METABOLISM 2017 © The Authors, Target acquired: …stke.sciencemag.org/content/sigtrans/10/468/eaag2298.full.pdf · Target acquired: Selective autophagy in cardiometabolic disease

SC I ENCE S I GNAL ING | R EV I EW

METABOL I SM

1Cardiovascular Division, Department of Medicine, Washington University School ofMedicine, St. Louis, MO 63110, USA. 2Department of Pathology and Immunology,Washington University School of Medicine, St. Louis, MO 63110, USA.*Corresponding author. Email: [email protected]

Evans et al., Sci. Signal. 10, eaag2298 (2017) 28 February 2017

2017 The Authors,

some rights reserved;

exclusive licensee

American Association

for the Advancement

of Science.

Dow

nloade

Target acquired: Selective autophagy incardiometabolic diseaseTrent D. Evans,1 Ismail Sergin,1 Xiangyu Zhang,1 Babak Razani1,2*

The accumulation of damaged or excess proteins and organelles is a defining feature of metabolic disease in nearlyevery tissue. Thus, a central challenge in maintaining metabolic homeostasis is the identification, sequestration, anddegradation of these cellular components, includingprotein aggregates,mitochondria, peroxisomes, inflammasomes,and lipid droplets. A primary route throughwhich this challenge ismet is selective autophagy, the targeting of specificcellular cargo for autophagic compartmentalization and lysosomal degradation. In addition to its roles in degradation,selective autophagy is emerging as an integral component of inflammatory and metabolic signaling cascades. In thisReview, we focus on emerging evidence and key questions about the role of selective autophagy in the cell biologyand pathophysiology of metabolic diseases such as obesity, diabetes, atherosclerosis, and steatohepatitis. Essentialplayers in these processes are the selective autophagy receptors, defined broadly as adapter proteins that both rec-ognize cargo and target it to the autophagosome. Additional domains within these receptors may allow integrationof information about autophagic flux with critical regulators of cellular metabolism and inflammation. Details regard-ing the precise receptors involved, such as p62 and NBR1, and their predominant interacting partners are justbeginning to be defined. Overall, we anticipate that the continued study of selective autophagy will prove to be in-formative in understanding the pathogenesis ofmetabolic diseases and toprovide previously unrecognized therapeu-tic targets.

d fro

on July 23, 2018

http://stke.sciencemag.org/

m
Introduction
Presently, the most widespread and fastest-growing threat to publichealth is the constellation of epidemiologically associated and mecha-nistically intertwined cardiometabolic diseases includingobesity, diabetes,nonalcoholic steatohepatitis, and associated cardiovascular complica-tions of atherosclerosis, myocardial infarction, and heart failure (1). Aconsistent, unifyingmechanism of metabolic dysfunction is an inabilityof cells to appropriately degrade proteins and organelles, resulting intheir accumulation (Fig. 1). The accumulation of these defective or ex-cess cellular components including protein aggregates, mitochondria,lipid droplets, peroxisomes, and inflammasomes not only representsorganelle-intrinsic failure but also perpetuates cellular dysfunctionthrough the buildup of potentially maladaptive signals, including reac-tive oxygen species (ROS), proinflammatory cytokines, and lipid inter-mediates. In turn, the importance of degradation is twofold: first, tolimit overt accumulation of dysfunctional or aging organelles, and sec-ond, to dampen the impact of signals generated by organelles (Fig. 2).Thus, a longstanding area of investigation lies in characterizing thesystems that cells use to identify and degrade dysfunctional cellularcomponents tomaintainmetabolic homeostasis. Two primarymachin-eries serve this goal: the ubiquitin-proteasome system (UPS) and au-tophagy. Although the UPS is pivotal to overall protein turnover andcell signaling, it primarily targets smaller and short-lived protein substratesand becomes overwhelmed in disease scenarios (2, 3). Autophagy canbe subdivided into several subtypes, which have been described else-where (4, 5). In macroautophagy, larger protein or organelle substratesare sequestered by a double-membrane phagophore, which extends toform amature autophagosome. The autophagosome then fuses with anacidic lysosome, where diverse hydrolases mediate degradation to yieldbasic biological substrates. Bulk macroautophagy refers to non-

selective mass degradation of cytosolic components under conditionssuch as starvation with the primary goal of recycling to generate nutri-ents. In contrast, selective autophagy is the coordinated tagging, autoph-agic sequestration, and degradation of specific protein and organellesubstrates (6). This specificity of degradationmakes selective autophagyimportant in limiting and correcting the cellular dysfunctions for whichno other pathway can compensate. In this Review, we explore how se-lective autophagy affects the physiology of cell and organismal metab-olism, giving particular consideration to its roles in modulating thepathogenesis of cardiometabolic disease.

Molecular Basis of Selective AutophagyThemolecular details of downstream events in autophagy after selectiv-ity is established, including phagophore assembly, membrane elonga-tion, and fusion with the lysosome for degradation can be found invarious comprehensive reviews and glossaries (4, 5, 7). The molecularbasis of selective autophagy lies in how specific organelles or proteinsare exclusively tagged and targeted to the autophagicmachinery (Fig. 2).Ubiquitination seems to be the predominant mechanism of tagging, isshared with proteasomal degradation, and features a complex set of E1,E2, and E3 ligases involved in conjugating proteins and organelles. Ubiq-uitin monomers also self-oligimerize into lysine-linked polyubiquitinchains. The particular bound ubiquitin lysine residue determines chainstructure and serves as a point of control for further specificity. Mono-ubiquitination, or polyubiquitination involving residues Lys27 andLys63, targets cargo to the autophagosome (as in macroautophagy) ordirectly to the lysosome (as in microautophagy) (6). In contrast, Lys48

polyubiquitin chains are a potent signal for proteasomal degradation(8). Therefore, the molecular events that dictate ubiquitination are cru-cial for conferring specificity at the initiation of selective autophagy.Next, a small but growing class of adapter molecules known as the se-lective autophagy receptors feature structural domains that recognizecargo, often through a ubiquitin-binding domain, and a means of initi-ating or binding to the autophagic machinery, such as a MAP1 light

1 of 17


CREDIT:A.KITTE

RMAN/SCIENCESIGNALING


chain 3 (LC3)binding domain. Prominent members of this familyinclude p62, neighbor of BRCA1 (NBR1), nuclear dot protein 52(NDP52), and optineurin (OPTN). Through combinatorial arrange-ments of different cargo, ubiquitination, and selective autophagy recep-tors, highly regulated and specific autophagy can be achieved. The exactcontributions of particular ubiquitination events and autophagy recep-tors in targeting organelles under various conditions are just beginningto be unraveled. Additionally, cases are emerging in which non-canonical means of selective cargo recognition (precision autophagy)or autophagic processing (chaperone-mediated or microautophagy)take precedence inmediating selective autophagic degradation. Overall,the impact of selective autophagy onmetabolic homeostasis and diseaseprogression can best be conceptualized at the level of the substrate,mainly specific organelles (Fig. 2). Directly implicating selective autoph-agy in physiological disease processes is experimentally daunting be-cause of the multifaceted roles of selective autophagy receptors in cellsignaling, redundancy, and the persistent lack of understanding of the


molecular processes involved.However, the combined evidence consist-ing of accumulation of dysfunctional organelles in disease states, pathol-ogies observed in selective autophagy-deficient disease models, andbeneficial effects of driving autophagy together make a compelling casefor its importance. We discuss how organelles important to metabolismare specifically processed through selective autophagy, howmodificationof selective autophagy in turn affects their function, and how these pro-cesses contribute to cardiometabolic disease pathology and protection.

AggrephagyProtein inclusions are formed when the cellular ability to degradeaggregation-prone, damaged or misfolded proteins is overwhelmed.Nascent, smaller protein aggregates are particularly cytotoxic due to apreponderance of reactive side chains (9). Thus, an additional adaptivecellular response to aggregate formation is the coalescence of such pro-tein masses into higher-order structures, including inclusion bodies

on July 23, 2018http://stke.sciencem

ag.org/D

ownloaded from

and the aggresome (10). Although classi-cally associated with neurodegenerativediseases, accumulation of protein aggre-gates has also been observed in nearlyevery cardiometabolic disease (Fig. 1). En-ter aggrephagythe degradation of pro-tein inclusions through selective autophagy.Aggrephagy is thought to have a near-monopoly on aggregate degradation dueto their size and complexity, and aggre-gates themselves halt proteasomal function(11). Several well-established moleculardetails (12) are informative to the under-standing of the procession of aggrephagy(Fig. 3) and are either confirmed or likelyto participate in the degradation of specif-ic inclusions found in cardiometabolicdiseases.

Newly formed aggregates of misfoldedor damaged proteins are first recognizedby members of the heat shock proteinfamily. E3 ubiquitin ligases including Cterminus of Hsc70-interacting protein(CHIP) or Parkin then cooperate withheat shock proteins to ubiquitinate cargo(13, 14). Cargo can take on two distincttypes of structures. Aggresomes are loca-lized to themicrotubule-organizing centerand depend on ordered cytoskeletal deliv-ery of substrates. A separate type of formedaggregate is microtubule-independent andreferred to as an aggresome-like induciblestructure or inclusion body. Ubiquitinatedsubstrate can be recognized by histone de-acetylase 6 (HDAC6) for microtubule-mediated delivery (14), or alternatively, aubiquitin-independent pathway involving(Bag3) cargo recognition can also route dam-aged proteins to the aggresome (15, 16).In any case, p62 is essential for aggregateformation (1719), and its ubiquitin-bindingdomain is required for this process and

HepatocyteProtein aggregationMitochondrial dysfunctionLipid accumulationPeroxisomal dysfunction

Atherosclerotic macrophageInflammasome activationProtein aggregationMitochondrial dysfunctionLipid accumulation

CardiomyocyteProtein aggregationMitochondrial dysfunctionLipid accumulation

celliAPP aggregationMitochondrial dysfunctionLipid accumulation

Skeletal myocyteMitochondrial dysfunctionLipid accumulation

AdipocytesLipid droplet accumulationMitochondrial dysfunctionPeroxisomal dysfunction

Fig. 1. Accumulation of dysfunctional organelles in cardiometabolic disease. Organelle accumulation and dys-function are common features of cardiometabolic diseases in nearly every tissue and cell type, including hepatocytes,macrophages, myocytes, cardiomyocytes, pancreatic islet b cells, and adipocytes. Specific players include lipid accumula-tion, mitochondrial dysfunction, protein aggregation, inflammasome activation, and peroxisome dysfunction.

2 of 17


CREDIT:A.KITTE



D

downstream sequelae (19, 20). Finally, p62 cooperates withNBR1 and adownstream adapter protein, ALFY (21), to mediate interaction ofmature aggregates with the autophagic machinery (Fig. 3) (12).

The role of aggrephagy in limiting protein aggregate toxicity is evi-dent in cardiometabolic disease states and even more so in autophagy-deficient disease models. In each case, it is worth considering thataggregate formation is in itself likely a protective response against pro-teotoxicity until clearance can occur. For example, Mallory-Denkbodies (MDBs; also known as Mallory bodies) are aggregates of theintermediate filament keratin enriched with ubiquitin and p62 andare classically found in alcoholic liver disease but have also been re-ported in the hepatocytes of patients with nonalchoholic fatty liver dis-ease (NAFLD) (22). MDB formation correlates with markers ofinflammation in human NAFLD and can be used as a histologicalmarker of disease progression to the more serious steatohepatitis (2325).Ultimately, a chief concern in the management of NAFLD is the even-tual progression to hepatocellular carcinoma. The accumulation of p62and MDB is a fundamental driver in this process (26) and is limited byautophagic removal (27). The specific players involved in aggrephagy of


MDBs remain elusive, although examination of MDB formation andhepatosteatotic phenotypes in the settings of autophagy and/or p62 de-ficiency has given preliminary insight. Broad disruption of autophagyconsistently leads to severe liver pathology (18, 28, 29) and buildup ofp62 inclusion bodies (18, 28), indicating a role for constitutive aggre-phagy in their clearance. Intriguingly, p62 deficiency partially abrogatesthe hepatic pathology induced by autophagy deficiency (18, 30), per-haps because of the many signaling roles of p62 (31). Several investiga-tions have unveiled further complexity in the contribution of p62signaling to the development of hepatocellular carcinoma, in whichp62 is proposed to be tumor-suppressive in liver stellate cells and onco-genic in hepatocytes (32, 33). p62 deficiency alone neither causes overthepatic dysfunction (18, 34) nor modulates removal of MDBs duringrecovery after drug-induced insult (35). Overall, it should be empha-sized that p62 is important for aggregate formation (35), has importantroles in cell signaling, and is cleared by autophagy but does not seem tobe a primary adapter recruiting the autophagy machinery to MDBs.Identification of the receptors specifically involved in MDB clearancewill be useful in understanding NAFLD pathophysiology.


ag.org/ow

nloaded from

AutophagosomeHealthy cell

Ub

Cardiometabolic disease

Inflammasomophagy

Fusion withlysosomes and

degradation

Lipophagy

Lipid droplet

Membranedisruption

Lipotoxicity

Inflammasomeactivation

ProteotoxicityIL-1

InflammasomePeroxisomeMitochondria Protein aggregate

AggrephagyPexophagyMitophagy

Toxic lipidintermediates

ROS

mtDNA

Detection withselective autophagyreceptors

Fig. 2. Selective autophagy degrades dysfunctional or excess organelles. Bottom: In cardiometabolic disease states, dysfunctional and/or excess organelles produce ad-verse signals thatmediate disease pathology. Mitochondria and peroxisomes produce ROS, andmitochondria additionally releasemitochondrial DNA (mtDNA) and incompletelyoxidized lipid intermediates. Activated inflammasomesproducemassive amountsof IL-1b. Lipiddroplets are a relatively safe storage site for neutral lipids; their saturation results inlipotoxicity, ectopic lipid deposition, andmembrane disruption. Protein aggregates are both inherently cytotoxic (proteotoxicity) and can activate inflammasomes as an exampleof pathological intraorganelle cross-talk. Top: Selective autophagy is a primary mode of degradation for each of these types of organelles and serves to both maintain intrinsicorganelle function and limit toxic by-products. Archetypal steps in selective autophagy include tagging of dysfunctional or excess cargo (for example, by ubiquitin), recognition byselective autophagy receptors (for example, p62), delivery to the autophagosome, and fusion with the lysosome for complete degradation.

3 of 17


CREDIT:A.KITTE



http://stke.sciencemD

ownloaded from

A second prominent role for aggrephagy in cardiometabolic dis-ease is in the clearance of the islet amyloid polypeptide (iAPP). iAPPis somewhat distinct from typical aggregate contents in that it self-oligomerizes into cytotoxic b sheets and amyloid fibrils. In type 2diabetes, iAPP accumulation initiates proteotoxicity and membranedisruption, leading to pancreatic b cell dysfunction (36). iAPP is alsofound in distal sites such as the heart, where it can induce oxidativestress and contractile dysfunction (37, 38). Rivera et al. have previouslyshown that iAPP itself induces autophagy-lysosomal dysfunction (39)and have suggested that p62-mediated activation of the transcriptionfactor nuclear respiratory factor2 (Nrf2) is protective in this setting(40). A crucial role for aggrephagy in iAPP homeostasis has beenestablished by a trio of studies showing that iAPP is protectivelysequestered in ubiquitin- and p62-enriched aggregates, which are subse-quently cleared by aggrephagy. Autophagy deficiency greatly exacerbatespancreatic b cell dysfunction and diabetes progression (4042). In addi-tion, Kim et al. have demonstrated that trehalose, an autophagy-inducingdisaccharide, can specifically induce clearanceof iAPPaggregates and res-cue diabetic pathology (41). In the context of atherosclerosis, we andothers have observed p62- and ubiquitin-tagged protein inclusions inmurine and human atherosclerotic lesions (3, 19, 43), and note that thesestructures are associated with plaque instability and acute coronary syn-dromes in human disease (44). The etiology and contents of these aggre-gates are mostly unknown, but candidates include extensively modifiedlow-density lipoprotein (LDL) andApoB (45,46), perhaps a result of failedlysosomal degradation after initial endocytic uptake (47). In contrast toothers findings in the liver (18), we have found that the formation ofp62 aggregates in autophagy-deficient macrophages is broadly protectiveand limits aggregate-induced inflammasome activation and overall ather-osclerotic progression (19). Overall, selective autophagy serves to limit thetoxicity of protein aggregates observed in many cardiometabolic diseases.


Autophagic Degradation of Inflammasomesby InflammasomophagyThe inflammasome is a large, oligomeric protein complex that producesinflammatory cytokines in a two-step process. Classic inflammatorypathways such as Toll-like receptor activation and subsequent nuclearfactor kB (NF-kB) signaling prime transcription and translation of theprecursor forms of interleukin-1b (IL-1b) and IL-18 peptides (proIL-1b and proIL-18). Second, damage-associated molecular patterns,such as cholesterol crystals and ROS, trigger inflammasome formation,which leads to caspase-1 activation and selective cleavage of the IL pro-peptides to active IL-1b and IL-18. These inflammasome products,which are generated primarily by macrophages (4851), act locallyand systemically on target cells to promote cell death, insulin resistance,and vascular dysfunction and are implicated in nearly every human car-diometabolic disease state (Fig. 1) (5255). A crucial aspect of inflam-masome regulationwas first detailed by Shi and colleagues, who showedthat inflammasomes are Lys63-polyubiquitinated on the ASC subunitand targeted to the autophagosome through p62 acting as the primaryselective autophagy receptor (56). This type of selective autophagy isalso activated by inflammatory signaling cascades and serves as a crucialmolecular brake to limit inflammation (56, 57). The modes of inflam-masome regulation by selective autophagy are ripe for investigation.Kimura et al. (58) describe a second, alternativemechanism of inflam-masome clearance through autophagy. The selective autophagy receptortripartite motif-containing 20 [TRIM20; also known as mediterraneanfever (MEFV)] recognizes inflammasome components including Nod-like receptor protein1 (NLRP1), NLRP3, and caspase-1 independentof ubiquitination, and uniquely serves as a platform to spur autophagythrough the recruitment of Beclin-1, ULK1, and ATG8 (58). This mech-anism contrasts with the binding of other autophagy receptors to ubiqui-tin and LC3 and is termed precision autophagy owing to the direct,

on July 23, 2018ag.org/

ubiquitin-independent recognition of spe-cific cargo (59, 60). Particularly fascinatingis the impaired autophagic clearance of in-flammasomes associated with commonmutations in the MEFV gene found in pa-tients with familial Mediterranean fever(58). This relatively common genetic dis-order results in IL-1b hypersecretion (61)and increased atherogenesis (62, 63).We be-lieve these studies collectively constitute suffi-cient evidence to distinguish autophagicclearance of inflammasomes from bulk au-tophagy and propose the term inflammaso-mophagy to refer to this selective process(Fig. 3).

Direct assessment of the role of inflam-masomophagy in commoncardiometabolicdiseases thus far is sparse, but studies in othersettings suggest the relevance of this processto interleukin secretion from macrophagesandprogressionofpathology inGaucherdis-ease (64), burn wound healing (65), andneuroinflammation (66). A broader role forautophagy (and selective autophagy) indam-pening cardiometabolic diseaseassociatedinflammasome activity is, however, wellestablished. In atherosclerosis, lack of autopha-gy in macrophages greatly potentiates disease

InflammasomophagyAggrephagy PexophagyInflammasomes

Precision autophagy

PeroxisomeProtein aggregatesand inclusions

Pex2

NBR1

Pex5

NBR1

ALFY

CHIP

p62

p62

?

Parkin

ULK1

ATG8

Beclin 1 TRIM20

Ubiquitin

LC3-PEInflammasome

Proteinaggregate

Fig. 3. Models of keymolecular eventsmediating pexophagy, aggrephagy, and inflammasomophagy. (A) Aggre-phagy: CHIP and Parkin are the main E3 ubiquitin ligases that target protein aggregates and inclusions. NBR1, p62, andALFY interactwith the autophagosome tomediate aggrephagy. (B) Inflammasomophagy: Inflammasomes are targeted forautophagic destruction through at least two routes. Amechanism termed precision autophagy involves direct recognitionof multiple inflammasome components by TRIM20 (also known as MEFV), which recruits key components of autophagymachinery such as Beclin, ULK1, and ATG8 (top). Alternatively, upon polyubiquitination of the ASC subunit by a yet-to-be-identified E3 ligase, the inflammasome is recognized by p62 for autophagy (bottom). (C) Pexophagy: Mammalian pexo-phagy primarily proceeds through Pex2-mediated ubiquitination of Pex5. NBR1 serves as the main selective autophagyreceptor for peroxisomes that mediates interaction with the autophagosome.

4 of 17




ag.org/D

ownloaded from

progression and is associated with increased IL-1b in the aorta and inperitoneal macrophages treated with cholesterol crystals (19, 67). Further,the effects of autophagy deficiency on inflammation are inflammasome-specific because other proinflammatory cytokines including tumor ne-crosis factor-a (TNFa) and IL-10 are unaffected by autophagy deficiency(67). In a parallel setting, IL-1b is increased in adipose tissuemacrophagesfrom obese autophagy-insufficient mice and in cultured macrophagestreated with palmitic acid (68). Third, macrophage autophagy deficiencyincreases susceptibility to liver fibrosis in amanner specifically dependenton increased IL-1b production (69). We have provided evidence thatdirectly implicates inflammasomophagy in cardiometabolic diseases(19). First, NLRP3 inflammasomes in macrophages treated with athero-sclerotic stimuli are ubiquitinated and tagged with p62. Deficiency ofp62, as predicted, results in inflammasome accumulation and increasedIL-1b secretion, which is likely and partially due to failure to clear in-flammasomes. p62 is a multifunctional scaffold protein that interactswith many metabolic and inflammatory cell signaling pathways. In thiscase, we have shown that inhibition of Nrf2, extracellular-regulatedkinase (ERK), p38, or NF-kB signaling does not affect the increased IL-1b secretion from p62-deficient macrophages. However, specific loss ofthe p62 ubiquitin-binding domain is sufficient to mimic total p62 defi-ciencys effects on IL-1b secretion, suggesting a primary importance ofthe roles of p62 in aggregating and clearing ubiquitin-tagged targets.Overall, the lack of the selective autophagy receptor p62 in macrophagesallows accumulation of inflammasomes and consequently increasesIL-1b secretion and atherogenesis in mice.

Complementary possibilities beyond direct clearance of inflamma-somes may partially explain why selective autophagy limits IL-1b pro-duction. First,mitochondrial ROS are a potent inflammasome-activatingdanger signal such that clearance of mitochondria (mitophagy) and as-sociated reductions in ROS may explain lower IL-1b (57, 70, 71). Con-versely, it should be noted that the interaction between mitochondriaand inflammasome activation is not unidirectional. Inflammasome acti-vation damages mitochondria, halts mitophagy, and initiates cell deathin a process termed pyroptosis (72). Zhong et al. have shown that acti-vation of the NF-kBinflammasome signaling is limited by concurrentNF-kB transcription of p62, which acts to induce mitophagy and limitinflammasome activation. This finding supports the concept that selec-tive autophagy acts as a brake on proinflammatory signal transduction.Additionally, proIL-1b is degraded through autophagy, potentiallylimiting the availability of this precursor substrate (73). However, weshould note that there appears to be a complex role for autophagy intargeting IL-1b itself. IL-1b lacks a signal sequence for classical secretionand is instead released through an alternative vesicular route (74). Au-tophagy has also been proposed to facilitate IL-1b trafficking throughthis mechanism, potentially leading to enhanced IL-1b secretion undercertain conditions (75, 76). The in vivo relevance and underlying mech-anisms require additional study, but a potential route involves heatshock protein-90 (Hsp90)mediated translocation of IL-1b into auto-phagic vesicle intermediates before secretion (76). Other nonautophagicmechanisms of release such as microvesicle shedding or overt disrup-tion of the plasmamembrane during cell death (77) are likely more rel-evant such that the net impact of autophagy deficiency on IL-1b is stillhyperproduction and secretion in most autophagy-deficient settings. Al-though reduced autophagy and increased inflammasome activity are in-dependently linked to the development of cardiometabolic disease,specific genetic or pharmacological manipulations are needed to pre-cisely determine the ways in which inflammasomophagy regulates IL-1b and influences downstream physiological sequelae.


PexophagyPeroxisomes aremultifunctional organelleswith underappreciated rolesin lipid metabolism and ROS signaling (78). Diverse classes of signalinglipids are degraded or synthesized at least in part by peroxisomes, in-cluding bile acids (79), ether lipids (80), and leukotrienes (81). Many ofthese compounds are ligands for nuclear hormone receptors that acti-vate transcription to profoundly alter cellular metabolism. Peroxisomesalso have major roles in ROS metabolism. Many peroxisomal enzymesproduce ROS as normalmetabolic by-products, which have homeostat-ic signaling roles, and the high peroxisomal abundance of the antioxi-dant enzyme catalase balances their impact. Peroxisomal ROS cross-talkwithmitochondrial and cytosolic ROS signaling (82) and are also thoughtto generally be in excess in cardiometabolic diseases (83). Peroxisome vol-ume is responsive to environmental stimuli such as increased fatty acidbioavailability, and damaging ROS production may be a consequenceof peroxisome biogenesis (83, 84). Aging, perhaps the most importantrisk factor for cardiometabolic diseases, also nearly universally featuresan accumulation of dysfunctional peroxisomes (85). Clearly, tight reg-ulation of peroxisome homeostasis has major implications for the de-velopment of cardiometabolic disease.

Pexophagy is the autophagic degradation of peroxisomes and is ex-clusively responsible for their clearance (86, 87). The most promisingmammalian ubiquitination target that facilitates pexophagy is the im-port receptor Pex5 (88). Zhang et al. (89) have identified a regulatorycascade inwhichROS stimulate the kinase ataxia telangiectasiamutated(ATM) to phosphorylate Pex5, which increases its ubiquitination. Intandem, ROS signal to ATM to inhibit mammalian target of rapamycincomplex 1 (mTORC1) at the peroxisome to broadly stimulate autoph-agy andprovide the autophagymachinery needed for pexophagy (89, 90).Nordgren et al. (91) propose a different monoubiquitination site onPex5 in mediating pexophagy but integrate this function with theprimary function of Pex5 in importing peroxisomal matrix proteins.They suggest that the dysfunctional matrix import that is observed indisease states (92, 93) stabilizes ubiquitinated Pex5 at the peroxisomalmembrane to favor pexophagy, consistent withmodels favored by com-putational analysis (94). E3 ubiquitin ligases present at the peroxisomethat could target Pex5 include Pex2, Pex10, and Pex12, and their actionmay be opposed by ubiquitin-specific protease 9x (USP9x)mediateddeubiquitination (95). Sargent et al. specifically propose Pex2 as the pre-dominant ubiquitin ligase for pexophagy during starvation and con-firmed Pex5 as a key target (Fig. 3) (96). Finally, what are the primaryselective autophagy adapter proteins that mediate pexophagy? Deosaranet al. (97) have established that NBR1, but not p62, is sufficient and nec-essary for mammalian pexophagy through its binding of ubiquitinatedPex5 (Fig. 3). The importance of NBR1 has been verified in physiologicalsettings including hypoxia-induced hepatocyte pexophagy (98) andstarvation-induced pexophagy (96). However, p62 is still found at ubi-quitinated peroxisomes andmay facilitate interaction of NBR1 with theperoxisome and/or autophagy machinery (89, 96, 97). Like mitochon-dria, peroxisomes are spatially dynamic organelles and undergo Pex11-mediated fission (99). This process is essential for pexophagy in yeastand occurs only in peroxisomes in contact with mitochondria (100).

Overall, it can be posited that pexophagy acts to limit peroxisomalb-oxidation capacity, signaling lipid generation, and ROS generation,especially in the case of dysfunctional peroxisomes. First, there are severalscenarios in which peroxisome biogenesis is thought to be a generallyadaptive response, and coincident pexophagymay limit or reverse inducedchanges. Characterization of drugs that stimulate peroxisome biogenesis(101) has led to the discovery of the peroxisomal proliferator-actived

5 of 17




ag.org/D

ownloaded from

receptor (PPAR) nuclear hormone receptor family, which are both thetargets of peroxisome-generated ligands (80) and inducers of peroxi-somebiogenesis amongmany other pleiotropic effects. ThePPARs havebeen widely targeted using the fibrate and thiazolidinedione (TZD)classes of drugs for their lipid-lowering, insulin-sensitizing properties,which are likely partially attributable to effects on peroxisomes (102, 103).Pexophagy is responsible for clearing peroxisomes upon withdrawal ofa peroxisome proliferator (87) and may act to maintain the integrity ofdrug-stimulated peroxisomes such that they are a sink rather than asource for ROS.

Pexophagy also likely participates in the regulation of adipocyte phe-notypes. Amajor focus of adipocyte biology involves the understandingof adipose tissue browning (the conversion of white adipocytes tobeige adipocytes, which are rich in energetically uncoupledmitochon-dria) to induce negative whole-body energy balance to treat metabolicdiseases (104). Peroxisome biogenesis is characteristic of adipose tissuebrowning and is likelynecessary tomeet themassiveb-oxidationdemandsof brown adipose tissue (105). Adipocyte-specific autophagy deficiencyprotects against diet-induced obesity and metabolic sequelae by inducinga brown adipocytelike phenotype (106). This effect has been attributed toimpairedmitophagy and accumulation ofmitochondria, but the accumu-lation of peroxisomes due to impaired pexophagy could also play a role.

Peroxisome-derived oxidative stress contributes to the pathology ofvarious cardiometabolic disease scenarios, including diabetic renal inju-ry (107), lipid-induced pancreatic b cell dysfunction (84), and aging(85). It will be of great interest to examine whether the in vitro findingsregarding peroxisomal fission and ATM-mediated sensing of ROStranslate to these settings. Beyond these preliminary discussions ofhow peroxisomes are involved in cardiometabolic diseases and regu-lated by pexophagy, several studies illustrate how disruption or elimina-tion of pexophagy is severely detrimental to metabolism. Ablation ofany of the key peroxisomal players in pexophagy (Pex2, Pex5, andPex11) results in severemetabolic perturbations (99, 108, 109), althoughunfortunately, the other essential roles of these proteins precludedrawing the strict conclusion that phenotypes arise from altered pexo-phagy. One last setting in which pexophagy is likely to be relevant is inthe liver, where peroxisomes are particularly abundant. Obesity inducedby a high-fat diet further enhances the biogenesis of ROS-producingperoxisomes in the liver (110). Predictably, liver-specific autophagy de-ficiency results in massive peroxisome accumulation (86), although ox-idative damage and steatosis are difficult to separate from the effects onmitophagy or lipophagy. Overall, more specific experimental manipu-lations of pexophagy are needed to fully understand its role in cardio-metabolic diseases with an emphasis on understanding howperoxisomesbecome dysfunctional and how the generation of peroxisome-specificROS and lipid species is altered.

MitophagyOf the organelles targeted for selective autophagy, the mitochondriahave the most straightforward impact on cellular energy metabolismand cardiometabolic disease phenotypes. Mitochondria produce ROSas a normal by-product of their essential roles in b-oxidation andoxidativephosphorylation to produce ATP. However, mitochondrial dysfunctionincluding ROS overproduction is implicated in cardiometabolic diseasepathologies including atherosclerosis (111), cardiomyopathy (112, 113),vascular endothelial dysfunction (114), and diabetes (115).MitochondrialROS adverselymodify lipids, proteins, andmitochondrial DNA, with thenet effect of impaired ATP production and propagation of further dys-


function. Mitochondrial DNA itself is also released into the cytoplasm,promoting inflammation and cell death (116, 117), and incompleteb-oxidation from lipid-overloadedmitochondria can generate toxic lip-id intermediates and consequent insulin resistance (Fig. 2) (118). Atten-tion has turned to restoring mitochondrial bioenergetics to limit theimpact of ROS and other toxic mitochondrial by-products.

Mitophagy is the selective autophagic clearance of mitochondria,and great progress has been made in understanding the various modesof its regulation and procession. First, it is important to conceptualizethe mitochondria as highly dynamic networked organelles whereinmorphology dictates function. Regular rounds of fission segregatemitochondria by membrane potential (119), constituting an integrativeset point for functionality that correlates positively with ATP-generatingcapacity and inversely with cumulative damage. Of the mitochondriasegregated by fission, those with adequate membrane potential re-fuse with the mitochondrial network, where elongated morphologymaximizes ATP synthesis efficiency. In contrast, fission-generatedmitochondrial fragments with borderline or low membrane potentialexist in a transient state with the potential for delayed fusion or autoph-agic degradation (120). Cellular nutrient excess, as found in cardiometa-bolic disease states, is also associated with fragmented mitochondria,which likely constitutes a protective response that favors both lower ox-idation efficiency to burn extra substrate and to steer damagedmitochondria toward degradation (120). From here, several integrativemolecular processes may mediate mitophagy. PTEN-induced putativekinase-1 (PINK1) mediates one mode of this next step; in healthymitochondria, it is constitutively imported in a membrane potentialdependent manner and degraded by the protease presenilin-associatedrhomboid-like (PARL) at the inner membrane (121). However, importis inhibited in dysfunctional mitochondria, leading to the accumulationof PINK1 at the outer membrane where its kinase activity recruits thelynchpin E3 ubiquitin ligase Parkin for mitophagy (Fig. 4) (122). Themitochondrial ubiquitination targets of Parkin are widespread, and it isgenerally not understood whether ubiquitination of any particular pro-tein carries special import (123). Further, Parkin-mediated ubiquitina-tion is opposed by the deubiquitinase USP30, serving as a last point ofcontrol to spare salvageable mitochondria from mitophagy (123, 124).Analysis of mitophagy inHeLa cells by several groups has suggested theparticular importance of the selective autophagy receptors OPTN andNDP52 (125, 126), whereas p62 seems to bemore critical formitochon-drial clustering (127). Lazarou et al. have identified a parallel mode ofPINK1-mediatedmitophagy in which PINK1 phosphorylates ubiquitinto recruit OPTN and NDP52, which in turn recruit the autophagy ini-tiation machinery including ULK1 (Fig. 4) (125). However, the univer-sality of these findingsmay be limited because p62 is amainmediator ofmitophagy in both macrophages (57) and cardiomyocytes (128). An-other mode of targeting mitochondria is through the selective autoph-agy receptor Bcl2-interacting protein 3 (BNIP3) and the relatedBNIP3L(also known as NIX) (129). The roles of BNIP3 and BNIP3L arecomplex because they are implicated both in apoptotic cell death byinducing mitochondrial dysfunction and in mediating protective mito-phagy (130, 131). Other additions to the growing family of proteins andpathways that mediate mitophagy include the Fanconi anemia proteins(132), FUNDC1 (133), and the selective autophagy receptor AMBRA1(134). The task to determine themitophagy pathways that are most rel-evant and targetable in cardiometabolic disease settings is ongoing andof great interest.

With the understanding that mitochondrial dysfunction is a ubiqui-tous feature of cardiometabolic diseases, many studies have begun to

6 of 17


CREDIT:A.KITTE




ag.org/D

ownloaded from

more directly characterize mitophagy as a modulator of pathology.First, decreases in skeletal muscle mass, function, and insulin sensitivityare implicated in the progression of metabolic disease such that themaintenance of mitochondrial function is particularly important in thishighly metabolically active tissue. Broad skeletal musclespecific au-tophagy deficiency results in the accumulation of aberrant mitochon-dria that coincides with atrophied myofibrils and losses in muscle massand function (135). The impact of this phenotype on systemic metab-olism is somewhat surprising; compensatory induction of the integratedstress response in muscle results in enhanced secretion of the mitokinefibroblast growth factor-21 (FGF-21) (136). In agreement with previousmurinemodels of FGF-21 induction (137), adipose tissue browning andincreased insulin sensitivity have been observed in muscle-specificautophagydeficientmice in the short term.However, these phenotypesare commonly observed in cachexic mice and must be weighed againstthe likely adverse long-term consequences of impaired muscle mito-phagy. Loss of estrogen receptor a (ERa) signaling, which potentlystimulates mitophagy in skeletal muscle, results in mitochondrialdysfunction, insulin resistance, and obesity inmice. These results impli-cate reduced skeletal mitophagy as a potential driver of insulin resistanceand adiposity in postmenopausal women (138). In a more direct charac-terization of skeletalmusclemitophagy,Drew et al. have identifiedHsp72as a mitochondrial stress sensor that cooperates with Parkin to inducemitophagy (139). Loss of Hsp72 or Parkin impairs skeletal muscle mito-phagy, insulin sensitivity, and fatty acid oxidation (139). Surprisingly,systemic Parkin-deficient mice overall are protected against high-fat dietinduced obesity and insulin resistance because Parkin promotes CD36-mediated lipid transport, which appears to outweigh the impact ofany mitophagy defects in the short term (140). Alternatively, the mito-phagy deficits induced by systemic Parkin knockout result in impairedglucose tolerance in the context of streptozotocin-induced type 1 diabe-tes (141). A second setting inwhichmitophagymayplay important pro-tective roles is in the vascular endothelium. Endothelial cells isolatedfrom diabetic patients feature fragmented mitochondria (142),increased oxidative stress (143), and impaired autophagic flux (144).Impaired autophagy (145) also coincides with increased mitochondrialoxidative stress (114) in the aging aortic endothelium. In either of thesesettings, pharmacological stimulation of autophagy, likely in great partthrough clearance of dysfunctionalmitochondria, reduces ROSproduc-tion and rescues vascular dysfunction (144, 145).

The role of mitophagy is perhaps best characterized in the setting ofcardiomyopathy. Deficiency of many mediators of mitochondrial dy-namics or mitophagy, including Opa1 (146), Mfn2 (147), and PINK1(148), results in overt heart failure. Montaigne et al. have shown thatcardiomyocytes isolated from human diabetic patients display reducedoxidative capacity, increased superoxide production, reduced ATG5,and fragmented mitochondria. Many of these markers of impaired mi-tophagy or mitochondrial dysfunction are associated with the severityof diabetes (as gauged by hemoglobin A1c levels) and reduced cardiaccontractile function (149).

In brown and beige adipocytes, mitophagy is again important formaintaining mitochondrial homeostasis, but it also reduces the abun-dance of metabolically beneficial uncoupled mitochondria. Mice withadipocyte-specific autophagy deficiency accumulate mitochondria inadipocytes and adopt a multilocular beige adipocytelike morphology(106, 150). Whereas these mice are resistant to typical sequelae of diet-induced obesity, a large portion of the phenotype may be due to vastlyincreased physical activity (106), and the mice are generally cachexicwith early mortality (150). Further, the interpretation of these studies


is complicated by the use of AP2-Cre to abrogate adipocyte autophagy;AP2 is now known to target many nonadipocyte tissues (151, 152). Ad-ditional studies have clarified the role of mitophagy in adipocytes.Altschuler-Keylin et al. have shown that upon withdrawal of classicalbeigeing stimuli, beige adipocytes rapidly reacquire a white adipocytephenotype in a mitophagy-dependent manner. A more precise murinemodel was used to characterize the effects of beige adipocyte-specificautophagydeficiency usingUCP1-Cre drivenATG12deletion. Thismodelrevealed no profound baseline phenotype, but prolonged b-adrenergicagonist induced adipocyte beigeing, enhanced energy expenditure,and reduced adiposity in response to high-fat diet (153). A second studyhas identified BNIP3 as a crucial mechanistic mediator of mitophagy inadipocytes.BNIP3 is a prominent target of PPARg, the transcription factorthat is the master regulator of adipogenesis and mediator of many phar-macological effects of TZDs. Further, adipocyte BNIP3 abundance wasincreased in several murine models of obesity, whereas BNIP3 deficiencypredisposed mice to insulin resistance and NAFLD. Mitochondria fromadipocytes from these mice are more elongated, have lower respiration,and have increased superoxide production (154). Last, p62 disruption inadipocytes results in compromisedmitochondrial structure and function,suggesting its involvement in adipocyte mitophagy. However, the obesityobserved in p62 deficiency is most likely due to defective b-adrenergicsignaling and mitochondrial biogenesis (34). Because mitophagy is adynamic process withmany putative and overlappingmediators, we arejust beginning to develop the genetic and pharmacological techniquesrequired to study its roles in disease. Given the widespread evidence formitochondrial dysfunction across many cardiometabolic disease set-tings and the essential role for mitophagy in clearing damagedmitochondria, this topic of investigation will continue to be particularlyfruitful.

Mitophagy

NDP52

PINK1

PINK1

BNIP3

USP30

OPTN

OPT

N

ULK1

Phosphogroup p62

p62

ND

P52

p62

Ubiquitin

LC3-PE

Parkin

Fig. 4. Model of key molecular events that mediate mitophagy.Mitophagy likelyproceeds through several complementary mechanisms. (Left) In response to mito-chondrial damage, PINK1 phosphorylates ubiquitin to directly enhance NDP52 andOPTN binding. NDP52 and OPTN recruit several components of the autophagy ma-chinery includingULK1 to initiatemitophagy. (Bottom) PINK1also activates the E3ubiq-uitin ligase Parkin, which targets many mitochondrial proteins. The deubiquitinaseUSP30 opposes Parkin to spare less-damaged mitochondria. Selective autophagyreceptors for mitophagy include NDP52, OPTN, and p62. (Right) BNIP3 family pro-teins in themitochondrial outer membrane directly mediatemitophagy by bindingto LC3. (Center) Polyubiquitin/p62 oligomers cluster damaged mitochondria to fa-vor mitophagy.

7 of 17




ag.org/D

ownloaded from

LipophagyIn a healthy physiological state, lipids are stored mainly in the form oftriglycerides and cholesterol esters inside lipid droplets, dynamic organ-elles that are most prominent in adipocytes. Dysfunction of the meta-bolic and storage capacities of adipocytes results in toxic spillover ofexcess lipids into the circulation and their storage as lipid droplets innearly every organ system (155). This ectopic accumulation is a primaryhallmark and driver of systemic insulin resistance and cardiometabolicpathology (156), and restoration of healthy oxidation and trafficking oflipids is a longstanding mechanistic target. Classic regulators of lipidmetabolism at the lipid droplet surface include neutral lipases such asneutral cholesterol ester hydrolase-1 (NCEH1), hormone-sensitive li-pase (HSL), and adipose triglyceride lipase (ATGL), and synthases suchas diglyceride acyltransferase (DGAT). Lipophagy, the selective autoph-agic degradation of intracellular lipids, is increasingly appreciated as aparallel, versatile contributor to lipid droplet homeostasis (Fig. 5).

Although little is known about the precise steps required to initiateand carry out autophagy at the lipid droplet, most studies thus far havefocusedonamodel inwhich lipiddroplets are sequestered throughautoph-agy anddelivered to the lysosome for lysosomal acid lipasemediated lipolysis(157159). To our knowledge, classic selective autophagy receptors such asp62 have not been detected at lipid droplets, and ubiquitination eventshave been linked only to regulatory degradation of particular lipid dropletproteins, often through the proteasome (160163). Nonetheless, lipo-phagy is selective because autophagosomes are found associated prefer-entially with lipid droplets under various conditions of lipid stress,distinguishing it frombulk autophagy (157159). Amolecularmediatorof autophagosome recruitment to lipid droplets is the small GTPaseRab7, which also regulates trafficking of autophagosomes to lysosomes,althoughmany structural aspects of these processes remain unclear (164).

Several studies highlight themechanisms by which selective autoph-agy regulates lipid droplets through nontraditional means. First, theneutral lipases ATGL andHSL have LC3-bindingmotifs that are crucialfor localization to lipid droplets and for overall lipolytic activity (165).These interactions exemplify the concept that many of the core autoph-agy structural proteins, including LC3, are themselves ubiquitin-likescaffolding proteins (166) with important functions beyond sequester-ing cargo. Second, under some conditions, autophagy also delivers lipidsto lipid droplets for subsequent neutral lipolysis and mitochondrial ox-idation (167, 168). Whether this process simply uses bulk autophagy ofmembrane lipids or can selectively target and degrade less essentialmembranes as substrates (169) is unknown.This observation additionallychallenges the prevailing notion that autophagy is purely a degradativemechanism at the lipid droplet and that the presence of LC3 at the lipiddroplet necessarily implies that degradative lipophagy is occurring. Third,lipolysis has been proposed to be tightly regulated by chaperone-mediated autophagy (CMA) (170). In this process, perilipin proteins(PLIN2/3),whichnormally coat the lipid droplet andblock access to neutrallipases, are tagged by the adapter heat shock cognate protein 70 (Hsc70),which promotes direct lysosome-associated membrane protein 2A(LAMP2A)mediated translocation into the lysosome. Thus, CMAdoes not directly deliver lipids to lysosomes but instead constitutes aselective autophagic modulation of lipid droplet physiology. Note thatlipid droplets are relatively stable and nontoxic storage sites for cho-lesterol esters and triglycerides. Labile species such as ceramides, freefatty acids, and diacyglycerol associated with metabolic disease (171)are more likely to mediate lipotoxicity. It is an intriguing possibilitythat autophagic processes could be a mechanism by which such lipidintermediates are delivered to lipid droplets.


Initial interest in the physiological implications of lipophagy arosefrom studies suggesting substantive roles of this process in lipid catab-olism, especially in the context of NAFLD. Singh et al. have shown thatin lipid-overloaded states, autophagic vesicles sequester lipid dropletsand deliver their contents to the lysosome for hydrolysis (158). Dys-function in hepatocyte autophagy leads to reduced b-oxidation, result-ing in lipid accumulation and hepatosteatosis (158, 172), although thisphenomenon may also be partially due to impaired mitophagy (86)rather than lipophagy per se. The limited data available suggest increasedabundance of autophagic markers in human NAFLD (22, 173), perhapsas a compensatory response to excess lipids. A second setting in whichlipophagy is critical for lipid catabolism is in foam cell macrophages,where massive accumulation of cholesterol esterrich lipid droplets is adefining event in atherosclerosis. Ouimet et al. have shown that macro-phage lipid droplets are delivered to lysosomes through autophagy, andcholesterol esters are then hydrolyzed through lysosomal lipolysis for ef-flux to high-density lipoprotein in an ABCA1-dependent manner (174).Consequently, it is not surprising that the functionality of lysosomesand associated acid lipases is crucial for lipophagy and disrupted in car-diometabolic pathology (175). Atherogenic lipids initiate lysosome dys-function that inhibits the degradation of all delivered autophagicsubstrates (43, 51, 176). Similarly, excess free fatty acids promote lyso-somal permeabilization and destabilization in human NAFLD (177).The severe liver steatosis and atherosclerosis observed in lysosomal acidlipase deficiency provide additional indirect evidence for lipophagy as arelevant route of cellular lipid catabolism (178180).

In contrast, the importance of lipophagic delivery to lysosomes is notuniversal. Adipose tissue macrophages activate lysosomal biogenesisduring obesity, most likely tometabolize and redistributemassive inter-nal and external lipid loads from stressed adipocytes (181, 182). Surpris-ingly, autophagy deficiency achieved by deletion of ATG5 or ATG7 inadipose tissue macrophages does not substantially disrupt lipid metab-olism or affect metabolic phenotypes, suggesting that lipophagy is not acritical route of lipid delivery to lysosomes even in this lipid-stressedsetting (67, 182). Could a predominance of CMA, non-ATG5/ATG7macroautophagy (183), or extracellular digestion through lysosomalexocytosis (184) explain these findings? Overall, many additionalquestions remain to be answered in understanding how impaired lipo-phagy may contribute to cardiometabolic disease and how it might betargeted to more appropriately traffic lipids. Key areas of investigationinclude how the autophagic machinery interacts with lipid droplets andhow lipophagy is regulated in a context-dependent manner to eitherdeplete or build lipid droplets in appropriate tissues.

Transcriptional Regulation and Integration of SelectiveAutophagy with Metabolic SignalingGiven the importance of selective autophagy in maintaining cellularhomeostasis, components of its molecular machinery are under tightregulation by stress-responsive andmetabolic transcription factors. Inturn, selective autophagy receptors often also serve as scaffolds thatinteract with and regulate many cell signaling pathways. Through thismechanism, their buildup can serve as a key monitored variable, sig-nifying failure or insufficiency of autophagic clearance.We briefly dis-cuss these key concepts and their relevance to cardiometabolic disease.

Transcriptional regulation of the molecular mediators of selectiveautophagy is poorly understood and seems to be mostly embeddedwithin broader transcriptional programs regulating autophagy-lysosomalbiogenesis and stress responses (Fig. 6). Among these factors are the

8 of 17


CREDIT:A.KITTE



http://stke.sciencD

ownloaded from

MITF-TFE family of transcription factors, which particularly targetautophagy-lysosomal genes, notably including p62 (43, 185). This controlis exerted primarily through the affinity of TFEB and TFE3 for co-ordinated lysosomal expression and regulation (CLEAR)motifs, subtypesof E-box elements in target promoter sequences (185187).Amajor pointof control for these transcription factors lies in their regulation by themTORC1 complex, a nutrient-sensitive signaling hub widely implicatedin metabolic regulation (188). Under basal conditions, mTORC1 phos-phorylates TFEB and/or TFE3 at the lysosome to facilitate binding to14-3-3 chaperones, which retain them in the cytosol, thereby limitingtranscriptional activity (189, 190). Upon conditions of starvation or ly-sosomal stress, mTORC1 activity is inhibited, allowing translocation ofTFEB and/or TFE3 to the nucleus to transcriptionally activate autoph-agy and lysosomal genes (187, 189191).

In the case of selective autophagy, details are just beginning toemerge on how this mTORC1-TFEB/TFE3 axis interacts with impor-tant players such as p62. TheMITF-TFE family of transcription factorstranslocate to the nucleus in a Parkin- or PINK1-dependent manner inresponse to mitochondrial stress, and the absence of TFEB or TFE3substantially impairs mitophagy (Fig. 6) (192, 193). In many diseaseand genetic autophagydeficient settings, p62 builds up as a result offailed autophagic clearance. However, complete genetic deficiency ofall MITF-TFE family members impairs mitochondrial stress-inducedp62 transcription and translation to such a degree that p62 abundancedoes not increase in response tomitochondrial stress despitemitophagydeficiency (192).

TFEB also exemplifies a second important concept in the transcrip-tional regulation of selective autophagy: the coupling of programs reg-ulating selective autophagy with those mediating biogenesis of theorganelles destined for degradation. For example, TFEB also transcription-


ally targetsperoxisome-proliferator activated receptor coactivator-1a (PGC-1a) (194), a transcriptional coactivator that orchestrates mitochondrialand peroxisomal biogenesis (105, 195). Another parallel mode of tran-scriptional regulation of selective autophagy involves the transcriptionfactor Nrf2, whichmediates antioxidant responses in response to ox-idative stress (Fig. 6). Nrf2 targets autophagy-related genes encodingp62, NDP52, and other autophagy components (196, 197). The mostintriguing aspect about this regulation is the feedback role of p62 onNrf2 activity. Under diverse conditions of impaired or insufficientselective autophagy, p62 accumulates on poorly degraded substrates(18, 19, 198), allowing p62 to bind to and neutralize a large portion ofthe cellular pool of Keap1, which would otherwise sequester Nrf2 forproteasomal degradation (199). This results in an increased availa-bility of free Nrf2, enabling nuclear translocation and transcriptionof selective autophagy targets (Fig. 6) (29, 200, 201). We postulatethat this mechanism is one of several feedback loops that respondtodysfunctional or insufficient selective autophagy to restore homeostasisand orchestrate cellular metabolism (Fig. 6). p62 also regulates and/or isa transcriptional target of several crucial metabolic signaling mediatorsincluding NF-kB (57, 202, 203), mTOR (32, 204, 205), PKC (203, 206),VDR/RXR (33), and ERK1 (207, 208). Unsurprisingly, several murinemodels of whole-body or tissue-specific p62 abrogation result in obesityand diabetes, and the precise tissues and physiological mechanismsinvolved are complex (34, 208, 209). The FOXO family of transcriptionfactors also regulate selective autophagy by targeting the genes encodingPINK1 (210), BNIP3 (211,212), andother autophagy genes. LikeNrf2 andTFEB, the FOXO transcription factors respond to awide range of stressesby promoting bothmitophagy and lipophagy (213). Ultimately, amatureunderstanding of selective autophagywill require further studies on its reg-ulation and integration with cellular metabolic signaling pathways.

on July 23, 2018em

ag.org/

Looking Forward: OutstandingFundamental Questions in theRoles of Selective Autophagy inMetabolic Disease andTherapeutic OpportunitiesIn this Review, we have described howdys-functional organelles are recognized andcleared by selective autophagy to main-tain cellular homeostasis and modulatecardiometabolic disease pathology. To putthese concepts into context, an importantopen question is whether the molecularmechanisms that establish the specificityof selective autophagy (such as ubiquitina-tionandselective autophagyreceptorbinding)are actually dysfunctional or rate-limitingindisease states.On theonehand,molecularmediators of these steps are clearly necessaryand important for limiting dysfunctionalorganelle accumulation as we have dis-cussed and evidenced by the maladaptivephenotypes of various cellular andmurineknockout models. However, no evidencethus far suggests a strict insufficiency incargo recognition and tagging. It will beimportant to testwhether enhancing theseearly steps in selective autophagy through

Lipophagy

Autophagosome

Neutrallipolysis

Autolysosome

Chaperone-mediatedautophagy

Lysosome

PLIN2/3ATGL

ATGL PLIN2/3PLIN2/3

PLIN2/3

PLIN2/3

Hsc70

Hsc70 PLIN2/3

Rab7

HSL

?

LC3-PE

LAMP-2A

Protease

Lipase

Fig. 5. Molecular mediators of lipophagy. (Left) Autophagosome recruitment to lipid droplets and downstreamautophagosome-lysosome interaction depend on the activity of the small GTPase Rab7. (Center) The neutral lipasesATGL and HSL contain LC3-binding domains that are required for recruitment to the lipid droplet, constituting amechanism for cross-talk between lipophagy and neutral lipolysis. (Right) PLIN2 and PLIN3 (PLIN2/3) proteins coatthe lipid droplet surface and block both lipophagy and neutral lipolysis. Chaperone-mediated autophagy (CMA) tar-gets PLIN2/3 using Hsc70 as an adapter, resulting in their direct translocation into lysosomes for degradation. Thisallows the machineries of both neutral lipolysis and lipophagy to access the lipid droplet.

9 of 17


CREDIT:A.KITTE



http://stke.D

ownloaded from

genetic overexpression models or specific drug targeting can promoteorganelle clearance for cardiometabolic benefit.We propose that appro-priate selective autophagy substrates are generally primed and readyfor autophagic sequestration and lysosomal degradation, but dys-function in these latter processes is more likely to be rate-limiting.For example, there may simply be a lack of autophagy initiation to ad-equately sequester cargo, the cargomay be too large, or the cargomay betoxic and disruptive to the function of the autophagosome and lyso-somes. Mere haploinsufficiency of ATG7 greatly accelerates the pro-gression of diabetic pathology by inducing inflammasomehyperactivity, mitochondrial dysfunction, and lipid accumulation(68), suggesting that the availability of autophagosome components eas-ily becomes limiting even with large obesity-induced compensatory in-creases (68). Conversely, global overexpression of ATG5 reducesadiposity, increases insulin sensitivity, and extends life span (214),and ATG7 overexpression has similar effects (172). Autophagic cargo,especially lipids, may be particularly toxic to autophagic and lysosomalmembranes and disrupt their function. For example, lysosomes losemembrane permeability and degradative function when loaded with li-pids as in atherosclerosis (43), and autophagosome and lysosomemem-brane dynamics and fusion are influenced by altered membrane lipidcomposition induced by high-fat diets (215).

These ideas raise several additional questions regarding the thera-peutic induction of autophagy to treat cardiometabolic disease. First,does broad induction of autophagy through various behavioral, genetic,or pharmacological means initiate a pattern of grossly nonselective bulkdegradation or is the tagging of dysfunctional organelles adequate tobias the induced autophagic machinery toward selective degradation


of this cargo? It is a concern that initiation of bulk autophagy will causecomplications due to degradation of off-target cellular components.However, even starvation, the prototypical stimulus for bulk autophagy,involves an ordered process of degradation (169, 216). Additionally, wehave discussed how several modes of selective autophagy are coupled totranscriptional programs of organelle biogenesis and that many of thesignaling cascades initiating autophagy also targetmediators of selectiveautophagy (Fig. 6) such that dysfunctional organelles may be priori-tized. Second, several groups have established that autophagy cannotproceedwith impaired lysosomes (217, 218). It remains unclearwhetherlysosomal dysfunctions such as impaired biogenesis, membrane per-meabilization, and insufficient acidity are truly a primary cause of au-tophagic failure and metabolic dysfunction. Could targeting lysosomaldysfunction serve as a better conceptual target in treatingmetabolic dis-ease? Finally, are specific molecular events (such as favoring phospho-rylation and binding of a particular selective autophagy receptor) viabletargets to enhance organelle clearance, or is organelle dysfunction toobroad for stimulating specific clearance pathways to be beneficial?

With these questions and concepts in mind, we briefly discuss pro-gress in stimulating autophagy to treat cardiometabolic disease with re-gard to the selective cargoes we have discussed (Table 1). First, exerciseand caloric restriction are highly effective frontline behavioral interven-tions to prevent and curb the progression of cardiometabolic disease.Both interventions stimulate autophagy, restore organelle homeostasis,and target many of the known transcriptional and posttranslationalmechanisms that regulate selective autophagy (219, 220). For example,Bag3- and p62-mediated selective autophagy is induced in humansperforming resistance training (221), and exercise-induced restoration

on July 23, 2018sciencem

ag.org/

ofmuscle glucose homeostasis is autophagy-dependent (222). Similarly, variousmodesof caloric restriction protect against stea-totic liver pathology (223) andmyocardialischemia-reperfusion injury throughmechan-isms associated with autophagy induction(224). How these behavioral interventionsregulate and whether their effects dependspecifically on selective autophagy are large-ly unexamined. Genetic manipulations offactors that control selective autophagy arealso of mechanistic and therapeutic inter-est. For example, uncovering how certaintranscription factors differentially targetmolecular mediators of selective autophagymay allow pharmacological compoundstargeting those pathways to be appropri-atelymatched to specific organelle dysfunc-tions. We highlight TFEB and TFE3 asparticularly attractive regulators of selec-tive autophagy because they target amech-anistically complete transcriptional programof selective autophagy, lysosomal biogenesis,and lipid catabolism. Overexpression ofTFEB and/or TFE3 in various tissue-specific manners protects against ather-osclerosis (43), steatohepatitis (194, 225),and type 2 diabetes (194, 225, 226). Ofcourse, the ultimate goal lies in pharmaco-logically targeting selective autophagy totreat cardiometabolic disease. Neutralizing

TFEB

TFEB

p62

Competitionfor binding

Mitochondrialstress

p62

Proteasomaldegradation

Induction ofselective autophagyEnhanced clearance

Autophagy

Stress

Nucleus

Nrf2

Keap1

Nrf2Nrf2

Nrf2

Keap1

Keap1

p62Keap1

p62

Selective autophagyand lysosomal genes

Keap1

Keap1

p62p62

p62

p62 p62

p62

p62

Proteinaggregate

Fig. 6. Transcriptional feedback control of selective autophagy. Several parallel feedback loops couple sensing oforganelle damage with transcriptional regulation of selective autophagy genes. Under basal conditions, Keap1 binds toand targets Nrf2 for proteasomal degradation. Protein aggregates and dysfunctionalmitochondria accumulate p62, whichbinds to and sequesters Keap1 to free Nrf2 and activate its transcriptional activity. Nrf2 targets many selective autophagygenes todegradep62-tagged cargo. Similarly, TFEB translocates to thenucleus in response tomitochondrial and lysosomalstresses to transcribe selective autophagy, mitochondrial, and lysosomal genes.

10 of 17



hD

ownloaded from

damaging organelle products such as ROS and inflammatory cytokineshas been extensively investigated, but conceptually, this approach isthought to be problematic because it also disrupts adaptive basal ROSand inflammatory signaling (227, 228) without proportionally restoringthe organelles that generate these signals. In contrast, pharmacologicallystimulating selective autophagy addresses the needs of both neutralizingexcess ROS and inflammatory signals and rescuing the upstream organ-elle dysfunction. To date, truly specific pharmacological manipulation ofa molecular mediator or organelle target of selective autophagy has notbeen therapeutically achieved. Nonetheless, this degree of precision maybe ineffective due to redundancy in clearance pathways and could causeoff-target effects because selective autophagy adapters have autophagy-independent roles. As discussed, broader stimulation of autophagycould be leveraged to bias clearance of already selectively tagged cargos,and inducing broader transcriptional regulators such as TFEB andTFE3 favors a comprehensive program of selective autophagic clear-ance. Various compounds can induce these types of responses. For ex-ample, mTORC1 inhibitors such as rapamycin induce TFEB nucleartranslocation and transcriptional activity (190) and also spur autophagyby relieving mTORC1 repression of ULK1, the primary kinase that in-itiatesmacroautophagy (229, 230). AlthoughmTOR inhibition has plei-otropic effects beyond inducing autophagy, it represents an importanttarget in treating cardiometabolic disease (231). Second, the naturallyoccurring disaccharide trehalose potently induces autophagy in vivoand is effective in treating NAFLD (68, 232), endothelial dysfunction


(145, 233), and diabetes (41, 68, 234). Because orally administered tre-halose does not efficiently cross the intestinal epithelium (235) andmaycause modest weight gain in humans due to its conversion to glucose(233), widespread use awaits means of improving its pharmacologicalproperties or exploring the use of closely related compounds (236).Third, the polyamine spermidine can stimulate autophagy by modulat-ing protein acetylation (237) and has shown promise in treating endo-thelial dysfunction (144, 238) and atherosclerosis (239) throughautophagy-related mechanisms. Of course, there are many moreautophagy-inducing compounds with therapeutic potential (240)andmore studies are required to better understand how theymay favorselective clearance of dysfunctional organelle cargo.

In summary, selective autophagy is a deterministic pathway thatmediates the clearance of diverse organelle cargoes widely implicatedin the progression of cardiometabolic disease. Key steps in thesecoordinated processes include tagging of cargo, recognition, interactionwith autophagic machinery, and lysosomal degradation. The precisemolecular mediators of selective autophagic clearance of inflamma-somes, lipid droplets, protein aggregates, peroxisomes, and mitochon-dria are somewhat distinct but perhaps most commonly shareubiquitination and p62- or NBR1-mediated recognition. We anticipatethat increased understanding of the regulation and interventional tar-geting of selective autophagy will prove fruitful in stemming the tide ofcardiometabolic disease.

on July 23, 2018ttp://stke.sciencem

ag.org/

REFERENCES AND NOTES1. G. A. Roth, M. D. Huffman, A. E. Moran, V. Feigin, G. A. Mensah, M. Naghavi, C. J. L. Murray,

Global and regional patterns in cardiovascular mortality from 1990 to 2013. Circulation132, 16671678 (2015).

2. C. Patterson, C. Ike, P. W. Willis IV, G. A. Stouffer, M. S. Willis, The bitter end: Theubiquitin-proteasome system and cardiac dysfunction. Circulation 115, 14561463(2007).

3. D. Versari, J. Herrmann, M. Gssl, D. Mannheim, K. Sattler, F. B. Meyer, L. O. Lerman,A. Lerman, Dysregulation of the ubiquitin-proteasome system in human carotidatherosclerosis. Arterioscler. Thromb. Vasc. Biol. 26, 21322139 (2006).

4. K. R. Parzych, D. J. Klionsky, An overview of autophagy: Morphology, mechanism, andregulation. Antioxid. Redox Signal. 20, 460473 (2014).

5. N. Mizushima, M. Komatsu, Autophagy: Renovation of cells and tissues. Cell 147,728741 (2011).

6. C. Kraft, M. Peter, K. Hofmann, Selective autophagy: Ubiquitin-mediated recognition andbeyond. Nat. Cell Biol. 12, 836841 (2010).

7. D. J. Klionsky, E. H. Baehrecke, J. H. Brumell, C. T. Chu, P. Codogno, A. M. Cuervo,J. Debnath, V. Deretic, Z. Elazar, E.-L. Eskelinen, S. Finkbeiner, J. Fueyo-Margareto,D. A. Gewirtz, M. Jttel, G. Kroemer, B. Levine, T. J. Melia, N. Mizushima,D. C. Rubinsztein, A. Simonsen, A. Thorburn, M. Thumm, S. A. Tooze, A comprehensiveglossary of autophagy-related molecules and processes (2nd edition). Autophagy 7,12731294 (2011).

8. P. Xu, D. M. Duong, N. T. Seyfried, D. Cheng, Y. Xie, J. Robert, J. Rush, M. Hochstrasser,D. Finley, J. Peng, Quantitative proteomics reveals the function of unconventionalubiquitin chains in proteasomal degradation. Cell 137, 133145 (2009).

9. M. S. Willis, C. Patterson, Proteotoxicity and cardiac dysfunctionAlzheimers disease ofthe heart? N. Engl. J. Med. 368, 455464 (2013).

10. R. R. Kopito, Aggresomes, inclusion bodies and protein aggregation. Trends Cell Biol. 10,524530 (2000).

11. E. J. Bennett, N. F. Bence, R. Jayakumar, R. R. Kopito, Global impairment of the ubiquitin-proteasome system by nuclear or cytoplasmic protein aggregates precedes inclusionbody formation. Mol. Cell 17, 351365 (2005).

12. T. Lamark, T. Johansen, Aggrephagy: Selective disposal of protein aggregates bymacroautophagy. Int. J. Cell Biol. 2012, 736905 (2012).

13. S. Murata, Y. Minami, M. Minami, T. Chiba, K. Tanaka, CHIP is a chaperone-dependent E3ligase that ubiquitylates unfolded protein. EMBO Rep. 2, 11331138 (2001).

14. J. A. Olzmann, L.-S. Chin, Parkin-mediated K63-linked polyubiquitination: A signal fortargeting misfolded proteins to the aggresome-autophagy pathway. Autophagy 4,8587 (2008).

Table 1. Therapeutic means of enhancing selective autophagy inmetabolic disease.

Treatment
Selective autophagy-related targets
identified thus farc

Impacts onardiometabolic diseases

Behavioralinterventions

Exercise
Bag3 (221), p62 (221),Nrf2 (241), TFEB (242),BNIP3 (243), and BRCA1
(244)

Broad protection

Caloric restriction(or fasting)

B
NIP3 (245), FOXO (245),TFEB (194, 246), Rab7
(164), p62 (247), and TFE3(187)

Broad protection

Genetic models

TFEB
p62(43, 185) and otherautophagy-lysosomalgenes (43, 185, 186) NAFLD (194), obesity/diabetes (194), and atherosclerosis (43)
TFE3
p62(192) among manyautophagy-lysosomal
genes (187, 192)

NAFLD (225, 248) and diabetes (225, 226)

Pharmacologicalstrategies

Trehalose
p62 (249) and othermeans of autophagy
modulation(145, 232, 249251)

NAFLD (68, 232),diabetes (41, 68, 234), andendothelial dysfunction

(145, 233)

Spermidine
ATM (252), PINK1 (252),and Parkin (252)
Atherosclerosis (239) andendothelial dysfunction

(144, 238)

11 of 17




ag.org/D

ownloaded from

15. M. Gamerdinger, S. Carra, C. Behl, Emerging roles of molecular chaperones and co-chaperones in selective autophagy: Focus on BAG proteins. J. Mol. Med. 89, 11751182(2011).

16. M. Gamerdinger, A. M. Kaya, U. Wolfrum, A. M. Clement, C. Behl, BAG3 mediateschaperone-based aggresome-targeting and selective autophagy of misfolded proteins.EMBO Rep. 12, 149156 (2011).

17. S. Pankiv, T. H. Clausen, T. Lamark, A. Brech, J.-A. Bruun, H. Outzen, A. vervatn,G. Bjrky, T. Johansen, p62/SQSTM1 binds directly to Atg8/LC3 to facilitatedegradation of ubiquitinated protein aggregates by autophagy. J. Biol. Chem. 282,2413124145 (2007).

18. M. Komatsu, S. Waguri, M. Koike, Y.-s. Sou, T. Ueno, T. Hara, N. Mizushima, J.-i. Iwata,J. Ezaki, S. Murata, J. Hamazaki, Y. Nishito, S.-i. Iemura, T. Natsume, T. Yanagawa,J. Uwayama, E. Warabi, H. Yoshida, T. Ishii, A. Kobayashi, M. Yamamoto, Z. Yue,Y. Uchiyama, E. Kominami, K. Tanaka, Homeostatic levels of p62 controlcytoplasmic inclusion body formation in autophagy-deficient mice. Cell 131, 11491163(2007).

19. I. Sergin, S. Bhattacharya, R. Emanuel, E. Esen, C. J. Stokes, T. D. Evans, B. Arif, J. A. Curci,B. Razani, Inclusion bodies enriched for p62 and polyubiquitinated proteins inmacrophages protect against atherosclerosis. Sci. Signal. 9, ra2 (2016).

20. C. Stumptner, A. Fuchsbichler, K. Zatloukal, H. Denk, In vitro production of Mallorybodies and intracellular hyaline bodies: The central role of sequestosome 1/p62.Hepatology 46, 851860 (2007).

21. M. Filimonenko, P. Isakson, K. D. Finley, M. Anderson, H. Jeong, T. J. Melia, B. J. Bartlett,K. M. Myers, H. C. G. Birkeland, T. Lamark, D. Krainc, A. Brech, H. Stenmark,A. Simonsen, The selective macroautophagic degradation of aggregated proteinsrequires the PI3P-binding protein Alfy. Mol. Cell 38, 265279 (2010).

22. Y. Fukuo, S. Yamashina, H. Sonoue, A. Arakawa, E. Nakadera, T. Aoyama, A. Uchiyama,K. Kon, K. Ikejima, S. Watanabe, Abnormality of autophagic function and cathepsinexpression in the liver from patients with non-alcoholic fatty liver disease. Hepatol. Res.44, 10261036 (2014).

23. K. Zatloukal, C. Stumptner, A. Fuchsbichler, H. Heid, M. Schnoelzer, L. Kenner, R. Kleinert,M. Prinz, A. Aguzzi, H. Denk, p62 Is a common component of cytoplasmicinclusions in protein aggregation diseases. Am. J. Pathol. 160, 255263 (2002).

24. K. Zatloukal, S. W. French, C. Stumptner, P. Strnad, M. Harada, D. M. Toivola, M. Cadrin,M. B. Omary, From Mallory to MalloryDenk bodies: What, how and why?Exp. Cell Res. 313, 20332049 (2007).

25. Y. Peng, B. A. French, B. Tillman, T. R. Morgan, S. W. French, The inflammasome inalcoholic hepatitis: Its relationship with MalloryDenk body formation. Exp. Mol. Pathol.97, 305313 (2014).

26. Y. Inami, S. Waguri, A. Sakamoto, T. Kouno, K. Nakada, O. Hino, S. Watanabe, J. Ando,M. Iwadate, M. Yamamoto, M.-S. Lee, K. Tanaka, M. Komatsu, Persistent activation of Nrf2through p62 in hepatocellular carcinoma cells. J. Cell Biol. 193, 275284 (2011).

27. R. Mathew, C. M. Karp, B. Beaudoin, N. Vuong, G. Chen, H.-Y. Chen, K. Bray,A. Reddy, G. Bhanot, C. Gelinas, R. S. DiPaola, V. Karantza-Wadsworth, E. White,Autophagy suppresses tumorigenesis through elimination of p62. Cell 137,10621075 (2009).

28. H.-M. Ni, B. L. Woolbright, J. Williams, B. Copple, W. Cui, J. P. Luyendyk, H. Jaeschke,W.-X. Ding, Nrf2 promotes the development of fibrosis and tumorigenesis in mice withdefective hepatic autophagy. J. Hepatol. 61, 617625 (2014).

29. M. Komatsu, H. Kurokawa, S. Waguri, K. Taguchi, A. Kobayashi, Y. Ichimura, Y.-S. Sou,I. Ueno, A. Sakamoto, K. I. Tong, M. Kim, Y. Nishito, S.-i. Iemura, T. Natsume, T. Ueno,E. Kominami, H. Motohashi, K. Tanaka, M. Yamamoto, The selective autophagy substratep62 activates the stress responsive transcription factor Nrf2 through inactivation ofKeap1. Nat. Cell Biol. 12, 213223 (2010).

30. A. Takamura, M. Komatsu, T. Hara, A. Sakamoto, C. Kishi, S. Waguri, Y. Eishi, O. Hino,K. Tanakaand, N. Mizushima, Autophagy-deficient mice develop multiple liver tumors.Genes Dev. 25, 795800 (2011).

31. J. Moscat, M. T. Diaz-Meco, p62 at the crossroads of autophagy, apoptosis, and cancer.Cell 137, 10011004 (2009).

32. A. Umemura, F. He, K. Taniguchi, H. Nakagawa, S. Yamachika, J. Font-Burgada, Z. Zhong,S. Subramaniam, S. Raghunandan, A. Duran, J. F. Linares, M. Reina-Campos, S. Umemura,M. A. Valasek, E. Seki, K. Yamaguchi, K. Koike, Y. Itoh, M. T. Diaz-Meco, J. Moscat,M. Karin, p62, upregulated during preneoplasia, induces hepatocellular carcinogenesisby maintaining survival of stressed HCC-initiating cells. Cancer Cell 29, 935948(2016).

33. A. Duran, E. D. Hernandez, M. Reina-Campos, E. A. Castilla, S. Subramaniam,S. Raghunandan, L. R. Roberts, T. Kisseleva, M. Karin, M. T. Diaz-Meco, J. Moscat,p62/SQSTM1 by binding to vitamin D receptor inhibits hepatic stellate cell activity,fibrosis, and liver cancer. Cancer Cell 30, 595609 (2016).

34. T. D. Mller, S. J. Lee, M. Jastroch, D. Kabra, K. Stemmer, M. Aichler, B. Abplanalp,G. Ananthakrishnan, N. Bhardwaj, S. Collins, S. Divanovic, M. Endele, B. Finan, Y. Gao,K. M. Habegger, J. Hembree, K. M. Heppner, S. Hofmann, J. Holland, D. Kchler,


M. Kutschke, R. Krishna, M. Lehti, R. Oelkrug, N. Ottaway, D. Perez-Tilve, C. Raver,A. K. Walch, S. C. Schriever, J. Speakman, Y.-H. Tseng, M. Diaz-Meco, P. T. Pfluger,J. Moscat, M. H. Tschp, P62 links b-adrenergic input to mitochondrial function andthermogenesis. J. Clin. Invest. 123, 469478 (2013).

35. P. Lahiri, V. Schmidt, C. Smole, I. Kufferath, H. Denk, P. Strnad, T. Rlicke, L. F. Frhlich,K. Zatloukal, p62/Sequestosome-1 is indispensable for maturation and stabilizationof Mallory-Denk bodies. PLOS ONE 11, e0161083 (2016).

36. J. Janson, R. H. Ashley, D. Harrison, S. McIntyre, P. C. Butler, The mechanism of isletamyloid polypeptide toxicity is membrane disruption by intermediate-sized toxicamyloid particles. Diabetes 48, 491498 (1999).

37. S. Despa, S. Sharma, T. R. Harris, H. Dong, N. Li, N. Chiamvimonvat, H. Taegtmeyer,K. B. Margulies, B. D. Hammock, F. Despa, Cardioprotection by controllinghyperamylinemia in a humanized diabetic rat model. J. Am. Heart Assoc. 3, e001015(2014).

38. S. Despa, K. B. Margulies, L. Chen, A. A. Knowlton, P. J. Havel, H. Taegtmeyer, D. M. Bers,F. Despa, Hyperamylinemia contributes to cardiac dysfunction in obesity anddiabetes: A study in humans and rats. Circ. Res. 110, 598608 (2012).

39. J. F. Rivera, T. Gurlo, M. Daval, C. J. Huang, A. V. Matveyenko, P. C. Butler, S. Costes,Human-IAPP disrupts the autophagy/lysosomal pathway in pancreatic b-cells: Protectiverole of p62-positive cytoplasmic inclusions. Cell Death Differ. 18, 415426 (2011).

40. J. F. Rivera, S. Costes, T. Gurlo, C. G. Glabe, P. C. Butler, Autophagy defends pancreatic bcells from human islet amyloid polypeptide-induced toxicity. J. Clin. Invest. 124,34893500 (2014).

41. J. Kim, H. Cheon, Y. T. Jeong, W. Quan, K. H. Kim, J. M. Cho, Y.-M. Lim, S. H. Oh, S.-M. Jin,J. H. Kim, M.-K. Lee, S. Kim, M. Komatsu, S.-W. Kang, M.-S. Lee, Amyloidogenicpeptide oligomer accumulation in autophagy-deficient b cells induces diabetes. J. Clin.Invest. 124, 33113324 (2014).

42. N. Shigihara, A. Fukunaka, A. Hara, K. Komiya, A. Honda, T. Uchida, H. Abe, Y. Toyofuku,M. Tamaki, T. Ogihara, T. Miyatsuka, H. J. Hiddinga, S. Sakagashira, M. Koike, Y. Uchiyama,T. Yoshimori, N. L. Eberhardt, Y. Fujitani, H. Watada, Human IAPP-induced pancreatic b celltoxicity and its regulation by autophagy. J. Clin. Invest. 124, 36343644 (2014).

43. R. Emanuel, I. Sergin, S. Bhattacharya, J. N. Turner, S. Epelman, C. Settembre, A. Diwan,A. Ballabio, B. Razani, Induction of lysosomal biogenesis in atheroscleroticmacrophages can rescue lipid-induced lysosomal dysfunction and downstreamsequelae. Arterioscler. Thromb. Vasc. Biol. 34, 19421952 (2014).

44. J. Herrmann, W. D. Edwards, D. R. Holmes, K. L. Shogren, L. O. Lerman, A. Ciechanover,A. Lerman, Increased ubiquitin immunoreactivity in unstable atherosclerotic plaquesassociated with acute coronary syndromes. J. Am. Coll. Cardiol. 40 19191927(2002).

45. W. Jessup, E. L. Mander, R. T. Dean, The intracellular storage and turnover ofapolipoprotein B of oxidized LDL in macrophages. Biochim. Biophys. Acta 1126, 167177(1992).

46. F. Ursini, K. J. A. Davies, M. Maiorino, T. Parasassi, A. Sevanian, Atherosclerosis: Anotherprotein misfolding disease? Trends Mol. Med. 8, 370374 (2002).

47. H. F. Hoff, N. Zyromski, D. Armstrong, J. ONeil, Aggregation as well as chemicalmodification of LDL during oxidation is responsible for poor processing inmacrophages. J. Lipid Res. 34, 19191929 (1993).

48. S. L. Masters, A. Dunne, S. L. Subramanian, R. L. Hull, G. M. Tannahill, F. A. Sharp,C. Becker, L. Franchi, E. Yoshihara, Z. Chen, N. Mullooly, L. A. Mielke, J. Harris, R. C. Coll,K. H. G. Mills, K. H. Mok, P. Newsholme, G. Nuez, J. Yodoi, S. E. Kahn, E. C. Lavelle,L. A. J. ONeill, Activation of the NLRP3 inflammasome by islet amyloid polypeptideprovides a mechanism for enhanced IL-1b in type 2 diabetes. Nat. Immunol. 11, 897904(2010).

49. M. E. Kotas, R. Medzhitov, Homeostasis, inflammation, and disease susceptibility.Cell 160, 816827 (2015).

50. B. Vandanmagsar, Y.-H. Youm, A. Ravussin, J. E. Galgani, K. Stadler, R. L. Mynatt,E. Ravussin, J. M. Stephens, V. D. Dixit, The NLRP3 inflammasome instigates obesity-induced inflammation and insulin resistance. Nat. Med. 17, 179188 (2011).

51. P. Duewell, H. Kono, K. J. Rayner, C. M. Sirois, G. Vladimer, F. G. Bauernfeind, G. S. Abela,L. Franchi, G. Nuez, M. Schnurr, T. Espevik, E. Lien, K. A. Fitzgerald, K. L. Rock,K. J. Moore, S. D. Wright, V. Hornung, E. Latz, NLRP3 inflammasomes arerequired for atherogenesis and activated by cholesterol crystals. Nature 464, 13571361(2010).

52. D. De Nardo, E. Latz, NLRP3 inflammasomes link inflammation and metabolic disease.Trends Immunol. 32, 373379 (2011).

53. R. Stienstra, J. A. van Diepen, C. J. Tack, M. H. Zaki, F. L. van de Veerdonk, D. Perera,G. A. Neale, G. J. Hooiveld, A. Hijmans, I. Vroegrijk, S. van den Berg, J. Romijn,P. C. N. Rensen, L. A. B. Joosten, M. G. Netea, T.-D. Kanneganti, Inflammasome is a centralplayer in the induction of obesity and insulin resistance. Proc. Natl. Acad. Sci. U.S.A.108, 1532415329 (2011).

54. S. L. Masters, E. Latz, L. A. J. ONeill, The inflammasome in atherosclerosis and type 2diabetes. Sci. Transl. Med. 3, 81ps17 (2011).

12 of 17




ag.org/D

ownloaded from

55. H.-M. Lee, J.-J. Kim, H. J. Kim, M. Shong, B. J. Ku, E.-K. Jo, Upregulated NLRP3inflammasome activation in patients with type 2 diabetes. Diabetes 62, 194204 (2013).

56. C.-S. Shi, K. Shenderov, N.-N. Huang, J. Kabat, M. Abu-Asab, K. A. Fitzgerald, A. Sher,J. H. Kehrl, Activation of autophagy by inflammatory signals limits IL-1b production bytargeting ubiquitinated inflammasomes for destruction. Nat. Immunol. 13, 255263(2012).

57. Z. Zhong, A. Umemura, E. Sanchez-Lopez, S. Liang, S. Shalapour, J. Wong, F. He,D. Boassa, G. Perkins, S. R. Ali, M. D. McGeough, M. H. Ellisman, E. Seki, A. B. Gustafsson,H. M. Hoffman, M. T. Diaz-Meco, J. Moscat, M. Karin, NF-kB restricts inflammasomeactivation via elimination of damaged mitochondria. Cell 164, 896910 (2016).

58. T. Kimura, A. Jain, S. W. Choi, M. A. Mandell, K. Schroder, T. Johansen, V. Deretic,TRIM-mediated precision autophagy targets cytoplasmic regulators of innate immunity.J. Cell Biol. 210, 973989 (2015).

59. T. Kimura, A. Jain, S. W. Choi, M. A. Mandell, T. Johansen, V. Deretic, TRIM-directedselective autophagy regulates immune activation. Autophagy 10.1080/15548627.2016.1154254 (2016).

60. T. Kimura, M. Mandell, V. Deretic, Precision autophagy directed by receptor regulators-emerging examples within the TRIM family. J. Cell Sci. 129, 881891 (2016).

61. A. Omenetti, S. Carta, L. Delfino, A. Martini, M. Gattorno, A. Rubartelli, Increased NLRP3-dependent interleukin 1b secretion in patients with familial Mediterranean fever:Correlation with MEFV genotype. Ann. Rheum. Dis. 73, 462469 (2014).

62. A. Acay, M. S. Ulu, A. Ahsen, G. Ozkececi, K. Demir, U. Ozuguz, S. Yuksel, G. Acarturk,Atherogenic index as a predictor of atherosclerosis in subjects with familialMediterranean fever. Medicina 50, 329333 (2014).

63. A. Akdogan, M. Calguneri, B. Yavuz, E. B. Arslan, U. Kalyoncu, L. Sahiner, O. Karadag,I. Ertenli, S. Kiraz, K. Aytemir, D. Akata, L. Tokgozoglu, A. Oto, Are familial MediterraneanFever (FMF) patients at increased risk for atherosclerosis? Impaired endothelial functionand increased intima media thickness are found in FMF. J. Am. Coll. Cardiol. 48, 23512353(2006).

64. E. Aflaki, N. Moaven, D. K. Borger, G. Lopez, W. Westbroek, J. J. Chae, J. Marugan,S. Patnaik, E. Maniwang, A. N. Gonzalez, E. Sidransky, Lysosomal storage and impairedautophagy lead to inflammasome activation in Gaucher macrophages. Aging Cell 15,7788 (2016).

65. H.-W. Chiu, C.-H. Chen, J.-N. Chang, C.-H. Chen, Y.-H. Hsu, Far-infrared promotes burnwound healing by suppressing NLRP3 inflammasome caused by enhanced autophagy.J. Mol. Med. 94, 809819 (2016).

66. Y. Yan, W. Jiang, L. Liu, X. Wang, C. Ding, Z. Tian, R. Zhou, Dopamine controls systemicinflammation through inhibition of NLRP3 inflammasome. Cell 160, 6273 (2015).

67. B. Razani, C. Feng, T. Coleman, R. Emanuel, H. Wen, S. Hwang, J. P. Ting, H. W. Virgin,M. B. Kastan, C. F. Semenkovich, Autophagy links inflammasomes to atheroscleroticprogression. Cell Metab. 15, 534544 (2012).

68. Y.-M. Lim

Documents

METABOLISM 2017 © The Authors, Target acquired: …stke.sciencemag.org/content/sigtrans/10/468/eaag2298.full.pdf · Target acquired: Selective autophagy in cardiometabolic disease