⁶⁸ includes participation of the same ubiquitin E3 ligase PRKN/PARK2/parkin and SQSTM1-like receptors (SLRs) in autophagic defense against Mycobacterium tuberculosis (Mtb)⁵⁷ and in mitophagy.⁵⁵ Recent studies continue to expand these parallels and have revealed a set of multilayered similarities between autophagic elimination of pathogens and removal of damaged organelles, including participation of E3 ligases (which mediate K48 ubiquitination, in addition to K63 ubiqutination, which is more classically associated with autophagy), ubiquitination-deubiquitination cycles possibly contributing to quality control, ubiquitin-directed VCP/p97 unfoldase/segregase action on damaged proteins, and proteasomal degradation of such substrates, all functioning in concert with or in advance of mitophagy^52–57 and autophagic elimination of damaged lysosomes or phagosomes.^{9, 58–61}

K48 ubiquitination can precede or act in parallel during elimination of damaged lysosomes⁶⁰ or mitochondria.^52–54 Both damaged lysosomes⁶⁰ and mitochondria⁵² engage VCP/p97, and, mitochondria (but not lysosomes, where deubiquitinases carry out deubiquitination of K48 linkages⁶⁰) engage the proteasome contributing to progression toward mitophagy.^52–54,56 In mitophagy, these activities expose the inner core of the organelle allowing the inner mitochondrial mitophagy receptor, PHB2 (prohibitin 2), to bind LC3-II and target mitochondria for autophagic degradation.⁵⁶ Similar K48 and proteasome requirements (with the SMURF1 E3 ligase generating K48 ubiquitin linkages) have been recently reported⁹ to act in combination with PARK2-K63-SLR to mediate autophagic targeting of Mtb phagosomes to the lysosome. These relationships are illustrated in Fig. 1.

Autophagy, galectins, and inflammation

In addition to the above autophagy-related processes that control inflammation, galectins (a group of cytosolic lectins) play a role in the detection of endomembrane injury and subsequent triggering of autophagy aimed at protection against sterile or infection-associated phagosomal^61,69–72 and lysosomal^58,59,61,67 membrane damage. Galectins can react to membrane damage and form intracellular puncta in response to lysosomal damaging agents such as polymers of Leu-Leu-OMe/LLOMe that poke membrane holes^67,73 and when cells take up inanimate objects such as latex beads⁶⁹ coated with transfection reagents.⁵⁹ Galectins recognize membrane damage mediated by bacterial secretory system effector proteins in phagosomes/vacuoles in cells infected with Shigella,⁷¹ Listeria,⁷⁰ Legionella,^74,75 Yersinia,⁷⁵ and Salmonella..^59,70,72 An in vitro role for LGALS8 (galectin 8) has been reported in autophagic control of Salmonella-containing vacuoles⁷² and viruses.^{76, 77}

Recent studies have placed LGALS8 at the crossroads of autophagy and inflammation.^{39, 78} First, LGALS8-marked membranes shield escaping bacteria from recruiting linear ubiquitin chain assembly complex/LUBAC, an E3 ligase that generates linear polyubiquitin chains, which, in addition to contributing to autophagic processes, activate IKBKG/NEMO and NFKB that in turn trigger inflammation.⁷⁸ Second, LGALS8 participates in secretion of IL1B in response to lysosomal damage.³⁹ These studies expand the effects of intracellular galectins to triggers and mechanisms of extracellular proinflammatory cytokine activation and delivery.

Another lectin, LGALS3 (galectin 3), plays a role in autophagic control of M. tuberculosis both in infected macrophages and in a mouse model of tuberculosis.⁶¹ LGALS3 controls autophagic responses to endomembrane (e.g. lysosomal and phagosomal) damage in cooperation with TRIM16 leading to the assembly of core ATG factors ATG16L1, ULK1 and BECN1.⁶¹ LGALS3 has been implicated in recognition of Legionella^{74, 75} and Yersinia⁷⁵ vacuolar damage, but connections with autophagy have not been investigated For the majority of galectins (i.e. with the exception of LGALS3⁶¹), their in vivo role remains to be established with respect to their control of autophagic functions in relationship to innate immunity and inflammation.

The complex interplay of autophagy and infection-associated inflammation in animals

All of the above processes are eventually linked to potential activation of proinflammatory cytokines. Thus, the selective autophagic removal of intracellular bacteria and endomembranous organelles is likely important not only for the direct benefits of removing unwanted cargo, but also for decreasing inflammasome and NFKB activation and reducing inflammation in tissues.

The links between defects in autophagy or autophagy genes (either core components of the machinery or targeting factors) and excessive inflammation have become evident from in vivo studies using murine models of Mtb infection. Increased parameters of lung inflammation^9,79,80 have been detected in Mtb-infected mice with defective Atg5 in the myeloid lineage (Atg5^fl/fl Lys2/LysM-Cre mice)^79,80 and whole-body smurf1 knockout mice.⁹ The types of cytokines detected in each case have some similarities, as IL17 is elevated in both Atg5^fl/fl Lys2-Cre⁷⁹ and whole-body smurf1 knockout mice.⁹ Neutrophilic infiltration and elevated IL1A/IL-1α or IL1B are detected in conditional knockout mice with the ATG5 defect in the myeloid lineage (Atg5^fl/fl Lys2-Cre mice)^{79, 80} but not in the study with smurf1 knockout mice, where mononuclear inflammatory cells dominate instead of neutrophils⁹ or in studies with myeloid-specific knockout of other core autophagy machinery components.⁸¹ Neutrophilic infiltration in Mtb-infected Atg5^fl/fl Lys2-Cre mice has been shown⁸¹ to be responsible for the early lethality of such mice,^79,80 and a neutrophil-driven transcriptional signature⁸² is considered to be a hallmark of active tuberculosis in humans. However, not all effects may be attributable to macroautophagy in the studies reported.⁸¹ An in vivo role for LGALS3 in protection against Mtb in a mouse model of tuberculosis has been reported, in keeping with its in vitro role in autophagic killing of Mtb,⁶¹ and this may also include regulation of inflammatory components secondary to phagosomal damage.

Type I interferon is another signature of active disease in Mtb-infected patients⁸² and is considered as a potentially counterproductive host response in bacterial infections (in contrast to its usual protective role in viral infections). However, it may also act as a balancer by suppressing excessive IL1B responses.⁸³ Bacterial DNA released from phagosomes harboring Mtb and Legionella pneumophila can be recognized by CGAS to induce type I interferon, proposed to be damaging or at least counterproductive for the host during bacterial infection.⁸⁴ However, mb21d1/CGAS knockout mice do not survive better as one might expect, but rather succumb to disease during the chronic phase of tuberculosis infection.⁸⁵ Perhaps the duality of the roles for MB21D1/CGAS in inducing type I IFN responses and connecting to the core autophagy apparatus⁸⁶ while sensing bacterial presence and targeting the autophagic response to control intracellular bacteria^84,85 may result in a net protective effect of MB21D1/CGAS against Mtb infection in mice.

While the widely held view is that defective autophagy results in excessive inflammation that contributes to disease pathology, there are some recent reports—where in the context of certain animal models of viral infection—that the “hyperinflammatory state” observed in the mice with myeloid-specific deletion of autophagy genes can be protective. For example, myeloid-specific deletion of several autophagy genes (e.g., Rb1cc1/Fip200, Atg5, Atg7, and Atg14) results in enhanced basal lung inflammation and resistance to influenza infection.⁸⁷ Similarly, myeloid-specific deletion of Rb1cc1/Fip200, Becn1, Atg14, Atg3, Atg5, Atg7, or Atg16l1 results in increased systemic inflammation (IFNG-dependent) in chronic herpesvirus infection, which prevents viral reactivation from latency.⁸⁸ These studies lead to the intriguing hypothesis that the normal homeostatic mechanisms that limit basal inflammation may actually promote certain viral infections. However, it should be cautioned that one cannot extrapolate from the effects of experimental deletion of autophagy genes prior to infection to effects of autophagy pathway inhibition on established infection.

Autophagy may reduce triggers of inflammation during viral infections

During viral infections, autophagy may reduce inflammation through mitophagy and consequent effects on inflammasome activation. For example, in influenza A infection, a NOD2-RIPK2-ULK1 pathway is important for mitophagy to prevent excessive inflammasome activation.⁸⁹ In addition, we speculate that the clearance of pathogen-associated molecular patterns (PAMPs) by selective autophagy of viruses (virophagy) may, in concert with damage-associated molecular patterns generated during infection, prevent inflammasome activation.

One potentially clinically relevant observation is that bone marrow failure (a condition triggered largely by excessive inflammasome activation) in children with FA syndrome is commonly triggered by viral infections; experimentally, in Fancc-deficient mice which are defective in both mitophagy and virophagy, bone marrow failure triggered by inflammasome activators can be prevented by anti-oxidants. Thus, the “perfect storm” for inflammatory diseases may be concurrent defects in mitophagy and pathogen autophagy, leading to cellular accumulation of PAMPs and damage-associated molecular patterns (including mitochondrial ROS), and the resulting inflammasome activation.

Conclusions

There is a growing list of host factors that function dually in autophagic and other responses to endomembrane damage and intracellular pathogens. This encompasses mitophagy and pathogen-induced autophagy (either bacterial- and/or viral-induced autophagy) including: PRKN/PARK2/parkin,^55,57 CALCOCO2/NDP52,^55,90,91 OPTN (optineurin),^55,92,93 TBK1,^{91,92,94–98} and SQSTM1,^{57,80,95,99,100} several of which have been previously reviewed by Randow and Youle¹⁰¹ and are updated here; and SMURF1,^6,9 FANCC,³³ and PEX13.^6,102 Similarly, autophagic and other homeostatic responses to lysosomal damage and damage to phagosomal/vacuolar membranes harboring bacteria or viruses, recognized and initiated by galectins, is an emerging example of organelle damage as a signal, parallel to that of mitochondrial damage. Specific galectins (LGALS3 and LGALS8) and other parts of the autophagic machinery respond to phagosomal,^61,69–72 endosomal, and lysosomal^58,59,61,67 membrane damage, including sterile damage^{58,59,61,67,69} or rupture imposed by bacteria^{59,61,70–72} ,^74,78 or viruses.^76,77

A further important cellular benefit in all of the above processes may be to coordinate the control of not only different damaged host membranes but also microbial PAMPs to prevent excessive inflammation. The roles of autophagy in immune regulation in inflammation independent of the processes discussed above, extend to its intrinsic effects on differentiation of immune cells, their polarization, and the function of immune networks, and include inflammatory M1 versus M2 macrophages, and Th1 versus Treg cells (not covered here).

In summary, autophagy represents an anti-inflammatory mechanism; it protects against endomembrane damage triggered by various agents of endogenous or infectious origin and prevents unnecessary or excessive inflammation. We propose that autophagy supports productive, and prevents unnecessarily over-exuberant, inflammatory responses, thus playing a balancing act intended to avoid excessive tissue damage and ensure a measured response. Defining the molecular and cellular aspects of this concept may allow us to harness this aspect of autophagy for clinical purposes. A better understanding of the connections between autophagy and the immune response may have broad applications as the pathology associated with numerous diseases involves some form of inflammation.

Funding Statement

HHS | NIH | National Institute of Allergy and Infectious Diseases (NIAID) [grant number AI42999], HHS | NIH | National Institute of General Medical Sciences (NIGMS) [grant number GM121176], HHS | NIH | National Institute of Allergy and Infectious Diseases (NIAID) [grant number AI109725], HHS | NIH | National Institute of Allergy and Infectious Diseases (NIAID) [grant number AI111935]

Acknowledgments

We thank numerous colleagues for their contributions to the field of immune and inflammatory manifestations of autophagy and apologize for the limited referencing defined by the scope of the review. Funding sources: NIH R01 AI042999, R01 AI111935, and 1P20GM121176-01 Center of Biomedical Research Excellence for Autophagy, Inflammation and Metabolism (AIM) in Disease to V.D. and NIH U19-AI109725 to B.L.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.