Genetic control of autophagy underlies pathogenesis of inflammatory bowel disease

Abstract

Autophagy contributes to cellular homeostasis in the face of nutrient deprivation and other cellular stresses. Cell type-specific functions for autophagy are critical in maintaining homeostasis at both the tissue level and at the whole-organism level. Recent work has highlighted the ways in which human genetic variants modulate autophagy to alter epithelial and immune responses in inflammatory bowel disease.

Regulation of autophagy

Autophagy is an evolutionarily conserved stress response pathway that directs cargoes to the lysosome, where they are degraded for subsequent recycling¹. Autophagy can be non-selective or selective; in the former case, autophagy functions as a bulk degradation pathway to help recycle nutrients and balance biosynthesis pathways. In contrast, non-selective autophagy involves the precise targeting of unwanted intracellular components such as protein aggregates, damaged organelles, intracellular pathogens, and lipid droplets. Autophagy is often considered a pro-survival mechanism with many benefits; however, prolonged, uncontrolled autophagy can lead to a type of cell death termed autosis². Therefore, the induction of autophagy and sustained autophagy must be regulated through both transcriptional and post-translational mechanisms^3–8. Emerging data have highlighted the multi-step regulation of post-translational modifications in controlling autophagy, including ubiquitination, phosphorylation, and acetylation⁹. These post-translational modifications are also critical in the induction of selective autophagy as well as cargo selection.

More than thirty core autophagy proteins mediate the autophagic cascade, which controls the pre-initiation, initiation, and elongation of the autophagosome^{10, 11}. Autophagy is first triggered by a pre-initiation complex composed of ULK1/2, ATG13, and FIP200, which together activate an initiation complex consisting of ATG14, Beclin 1, VPS34, and VPS15 to generate the isolation membrane of a nascent autophagosome. Next, two ubiquitin-like ATG conjugation systems (ATG12 and LC3/ATG8) control elongation of the isolation membrane into a double membrane-bound autophagosome containing cargo for degradation. Once elongation is complete, SNARES such as syntaxin 17 mediate the fusion of autophagosomes and lysosomes, where lysosomal enzymes degrade the autophagosomal cargo^12–14. Until recently, it was widely accepted that genetic deletion or loss of any of the core ATG genes abolishes autophagosome formation and autophagic function. However, Tsuboyama et al. recently demonstrated using live imaging that the ATG conjugation machinery is not essential for autolysosome formation¹⁵. Additionally, Nguyen et al. recently demonstrated that LC3s and GABARAPs are not required for autophagosome formation¹⁶. These recent findings create a new paradigm for our understanding of the role of individual autophagy proteins in the autophagic cascade and prompt the re-evaluation of published data involving deletion of a single ATG gene.

Tsuboyama et al.’s new insights could also be useful in understanding the contribution of individual ATG proteins to autophagy as well as their roles in non-canonical autophagy and autophagy-independent pathways that are important for disease. Non-canonical autophagy has been shown to be critical in limiting inflammation in response to fungal pathogens and dying cells as well as in the response to viral and parasitic pathogens^17–20, including LC3-associated phagocytosis (LAP). In LAP, specific components of the autophagy machinery conjugate LC3 to phagosomal membranes to aid in the clearance phagocytosed fungal pathogens^{21, 22}. Cargoes degraded by LAP are engulfed in a single phagosomal membrane that requires the protein rubicon but does not require the autophagy pre-initiation complex, thus distinguishing LAP from canonical autophagy^{21, 22}. In mice, defects in LAP lead to an autoimmune-like response to dying cells and uncontrolled inflammation¹⁸. Additionally, LRRK2, a gene associated with Parkinson’s disease, Crohn’s disease, and leprosy has recently been shown to negatively regulate a non-canonical autophagy pathway, suggesting that dysregulation of this pathway may directly influence disease susceptibility ²³.

Recent studies have also highlighted the autophagy-independent roles of ATG genes²⁰. For example, ULK1/2 is able to phosphorylate SEC16A and thus directly regulate ER export^{24, 25}. Additionally, ATG5 plays a unique role in the cell-specific response to Mycobacterium tuberculosis infection, although the underlying mechanism is yet to be defined²⁶. These data highlight how specific stimuli can regulate non-canonical autophagy and, in turn, how key autophagy genes likely have many unrecognized roles in regulating immunity.

An overview of autophagy in the intestine

The intestine must contend with constant stress, including changes in nutrients, microbial interactions, and intestinal injury. The regulation of nutrient absorption, fluid, and ion balance requires tight control of a number of intracellular processes. Nutrient and ion levels fluctuate with nutrient intake, and control of autophagy is one way cells rapidly adjust to these fluctuations. For example, studies have shown that decreased levels of vitamin D receptor or vitamin D can lead to decreased levels of ATG16L1, resulting in impaired autophagy and dysbiosis^{27, 28}. Impaired autophagy in the setting of amino acid starvation can trigger intestinal inflammation through inflammasome activation and production of reactive oxygen species (ROS)²⁹. Therefore, maintaining intestinal homeostasis is largely a function of controlling stress responses so as to avoid deleterious inflammation while limiting pathogenic outcomes. Autophagy is therefore required for intestinal homeostasis, the response to intestinal injury, and repair.

An additional major stress for epithelial cells and immune cells in gut tissue is the need for these cells to constantly adjust to the close juxtaposition of a large load of microbes. A critical feature of intestinal homeostasis is that cells must maintain tolerance to harmless microbes while retaining the capacity to break this tolerance in order to mount an effective immune response to pathogenic microbes. To accomplish this balance, gut resident immune cells are primed for a quick response, resulting in a level of low-grade inflammation at all times. Autophagy functions at multiple steps in this process, including microbial monitoring, maintaining epithelial barrier integrity, direct intracellular pathogen clearance, and regulating inflammation^{30, 31}. In turn, microbes and microbial products can directly alter the intestinal ecosystem and can directly influence autophagy^32–37.

Autophagy is also important in controlling the development, differentiation, and function of immune subsets in the intestine. Numerous studies have carefully dissected the contribution of autophagy pathways to intestinal T cell development and function^{30, 38}. Recent studies have also evaluated the relationship between autophagy and metabolism to understand how metabolic changes influence T cell survival^{39, 40}. It is clear that defects in autophagy preferentially disrupt regulatory T cell development and function in the intestine^{39, 40}. In B cells, autophagy is required for plasma cell function and antibody responses^{41, 42}. In dendritic cells, autophagy controls how antigens are processed and presented for antigen presentation⁴³. In macrophages, autophagy limits inflammasome-mediated cytokine production and controls LAP-mediated elimination of pathogens⁴⁴. Innate lymphocyte cells (ILCs) are groups of innate cells that orchestrate cellular crosstalk in the intestine and other tissues to modulate inflammatory responses⁴⁵. It has been recently demonstrated that autophagy is required for ILC development and function⁴⁶. Together these data highlight the fundamental role for autophagy in balancing immunity and promoting survival of key immune cells during the immune response (Figure 1).

Figure 1. Pleiotropic effects of autophagy in the intestine.

Non-selective or bulk autophagy underlies cellular homeostasis in the intestine. Additionally, perturbations to non-canonical autophagy and cargo-selective autophagy result in cell type-specific alterations. This schematic illustrates how disruptions to autophagy, selective autophagy, or non-canonical autophagy can impact intestinal epithelial cells as well as both innate and adaptive immune cells. Furthermore, it is likely that autophagy controls the differentiation of rare cell types that have yet to be defined.

Selective autophagy is critical in the response to pathogens as well as in cellular homeostasis. Numerous forms of selective autophagy have been identified that target infectious microbes such as bacteria (antibacterial autophagy), viruses (virophagy), and parasites (parasitophagy) as well as damaged or malfunctioning organelles such as lysosomes (lysophagy), endoplasmic reticulum (ERophagy), endosomes, and mitochondria (mitophagy)⁴⁷. Additionally, autophagy-mediated turnover of lipids (lipophagy), aggregated protein (aggrephagy), and ferritin (ferritinophagy) have also been shown to be important in maintaining cellular physiology, and these pathways might have important cell type-specific consequences in the gut^{48, 49}. For example, lysophagy is involved in sensing and clearing membrane damage in order to limit injury^{50, 51}. Receptors for these processes have been recently identified; however, there are likely a multitude of receptors for selective autophagy to create substrate selectivity and many receptors remain to be discovered^{47, 48, 52–54}.

Autophagy in response to epithelial injury

Epithelial cells in the intestine form a physical barrier between the host and the luminal environment. Cellular junction proteins control paracellular permeability, allowing macromolecules such as nutrients to pass but limiting the dissemination of microbes⁵⁵. Autophagy impinges upon many regulatory steps in maintaining barrier function, including cell type-specific functions. Epithelial barrier integrity can be directly controlled by autophagy through the degradation of the tight junction protein claudin-2⁵⁶. In Caenorhabditis elegans, bacterial pore-forming toxin stimulates the transcription factor HLH-30 (TFEB), which induces autophagy, lysosomal, and inflammatory genes to control bacterial infection and maintain host tolerance^{57, 58}. Both junctional integrity and the direct clearance of bacteria through selective autophagy are considered to be the mechanisms underlying this control. Autophagy therefore plays a role in tissue remodeling, immunity, and inflammation.

Autophagy is also critical for the cell type-specific control of specialized secretory intestinal epithelial cells, including Paneth cells and goblet cells, which together enforce the barrier between intestinal epithelial cells and the commensal microbiota. Paneth cells are located at the base of the crypts in the small intestine and contain large secretory granules that secrete large quantities of antimicrobial peptides (AMPs). Goblet cells secrete mucins, which form the mucous layer in the intestine. Defects in autophagy affect the development and function of both of these cell types: for example, defects in autophagy disrupt secretary granule morphology and secretion in both Paneth cells and goblet cells, resulting in changes in the inflammatory microenvironment^59–62. Autophagy mediates polarized secretion of cytokines and lysosomal contents in highly secretory cell types, suggesting that autophagy controls common vulnerabilities in these cell types^{63, 64}. The precise links between degradative autophagy and autophagy-mediated control of secretion are still being defined, although recent reports suggest specialized cargo receptors and SNAREs are involved⁶⁵.

Autophagy and the host response to bacteria

Core autophagy genes in intestinal epithelial cells are central to the control of the dissemination of Salmonella enterica serovar Typhimurium infection in vivo^{66, 67}. Antibacterial autophagy represents an important defense mechanism against the intracellular pathogens S. Typhimurium, Shigella flexneri, and Listeria monocytogenes. To trigger antibacterial autophagy, intracellular bacteria must be sensed and then marked for degradation by post-translational modifications and/or recruitment of adaptor proteins such as NDP52, p62, optineurin, and NBR1⁶⁸. These adaptors bring marked cargo into close proximity with the autophagy machinery through LC3-interacting region (LIR) motifs, which mediate the interaction with ATG8 family proteins. Some intracellular bacteria express type III secretion systems that allow bacteria to escape into the cytosol, triggering membrane damage resulting in the exposure of host glycan “eat me” signals. These signals recruit galectin-8 and TBK1, which in turn can recruit autophagy adapter proteins^{69, 70}. Ubiquitination plays a key role in bacterial targeting and a role has recently been emerging for specific E3 ligases in modulating substrate specificity in the autophagy pathway^{71, 72}. Several E3 ubiquitin ligases including LRSAM1, Parkin, Smurf1, RNF166, and some TRIM proteins have each been shown to ubiquitinate bacteria or bacteria-associated proteins to drive antibacterial autophagy^73–76. Various TRIM proteins have also been implicated in antibacterial autophagy as well as other types of selective autophagy that have a direct effect on inflammation^{74, 77, 78}.

Bacteria have evolved several mechanisms to cope with the selective pressure of host autophagy, underscoring the importance of autophagy in bacterial defense. Typically these mechanisms involve the expression of bacterial proteins that mediate escape from the autophagy machinery or evasion from bacterial targeting^79–81. Some pathogens have even evolved ways to exploit the autophagy machinery for bacterial growth. Bhogaraju et al. recently reported a unique evasion mechanism by Legionella pneumophila to suppress autophagy. In this process, a L. pneumophila effector protein promotes arginine phosphoribosylation of ubiquitin, resulting in a disruption of autophagy as well as other ubiquitin-dependent cellular processes⁸². Taken together, these studies highlight the multi-layered defense network that serves to target bacteria for degradation to protect the host in the event of bacterial evolution and escape.

Cross-regulation of selective autophagy pathways

Although core autophagy proteins are highly conserved, there is much functional diversity in proteins involved in the targeting and recognition of selective autophagy substrates such as damaged mitochondria or pathogens including bacteria and viruses. Mitophagy is a selective form of autophagy that eliminates damaged mitochondria or helps to reduce the number of mitochondria when cellular energy demands are low⁸³. Mitochondria function to boost innate responses to bacteria and viruses through generation of ROS and by serving as key scaffolds for some pattern recognition receptors in immune cells. In the intestine of C. elegans, mitochondrial damage triggers mitophagy through an innate immune pathway; this pathway is neuroprotective, demonstrating that effects of mitophagy are not cell-autonomous⁸⁴. Recent studies have identified a number of proteins that function in mitophagy, antibacterial autophagy, and/or virophagy (autophagic elimination of viruses), suggesting that these processes share signaling components. The E3 ubiquitin ligase Parkin is critical for mitophagy and has recently been shown to be recruited to intracellular M. tuberculosis and S. Typhimurium to promote autophagy^{54, 74, 85}. The E3 ligase SMURF1 and the peroxisomal protein PEX13 are also critical for mitophagy, virophagy, and antibacterial autophagy^{76, 86, 87}. The kinase TBK1 has been also recently been shown to be actively recruited to bacteria and damaged mitochondria; once there, TBK1-dependent phosphorylation recruits downstream adaptors required for antibacterial autophagy and mitophagy^{53, 69}. These data highlight the strong overlap between xenophagy and mitophagy.

Autophagy and cellular stress responses

Recent data have revealed strong connections between autophagy and other stress response pathways such as ROS, antioxidant response, DNA repair, ER stress, and inflammation. The adaptor protein p62 is recruited to defective mitochondria and intracellular bacteria and is required for Parkin-dependent mitophagy and pathogen-specific antibacterial autophagy^{74, 88}. The NRF2-KEAP1 pathway upregulates anti-inflammatory and cytoprotective genes in order to protect from oxidative stress⁸⁹. Recently, p62 has been shown to link autophagy with the NRF2-KEAP1 pathway by mediating selective autophagy of KEAP1, thereby promoting NRF2 activation⁹⁰. This finding suggests a mechanism whereby selective autophagy increases a secondary protective stress response pathway.

A recent study demonstrated that the Fanconi anemia complementation group C (FANCC) protein interacts with Parkin to execute mitophagy, and loss of FANCC increases IL-1β and ROS production⁹¹.This finding is consistent with studies showing that defective mitophagy directly contributes to increased IL-1β secretion⁹². Together these data suggest that impaired selective autophagy in the intestine disrupts direct bacterial clearance as well as mitophagy, with direct implications for inflammation.

Significant attention has also been given to the interplay between ER stress and autophagy in the intestine⁹³. Paneth cell-specific deletion of XBP1, a transcription factor that controls the unfolded protein response, is sufficient to induce Crohn’s disease-like ileitis⁹⁴. Additionally, ER stress and autophagy were shown to function as compensatory pathways in intestinal epithelial cells; loss of both pathways resulted in more severe disease⁹⁴. These data underscore the critical role of stress response pathways in the intestine and highlight the crucial role for autophagy in coordinating these responses.

Autophagy and inflammatory bowel disease

Inflammatory bowel disease (IBD) is a complex disease involving chronic inflammation of the gastrointestinal tract, including Crohn’s disease and ulcerative colitis. Over the last decade, there has been enormous progress in the discovery of numerous genetic variants that increase or decrease the risk of IBD. These genetic variants have provided a roadmap to understanding key cell types and pathways controlling intestinal homeostasis^95–99. The connection between autophagy and IBD was first realized through the identification of a common coding polymorphism in the core autophagy gene ATG16L1 (ATG16L1 T300A) that confers increased risk to Crohn’s disease^{98, 100}. Since this discovery, other genes associated with Crohn’s disease susceptibility, including IRGM, have also been associated with antibacterial autophagy and the clearance of intracellular pathogens, underscoring the importance of this pathway in IBD¹⁰¹.

Multiple studies have investigated the contribution of autophagy to IBD susceptibility and chronic disease¹⁰². Studies of Crohn’s disease patients who express the ATG16L1 T300A polymorphism have demonstrated increases in proinflammatory cytokines in response to specific stimuli¹⁰³. In addition, defects in the size and number of Paneth cell granules are also found in patients harboring ATG16L1 T300A, suggesting that these cell types are highly susceptible to perturbations in autophagy and highlighting the role for autophagy in Paneth cell development and function¹⁰⁴. In patients with quiescent disease, Paneth cells exhibit higher levels of ER stress when the ATG16L1 T300A polymorphism is present, further supporting a role for both ER stress and autophagy defects in disease pathogenesis^{105, 106}. Additionally, recent studies suggest that accumulation of IRE1α in autophagy-deficient mice and in individuals homozygous for the T300A promotes intestinal inflammation¹⁰⁷.

Studies of ATG16L1 T300A mice have shown that this polymorphism confers impaired antibacterial autophagy and defects in Paneth cells and goblet cells similar to those reported in human studies^{60, 61}. Recent studies have also found that microbiome-derived outer membrane vesicles, which are often shed in the intestinal lumen, are inefficient at inducing regulatory T cells to suppress inflammation and promote microbial tolerance in mice and humans that harbor the ATG16L1 T300A polymorphism¹⁰⁸. Finally, ATG16L1 T300A has also been shown to control dendritic cell migration through RAC1 hyperactivation¹⁰⁹. Thus a single nucleotide polymorphism in a core autophagy gene can exert cell type-specific effects as well as alterations to host-microbe dynamics, and these effects can cooperate to exacerbate inflammation and impair antibacterial responses.

The mechanistic basis for disruption of autophagy by ATG16L1 T300A was found to be increased caspase 3-mediated cleavage of ATG16L1 T300A, resulting in decreased protein levels and disruption of selective autophagy^{60, 61}. Additionally, the T300A polymorphism has been shown to disrupt unconventional autophagy mediated by TMEM59¹¹⁰. Calpain-mediated cleavage of ATG5 and beclin 1 has also been shown to be increased in patients with colitis, suggesting that autophagy protein stability in the intestine may underlie the inflammation that contributes to IBD¹¹¹. Finally, recent studies have identified a role for IKKα in the phosphorylation and stabilization of ATG16L1 to prevent caspase-dependent cleavage in response to extracellular stimuli, further highlighting the important of ATG16L1 stabilization during inflammation¹¹².

Additional genes within IBD susceptibility loci have also recently been shown to control autophagy. The MTMR3 IBD risk allele decreases pattern recognition receptor-mediated autophagy, thereby increasing IL-1β secretion and NFκB signaling¹¹³. Moreover, an siRNA screen of IBD-associated genes in epithelial cells found 31 IBD-associated genes that function in antibacterial autophagy (as measured by LC3-bacteria colocalization) or intracellular pathogen defense (as measured by intracellular bacterial replication). The discovery of autophagy-dependent roles for these genes in bacterial defense strengthens the critical role for autophagy in IBD and identifies important new functionally related pathways that likely contribute to pathogenesis. As discussed above, this study identified roles for IBD risk genes PEX13 and SMURF1 in antibacterial autophagy, expanding their function beyond mitophagy⁸⁷. PEX13 may play autophagy-independent roles in pathogen defense, as infection has been shown to induce innate signaling from peroxisomes¹¹⁴. RUFY4 positively regulates autophagy in response to IL-4 indicating the relationship between immune activation and autophagy stimulation¹¹⁵. C1orf106, a protein of unknown function encoded by an IBD risk gene, was shown in large-scale binding datasets to associate with NDP52 as well as TRIM27, an E3 ligase that regulates degradation of NOD2, suggesting a potential mechanism for a role in antibacterial autophagy^{116, 117}. Finally, the IBD risk gene SLC39A8 encodes a solute carrier that transports manganese and zinc into cells as an electroneutral metal/bicarbonate symporter¹¹⁸. The antibacterial autophagy adapter proteins NDP52, p62, optineurin, and NBR1 each contain a zinc finger motif that requires coordination of a zinc ion, and this domain in NDP52 is required for ubiquitin binding and recognition^{69, 119, 120}. It is tempting to speculate that impaired zinc transport associated with polymorphisms in SLC39A8 might alter antibacterial autophagy with direct consequences on microbial composition^{121, 122}. These genes as well as others are summarized in Table 1. These findings underscore the role of autophagy in intestinal homeostasis and highlight the myriad pathways that converge to perturb autophagy.

Table 1.

Genes within IBD risk loci that function in autophagy-related pathways and bacterial defense

Gene	Function in autophagy-related pathways or bacterial defense	References
GPR65	autophagy and lysosomal function	87
SMURF1	mitophagy, virophagy, antibacterial autophagy	86, 87, 132
PEX13	mitophagy, virophagy, antibacterial autophagy	11, 86, 87
RUFY4	autophagy and endosomal trafficking in immune cells	115
C1orf106	interacts with NDP52, function unknown	116
UTS2	UTS2R ligation by UTS2 impairs autophagosome formation	133
TPPP	promotes unconventional autophagy-dependent secretion	134
UBA7	inhibits autophagy through ISGylation	135
SLC22A4	upregulated in response to infection with Shigella in vivo
LITAF	upregulates autophagy	136
ATG16L1	core autophagy gene	137
IRGM	antibacterial autophagy	138, 139

Recent studies also elucidated a role for GPR65, a H⁺-sensing G protein-coupled receptor, in maintaining lysosomal function^{87, 123}. A number of IBD-associated polymorphisms are located within the GPR65 locus, including one polymorphism, I231L, that results in a missense mutation. Loss of GPR65 or expression of GPR65 I231L resulted in increased lysosomal pH, impaired antibacterial autophagy, and accumulation of aberrant lysosomes⁸⁷. Gpr65^−/− mice were more susceptible to the murine intestinal pathogen Citrobacter rodentium, failing to clear the infection and exhibiting more severe epithelial damage and increased inflammatory cell recruitment⁸⁷. Together, these data underscore the role of GPR65 in lysosomal biology and elucidate the lysosome as a critical regulator of immune function and epithelial homeostasis in the intestinal environment.

Expression of GPR65 is largely restricted to the intestine and immune cells, allowing a distinction between patients with lysosomal storage diseases due to mutations in broadly expressed lysosomal genes in contrast to patients with GPR65 polymorphisms that reduce GPR65 function. Hermansky-Pudlak syndrome is a rare disorder resulting from mutations in genes involved in the formation and trafficking of lysosome-related organelles, and a subset of patients with Hermansky-Pudlak syndrome develop Crohn’s disease-like colitis¹²⁴. A subset of patients with Niemann-Pick disease type C1, a lysosomal lipid storage disorder caused by mutations in NPC1, also develop Crohn’s disease-like symptoms¹²⁵. Impaired antibacterial autophagy is also observed with mutations in NPC1¹²⁵. These data suggest that changes in lysosomal function, particularly in immune cells or intestinal epithelial cells, can alter susceptibility to microbes at the host-commensal interface. Additionally, inflammatory signaling can be amplified when endosomes or phagosomes formed after microbial recognition are not properly cleared through the lysosome, providing another mechanism for sustaining a pro-inflammatory feedback loop in the setting of microbial stimuli¹²⁶.

Perspectives

Dysregulation of autophagy contributes to IBD susceptibility¹²⁷. Recent studies on the roles of Zellweger syndrome spectrum disorder gene PEX13 and the Joubert syndrome gene INPP5E in autophagy suggest that autophagy dysregulation may underlie a number of as-yet-unrecognized diseases^{128, 129}. Disruption of autophagy can also directly contribute to uncontrolled inflammation. These data suggest that small molecules that induce autophagy may hold broad therapeutic potential. However, the cross-regulation of autophagy and other cellular stress response pathways make selective targeting of autophagy challenging in vivo. Additionally, more mechanistic insights into the pathways that contribute to disease will be necessary. In the case of patients with the GPR65 I231L polymorphism, a robust autophagy inducer may exacerbate disease and only serve to increase the burden on poorly functional lysosomes. There is an urgent need for better tools to study primary and secondary effects of novel genes and small molecules on autophagy. Recent studies have described new methods to more precisely evaluate autophagic flux and to study the effects of ATG gene deletion in multiple cell types in vivo^{130, 131}. These tools will be helpful in elucidating the autophagy-dependent effects of genes and small molecules in the control of intestinal homeostasis and IBD.