Proteostasis and Aging

perspective Focus on AGI N G

Proteostasis and aging

Susmita Kaushik & Ana Maria Cuervo

Accumulation of intracellular damage is an almost universal of the proteostasis networks in invertebrates and in mammals hallmark of aging. An improved understanding of the systems extend their lifespans and healthspans1,3,4. that contribute to cellular protein quality control has shed light A detailed description of each of the systems that maintains cellular on the reasons for the increased vulnerability of the proteome proteostasis is beyond the scope of this Perspective. Here we focus to stress in aging cells. Maintenance of protein homeostasis, on the recognized importance of the loss of proteostasis in aging. or proteostasis, is attained through precisely coordinated We first describe specific characteristics of the proteostasis networks systems that rapidly correct unwanted proteomic changes. that make these systems vulnerable to the chronic stress that is often Here we focus on recent developments that highlight the associated with aging. Then we touch upon newly identified dimen- multidimensional nature of the proteostasis networks, which sions of the proteostasis networks, beyond the relatively well-known allow for coordinated protein homeostasis intracellularly, cellular cytoplasm, that have transformed the ways in which we in between cells and even across organs, as well as on think about proteostasis. This includes concepts such as organelle how they affect common age-associated diseases when they proteostasis, organ or tissue proteostasis and even organism pro- malfunction in aging. teostasis networks that help to integrate a coordinated proteostatic response throughout the whole body. We argue that this integrated Protein homeostasis, or proteostasis, is assured through the coordi- view of proteostasis is of utmost relevance to further understand the nated action of intricate cellular systems—the proteostasis networks. basis and consequences of a loss of proteostasis in aging and that it Under normal conditions, these systems rapidly sense and rectify has generated considerable interest as a yet-unexplored therapeutic disturbances in the proteome to restore basal homeostasis1. During target for the treatment of age-related diseases. stress, similar systems preserve proteome solubility and functionality by bringing it to an altered point of proteostasis balance that takes into Aging of the components of proteostasis systems consideration the stress-induced cellular changes2. The main players in proteostasis maintenance are chaperones and Although the robustness and adaptability of the proteostasis net- two proteolytic systems, the ubiquitin-proteasome and the lysosome-

Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature works is remarkable, if stressors are chronic, the proteostasis balance autophagy systems (Fig. 1). These components decide the fate of becomes difficult to maintain, and proteotoxicity develops3. With unfolded proteins: whether they will refold into their original stable age, the ability of many cells and organs to preserve proteostasis conformation or whether they will instead be eliminated from the under resting and stress conditions is gradually compromised4. Loss cell through proteolysis8. npg of proteostasis is part of the pathogenesis of many human patholo- gies, including neurodegenerative diseases such as Alzheimer’s disease Chaperones. Chaperones assist proteins through each of the different or Parkinson’s disease3. It is not a coincidence that many of these conformational changes that they undergo during their lifetime, which diseases—generically known as proteinopathies or protein confor- include de novo folding, assembly and disassembly, transport across mational diseases—are regarded as age-related disorders, given that membranes and targeting for degradation9. The need for chaperones the physiological deterioration of the proteostasis networks with age originates from the crowded environment in the cytoplasm as well as is an important aggravating factor in these diseases1,4,5. in the lumen of most organelles. An important aspect of chaperone Numerous lines of evidence support a tight relationship between functioning is the molecules’ ability to integrate multiple cellular cues proteostasis and healthy aging. Although a gradual loss of proteos- in order to decide the fate of cellular proteins that have lost their stable tasis can be detected in most organisms as they age, the longest- conformations. Thus, even for proteins that have experienced the same living species have been shown to have more stable proteomes6 degree of unfolding, chaperones will assist them to either refold or (comprised of cellular proteins that are more resistant to damage), degrade, depending on the feasibility of the cell doing one or the other and, for example, in the case of the long-lived naked mole rat, pro- at that particular time. Factors such as cellular ATP content (because teome stability correlates with enhanced activity in the proteostasis substrate binding and release requires ATP hydrolysis), as well as over- systems7. Furthermore, interventions that modulate the activity all chaperone availability, may also contribute to this final decision. Once a commitment to degradation is made, chaperones often also

Department of Developmental and Molecular Biology, Institute for Aging Studies, decide the proteolytic pathway that each misfolded protein will follow. Albert Einstein College of Medicine, New York, New York, USA. Correspondence For example, HSC70, a constitutive cellular chaperone, can target pro- should be addressed to A.M.C. ([email protected]) teins for degradation. The two systems most commonly used for this Received 20 September; accepted 2 November; published online 8 December breakdown of proteins are the proteasome and autophagy systems. 2015; doi:10.1038/nm.4001 The proteasome, a multi-subunit protease that is most abundant in

1406 VOLUME 21 | NUMBER 12 | DECEMBER 2015 nature medicine perspective

Figure 1 Changes with age in intracellular proteostasis systems. Chaperones and two proteolytic systems, the ubiquitin proteasome UPS activity UPS system (UPS) and autophagy, take care of Ubiquitination maintenance of intracellular proteostasis. Subunit levels Exosome Chaperones (blue, yellow and gray circles) assist Assembly Chaperones de novo synthesized proteins and unfolded proteins to reach their folded stable status. Proteasome If folding is not possible, chaperones target the unfolded protein for degradation by the proteasome (often after ubiquitination) or in lysosomes. Single soluble proteins can reach the CMA activity Multivesicular body or L2A levels lysosomal lumen through a membrane transporter late endosome CMA in chaperone-mediated autophagy (CMA). Once misfolded proteins organize into oligomers HSC70 or insoluble aggregates, the only options for their elimination from the cytosol are either by Protein Folded protein Misfolded protein L2A degradation in lysosomes through macroautophagy synthesis Lysosome (MA) or expulsion outside the cell by means of Macroautophagy small vesicles (exosomes). Red boxes indicate Chaperone changes with age in different steps or components depletion of the intracellular proteostasis networks. Ribosome APG-LYS, autophagosome-lysosome; HSC70, heat-shock cognate protein of 70kDa; L2A, Aggregate lysosome-associated membrane protein type 2A. Autophagosome Aggregates MA activity Cargo recognition Phagophore induction the cytosol—although it can also be detected APG-LYS fusion

in the nucleus—is responsible for the rapid Debbie Maizels/Nature Publishing Group degradation of proteins that are often tagged with covalently attached strands of the small protein ubiquitin10,11. the substrate protein can also interfere with the chaperone’s ability Proteins can also be degraded in lysosomes through a process known to recognize its target. For example, the accumulation with aging as autophagy. Different types of autophagy have been identified of advanced glycation end-products through non-enzymatic modi- (macroautophagy, microautophagy and chaperone-mediated fications on long-lived proteins interferes with normal chaperone autophagy), and the choice of which to use depends on how the function. This type of modification is amenable to repair by enzymes proteins are identified and delivered to lysosomes12 (Fig. 1). The such as methionine sulfoxide reductase, but the abundance of these association of other chaperones and co-chaperones, such as CHIP enzymes decreases with age, further contributing to an accumulation or BAG3, with HSC70 determines HSC70-mediated targeting of of altered proteins that are unrecognizable to chaperones20. Cells proteins through degradation by the proteasome and macroautophagy, may accommodate some of these age-related changes by modifying

Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature respectively. Properties in the cargo protein also contribute to the the pool of chaperones involved in proteostasis. Thus, although the selection of the degradation pathway. For example, although single HSP70 and HSP90 heat-shock protein families of chaperones have unfolded proteins can be eliminated through almost any degradation well-recognized roles in proteome balance under normal conditions, pathway, once multiple proteins organize into oligomeric complexes recent studies in nematodes highlight a prominent function of small npg or aggregates, they can only be degraded by means of a selective form heat-shock proteins in the preservation of proteostasis in aging; of macroautophagy known as aggrephagy13, or chaperone-assisted such small heat-shock proteins trap excess cytosolic proteins into selective autophagy if mediated by HSC70 (ref. 14). The exposure protective aggregates21,22. of specific regions in the amino acid sequence of the protein or of degradation tags, acquired through posttranslational modification Autophagy and proteasome activity. Age-related changes in prote- (i.e., ubiquitination, acetylation, etc.) also determine the degrada- ostasis are not restricted to chaperones. Autophagy and proteasome tion pathway15. activity both decrease with aging3,23, but they do so to a lesser degree Many age-related cellular changes can influence chaperoning in association with healthy aging and longevity, such as in centenar- activities. Thus, poor cellular energetics—a feature characteristic of old ians and long-lived animal species (i.e., the naked mole rat)7. Multiple organisms owing to reduced mitochondrial function and dysregula- types of interventions support the idea that diminishing the proteo- tion of lipid and glucose metabolism, etc.16,17—often limit the amount toxic load during aging can improve lifespan or healthspan (Table 1). of available ATP. These differences in the bioavailability of ATP with Promoting proteasome or autophagy activity via the overexpres- age could be responsible for the repression of the ATP-dependent sion of proteasome subunits or essential autophagy genes increases chaperones and the induction of ATP-independent chaperones that lifespan and confers resistance to stress in Saccharomyces cerevisiae, have been recently identified in the aging brain18. Similarly, the frac- Caenorhabditis elegans and Drosophila melanogaster24,25. Information tion of chaperones that is available for cargo recognition can also on such interventions in mammals is just beginning to emerge. For become markedly reduced with age. For example, sustained chronic example, whole-body overexpression of the essential autophagy gene stressors, such as the continued presence of a metastable protein, have Atg5 in mice revealed anti-aging phenotypes and a lifespan exten- been shown to act as a ‘sink’ for chaperones19. The loss of chaperone sion of about 20% (ref. 26). Interestingly, most of the interventions function and a reduction in availability further aggravate the problems that slow down aging in experimental models are associated with with protein quality control. Undesired age-related modifications in improved proteostasis, and in many instances, these interventions

nature medicine VOLUME 21 | NUMBER 12 | DECEMBER 2015 1407 perspective

demonstrated autophagy-activating pro Table 1 Examples of interventions that increase lifespan by reducing proteotoxicity perties. For example, calorie restriction, Proteostasis system Organism Genetic or pharmacologic intervention Reference rapamycin, metformin, resveratrol and Chaperones C. elegans HSP16 overexpression 91 spermidine, which are well known for C. elegans CR 92 their ability to extend lifespan and/or health- Protein repair enzymes D. melanogaster Protein repair methyltransferase overexpression 93 span, have all been proven to directly activate D. melanogaster Methionine sulfoxide reductase overexpression 94 autophagy, although probably through Ubiquitin-proteasome S. cerevisiae Loss of Ubr2 (negative regulator of UPS transcription) 95 25 different mechanisms . Because at least system C. elegans Rpn6 overexpression 96 some of these interventions could also C. elegans CR 97 affect protein homeostasis through their D. melanogaster FOXO overexpression 98 effects on chaperone levels and protein syn- D. melanogaster Rpn11 overexpression 99 thesis, the contribution of the autophagy- M. musculus IGF1 overexpression 100 activating properties to their overall influ- M. musculus CR 24 ence on lifespan is currently under study. Autophagy C. elegans IGF signaling loss 101 Furthermore, protein degradation—and C. elegans CR 102 proteostasis in general—can also be positively C. elegans Resveratrol 103 influenced by known non-pharmacological C. elegans Spermidine 104 inducers. The activating effect of physical D. melanogaster ATG8a overexpression 105 exercise on the activity of chaperones, on the D. melanogaster Spermidine 104 proteasome and on some types of autophagy D. melanogaster Rapamycin 106 has been well documented27–30. Similarly, M. musculus ATG5 overexpression 26 dietary interventions can also positively influ- Organelle S. cerevisiae Deletion of UPR genes 107 ence proteostasis in most cases, through the S. cerevisiae CR 108 activation of protein degradation or through S. cerevisiae CR 109 changes in the balance between protein syn- C. elegans Increased XBP-1 and IRE-1 in daf-2 mutants 110 Activated ER UPR and mtUPR 111 thesis and degradation. Thus, in addition to C. elegans C. elegans CR 112 starvation—one of the best-known inducers D. melanogaster CR 113 of autophagy—common dietary components D. melanogaster Hsp22 overexpression 114 such as olive oil, vitamins and even coffee D. melanogaster Overexpression of mtUPR genes 115 have also been shown to have stimulatory Abbreviations: CR, caloric restriction; ER, endoplasmic reticulum; UPR, unfolded protein response; properties with respect to different protein mtUPR, mitochondrial UPR quality systems31,32. Compared with the many studies highlighting the possible anti- often been considered to be a uniform entity in which most proteins aging value of improving protein degradation, the potential of improv- undergo similar changes. But is that true? Do all proteins have similar ing protein repair has been poorly explored. One limitation to the susceptibility to cellular aggressors? Are there subgroups of proteins

Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature latter approach is limited knowledge about repair enzymes and how whose members are more prone to losing their stable conformations they change with age. A recent study observed lifespan extension in than are others? Potentially, there are also differences in the time-course flies that overexpressed methionine sulfoxide reductase, the enzyme in which different proteins escape homeostasis. Recent studies in that reduces oxidized methionine20; this provides new support for fur- C. elegans have analyzed these and related questions by monitoring more npg ther exploring the anti-aging value that manipulating levels of these than 5,000 proteins during the worm lifespan. Contrary to predictions, repair enzymes could have. the proteome changes with age were far from ‘subtle’, although they were Interestingly, recent studies suggest that the functional decline of gradual. A pronounced increase in protein abundance seems to be the the proteostasis networks may occur earlier than anticipated in the main reason behind proteome imbalance, loss of stoichiometry and lifespan of an organism. In fact, studies in C. elegans have challenged protein aggregation22. Similar future studies on the aging mammalian the idea that proteostasis failure results from the gradual accumula- proteome should help to clarify the universality of the central role of tion of cellular anomalies, instead pointing to programmed events changes in protein abundance in the loss of proteostasis with age. at an early age as being responsible for the proteostasis collapse later One of the recent advances in our understanding of cytosolic protein life33. At present, such observations are limited to invertebrates. If ostasis is the realization that proteolytic systems function in a coordi- this finding were to hold true in mammals, however, it would become nated manner. Proteasome and autophagy share substrates, effectors important to identify this early time frame and to determine whether and regulators, which allows for continuous cross talk between these it is common for the whole organism or whether there are organ- pathways and compensation for one another36. This compensation is specific time differences; understanding these factors would help advantageous for cells in disease conditions or in aging when one of to correctly tailor interventions that could ameliorate the loss of the proteolytic systems ceases to work properly. For example, a grow- proteostasis with age or in age-related diseases. ing number of reports illustrates how a loss of proteostasis resulting from blockage of the proteasome is prevented by an upregulation of Intracellular proteostasis: cytosolic and organelle protein macroautophagy37. Similarly, cells react to the inhibition of one form quality control of autophagy by activating a different one38,39. These pathways are Cytosolic proteostasis. Most advances in our understanding of cellular never redundant, and the loss of one of them becomes evident once proteostasis and its changes with age arose from the study of the cytosolic cells are exposed to stress; however, at least under basal conditions, proteostasis system34,35. In these studies, the cytosolic proteome has these compensatory activities are able to preserve homeostasis36.

1408 VOLUME 21 | NUMBER 12 | DECEMBER 2015 nature medicine perspective

Such coordinated functioning of the protein-degradation pathways states—handled by the endoplasmic reticulum (ER) creates a need should be kept in mind when interpreting results from the experimen- for robust systems of protein-quality surveillance in this organelle44. tal inhibition of one of these pathways. Most of this compensation A three-arm transcriptional program known as the unfolding protein has been documented in cultured cells, but recent studies support its response (UPR)—activated in response to a loss of proteostasis in occurrence in whole organisms as well. Autophagy is compensatorily the ER—and a dedicated retrotranslocation system, which allows for activated in the rat hippocampus in response to proteasomal inhibi- the rerouting of unfolded luminal and membrane proteins toward tion40; similarly, mice in which chaperone-mediated autophagy in the the cytosol for degradation, together ensure ER homeostasis44. The liver has been blocked display a robust activation of macroautophagy UPR is initiated by three ER transmembrane proteins, IRE1, PERK and the proteasome system, both of which contribute to proteostasis and ATF6, which subsequently induce the expression of several genes in this organ41. However, compensation among proteolytic systems involved in protein processing and maturation in order to counteract seems to be organ-dependent. For example, this compensation of the ER stress and to restore the proteostatic balance (Fig. 2). If the chaperone-mediated autophagy failure by macroautophagy described amount of unfolded proteins in the ER exceeds the ability of the UPR in the liver is not observed in the retina42. In addition, for reasons to return the ER proteome to homeostasis, retrotranslocation is acti- still unknown, aging has a negative effect on the cross talk between vated and this results in the arrival of the unfolded protein pool into proteolytic pathways. For example, in aged mice, the compensatory the cytosol45. The proteasome was thought to be the only degrada- upregulation of some of these proteolytic pathways is lost40,41. The tion pathway involved in this process, but recent studies support the failure with age of signaling pathways such as the insulin-like growth importance of a selective form of autophagy, now termed ER-phagy, factor 1 (IGF1) pathway (in the case of proteasome-autophagy cross for the maintenance of ER homeostasis. In this case, degradation talk)40, or of the transcription factor EB (TFEB) pathway (in cross includes not only unfolded proteins but also whole regions of the ER- talk between macroautophagy and chaperone-mediated autophagy)41, limiting membrane, where the ER-resident protein FAM134B local- have been implicated in the loss of these compensatory abilities in izes and acts as a receptor to facilitate the degradation of the ER46. aged organisms. Interestingly, chronic, moderate blockage of some The tight connections between ER homeostasis and these two subunits of the proteasome improves the ability of this system to cytosolic proteolytic systems make the ER vulnerable to age-related respond to an acute proteotoxic insult43. This rebalancing of prote- changes in the activity of these systems. In addition, sustained ER ostasis in response to moderate blockage is conserved from yeast to stress and/or inappropriate response to this stress are common humans43, but the underlying mechanisms that enable the cross talk characteristics of aging and of many chronic age-related disorders and the contribution of the cross talk to these changes need further (Table 2). This justifies recent efforts to chemically attenuate such investigation.

Organelle proteostasis. Despite the empha- Endoplasmic reticulum sis of a vast part of the literature on cytosolic proteostasis, the importance of protein homeostasis inside organelles and the existence of organelle-specific proteostasis mechanisms are now well accepted (Fig. 2). The high FAM134B IRE1

Figure 2 Organelle proteostasis networks. Schematic of the mechanisms that preserve ATF6 XBP1 Phagophore PERK npg proteostasis in the endoplasmic reticulum (ER) ERAD and in mitochondria. The ER and mitochondria Autophagosome ER-phagy can undergo degradation as a whole organelle ATF4 Ubiquitin through specialized forms of autophagy, ER-phagy (assisted by the recently identified UPR FAM134B protein) and mitophagy, respectively. In addition, these organelles have their own ER proteostasis systems that are activated by chaperones the presence of unfolded proteins. In the ER, these unfolded proteins activate the Lysosome Nucleus unfolded protein response (UPR) that has three arms (mediated by PERK and ATF4, IRE1 and XBP1, and by ATF6). Activation of Proteasome the UPR attenuates protein translation and

enhances the expression of ER chaperones. Mitophagy If this response is not sufficient, unfolded proteins are retrotranslocated for degradation mtUPR in the cytoplasm by the proteasome (ERAD). ClpXP The mitochondrial UPR (mtUPR) is similarly protease activated in response to proteotoxic stress to Mitochondrial enhance chaperone content in mitochondria and chaperones to attenuate translation. Unfolded mitochondrial Autophagosome proteins are cleaved by the ClpXP protease Mitochondrion into small peptides that upon translocation Phagophore

into the cytosol activate the mtUPR. Debbie Maizels/Nature Publishing Group

nature medicine VOLUME 21 | NUMBER 12 | DECEMBER 2015 1409 perspective

sustained ER stress. Most of the small molecules that are currently under enhancement of ER-phagy. Given that overexpression of FAM134B development or in testing for the treatment of age-related disorders causes ER fragmentation and lysosomal degradation46, FAM134B such as neurodegeneration or metabolic disorders target components could become a suitable target of specific ER-phagy activators. of the UPR47,48. However, recent findings in worms have revealed that The impact of aging on recently identified ER proteostasis mecha- the UPR can also be activated by lipid disturbances alone (without nisms remains largely unexplored. For example, ER proteostasis can disturbed proteostasis)49, opening the tantalizing possibility of using be induced—even before the UPR is activated—via the delivery of dietary changes to modulate this response in situations of ER proteo- misfolded proteins to the cell surface, after which they can be inter- toxicity. Under these conditions, stimulation of autophagy could also nalized and degraded in lysosomes50. A better understanding of the be valuable for alleviating the overloaded stress. Until now, however, molecular effectors of this rapid ER stress–induced export (RESET) most compounds undergoing testing have resulted in global activation could provide additional targets to alleviate the loss of ER proteostasis of the autophagy process rather than the more desirable selective in aging and some of the chronic age-related diseases.

Table 2 Changes in components of the proteostasis networks with age and in some age-related diseases Proteostasis system Change Aging condition or age-related disease Cytosol Chaperones Malfunctions Chaperone depletion Huntington’s disease, Parkinson’s disease, ataxia, amyotrophic lateral sclerosis Change in chaperone type Brain aging, Alzheimer’s disease, Parkinson’s disease, Huntington’s disease Specific mutations α-Crystallin Early onset cataracts, desmin-related myopathy, cardiomyopathy DNAJB6 Dominantly inherited myopathy HSC70 Cardiovascular disease HSJ1 Distal hereditary motor neuropathy HSP22 Charcot-Marie-Tooth disease HSP27 Charcot-Marie-Tooth disease Sacsin Spastic ataxia Ubiquitin-proteasome system Malfunctions Defective proteasome activity Aging, Parkinson’s disease, Huntington’s disease Proteasome damage Aging Proteasome assembly changes Aging, prion diseases, autoinflammatory disease Ubiquitination defects Aging lens, retinal degeneration, muscle degeneration, Huntington’s disease Specific mutations Ataxin-3 Machado-Joseph disease PSMB8 Nakajo-Nishimura syndrome Ubiquilin-2 Amyotrophic lateral sclerosis UCHL1 Parkinson’s disease Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature VCP/p97 (ERAD) Paget’s disease, Frontotemporal dementia Autophagy Malfunctions Inefficient induction Aging npg Reduced autophagosome clearance Aging, Alzheimer’s disease, Parkinson’s disease Defective cargo recognition Huntington’s disease Specific mutations ATG16L1 Crohn’s disease LAMP2A Cardiovascular disease, myopathy p62 Amyotrophic lateral sclerosis, Paget’s disease Parkin (mitophagy) Parkinson’s disease PINK1 (mitophagy) Parkinson’s disease Presenilin-1 Familial Alzheimer’s disease Organelles Intrinsic proteostasis mechanisms Malfunctions ER stress response Alzheimer’s disease, Parkinson’s disease, cardiovascular disease, metabolic syndrome, fibrosis, type 2 diabetes, familial insomnia, atherosclerosis, arthritis, cancer ERAD Tauopathies, polyglutamine diseases, rheumatoid arthritis mtUPR Premature aging phenotype, shortened lifespan Specific mutations PDI (ER) Amyotrophic lateral sclerosis VCP/p97 (ER) Paget’s disease, Frontotemporal dementia HSP60 (mitochondria) Hereditary spastic paraplegia, neurodegenerative disorder linked to brain hypomyelination and leukodystrophy Mortalin (mitochondria) Parkinson’s disease Parkin (mitophagy) Parkinson’s disease

1410 VOLUME 21 | NUMBER 12 | DECEMBER 2015 nature medicine perspective

Remarkable progress has also been made in understanding the are and how these pathways work remain to be elucidated. Also poorly mechanisms behind mitochondrial proteostasis (Fig. 2). Maintenance understood is the occurrence of protein breakdown in the nucleo- of mitochondrial proteostasis is particularly challenging because of plasm. Even though the proteasome has been shown to be responsible the abundance of reactive oxygen species (ROS) in this compartment. for the degradation of oxidatively damaged histones and of surplus This probably explains why damaged portions of the mitochondrial ribosomal proteins62,63, it is not clear whether that degradation occurs network that succumb to severe proteotoxicity are often segregated in the nucleus or only once these proteins have been transported from the network and eliminated as a whole organelle. The most com- to the cytosol. In the case of pathogenic proteins that aggregate in mon way of getting rid of malfunctioning mitochondria is through the nucleoplasm, there is strong evidence supporting proteasome autophagic degradation (mitophagy)51. The molecular complexity of degradation in situ64. Currently, little is known about the possible this process, the variety of mitophagy types that coexist in the cell changes in nuclear proteostasis during aging. Are there subsets and the recently identified alternative ways by which the lysosomal of nuclear proteins—as is the case in the cytosol—that are more system can contribute to mitochondria degradation independent vulnerable to a loss of proteostasis in aging? Do the sensing mecha- of the autophagy machinery51 all highlight our relatively limited nisms of nuclear proteotoxicity fail with age? Could part of the loss knowledge about mitochondrial clearance from the cell. However, of proteostasis with age in the cytoplasm originate from an increase numerous studies have reported the malfunctioning of mitophagy and in the spillover from the nucleus of altered proteins? thus mitochondrial clearance in aging and in age-related disorders In this respect, little is known about the rules that govern communi- (Table 2), in particular those involving neurodegeneration, making cation between the cytoplasm and organelle proteostasis networks or the study of these processes a high priority in these fields. whether they change in old organisms. Recently, the single-stranded Although mitochondrial quality control is often tightly linked to the DNA-binding mitochondrial protein SSBP1 was shown to translocate proteolytic cytosolic systems, this organelle, similarly to the ER, con- to the nucleus in response to heat shock to modulate the heat shock tains its own set of chaperones and proteases. The mitochondrial UPR factor protein 1 (HSF1)-dependent expression of cytosolic, nuclear and (mtUPR) is activated when unfolded or misfolded proteins accumulate mitochondrial chaperones—the first known example of intercompart- in the mitochondria and aggregate52,53. Recent studies have shown that mental regulation of proteostasis65. We still do not know whether there communication of the mtUPR with the nucleus can induce the expres- is a hierarchical order whereby preservation of proteostasis in some sion of nuclear genes encoding for mitochondrial proteins in response compartments is prioritized over others, or whether poorly handled to impaired proteostasis54,55. The molecular players involved in mtUPR proteotoxicity in one cellular compartment is always systematically are the subjects of intensive investigation in aging research because the fed into another in a joint attempt to restore overall cellular prote- modulation of the mtUPR has been shown to have lifespan-extension ostasis. Additionally, mechanisms that preserve quality control and properties in multiple organisms56. Some of the proteins connected proteostasis in other cellular compartments have been poorly explored. with the mtUPR, such as the mitochondrial deacetylase SIRT3, are Turnover of whole organelles, rather than the selective degradation of also well-known lifespan modulators whose deficiencies have been only specific organelle–resident proteins, may be favored for cellular linked to a higher incidence of age-related diseases (neurodegen- compartments such as the Golgi, endosomes or even lysosomes, for eration, metabolic syndrome and cancer) (Table 2)57. Furthermore, which full sequestration and degradation through a form of selective age-related changes in some of the recently identified components of autophagy (lysophagy) has been described66. The dependence of these the mtUPR have also been described. For example, the expression of processes on autophagic effectors may make them subject to the same 23

Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature SIRT7, a component of the regulatory branch of the mtUPR, is reduced age-dependent malfunctioning described for autophagy . in aged hematopoietic stem cells, and its overexpression in this setting improved the regenerative capability of these cells58. Intercellular proteostasis: integrating organ and tissue responses Quality control in the nuclear compartment is not as well Most studies to date have approached the analysis of age-related npg understood. Historically, the nucleus has been viewed as less vul- changes in proteostasis as a cell-autonomous problem. However, nerable to proteotoxicity than other organelles because the complex growing evidence supports the existence of regulatory, intercellular nuclear envelope and tightly regulated nuclear transport systems limit proteostasis networks that help in the coordinated response of tissues contact between nuclear material and the crowded pro-aggregating and organs to proteotoxic insults. cytoplasm milieu. However, it is now well accepted that chaperones, Whereas the effectors of intercellular proteostasis, chaperones and ubiquitin ligases and the proteasome present in the nucleoplasm all proteolytic systems are, for the most part, the same ones described for contribute to nuclear proteostasis59. intracellular proteostasis, the key question for coordinated proteosta- Interestingly, some of the proteins that contribute to nuclear pro- sis between cells is how signals of proteostasis stress are transmitted tein quality reside in the nucleus whereas others are shuttled from from one cell to another. Some of the communication channels may the cytosol or ER into the nucleus in times of proteotoxic stress60. be the same ones normally used by cells for intercellular communi- This last group transiently co-localizes to nuclear aggregates of mutant cation. For example, the coupling of intercellular gap junctions to huntingtin, ataxin-1 or TAR DNA-binding protein 43 (TDP43)— the activation of autophagy has been demonstrated by showing that proteins associated, respectively, with Huntington’s disease, spinoc- connexins—the main components of gap junctions—also function as erebellar ataxia 1 and amyotrophic lateral sclerosis61. This transport endogenous inhibitors of autophagy67. of chaperones to the nucleus in times of stress points to a sensing However, intercellular proteostasis is not limited to the transmission mechanism that generates a signal to induce entry into the nucleus of stress signals across cells to boost activation of their intrinsic of protein quality components to assist in restoring proteostasis. It proteostasis machinery. Components of the proteostasis network, is possible that this signal also attracts modifiers that tag proteins and even toxic proteins, can be transferred from one cell to another for degradation because certain ubiquitin ligases, such as UHRF2 with the ultimate goal of preserving whole-organ proteostasis (Fig. 3). and PML4, are known to recognize mutant polyglutamine proteins For example, studies in D. melanogaster expressing aggregation- localized to the nucleus. What the nuclear sensors of proteotoxicity prone proteins found that experimentally increasing chaperones

nature medicine VOLUME 21 | NUMBER 12 | DECEMBER 2015 1411 perspective

Figure 3 Schematic of possible mediators of intercellular proteostasis: exosomes and nanotubes. It is feasible that bidirectional transfer allows for the replenishment of proteostasis effectors from the cell with more robust proteostasis to one with weaker proteostasis, Proteasome Lysosomes Chaperones and transport of protein aggregates and damaged proteins in the Nanotubes reverse direction.

(Hsp70 and Hsp40) in a group of cells improved proteostasis in other cells in the same tissue, thereby demonstrating non-cell-autonomous maintenance of proteostasis68. Exosomes have emerged as important organelles in cellular Exosomes intercommunication69. Originating from the invagination of the endosomal membrane, these small vesicles trap samples of the cytosol that are then released to the extracellular environment upon the fusion of multivesicular endosomes with the plasma membrane. Exosomes can act as vehicles for the exchange of chaperones and maybe for other proteostasis effectors (Fig. 3). For example, chaperones can be

detected in exosomes; their abundance increases under conditions of Weak Strong proteotoxicity; and adding chaperone-containing exosomes to cul- proteostasis proteostasis tured cells expressing aggregation-prone proteins decreases inclusion networks networks body formation, demonstrating that exosomes can be an efficient Protein aggregates Amyloid 68 mechanism for the transfer of chaperones between cells . Debbie Maizels/Nature Publishing Group It would be interesting to evaluate how the recently described decline of multivesicular endosomal proteostasis with age70 could of intracellular proteostasis mechanisms—a common development affect the transferring abilities of exosomes in old organisms. In this in these diseases—requires future investigation. It is still not known respect, recent studies have compared neuronally derived exosomes whether this mechanism of propagation increases in physiological from individuals diagnosed with Alzheimer’s disease before the onset aging and contributes to the whole organ’s disturbance of proteostasis, of symptoms and up to 10 years after diagnosis with those from non- but the role of prion-like propagation in neurodegenerative diseases affected individuals. Results showed consistently higher exosomal of aging, including in Alzheimer’s disease and other tauopathies, as levels of ubiquitinated proteins and lysosomal proteins, but lower well as in Parkinson’s, Creutzfeldt-Jakob and Lou Gehrig’s diseases, as abundance of HSC70 in the exosomes of those with Alzheimer’s well as in tauopathies, is gaining acceptance. Because the majority of disease71. Studies such as this one highlight the possible value of these neurodegenerative diseases are sporadic and late in onset, it has exosomal proteins as biomarkers of disease spreading, but they also been proposed that prion accumulation occurs throughout life, and reinforce the need for a better understanding of exosome biology as once it exceeds some critical threshold, prion spreading triggers global a way to elucidate changes in proteostasis across cells with age and in neurological dysfunction78. Although still not explored in detail, it is age-related diseases. also possible that nanotubes participate in the spreading of protein

Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature Another mechanism of the intercellular transfer of materials that is aggregates; they have been implicated in transferring exogenous and gaining considerable attention is the use of tunneling nanotubes that endogenous prions between infected and naïve neuronal cells, and, allow for the direct exchange of cellular components between non- in Creutzfeldt-Jakob disease79, from bone marrow–derived dendritic adjacent cells that are separated by distances greater than 100 µm. cells in the periphery to primary neurons. npg The material exchanged through nanotubes ranges from regulatory Overall, this transfer of chaperones, proteases and protein aggregates miRNAs to whole organelles, such as lysosomes, mitochondria, endo- across cells may result from the heterogeneity of the robustness of the somes and lipid droplets (Fig. 3)72–74. Because tunneling nanotubes proteostasis systems inside organs. Cell-type and regional differences between neural stem cells and brain microvascular endothelial cells in the abundance and activity of the proteasome or the autophagy- have been shown to provide neuroprotection75, it is appealing to think lysosomal system have been documented in different organs and that the transfer of, for example, lysosomes from one cell to another have been correlated with variances in their proteostasis capabilities. could help to sustain high levels of autophagy without the need for These regional differences in proteostasis could determine regional lysosomal biogenesis in the stressed cell. susceptibility to disease. For example, the activity of chaperone- Besides the transfer of components of the proteostasis networks mediated autophagy, a type of autophagy shown to contribute to the across cells, the idea that pathogenic proteins could transfer across degradation of α-synuclein, has been found to be lower in aggregation- cells in a prion-like manner has gained momentum as a possible prone regions of the brain in both control animals and animals model for the propagation of neurodegenerative diseases. In this expressing the pathogenic form of α-synuclein80. In this context, model, aggregated proteins are released from the affected cells and intercellular proteostasis could initially serve to sustain proteosta- travel to nearby healthy cells where they act as seeds for aggregate sis in the weakest regions by transferring aggregates out of the brain formation (Fig. 3). Several aggregation-prone proteins involved in region toward regions with a more robust proteostasis and trafficking neurodegenerative diseases (α-synuclein, tau, β-amyloid and super- the components of the proteostasis networks in the reverse direction oxide dismutase 1) have been detected in exosomes76,77. Consequently, (Fig. 3). The growing repertoire of fluorescent and radiometric probes exosomes could have tremendous potential as biomarkers for the available for the study of different proteostasis events (protein folding, diagnosis and prognosis of these diseases. Whether the spreading of proteasome and autophagy activities, the UPR response, etc.) in real protein aggregates via exosomes is a primary feature in these diseases time, and the newly developed single-cell image tracking systems81,82 or a ‘way out’ for aggregated proteins that are secondary to blockage should allow researchers to perform detailed time-course analysis of

1412 VOLUME 21 | NUMBER 12 | DECEMBER 2015 nature medicine perspective

aggregopathies in the near future. This type of analysis could help to a protein that mediates neurotransmitter release in neurons, was identify tissue areas of proteostasis weakness, possible changes with one of the first components of neuroendocrine signaling to be age in the proteostatic strength of these areas and the contribution implicated in distant UPR activation in worms84. As expected given of nearby organ regions with better proteostasis to their response to the role of such a mediator, UNC-13 mutant worms also displayed toxic insults. Also pending is a better understanding of the impact that reduced longevity. Secretion of a UPR component, the ER chaperone aging could have on the mechanisms that mediate intercellular prote- ERDJ3, has also been shown to contribute to extracellular proteos- ostasis. Do transferring mechanisms, kinetics of transference or the tasis86, although whether or not this chaperone could also influ- amount of chaperones and proteases transferred between cells change ence proteostasis in distant tissues requires further investigation. with age? Do aging cells frequently use outside ‘dumping’ of protein Interestingly, the search for hormone-like factors that mediate aggregates in response to overloading of their intrinsic proteostasis proteostasis at a distance comes at a time when the aging research machinery? Could exosome-based transfer be effective in boosting community is fully invested in the search for serum circulating factors proteostasis in old organs? with anti-aging properties. The concept originated from early studies showing the rejuvenating effect—at the stem cell level, at least—of Tele-proteostasis: integrating distant proteostasis networks administering serum from young mice to aged mice87 and has been The realization that proteostasis networks in different organs recently extended as a protective mechanism against age-related neu- coordinate with each other and that changes in one of them affect rodegeneration88. Now that some of the circulating factors respon- the functioning of the others has been one of the most exciting sible for these beneficial properties have been identified, a burning developments in the field of proteostasis in recent years. This ‘tele- question is whether part of their positive effect is exerted through proteostasis’ (as we term proteostasis coordinated from a distance the regulation of proteostasis and whether they could be some of through a combination of cell autonomous and non-autonomous the sought-after mediators of tele-proteostasis. Furthermore, caloric mechanisms) allows, for example, for the induction of a systemic restriction, an effective anti-aging intervention, has been extensively heat-shock response by activating the heat-shock protein response reported to improve proteostasis, and serum from caloric-restricted in neuronal cells alone62. animals is known to delay senescence and improve stress resistance One of the first observations of interorgan regulation of proteos- of cultured cells89. Both of these facts support the idea that circulat- tasis came from studies in worms expressing, in an inducible man- ing factors under these conditions contribute to the regulation of ner, proteins prone to misfolding that trigger the expression of the multiple-organ proteostasis. Whether the circulating mediators of HSP90 chaperone to assist with protein refolding. Interestingly, such tele-proteostasis are universal and independent of the cellular condi- HSP90 expression was observed not only in the cells expressing the tions, or whether specific subsets of factors are used in response to misfolded protein—in that case, muscle cells—but also in cells that different insults, however, requires further investigation. did not express the misfolded protein83. In support of the idea that Given that this is the most newly understood aspect of proteosta- this distal proteostasis response was of value for handling protein sis, many additional interesting questions remain under the general misfolding in the affected cells, expression of HSP90 when restricted umbrella of factors in the circulation that influence aging. Is there a to neuronal or intestinal cells was still effective in resolving the aggre- role in interorgan proteostasis for inflammatory molecules, which are gation problems in muscle cells83. These studies provided experimen- produced in abundance in chronic inflammatory processes in aging? tal support that modulating the proteostasis responses in one organ Old organisms also undergo major metabolic changes; do circulating 83

Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature elicits a similar response in a distant tissue . nutrients influence the coordination of proteostatic networks in dif- Shortly afterwards, it was demonstrated that tele-proteostasis ferent organs? And of course, the question that we all are considering: applies to both cytosolic and organelle proteostasis, and also, that can manipulation of proteostasis in one organ be used for therapeutic the robustness of tele-proteostasis is an important determinant of purposes in a distant one? It would be remarkable if some of the severe npg longevity. In worms that have been genetically modified to have a neurodegenerative diseases that affect the elderly could be treated diminished systemic UPR (proven to be short lived), restoration of with agents that target proteostasis in peripheral organs. the missing UPR component in neurons and intestinal cells only was sufficient to increase the worms’ longevity84. In fact, neuronal res- Conclusions toration of the missing UPR component activated the UPR in distal Overwhelming evidence supports the maintenance of cellular pro- intestinal cells and promoted ER stress resistance in aged worms. teostasis as one of the key processes in ensuring longevity. A better Although most of the discoveries in non-cell-autonomous prote- understanding in recent years of the intracellular systems that contrib- ostasis have been made in invertebrates, evidence of a similar proc- ute to protein quality control has formed the conceptual framework ess in mammals is starting to emerge. Constitutive expression of a for several successful attempts to improve cellular proteostasis to delay component of the UPR in the hypothalamic neurons in mice also signs of aging and the progression of age-related diseases in some activated expression of this component in the liver, supporting cou- experimental models. Although many proof-of-principle interventions pling between ER stress in neurons and hepatocytes85. In this case, the have so far relied on genetic manipulations, which are of questionable consequences of such tele-proteostasis go beyond the maintenance of applicability in elderly patients, there are ongoing investigations using proteome stability in the distant organ; they also include important chemical targeting of chaperones or the proteolytic systems in the con- changes in hepatic metabolism that resonate in organismal energet- text of adult-onset neurodegeneration90. However, the pharmacopeia ics. Thus, ER stress in hypothalamic neurons is coupled to increased available for proteostasis modulation is still limited. The identification hepatic insulin sensitivity and also to suppressed glucose production of new types of proteostasis (intercellular and interorgan) should help in this organ85. to expand the number of possible targets with therapeutic potential. Since the discovery of this cell non-autonomous activation of the Furthermore, consideration should be given to the effects of non- chaperone response and of the UPR and their links to longevity, a pharmacological interventions that are known to affect aging, such as race has begun to identify the systemic mediators involved. UNC-13, exercise, diet, social interactions, behavior modifiers, etc.

nature medicine VOLUME 21 | NUMBER 12 | DECEMBER 2015 1413 perspective

As a rapidly evolving field, proteostasis is in need of technological 29. He, C. et al. Exercise-induced BCL2-regulated autophagy is required for muscle glucose homeostasis. Nature 481, 511–515 (2012). advances that allow for better read-outs. Different probes are now 30. Jamart, C. et al. Modulation of autophagy and ubiquitin-proteasome pathways available for studying separate proteostasis events, but to gain an during ultra-endurance running. J. Appl. Physiol. 112, 1529–1537 (2012). integrated understanding of organism proteostasis and its changes 31. Katsiki, M., Chondrogianni, N., Chinou, I., Rivett, A.J. & Gonos, E.S. The olive constituent oleuropein exhibits proteasome stimulatory properties in vitro and with age, we need methods that enable real-time and organ-spanning confers life span extension of human embryonic fibroblasts. Rejuvenation Res. monitoring of proteostasis in whole organisms. Beyond technological 10, 157–172 (2007). advances, the other important change—fortunately, already under 32. Salomone, F. et al. Coffee enhances the expression of chaperones and antioxidant proteins in rats with nonalcoholic fatty liver disease. Transl. Res. 163, 593–602 way—is a conceptual adjustment to the idea that proteostasis is not a (2014). cellular problem and that ‘it takes an organism’ to maintain it. 33. Ben-Zvi, A., Miller, E.A. & Morimoto, R.I. Collapse of proteostasis represents an early molecular event in Caenorhabditis elegans aging. Proc. Natl. Acad. Sci. USA 106, 14914–14919 (2009). COMPETING FINANCIAL INTERESTS 34. Iram, A. & Naeem, A. Protein folding, misfolding, aggregation and their The authors declare no competing financial interests. implications in human diseases: discovering therapeutic ways to amyloid- associated diseases. Cell Biochem. Biophys. 70, 51–61 (2014). Reprints and permissions information is available online at http://www.nature.com/ 35. Kim, Y.E., Hipp, M.S., Bracher, A., Hayer-Hartl, M. & Hartl, F.U. Molecular chaperone functions in protein folding and proteostasis. 82, reprints/index.html. Annu. Rev. Biochem. 323–355 (2013). 36. Park, C. & Cuervo, A.M. Selective autophagy: talking with the UPS. Cell Biochem. 1. Vilchez, D., Saez, I. & Dillin, A. The role of protein clearance mechanisms in Biophys. 67, 3–13 (2013). organismal ageing and age-related diseases. Nat. Commun. 5, 5659 (2014). 37. Korolchuk, V.I., Menzies, F.M. & Rubinsztein, D.C. Mechanisms of cross-talk 2. Roth, D.M. & Balch, W.E. Modeling general proteostasis: proteome balance in between the ubiquitin-proteasome and autophagy-lysosome systems. FEBS Lett. health and disease. Curr. Opin. Cell Biol. 23, 126–134 (2011). 584, 1393–1398 (2010). 3. Morimoto, R.I. & Cuervo, A.M. Proteostasis and the aging proteome in health and 38. Kaushik, S., Massey, A., Mizushima, N. & Cuervo, A.M. Constitutive activation of disease. J. Gerontol. A Biol. Sci. Med. Sci. 69 (suppl. 1), S33–S38 (2014). chaperone-mediated autophagy in cells with impaired macroautophagy. Mol. Biol. 4. Labbadia, J. & Morimoto, R.I. The biology of proteostasis in aging and disease. Cell 19, 2179–2192 (2008). Annu. Rev. Biochem. 84, 435–464 (2015). 39. Massey, A.C., Kaushik, S., Sovak, G., Kiffin, R. & Cuervo, A.M. Consequences of 5. Labbadia, J. & Morimoto, R.I. Proteostasis and longevity: when does aging really the selective blockage of chaperone-mediated autophagy. Proc. Natl. Acad. Sci. begin? F1000Prime Rep. 6, 7 (2014). USA 103, 5805–5810 (2006). 6. Treaster, S.B. et al. Superior proteome stability in the longest lived animal. 40. Gavilán, E. et al. Age-related dysfunctions of the autophagy lysosomal pathway Age (Dordr) 36, 9597 (2014). in hippocampal pyramidal neurons under proteasome stress. Neurobiol. Aging 36, 7. Pérez, V.I. et al. Protein stability and resistance to oxidative stress are determinants 1953–1963 (2015). of longevity in the longest-living rodent, the naked mole-rat. Proc. Natl. Acad. 41. Schneider, J.L. et al. Loss of hepatic chaperone-mediated autophagy accelerates Sci. USA 106, 3059–3064 (2009). proteostasis failure in aging. Aging Cell 14, 249–264 (2015). 8. Tyedmers, J., Mogk, A. & Bukau, B. Cellular strategies for controlling protein 42. Rodríguez-Muela, N. et al. Balance between autophagic pathways preserves retinal aggregation. Nat. Rev. Mol. Cell Biol. 11, 777–788 (2010). homeostasis. Aging Cell 12, 478–488 (2013). 9. Feldman, D.E. & Frydman, J. Protein folding in vivo: the importance of molecular 43. Tsvetkov, P. et al. Compromising the 19S proteasome complex protects cells from chaperones. Curr. Opin. Struct. Biol. 10, 26–33 (2000). reduced flux through the proteasome. eLIFE 4 doi:10.7554/eLife.08467 10. Navon, A. & Ciechanover, A. The 26 S proteasome: from basic mechanisms to (2015). drug targeting. J. Biol. Chem. 284, 33713–33718 (2009). 44. Walter, P. & Ron, D. The unfolded protein response: from stress pathway to 11. Tanaka, K. The proteasome: overview of structure and functions. Proc. Jpn. Acad. homeostatic regulation. Science 334, 1081–1086 (2011). Ser. B Phys. Biol. Sci. 85, 12–36 (2009). 45. Brodsky, J.L. Cleaning up: ER-associated degradation to the rescue. Cell 151, 12. Kaushik, S. & Cuervo, A.M. Chaperones in autophagy. Pharmacol. Res. 66, 1163–1167 (2012). 484–493 (2012). 46. Khaminets, A. et al. Regulation of endoplasmic reticulum turnover by selective 13. Lamark, T. & Johansen, T. Aggrephagy: selective disposal of protein aggregates autophagy. Nature 522, 354–358 (2015). by macroautophagy. Int. J. Cell Biol. 2012, 736905 (2012). 47. Moreno, J.A. et al. Oral treatment targeting the unfolded protein response prevents 14. Arndt, V. et al. Chaperone-assisted selective autophagy is essential for muscle neurodegeneration and clinical disease in prion-infected mice. Sci. Transl. Med.

Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature maintenance. Curr. Biol. 20, 143–148 (2010). 5, 206ra138 (2013). 15. Stolz, A., Ernst, A. & Dikic, I. Cargo recognition and trafficking in selective 48. Cao, S.S. & Kaufman, R.J. Targeting endoplasmic reticulum stress in metabolic autophagy. Nat. Cell Biol. 16, 495–501 (2014). disease. Expert Opin. Ther. Targets 17, 437–448 (2013). 16. Ma, Y. & Li, J. Metabolic shifts during aging and pathology. Compr. Physiol. 5, 49. Hou, N.S. et al. Activation of the endoplasmic reticulum unfolded protein response 667–686 (2015). by lipid disequilibrium without disturbed proteostasis in vivo. Proc. Natl. Acad. npg 17. Ritz, P. & Berrut, G. Mitochondrial function, energy expenditure, aging and insulin Sci. USA 111, E2271–E2280 (2014). resistance. Diabetes Metab. 31 (spec. no. 2), 5S67–5S73 (2005). 50. Satpute-Krishnan, P. et al. ER stress-induced clearance of misfolded GPI-anchored 18. Brehme, M. et al. A chaperome subnetwork safeguards proteostasis in aging and proteins via the secretory pathway. Cell 158, 522–533 (2014). neurodegenerative disease. Cell Rep. 9, 1135–1150 (2014). 51. Lemasters, J.J. Variants of mitochondrial autophagy: Types 1 and 2 mitophagy 19. Yu, A. et al. Protein aggregation can inhibit clathrin-mediated endocytosis and micromitophagy (Type 3). Redox Biol. 2, 749–754 (2014). by chaperone competition. Proc. Natl. Acad. Sci. USA 111, E1481–E1490 52. Heo, J.M. et al. A stress-responsive system for mitochondrial protein degradation. (2014). Mol. Cell 40, 465–480 (2010). 20. Vanhooren, V. et al. Protein modification and maintenance systems as biomarkers 53. Braun, R.J. et al. Accumulation of basic amino acids at mitochondria dictates of ageing. Mech. Ageing Dev. doi:10.1016/j.mad.2015.03.009 (2015). the cytotoxicity of aberrant ubiquitin. Cell Rep. 10, 1557–1571 (2015). 21. Morrow, G., Samson, M., Michaud, S. & Tanguay, R.M. Overexpression of the 54. Jovaisaite, V. & Auwerx, J. The mitochondrial unfolded protein response- small mitochondrial Hsp22 extends Drosophila life span and increases resistance synchronizing genomes. Curr. Opin. Cell Biol. 33, 74–81 (2015). to oxidative stress. FASEB J. 18, 598–599 (2004). 55. Haynes, C.M. & Ron, D. The mitochondrial UPR - protecting organelle protein 22. Walther, D.M. et al. Widespread proteome remodeling and aggregation in aging homeostasis. J. Cell Sci. 123, 3849–3855 (2010). C. elegans. Cell 161, 919–932 (2015). 56. Jensen, M.B. & Jasper, H. Mitochondrial proteostasis in the control of aging and 23. Rubinsztein, D.C., Marino, G. & Kroemer, G. Autophagy and aging. Cell 146, longevity. Cell Metab. 20, 214–225 (2014). 682–695 (2011). 57. McDonnell, E., Peterson, B.S., Bomze, H.M. & Hirschey, M.D. SIRT3 regulates 24. Chondrogianni, N., Georgila, K., Kourtis, N., Tavernarakis, N. & Gonos, E.S. 20S progression and development of diseases of aging. Trends Endocrinol. Metab. 26, proteasome activation promotes life span extension and resistance to proteotoxicity 486–492 (2015). in Caenorhabditis elegans. FASEB J. 29, 611–622 (2015). 58. Mohrin, M. et al. Stem cell aging. A mitochondrial UPR-mediated metabolic 25. Madeo, F., Zimmermann, A., Maiuri, M.C. & Kroemer, G. Essential role for checkpoint regulates hematopoietic stem cell aging. Science 347, 1374–1377 autophagy in life span extension. J. Clin. Invest. 125, 85–93 (2015). (2015). 26. Pyo, J.O. et al. Overexpression of Atg5 in mice activates autophagy and extends 59. Shibata, Y. & Morimoto, R.I. How the nucleus copes with proteotoxic stress. Curr. lifespan. Nat. Commun. 4, 2300 (2013). Biol. 24, R463–R474 (2014). 27. Morton, J.P., Kayani, A.C., McArdle, A. & Drust, B. The exercise-induced stress 60. Andersen, J.S. et al. Nucleolar proteome dynamics. Nature 433, 77–83 (2005). response of skeletal muscle, with specific emphasis on humans. Sports Med. 39, 61. Janer, A. et al. PML clastosomes prevent nuclear accumulation of mutant ataxin- 643–662 (2009). 7 and other polyglutamine proteins. J. Cell Biol. 174, 65–76 (2006). 28. Ulbricht, A. et al. Induction and adaptation of chaperone-assisted selective 62. Ullrich, O. et al. Poly-ADP ribose polymerase activates nuclear proteasome to autophagy CASA in response to resistance exercise in human skeletal muscle. degrade oxidatively damaged histones. Proc. Natl. Acad. Sci. USA 96, Autophagy 11, 538–546 (2015). 6223–6228 (1999).

1414 VOLUME 21 | NUMBER 12 | DECEMBER 2015 nature medicine perspective

63. Lam, Y.W., Lamond, A.I., Mann, M. & Andersen, J.S. Analysis of nucleolar protein 90. Pratt, W.B., Gestwicki, J.E., Osawa, Y. & Lieberman, A.P. Targeting Hsp90/Hsp70- dynamics reveals the nuclear degradation of ribosomal proteins. Curr. Biol. 17, based protein quality control for treatment of adult onset neurodegenerative 749–760 (2007). diseases. Annu. Rev. Pharmacol. Toxicol. 55, 353–371 (2015). 64. Iwata, A. et al. Intranuclear degradation of polyglutamine aggregates by the 91. Walker, G.A. & Lithgow, G.J. Lifespan extension in C. elegans by a molecular ubiquitin-proteasome system. J. Biol. Chem. 284, 9796–9803 (2009). chaperone dependent upon insulin-like signals. Aging Cell 2, 131–139 (2003). 65. Tan, K. et al. Mitochondrial SSBP1 protects cells from proteotoxic stresses by 92. Steinkraus, K.A. et al. Dietary restriction suppresses proteotoxicity and enhances potentiating stress-induced HSF1 transcriptional activity. Nat. Commun. 6, 6580 longevity by an hsf-1-dependent mechanism in Caenorhabditis elegans. Aging (2015). Cell 7, 394–404 (2008). 66. Hung, Y.H., Chen, L.M., Yang, J.Y. & Yang, W.Y. Spatiotemporally controlled induction 93. Chavous, D.A., Jackson, F.R. & O’Connor, C.M. Extension of the Drosophila of autophagy-mediated lysosome turnover. Nat. Commun. 4, 2111 (2013). lifespan by overexpression of a protein repair methyltransferase. Proc. Natl. Acad. 67. Bejarano, E. et al. Connexins modulate autophagosome biogenesis. Nat. Cell Biol. Sci. USA 98, 14814–14818 (2001). 16, 401–414 (2014). 94. Ruan, H. et al. High-quality life extension by the enzyme peptide methionine 68. Takeuchi, T. et al. Intercellular chaperone transmission via exosomes contributes sulfoxide reductase. Proc. Natl. Acad. Sci. USA 99, 2748–2753 (2002). to maintenance of protein homeostasis at the organismal level. Proc. Natl. Acad. 95. Kruegel, U. et al. Elevated proteasome capacity extends replicative lifespan in Sci. USA 112, E2497–E2506 (2015). Saccharomyces cerevisiae. PLoS Genet. 7, e1002253 (2011). 69. Lo Cicero, A., Stahl, P.D. & Raposo, G. Extracellular vesicles shuffling intercellular 96. Vilchez, D. et al. RPN-6 determines C. elegans longevity under proteotoxic stress messages: for good or for bad. Curr. Opin. Cell Biol. 35, 69–77 (2015). conditions. Nature 489, 263–268 (2012). 70. Cannizzo, E.S. et al. Age-related oxidative stress compromises endosomal 97. Depuydt, G. et al. Reduced insulin/insulin-like growth factor-1 signaling and proteostasis. Cell Rep. 2, 136–149 (2012). dietary restriction inhibit translation but preserve muscle mass in Caenorhabditis 71. Goetzl, E.J. et al. Altered lysosomal proteins in neural-derived plasma exosomes elegans. Mol. Cell. Proteomics 12, 3624–3639 (2013). in preclinical Alzheimer disease. Neurology 85, 40–47 (2015). 98. Demontis, F. & Perrimon, N. FOXO/4E-BP signaling in Drosophila muscles 72. Astanina, K., Koch, M., Jungst, C., Zumbusch, A. & Kiemer, A.K. Lipid droplets regulates organism-wide proteostasis during aging. Cell 143, 813–825 (2010). as a novel cargo of tunnelling nanotubes in endothelial cells. Sci. Rep. 5, 11453 99. Tonoki, A. et al. Genetic evidence linking age-dependent attenuation of the 26S (2015). proteasome with the aging process. Mol. Cell. Biol. 29, 1095–1106 (2009). 73. Burtey, A. et al. Intercellular transfer of transferrin receptor by a contact-, Rab8- 100. Crowe, E., Sell, C., Thomas, J.D., Johannes, G.J. & Torres, C. Activation of dependent mechanism involving tunneling nanotubes. FASEB J. 29, 4695–4712 proteasome by insulin-like growth factor-I may enhance clearance of oxidized (2015). proteins in the brain. Mech. Ageing Dev. 130, 793–800 (2009). 74. Wang, X. & Gerdes, H.H. Transfer of mitochondria via tunneling nanotubes rescues 101. Meléndez, A. et al. Autophagy genes are essential for dauer development and apoptotic PC12 cells. Cell Death Differ. 22, 1181–1191 (2015). life-span extension in C. elegans. Science 301, 1387–1391 (2003). 75. Wang, X. et al. Rescue of brain function using tunneling nanotubes between 102. Jia, K. & Levine, B. Autophagy is required for dietary restriction-mediated life neural stem cells and brain microvascular endothelial cells. Mol. Neurobiol. span extension in C. elegans. Autophagy 3, 597–599 (2007). doi:10.1007/s12035-015-9225-z (2015). 103. Morselli, E. et al. Caloric restriction and resveratrol promote longevity through the 76. Agosta, F., Weiler, M. & Filippi, M. Propagation of pathology through brain Sirtuin-1-dependent induction of autophagy. Cell Death Dis. 1, e10 (2010). networks in neurodegenerative diseases: from molecules to clinical phenotypes. 104. Eisenberg, T. et al. Induction of autophagy by spermidine promotes longevity. CNS Neurosci. Ther. 21, 754–767 (2015). Nat. Cell Biol. 11, 1305–1314 (2009). 77. Russo, I., Bubacco, L. & Greggio, E. Exosomes-associated neurodegeneration and 105. Simonsen, A. et al. Promoting basal levels of autophagy in the nervous system progression of Parkinson′s disease. Am. J. Neurodegener. Dis. 1, 217–225 enhances longevity and oxidant resistance in adult Drosophila. Autophagy 4, (2012). 176–184 (2008). 78. Prusiner, S.B. Biology and genetics of prions causing neurodegeneration. Annu. 106. Bjedov, I. et al. Mechanisms of life span extension by rapamycin in the fruit fly Rev. Genet. 47, 601–623 (2013). Drosophila melanogaster. Cell Metab. 11, 35–46 (2010). 79. Gousset, K. et al. Prions hijack tunnelling nanotubes for intercellular spread. Nat. 107. Labunskyy, V.M. et al. Lifespan extension conferred by endoplasmic reticulum Cell Biol. 11, 328–336 (2009). secretory pathway deficiency requires induction of the unfolded protein response. 80. Malkus, K.A. & Ischiropoulos, H. Regional deficiencies in chaperone-mediated PLoS Genet. 10, e1004019 (2014). autophagy underlie alpha-synuclein aggregation and neurodegeneration. Neurobiol. 108. Barros, M.H., Bandy, B., Tahara, E.B. & Kowaltowski, A.J. Higher respiratory Dis. 46, 732–744 (2012). activity decreases mitochondrial reactive oxygen release and increases life span 81. Liu, Y., Zhang, X., Chen, W., Tan, Y.L. & Kelly, J.W. Fluorescence turn-on folding in Saccharomyces cerevisiae. J. Biol. Chem. 279, 49883–49888 (2004). sensor to monitor proteome stress in live cells. J. Am. Chem. Soc. 137, 109. McCormick, M.A. et al. A comprehensive analysis of replicative lifespan in 4,698 11303–11311 (2015). single-gene deletion strains uncovers conserved mechanisms of aging. Cell Metab. 82. Arrasate, M. & Finkbeiner, S. Automated microscope system for determining factors 22, 895–906 (2015). that predict neuronal fate. Proc. Natl. Acad. Sci. USA 102, 3840–3845 (2005). 110. Henis-Korenblit, S. et al. Insulin/IGF-1 signaling mutants reprogram ER stress Nature America, Inc. All rights reserved. America, Inc. © 201 5 Nature 83. van Oosten-Hawle, P., Porter, R.S. & Morimoto, R.I. Regulation of organismal response regulators to promote longevity. Proc. Natl. Acad. Sci. USA 107, proteostasis by transcellular chaperone signaling. Cell 153, 1366–1378 (2013). 9730–9735 (2010). 84. Taylor, R.C. & Dillin, A. XBP-1 is a cell-nonautonomous regulator of stress 111. Shore, D.E., Carr, C.E. & Ruvkun, G. Induction of cytoprotective pathways is resistance and longevity. Cell 153, 1435–1447 (2013). central to the extension of lifespan conferred by multiple longevity pathways.

npg 85. Williams, K.W. et al. Xbp1s in Pomc neurons connects ER stress with energy PLoS Genet. 8, e1002792 (2012). balance and glucose homeostasis. Cell Metab. 20, 471–482 (2014). 112. Chen, D., Thomas, E.L. & Kapahi, P. HIF-1 modulates dietary restriction-mediated 86. Genereux, J.C. et al. Unfolded protein response-induced ERdj3 secretion links lifespan extension via IRE-1 in Caenorhabditis elegans. PLoS Genet. 5, e1000486 ER stress to extracellular proteostasis. EMBO J. 34, 4–19 (2015). (2009). 87. Conboy, M.J., Conboy, I.M. & Rando, T.A. Heterochronic parabiosis: historical 113. Zid, B.M. et al. 4E-BP extends lifespan upon dietary restriction by enhancing perspective and methodological considerations for studies of aging and longevity. mitochondrial activity in Drosophila. Cell 139, 149–160 (2009). Aging Cell 12, 525–530 (2013). 114. Kim, H.J., Morrow, G., Westwood, J.T., Michaud, S. & Tanguay, R.M. Gene 88. Katsimpardi, L. et al. Vascular and neurogenic rejuvenation of the aging mouse expression profiling implicates OXPHOS complexes in lifespan extension of flies brain by young systemic factors. Science 344, 630–634 (2014). over-expressing a small mitochondrial chaperone, Hsp22. Exp. Gerontol. 45, 89. de Cabo, R. et al. Serum from calorie-restricted animals delays senescence and 611–620 (2010). extends the lifespan of normal human fibroblasts in vitro. Aging (Albany, NY) 7, 115. Owusu-Ansah, E., Song, W. & Perrimon, N. Muscle mitohormesis promotes longevity 152–166 (2015). via systemic repression of insulin signaling. Cell 155, 699–712 (2013).

nature medicine VOLUME 21 | NUMBER 12 | DECEMBER 2015 1415