‘Nodophagy’

Abstract

Autophagy is a homeostatic pathway that processes and recycles damaged organelles and other cytoplasmic contents. While studies have implicated autophagy in the immune response to infection, the understanding of how the autophagic machinery specifically targets intracellular pathogens has remained elusive. Two recent studies have uncovered an autophagy-mediated immune response to bacteria through their detection by Nod receptors. In particular, Nod1 and Nod2 recruit the autophagic protein ATG16L1 to the plasma membrane at the bacterial entry site to promote an autophagy-dependent elimination of bacteria. In addition, Nod2 and ATG16L1 synergize to initiate an adaptive immune response to bacterial invasion by enhancing major histocompatibility complex (MHC) class II antigen presentation. These findings link two Crohn disease-associated susceptibility genes and reveal that cells expressing the risk-associated variants of ATG16L1 are defective in autophagy-mediated bacterial handling and antigen presentation. This could lead to bacterial persistence and contribute to the pathogenesis of the disease.

Introduction

Many chronic inflammatory disorders, while multi-factorial, share common broad etiological influences, including environmental, genetic, microbial and host immune factors. Recent international collaborative efforts between clinicians and scientists, coupled with emerging sophisticated technologies and data analysis methodologies, have given rise to significant advances in our understanding of mammalian physiology and the pathobiology of such diseases. The idiopathic inflammatory bowel diseases are examples of such chronic debilitating conditions and are intriguing models of chronic inflammatory disease, because they are generally manifest at the very interface between host and environment, the intestinal mucosa.¹

Clinical Aspects of Inflammatory Bowel Diseases

The idiopathic inflammatory bowel diseases (IBD) are a group of chronic debilitating disorders affecting the intestine of children and adults. While ulcerative colitis (UC) is a continuous mucosal inflammatory disorder of the colon, Crohn disease (CD) may affect any segment of the intestine and is generally notable for more transmural, discontinuous ulceration with granuloma formation. UC and CD are considered as subtypes that lie at different ends of the IBD spectrum. Certain characteristics are used to differentiate UC from CD in clinical practice, although there remains considerable overlap of current treatment regimens between these conditions.² Although the precise causes of IBD remain elusive, a widely accepted disease model proposes that in genetically susceptible individuals and under microbial and environmental influences, an aberrant immune response arises and/or persists as chronic intestinal inflammation.³ Several aspects of intestinal mucosal dysfunction have been implicated in the etiology of IBD, including loss of mucosal integrity, intolerance toward the intestinal microbiota, abnormal immune sensing, signaling, differentiation and processing of either self and non-self antigens.³ Gene polymorphisms resulting in permissive innate immune responses to invasive micro-organisms and/or their motifs facilitate microbial replication and loss of the functional integrity of the epithelial barrier. Subsequent pathological access to sub-epithelial tissues may lead to local chronic inflammatory activation, microbial dissemination and systemic inflammatory responses. The parallel inability of myeloid cells to respond thoroughly to this microbial insult, again owing to these same polymorphisms, further predisposes the host to chronic, albeit low grade, infection and inflammation. The last decade has seen a wealth of robust genetic studies of IBD linking heretofore unappreciated pathways with IBD pathogenesis (Table 1).^4–8 The genetic convergence of both autophagy and pattern recognition receptor function on CD susceptibility is one example of an intriguing opportunity to study whether these seemingly disparate pathways also share functional links that contribute to CD pathogenesis. Their clinical implication may be more far- reaching, as the roles of both Nod2 and autophagy continue to expand beyond IBD to include graft-versus-host disease, sepsis and malignancies to name but a few.^9–13

Table 1.

Key immune pathways implicated in IBD pathogenesis from genetic studies

Pathway	Gene/Allele
Pattern recognition receptors
Nod-like receptor family	NOD2, NLRP3, NOD1*
Toll-like receptor family	TLR4, TLR9
IL-12/IL-23-Th17 pathway	IL23R, IL12B
JAK-STAT pathway	AK2, STAT3
Autophagy	ATG16L1, IRGM, LRRK2, ATG5*
Major histo-compatibility complex	HLA-DRB1, HLA-B21, HLA-DR8, HLA-DR10
Tumor necrosis factor pathway	TNFSF15, TNFRSF6B, TRAIP, TNFAIP3, TNFR
Protein tyrosine phosphatases	PTPN2, PTPN22
IL-10 Family	IL10, 1L19, IL10RA, IL10RB
IFN	IFNG, IL26
C-C chemokine ligands and receptors	CCR6, CCL2, CCL7, CCL11
Phagosome function	NCF4
Immunoglobulin superfamily	CTLA4

^*tentative candidate.

Genetic Discoveries Guide Pathogenesis Research

Genome-wide association studies (GWAS) have identified several susceptibility genes for IBD due to single nucleotide polymorphisms (SNPs) in their sequence. In particular, NOD2 was the first gene associated with susceptibility to CD.¹⁴ Among the three main SNPs identified in NOD2 (SNP8, 12, 13), the L1007fs frameshift mutation (SNP13) caused by a cytosine insertion (3020insC) displays the strongest association with CD (Fig. 1).^14–16 In addition to CD, three missense mutations (R334W, R334Q and L469F) in the nucleotide-binding domain (NBD) of Nod2 confer susceptibility to another granulomatous disorder known as Blau syndrome (BS) (Fig. 1).¹⁷ Nod1 and Nod2 are intracellular sensors of bacterial peptidoglycan that belong to a large family of germ line-encoded pattern-recognition receptors (PRRs), similar to the Toll-like receptors (TLRs) and retinoic acid-inducible gene I (RIG-I)-like helicase receptors (RLRs).¹⁸ Upon recognition of their cognate ligands, PRRs trigger downstream pathways that lead to an inflammatory response with secretion of cytokines, chemokines and antimicrobial peptides that culminate with the removal of the insult. Nod1 and Nod2 sense γ-D-glutamyl-meso-diaminopimelic acid (iE-DAP) and muramyl dipeptide (MDP), respectively and recruit the serine/threonine kinase RIP2 to mount an inflammatory response largely dependent on NFκB and a Th2 driven adaptive response.^19,20 Although Nod receptors are mostly known as sensors of peptidoglycan-derived moieties from bacterial sources, a recent study also involves Nod2 in the antiviral response by triggering activation of interferon regulatory factor-3 (IRF3) and production of interferonβ (IFN) through direct sensing of ssRNA.²¹ In addition, it has been reported that viral infection is one of many environmental factors that contribute to gut inflammation in CD.²² Thus, the strong association of NOD2 gene polymorphisms with CD suggests a crucial role of this receptor in coordinating the innate and adaptive antimicrobial defense in CD pathogenesis.

Figure 1.

Schematic diagram of Nod2 structure. Nod2 contains two N-terminal caspase-recruitment domains (CARD1 and 2) followed by a central nucleotide-binding domain (NBD), flanked on its C-terminus by a leucine-rich repeat (LRR) domain. The P-loop containing the ATP-binding motif is located in the NBD domain. Numbers correspond to amino acid residues. The three main Crohn disease (CD)-associated SNPs (SNP8, 12 and 13) in C-terminal and the Blau syndrome (BS)-associated mutations (R334W, R334Q and L469F) in the NBD domain are indicated by arrows.

The era of GWAS has more recently heralded the dawn of autophagy as another fundamental element in IBD pathogenesis. Polymorphisms in the autophagy genes ATG16L1 and IRGM are among the most strongly associated and reproducible novel CD susceptibility gene variants.^4,6,23,24 The SNP rs2241880 resulting in substitution T300A at the amino acid level in the ATG16L1 gene was identified as an important risk factor for CD.^4,5 ATG16L1 is one of two mammalian homologs of the yeast autophagy protein Atg16 that was first identified in Saccharomyces cerevisiae.²⁵ ATG16L1 contains an N-terminal ATG5 binding domain, a coil-coiled domain that mediates homomultimerization and a C-terminal region with seven WD repeats of unknown function that is not present in yeast Atg16.²⁶ ATG16L1 forms a multimeric complex with the Atg12-Atg5 conjugate and was shown to be essential for autophagy as it specifies the site for LC3 conjugation during autophagosomes formation.²⁶ Besides its well-documented role in cell homeostasis, autophagy was identified as a host defense mechanism against intracellular pathogens.^27–29 Indeed, autophagy participates in the direct elimination of microbes through the fusion of microbe-containing autophagosomes with degradative lysosomal compartments. Recent studies have also underscored the contribution of autophagy to adaptive immunity by increasing MHC class II antigen presentation of cytosolic and nuclear antigens.^30–33

Despite the contribution of Nod-mediated bacterial recognition and autophagy to both innate and adaptive immune responses to infection, it is not clear how bacterial detection by Nod receptors is linked to their role in adaptive immunity and how bacteria, as opposed to self antigens, are specifically targeted by the autophagic machinery. The strong association of NOD2 and ATG16L1 polymorphisms with CD and their combined increased risk for the disease suggest an intricate interplay of these two factors in the biology of CD.²³ Here, we provide new insights into the functional and clinical implications of NOD2 and ATG16L1 polymorphisms on the pathogenesis of CD. We also review the significant findings of two recent studies by Cooney et al. and our group that link Nod2 and ATG16L1 in the autophagy-mediated response to bacterial invasion.^34,35

NOD2 Polymorphisms in Crohn Disease

Landmark genetic linkage studies in 2001 identified that NOD2 gene polymorphisms markedly increase susceptibility to CD, with homozygosity for the major risk alleles conferring an up to four-fold increased risk of CD.^6,14,15,36 This strong genetic association positioned anti-microbial innate immune defense as a pillar of CD disease pathogenesis. The key susceptibility frameshift mutation in NOD2 results in a truncated protein which no longer efficiently senses its ligand, the microbial motif, MDP.³⁷ Patients harboring NOD2 variants have dysregulated immune responses and often have an ileal, inflammatory, fibrosing phenotype of CD.³⁸

How NOD2 variants contribute to CD has been a question of intense scrutiny in recent years that remains unanswered, notwithstanding the myriad of data generated from both human and murine translational studies. In vitro activation of Nod2 leads to transcription of multiple genes, often mediated through NFκB activation and mitogen-activated protein kinase signaling, which results in the production of pro-inflammatory mediators and antimicrobial molecules.³⁹ These in turn set in motion the recruitment of immune cells to effect the immune response. Intestinal NOD2 expression is most intense in the Paneth cells of the small intestinal crypts, but also in myeloid, epithelial and myofibroblast cells; whether the Nod2fs protein shares this expression profile is unclear.⁴⁰ NOD2 knockout mice (NOD2^-/-) are unresponsive to MDP, secrete less antimicrobial peptides and are more susceptible to certain bacterial infections via the oro-gastric route.⁴¹ Wild-type NOD2 has been linked with anti-inflammatory activity that may curb TLR-driven inflammation—loss of such NFκB regulation in CD could in turn perpetuate the inflammatory response.⁴² Some have found it difficult to reconcile the observations that while patients with active CD have elevated levels of circulating proinflammatory cytokines, such as IL-1β, IL-6 and TNFα, myeloid cells of those with the frameshift mutation display diminished cytokine responses to MDP stimulation in vitro.^16,43,44 Monocytes from CD patients with the Nod2fs were also recently shown to have reduced production of the anti-inflammatory IL-10 in response to bacterial stimulation and loss of TLR-NOD2 synergy; ectopic expression of the Nod2fs protein was shown to interfere with transcription of the IL10 promoter, a proposed mechanism to account for the reduced IL-10 secretion.⁴⁵ A somewhat contrasting hypothesis of Nod2 dysfunction in CD has arisen from the ‘knock-out, knock-in’ Nod2fs murine model, in which a mutation homologous to the human NOD2fs mutation was introduced.⁴⁶ In contrast to human cells bearing the NOD2fs mutation, NFκB activation and IL-1β secretion were enhanced following MDP treatment of knock-in macrophages and dextran sodium sulphate (DSS) treatment of these mice resulted in more severe inflammation, ulceration and increased colonic tissue levels of IL-6 and IL-1β in comparison to their wild-type counterparts. Ultimately, whether NOD2 polymorphisms result in an absolute loss or gain of function or in a manifestation of mixed functional outcomes remains to be reconciled.

ATG16L1 Polymorphism in Crohn Disease

Genetic analysis of over 5 × 10⁵ single nucleotide polymorphisms (SNPs) from thousands of Caucasian patients and controls, with confirmatory replication in multiple independent cohorts, have led to the identification of up to 100 IBD susceptibility genes.^6,47 Polymorphisms in the autophagy genes ATG16L1 and IRGM are among the most strongly associated and reproducible novel CD susceptibility gene variants.^4,6,23,24 The non-synonymous SNP in ATG16L1 is the causal risk variant in this gene and results in an amino acid substitution at position 300, from threonine (300T; protective) to alanine (300A; risk). Functional studies have recently advanced our understanding of the ATG16L1 polymorphism in the context of both its cellular biology and its translation from bench to bedside. Gene silencing and knock-down experiments have underlined the crucial role of ATG16L1 in autophagy-mediated clearance of Salmonella typhimurium.^5,48 Recently, adherent invasive E. coli (AIEC), which are highly present in the mucosa of CD patients,^49–52 were shown to induce autophagy in epithelial cell lines in an ATG16L1-dependent manner.⁵³ Interestingly, reconstitution experiments performed in those studies showed that the CD-associated ATG16L1 300A variant could not fully restore the ATG16L1 activity in autophagy-dependent anti-bacterial experiments.

Exciting in-vivo data from two different ATG16 murine models gave further credence to autophagy as a key pathway in intestinal health and in IBD pathogenesis. Saitoh et al. described aberrant IL-1β and IL-18 secretion in macrophages expressing a truncated form of ATG16L1, following stimulation with LPS and inoculation with non-invasive Gram-negative, commensal-type bacteria.⁵⁴ Chimeric mice with ATG16L1-deficient haematopoietic cells had increased susceptibility to DSS-induced colitis. In an alternative murine model, mice with either hypomorphic expression of ATG16L1 protein (ATG16L1^HM) or reduced intestinal epithelial ATG5 or ATG7 expression harbored Paneth cell-specific abnormalities including degenerating mitochondria, loss of lysozyme granule integrity and absence of apical microvilli.^55,56 In a key clinical correlation, related Paneth cell abnormalities were described in patients with CD who were homozygous for the ATG16L1 risk vs. protective allele. Transcriptional profiling analysis also revealed enriched transcripts for the adipocytokines, leptin and adiponectin, similar to those reported previously in patients with CD.^57,58 More recently, Cadwell et al. reported that Paneth cell abnormalities and gut inflammation observed in ATG16L1-deficient mice were dependent on persistent infection of these mice with norovirus found in conventional barrier living conditions but also on the presence of commensal bacteria in the gut. Notably, they showed that DSS-induced colitis in ATG16L1^HM mice pre-infected with the persistent strain of norovirus (MNV CR6) exhibited multiple hallmarks of human Crohn disease.²² The above findings underscore the importance of autophagy pathway integrity to normal Paneth cell function and secretion. Indeed, two recent studies have suggested a possible link between the autophagy pathway and the unconventional protein secretion pathway in yeast species.^59,60 In both model systems, gene silencing of indispensible autophagy genes was associated with impaired secretion of specific peptides and suggested that evasion of vacuole (lysosome) fusion by trafficking autophagosomes facilitated transport of specific proteins to the plasma membrane. Paneth cell dysfunction and aberrant cytokine responses were also previously described in IBD patients harbouring NOD2 polymorphisms.^40,61–65 Not only have the autophagy and NLR pathways converged on Paneth cell dysfunction, recent reports have presented a new perspective on how these pathways are fundamentally linked and implicated in the biology of CD. Two independent studies by Cooney et al. and our group have shown that engagement of Nod2 with its ligand, MDP, is capable of inducing autophagy in a manner dependent on NOD2 and ATG16L1 genotype.^34,35

Nod1 and Nod2 Control Autophagy through the Recruitment of ATG16L1

In recent years, several studies have investigated a more specific role of the autophagy-mediated defense against pathogens through the recognition of PAMPs by PRRs. Stimulation of a number of TLRs (such as TLR1, TLR3, TLR4, TLR5, TLR6 and TLR7) with their cognate ligands were previously shown to activate autophagy as a defense mechanism capable of directly eliminating intracellular pathogens in certain situations.^66–69 It has also been reported that TLR4 physically interacts with the autophagy molecule Beclin-1 to promote LPS-induced autophagosomes formation.⁶⁹ In contrast, until recently, no studies had clearly demonstrated a link between autophagy and cytosolic PRRs such as RLRs and NLRs although many intracellular microbes induced autophagy.^27,29,70 In our in vitro experiments we show that stimulation of myeloid and non-myeloid cells with bacterial PG fragments (Nod ligands) induced an upregulation in the number of autophagosomes as evidenced by the presence of LC3 puncta and LC3I to LC3II conversion. Similarly, peritoneal injections of Nod agonists (FK565 for Nod1 and MDP for Nod2) in mice resulted in an increase in autophagosomes formation in peritoneal macrophages isolated from those mice. We further confirmed a role of this Nod-dependent induction of autophagy in the control of bacterial proliferation using Shigella flexneri and Listeria monocytogenes as models. Indeed, in Nod1- and Nod2-deficient cells we observed a decrease of bacteria-containing LC3⁺ compartments, which correlated with an increase in bacterial burden. Similarly, a recent study shows that Nod2 stimulation enhanced the autophagy-mediated killing of Salmonella through a functional interaction between Nod2 and ATG16L1.⁷¹ Another significant observation in our study was that a fraction of ATG16L1 is present at the plasma membrane in physiological conditions. ATG16L1 protein was previously shown to localize to isolation membranes in a multimeric complex with the Atg12-Atg5 conjugate and specify the site for LC3 conjugation during autophagosome formation.²⁶ The importance of ATG16L1 in the localization of the autophagic machinery was further underscored in another study showing accumulation of LC3 at the plasma membrane following ectopic expression of ATG16L1 at this site.⁷² On the other hand, although described as cytosolic receptors, Nod1 and Nod2 were also shown to localize to the plasma membrane.^73–75 In an effort to understand the mechanism underlying the induction of autophagy by Nod proteins we investigated a potential interaction or colocalization of Nod receptors and ATG16L1 at the cell surface. Using coimmunoprecipitation experiments and immunofluorescence we demonstrated that the Nod proteins could directly interact with ATG16L1 and that the presence of ATG16L1 at the plasma membrane is dependent on Nod receptors. Notably, we showed that during infection, Nod proteins recruit ATG16L1 to the plasma membrane at the site of bacterial entry suggesting that Nod-mediated recognition of bacteria at the entry foci is an important step in the formation of bacterial-targeted autophagosomes (Fig. 2). As mentioned above, NOD2 and ATG16L1 are two of the most important genetic susceptibility factors associated with CD. Of significance, we showed that the most prevalent NOD2 mutation in patients with CD (L1007fsinsC-Nod2) resulted in a truncated Nod2 protein that not only failed to localize to the membrane as previously reported,^74,76 but also retained ATG16L1 in the cytosol resulting in impaired wrapping of invading bacteria by autophagosomes (Fig. 2). In contrast, while cells from donors homozygous for the CD risk-associated ATG16L1 allele (ATG16L1*300A) showed impaired induction of autophagy when stimulated with MDP or Gram-positive peptidoglycan, no defect was observed in the interaction and the colocalization of the ATG16L1 T300A risk variant with Nod receptors at the membrane. This result suggests that the CD-associated polymorphism of ATG16L1 might affect the Nod-mediated autophagic response at a later stage in the process of autophagy.

Figure 2.

In fibroblasts, Nod2 senses bacteria at the plasma membrane and recruits the autophagic machinery through direct interaction with ATG16L1. This leads to the autophagy-dependent elimination of the bacteria after fusion of bacteria-containing autophagosomes with degradative lysosomal compartments. The CD-associated variant Nod2 1007fs cannot detect bacteria at the plasma membrane resulting in a defect in the recruitment of ATG16L1 and the autophagic response. While the CD-associated variant ATG16L1*300A can be recruited at the membrane by Nod2, it further negatively affects the autophagic pathway. In parallel, Nod2 activation induces an inflammatory response to bacterial invasion through the classical RIP2/NFκB pathway that is dispensable for Nod2 autophagy at the membrane.

The most classical feature of the innate immune response to microbes is the detection of PAMPs by PRRs followed by the activation of signaling pathways that ultimately lead to an inflammatory response.⁷⁷ However, the signaling pathways leading to autophagy activation downstream of PRRs remain poorly understood. TLR stimulation results in the activation of NFγB and AP-1 by the MyD88-dependent pathways and/or type I interferons (IFNs) by TRIF-dependent pathways.¹⁸ In the context of autophagy, a recent study has shown that TLR-dependent induction of autophagy may involve TRIF- and MyD88-dependent recruitment of Beclin 1, a key inducer of autophagosome formation.⁶⁹ Upon activation, Nod1 and Nod2 are known to recruit adaptor protein RIP2 through homotypic CARD-CARD interactions. RIP2 in turn, interacts with the IKK subunit IKKγ (also called NEMO), promoting its ubiquitylation and subsequent degradation of IγBα, that lead to NFγB activation (Fig. 2).⁷⁸ Since RIP2 was shown to be crucial for downstream signaling of Nod receptors,^41,79 we sought to examine its role in the induction of autophagy. Using RIP2- and NEMO-deficient mouse embryonic fibroblasts (MEFs), we showed no contribution of these factors in stress- or bacterial-mediated autophagy. This suggests that Nod receptors can bypass the downstream signaling by directly recruiting the autophagy machinery to the plasma membrane through ATG16L1, therefore allowing the cell to deal with the infection in what seems to be a fast innate immune response (Fig. 2).

Nod2 and ATG16L1 Synergize in Autophagy-Mediated Adaptive Immunity

While we showed a synergy between Nod receptors and ATG16L1 in the recognition and wrapping of bacteria at an early time of the autophagy process, another complementary study further extend the interplay between Nod2 and ATG16L1 to the initiation of an adaptive immune response.³⁵

Antigens presented on MHC class II are typically exogenous proteins that are phagocytosed by the antigen-presenting cell (APC) or endogenous proteins that reside in the secretory system. However some cytosolic self or non-self antigens were also shown to be presented by MHC class II.^80,81 Although autophagy was long thought to be a possible mechanism involved in this process,⁸² recent studies have really confirmed its role in the delivery of cytosolic and nuclear antigens onto MHC class II molecules.^30–33 However, how intracellular microbes, as opposed to self antigens, are specifically recognized and targeted to the autophagy machinery for MHC II presentation remains elusive. The recent study by Cooney et al. gives new insights into how intracellular bacteria are specifically directed to the autophagic machinery through Nod2 recognition for antigen processing and presentation by MHC class II molecules. They show that stimulation of dendritic cells (DCs) with the Nod2 ligand MDP increased the level of MHC class II surface expression as a result of redistribution from within the cell. This increase in HLA-DR surface expression was due to a relocalization of LC3 to HLA-DM-containing compartments. This suggests that Nod2 stimulation not only induces the formation of autophagosomes but also direct them for fusion with specific MHC class II compartments (Fig. 3). To further confirm the role of Nod2 in autophagy-mediated MHC class II antigen presentation they performed CD4⁺ T-cell proliferation assays following infection with Salmonella typhimurium and showed that Nod2- and ATG16L1-deficient DCs exhibited reduced antigen presentation capacity. In line with our observations, they show that DCs from individuals with the CD-expressing variants NOD2 (L1007fsinsC-Nod2) and ATG16L1 (ATG16L1 T300A) were defective in MDP-induced autophagy but also in MHC class II processing and in CD4⁺T cell antigen specific responses (Fig. 3). In addition, Nod2 variant-expressing DCs infected with Salmonella or an adherent-invasive E. coli derived from the mucosa of individuals with CD, showed reduced colocalization of bacteria with lysosomes which is correlated with an increase in bacterial proliferation. Similar results were obtained in our Shigella infection experiments using Nod1-and Nod2-deficient cells or BMDMs isolated from mice homozygous for the Nod2fs mutation (L980fs), which corresponds to the L1007fsinsC mutation in humans.³⁴

Figure 3.

In dendritic cells, sensing of bacteria by Nod2 induces the engulfment of bacteria into autophagosomes that ultimately fuse with MHC class II containing compartments leading to MHC class II presentation of bacterial antigens to CD4⁺ T cells. In addition, the RIP2-dependent activation of NFκB and MAPKs by Nod2, negatively and/or positively control Nod2-mediated autophagy in those cells. The CD-expressing variants Nod2 1007fs and ATG16L1*300A negatively affect Nod2 autophagy resulting in a defect in MHC class II presentation.

Overall, this study by Cooney et al. supports our results showing the crucial role of the interplay between Nod2 and ATG16L1 in autophagy-mediated clearance of intracellular bacteria by lysosomal degradation. In addition, they bring new insights into the contribution of the Nod2-mediated autophagy to the adaptive immune response to infection through the MHC class II antigen presentation by DCs (Fig. 3).

Nod-Mediated Autophagy and Downstream Signaling

In contrast to our study, Cooney et al. support that Nod2-mediated autophagy is dependent on the downstream factor RIP2 as evidenced by a decrease in autophagosomes formation in RIP2-deficient DCs. This discrepancy could be explained by the type of cells used in both studies. Using epithelial cells (HeLa) and fibroblasts (MEFs) as our main working models, we provide a mechanism were bacteria are directly recognized at the plasma membrane by Nod receptors leading to the recruitment of ATG16L1 in a RIP2-independent manner (Fig. 2). We propose that in such cells, the initiation of autophagy at the membrane early after infection is critical for the complete autophagic response to occur. In a direct correlation with intestinal epithelial cells that are involved in the innate defense against gut infection, the ability to bypass the RIP2/NFκB pathway would be crucial for a rapid autophagy-dependent elimination of bacteria, as the absence of Nod at the membrane, but not the deficiency in RIP2, results in a defect in bacterial autophagy. Further studies using intestinal cells involved in innate host defense such as Paneth cells would be valuable in regard to this model. Interestingly, the fact that NF.B activation was previously shown to inhibit TNFα-induced autophagy further support our RIP2/NFκB-independent model of Nod-mediated autophagy.⁸³ In professional APCs such as DCs, autophagy plays an important role in adaptive immunity through MHC class II antigen presentation. Therefore, we can assume that the recruitment of the autophagic machinery to the cell surface is not a prerequisite for bacterial autophagy in DCs. The Nod-dependent engulfment of bacteria in autophagosomes would more likely take place in the cytosol at later times of infection under the control of RIP2-dependent signaling as reported by Cooney et al. (Fig. 3). Interestingly, Cooney et al. show that Nod2-mediated autophagy is affected by SB203580, an inhibitor of the MAPK p38, which was recently also shown to also inhibit RIP2.⁸⁴ Although SB203580 could be acting directly on RIP2 and inhibits Nod2-dependent autophagy thereby supporting the data obtained in RIP2-deficient DCs, this result also suggests a possible involvement of p38 as an inducer of Nod2 autophagy downstream of RIP2. The inflammatory response induced by Nod receptors is largely dependent on RIP2/NFκB but other studies have demonstrated that Nod1 and Nod2 can also activate the MAPKs p38, ERK and JNK in a RIP2-dependent manner.^85–87 While NFκB is a negative regulator of autophagy,⁸³ MAPKs have varying effects. ERK and p38 activation affect the maturation step of autophagy, playing a positive and a negative role, respectively.⁸⁸ Activation of p38 inhibits starvation-induced autophagy through a direct competition with mammalian Atg9 for binding to a p38-interacting protein (p38IP).⁸⁹ In contrast, other studies have reported a direct or indirect role of p38 as a positive regulator of autophagy. ER stress induces cell death and autophagy through p38 in human gingival cells.⁹⁰ p38 can also activate NADPH oxidase to generate reactive oxygen species (ROS),⁹¹ the latter being known as autophagy inducers.^83,92 Another MAPK, JNK, has a clearly positive role in activating autophagy by phosphorylating Bcl-2 thus releasing Beclin 1 from inhibitory complexes with Bcl-2.⁹³ In the light of those studies, we propose that the Nod-mediated autophagic response will most likely depend on the nature, strength, timing and duration of the pathways activated, to the cell type involved and/or other regulatory factors. In this regard, a recent study shows a differential effect of the E3 ligase ITCH on Nod2-dependent signaling pathways through polyubiquitination of RIP2. While ITCH is required for optimal NOD2:RIP2-induced p38 and JNK activation, it inhibits NOD2:RIP2-induced NFκB activation.⁹⁴ Interestingly, ITCH localizes in endocytic compartments and is thought to affect Nod2 signaling after MDP uptake by macrophages.⁹⁵ This is one example of how a factor, depending on its localization and the timing of the stimulus, can orientate the Nod response towards a specific pathway that can ultimately lead to autophagy activation. The fact that autophagy is also known to be regulated by several ubiquitin ligases⁹⁶ opens some new areas that could be investigated to help deciphering the signaling events that link Nod receptors to autophagy.

Conclusion

The two primary studies highlighted above provide the first glimpses of a heretofore unappreciated dimension of our immune defence—namely the Nod-mediated autophagic axis—and how both RIP2-dependent and -independent mechanisms can be differentially used by cells to mount responses to a microbial insult. The significant challenge for patients and their clinicians is that their IBD remains incurable and current medical treatments are generally limited to modulating inflammation and promoting mucosal healing. Novel therapeutic strategies are likely to flow from a better understanding of the immune, microbial, environmental and genetic elements that contribute to IBD.