INTRODUCTION

      Aging research explores the decline in function of organisms during adulthood. Since 2013, when the first edition of the hall-marks of aging was published in Cell, 1 close to 300,000 articles dealing with this subject have been published, which is as many as during the preceding century. Hence, time has become ripe for a new edition of the hallmarks of aging incorporating the main knowledge obtained a decade on.

      The distinction among ‘‘hallmarks’’ is intrinsically diffuse, since they interact and are not independent of each other. Therefore, their classification is inevitably arbitrary, but we proposed three criteria that must apply for each hallmark of aging: (1) the time dependent manifestation of alterations accompanying the aging process, (2) the possibility to accelerate aging by experimentally accentuating the hallmark, and—most decisively—(3) the oppor tunity to decelerate, halt, or reverse aging by therapeutic inter ventions on the hallmark. Rather than elaborating a compendium of age-associated alterations, we shall focus on the molecular, cellular, and systemic processes mechanistically accounting for their manifestation. That said, both in laboratory animals and in human medicine, objective quantification of morphological and functional decline affecting the aging organism is essential to measure biological aging. Indeed, disparity between biological and chronological age can reflect the efficacy of age-accelerating or -decelerating manipulations that evaluate the contribution of a given hallmark to the aging process. For this reason, standardized physiological measurements (e.g., respirometry to measure basal and maximal energy expenditure), functional tests (e.g., at the sensory, psychomotor, and cognitive levels), and ever more sophisticated ‘‘omics’’ technologies (e.g., genomics, epigenomics, transcriptomics, proteomics, and metabolomics), often applied at the single-cell level, are instrumental for evaluating the spatiotemporal patterns of health degradation and the (in)efficacy of anti-aging strategies.

       In 2013, we suggested nine molecular, cellular, and systemic hallmarks of aging: DNA instability, telomere attrition, epigenetic alterations, loss of proteostasis, deregulated nutrient-sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion, and altered intercellular communication.1 Recent research has confirmed and extended the importance of all these hallmarks. They have withstood scrutiny by tens of thousands of aging researchers, but they require an update to deal with the discoveries of the last decade. For example, in 2013, much of the evidence on anti-aging interventions was limited to non-mammalian model organisms including yeast, nematodes, and fruit flies. Fortunately, experiments involving mice (and in some cases, non-human primates) have now corroborated the validity of most of these hallmarks in mammals. Of note, human age-related diseases have statistically higher chances to co-occur and to share genomic characteristics when they are causally linked to the same hallmark rather than to different hallmarks,2 clinically validating the approach that we have chosen.

      Besides the necessary update of the previous hallmarks, we have also introduced some reorganizations and included the following three additional hallmarks of aging: disabled macroautophagy, chronic inflammation, and dysbiosis. Disabled macroautophagy was initially treated as a special case of loss of proteostasis. However, macroautophagy does not only affect proteins but can target entire organelles and non-proteinaceous macromolecules, justifying its discussion as a separate entity. Moreover, we considered that the final hallmark that we listed in 2013, altered intercellular communication, was too vast, requiring a separate discussion of chronic inflammation and age-associated dysbiosis (Figure 1).

Figure 1. The hallmarks of aging The scheme compiles the 12 hallmarks of aging proposed in this work: genomic instability, telomere attrition, epigenetic alterations, loss of proteostasis, disabled macroautophagy, deregulated nutrient-sensing, mitochondrial dysfunction, cellular senescence, stem cell exhaustion, altered intercellular communication, chronic inflammation, and dysbiosis. These hallmarks are grouped into three categories: primary, antagonistic, and integrative.

GENOMIC INSTABILITY

      Genome integrity and stability are pervasively threatened by exogenous chemical, physical, and biological agents, as well as by endogenous challenges such as DNA replication errors, chromosome segregation defects, oxidative processes, and spontaneous hydrolytic reactions. The wide range of genetic lesions caused by these extrinsic or intrinsic sources of damage include point mutations, deletions, translocations, telomere shortening, single- and double-strand breaks, chromosomal rearrangements, defects in nuclear architecture, and gene disruption caused by the integration of viruses or transposons. All these molecular alterations and the resulting genomic mosaicism may contribute to both normal and pathological aging.3 Accordingly, organisms have evolved a complex array of DNA repair and maintenance mechanisms to deal with the damage inflicted to nuclear and mitochondrial DNA (mtDNA) and to ensure the appropriate chromosomal architecture and stability. These DNA repair networks lose efficiency with age, which accentuates the accumulation of genomic damage and the ectopic accumulation of DNA in the cytosol4 (Figure 2A).

Figure2.Loss of cellular integrity caused by genomic instability, telomere attrition, and epigenetic alterations (A)Endogenous o rexogenous agents cause avariety of DNA lesions that contribute to both normal and pathological aging. Such lesions can by repaired by a variety of mechanisms that lose efficiency with age. Excessive DNA damage, insufficient DNA repair, alterations in nuclear architecture, and telomere attrition favor the aging process. BER, base excision repair; HR, homologous recombination; NER, nucleotide excision repair; NHEJ, non-ho-mologous end joining; MMR, mismatch repair; SAC, spindle assembly checkpoint; TERT, telomerase reverse transcriptase; TLS, trans-lesionsyn-thesis. (B)Changes in theacetylation and methylation of DNA or histones, as well as in levels or activity of chromatin-associated proteins or of non-coding RNAs (ncRNAs)induce epigenetic changes that contribute to the aging process. The red portions of the hourglasses indicateage-associated alterations and the blue portions strategies for avoiding them.

Nuclear DNA

      Cells from aged humans and model organisms accumulate somatic mutations at nuclear DNA.5 Other forms of damage, such as chromosomal aneuploidy and copy-number variations, are also associated with aging. All these DNA alterations may affect essential genes and transcriptional pathways, resulting in dysfunctional cells that may finally compromise tissue and organismal homeostasis. This is especially relevant when DNA damage impacts on stem cells, hampering their role in tissue renewal or leading to their exhaustion, which in turn promotes aging and increases susceptibility to age-related pathologies.6,7 The mutational burden in histologically normal human tissues is remarkable. For example, normal esophageal epithelium cells from young individuals already display hundreds of mutations and may carry more than 2,000 mutations per cell by middle age. 8 The accumulation of DNA mutations throughout life is likely tolerated because of the excessive energetic cost of the complete repair of all genomic damages caused by exogenous and endogenous challenges. Consequently, cells favor survival over genomic integrity.9 These data also suggest that similar to carcinogenesis, driver mutations alone may not be sufficient to accelerate aging because they require a permissive microenvironment created by non-mutagenic promoting factors to become penetrant.10

      Comparative analysis of the mutational landscape across mammalian species has shown that species-specific somatic mutation rate is inversely correlated with lifespan.11 To date, there is no clear evidence that the normal rate of mutation fixation is responsible for aging, but numerous studies have shown that DNA repair deficiencies have the potential to cause aging. Thus, alterations in DNA repair mechanisms accelerate aging in mice and underlie several human progeroid syndromes. 12 Conversely, transgenic mice overexpressing the mitotic checkpoint kinase BubR1 exhibit an extended healthy lifespan13 (Table 1). Moreover, studies in humans and other long-lived species have shown that enhanced DNA repair mechanisms coevolve with increased longevity.14 Sirtuin-6 (SIRT6) may play a major role in this differential reparative efficiency across species. Overexpression of SIRT6 in mice reduces genomic instability, improves double-strand break repair, and extends lifespan15 (Table 1), although other explanations, such as improved glucose metabolism and restoration of energy homeostasis, have been proposed to explain the prolongevity effects of SIRT6.16 Notably, recent work has shown that small-molecule activation of 8-oxoguanine DNA glycosylase 1 increases oxidative DNA damage repair and may have therapeutic applications in the context of aging and other processes linked to excessive oxidative damage.17 These findings suggest that interventions aimed at reducing the mutational load of nuclear DNA or at enhancing or rerouting its repair mechanisms may delay aging and the onset of age-related diseases, but further causal evidence in this regard is still missing.

Mitochondrial DNA

      Genomic instability affecting mtDNA may contribute to aging and age-related pathologies.96 mtDNA is strongly impacted by aging-associated mutations and deletions due to its high replicativeindex, the limited efficiency of its repair mechanisms, its oxidative microenvironment, and the lack of protective histones embracing this small DNA molecule. Somatic mtDNA alterations increase across human tissues during aging, but it remains unclear whether this increase truly impacts the aging process at the functional level. The causal implication of mtDNA mutations in driving aging has been difficult to assess because of ‘‘heteroplasmy,’’ which implies the co-existence of mutated and wild-type genomes within the same cell. However, deep-sequencing of aged cells revealed that their mtDNA mutational load may substantially increase through clonal expansion events.97 The accelerated expansion of mitochondrial mutations with age has also been observed in both primate oocytes and somatic tissues,98 as well as in lympho blasts from patients with neurodegenerative diseases.99 Of note, ultra-sensitive sequencing indicates that most mtDNA mutations in aged cells arise from replication errors caused by mtDNA polymerase g rather than from oxidative stress.96

      Preliminary evidence that mtDNA mutations might be directly involved in aging and age-related pathologies was provided by human disorders that are caused by mtDNA damage and partially phenocopy aging.100 Further causative evidence has arisen from studies on mice deficient in DNA polymerase γ that exhibit accelerated aging and reduced lifespan associated with deletions rather than point mutations in mtDNA101 (Table 1). Overall, these data suggest that the avoidance, attenuation, or correction of mtDNA mutations might contribute to extend healthspan and lifespan. Nevertheless, as in the case of nuclear DNA mutations, experimental evidence demonstrating deceleration of aging by gain of function in mtDNA repair mechanisms is still largely missing.

Nuclear architecture

      Defects in the nuclear lamina, which constitutes a scaffold for tethering chromatin and protein complexes, can generate genome instability.102 Accelerated aging syndromes such as the Hutchinson-Gilford and the Ne´ stor-Guillermo progeria syndromes (HGPS and NGPS, respectively) are caused by mutations in genes LMNA and BANF1 encoding protein components of nuclear lamina. Alterations of the nuclear lamina and production of an aberrant prelamin A isoform called progerin are also characteristics of normal human aging, and lamin B1 levels decline during cellular senescence.18 Animal and cellular models have facilitated the identification of the response mechanisms and stress pathways elicited by nuclear lamina aberrations caused by aging and progeria, including activation of tumor suppressor protein p53 (TP53), deregulation of the somatotrophic axis, and attrition of adult stem cells.18

      The causal implication of nuclear lamina abnormalities in premature aging has been corroborated by the observation that decreasing prelamin A or progerin levels delays the onset of progeroid features and extends lifespan in mouse models of HGPS. This can be achieved by systemic injection of antisense oligonucleotides, farnesyltransferase inhibitors, a combination of statins and aminobisphosphonates, restoration of the somatotrophic axis, or blockade of NF-kB signaling.103 Some of these interventions have been already approved for use in progeria patients.104 Moreover, gene editing strategies have been recently developed to correct LMNA mutations in cells from HGPS patients and in animal models of this disease.105,106 Hopefully, these approaches will be clinically implemented for the future treatment of progeria, but to date, no evidence is available showing that reducing progerin would delay normal aging.

TELOMERE ATTRITION

      DNA damage at the end of chromosomes (telomeres) contributes to aging and age-linked diseases.107 Replicative DNA polymerases are unable to complete the copy of telomere regions of eukaryotic DNA. Accordingly, after several rounds of cell division, telomeres undergo a substantial shortening that induces genomic instability and finally leads to either apoptosis or cell senescence. These deleterious effects can be prevented by the reverse-transcriptase activity of telomerase, an active ribonucleoprotein that elongates telomeres to maintain their adequate length.108,109 However, most mammalian somatic cells do not express telomerase, which leads to the progressive and cumulative erosion of telomere sequences from chromosome ends throughout life. There are several examples in which telomere attrition attenuates carcinogenesis through limiting the replicative lifespan of malignant cells. Hence, in contrast to genomic instability which unambiguously favors oncogenesis, telomere attrition may antagonize malignancy. For this reason, we consider telomere attrition as a hallmark of aging that is separable from genomic instability.110

      Telomerase deficiency in humans is associated with premature development of diseases such as pulmonary fibrosis, aplastic anemia, and dyskeratosis congenita, all of which hamper the regenerative capacity of the affected tissues.111 Telomere shortening is also observed during normal aging in many different species, including humans and mice.112 The telomeric attrition rate is influenced by age, genetic variants, lifestyle, and social factors; depends on the proliferative activity of the affected cells; and predicts lifespan in a wide variety of species.112 Telomere uncapping can also result from deficiencies in shelterins, a group of proteins that block the DNA damage response at chromosome ends and modulate telomere length. Several loss-of-function models for shelterin components indicate a decline of tissue regenerative capacity and accelerated aging, even in the presence of telomeres with a normal length.113

      Genetically modified animal models have revealed causal links between telomere attrition, cellular senescence, and organismal aging. Mice with shortened or lengthened telomeres exhibit decreased or increased lifespan, respectively.19 Notably, the premature aging of telomerase-deficient mice can be reverted when telomerase is genetically reactivated20 (Table 1). Moreover, normal aging can be delayed in mice by pharmacological activation or systemic viral transduction of telomerase,21 whereas mice with hyperlong telomeres show increased lifespan and metabolic health improvement22 (Table 1). Likewise, mice engineered to maintain physiological levels of telomerase in adult neurons preserve the survival of these cells and maintain cognitive function in Alzheimer’s disease models23 (Table 1). Thus, aging can be modulated by telomerase activation.

Telomerase activation to decelerate aging and treat telomere diseases

      In humans, many studies have provided evidence for causal associations between short telomere length and age-related diseases. 114 In particular, generation of mouse models with short telomeres has demonstrated that telomeric attrition is at the origin of telomere syndromes115 and prevalent age-associated diseases, such as pulmonary and kidney fibrosis.24,116 These links between telomere dynamics and organismal aging have resulted in the design of new interventions to delay aging and age-related diseases. As an example, telomerase activation using a gene therapy strategy has shown therapeutic effects on mouse models of pulmonary fibrosis and aplastic anemia. 24,25

EPIGENETIC ALTERATIONS

      The large variety of epigenetic changes that contribute to aging include alterations in DNA methylation patterns, abnormal posttranslational modification of histones, aberrant chromatin remodeling, and deregulated function of non-coding RNAs (ncRNAs) (Figure 2B). These regulatory and often reversible changes impact on gene expression and other cellular processes, resulting in the development and progression of several age-related human pathologies, such as cancer, neurodegeneration, metabolic syndrome, and bone diseases. A vast array of enzymatic systems is involved in the generation and maintenance of epigenetic patterns. These enzymes include DNA methyltransferases, histone acetylases, deacetylases, methylases, and demethylases, as well as protein complexes implicated in chromatin remodeling or in ncRNA synthesis and maturation.

DNA methylation

      The human DNA methylation landscape accumulates multiple changes with the passage of time.117 Early studies described an age-associated global hypomethylation, but further analyses revealed that specific loci, including those of several tumor suppressor genes and Polycomb target genes, are hypermethylated with age. Cells from patients and mice with progeroid syndromes also exhibit DNA methylation changes that partially recapitulate those found in normal aging.118 The functional consequences of most of these age-related epimutations are uncertain, as the majority of changes affect introns and intergenic regions.119

      Epigenetic clocks based on DNA methylation status at selected sites have been introduced to predict chronological age and mortality risk as well as to evaluate interventions that may extend human lifespan.119 This has been demonstrated with protocols aimed at thymus regeneration, which resulted in improved risk indices for many age-related diseases and a mean epigenetic age approximately 1.5 years less than baseline after 1 year of treatment. Moreover, predictions of human morbidity and mortality showed a 2-year decrease in epigenetic versus chronological age, which persisted 6 months after discontinuing treatment.27 Likewise, a-ketoglutarate supplementation for 7 months turned back the epigenetic clock by 8 years.26 In summary, DNA methylation changes are associated with aging, but there is no definitive evidence that they actually cause aging. Further studies will be necessary to demonstrate that defective maintenance of DNA methylation produces accelerated aging and that improved fidelity in maintenance of DNA methylation patterns extends longevity. It will be also necessary to identify the molecular drivers responsible for the modulation of changes occurring in the aged human methylome.

Histone modifications

      Global loss of histones and tissue-dependent changes in their post-translational modifications are also closely linked to aging. Increased histone expression extends lifespan in Drosophila, 120 whereas increased histone H4K16 acetylation or H3K4 trimethylation and decreased levels of H3K9 or H3K27 trimethylation are found in fibroblasts from aged individuals and progeroid patients. These histone modifications can lead to transcriptional changes, loss of cellular homeostasis, and age-associated metabolic decline.121 Of note, loss of heterochromatic marks at telomeres has been shown to lead to telomere lenthening. 122

     Histone demethylases modulate lifespan by targeting components of key longevity routes such as the insulin/insulin growth factor-1 (IGF-1) signaling pathway. Other histone-modifying enzymes such as members of the SIRT family of protein deacetylases and ADP-ribosyltransferases also contribute to healthy aging.29 Transgenic overexpression of SIRT1 improves genomic stability and metabolic efficiency during aging in mice, although without increasing longevity.29 Overexpression of mitochondrial SIRT3 reverses the regenerative capacity lost in aged hematopoietic stem cells (HSCs) and can mediate the beneficial effects of dietary restriction in longevity.123 Similarly, Sirt6 ablation in mice results in accelerated aging,124 whereas Sirt6 overexpression extends lifespan.16 The underlying mechanisms derive from the fact that Sirt6 is a multitask protein with ability to interconnect chromatin dynamics with metabolism and DNA repair.125 Finally, Sirt7 deficiency induces global genomic instability, metabolic dysfunctions, and premature aging. 29 Together, these findings are consistent with the idea that a decrease in deacetylase activity would result in chromatin relaxation, increased exposure to DNA damaging agents, and enhanced genomic instability.126 Conversely, genetic inactivation of the histone acetyltransferase KAT7 in human stem cells decreases histone H3K14 acetylation and alleviates cell senescence features.28 Moreover, intravenous injection of lentiviral vectors encoding Cas9/sg-Kat7 ameliorates hepatocyte senescence and liver aging and extends lifespan in both normal and progeroid mice.28 Inhibitors of histone acetyltransferases also ameliorate the premature aging phenotype and extend lifespan of progeroid mice, whereas histone deacetylase activators promote longevity in part via upregulation of SIRT1 activity. 127 Together, these findings suggest that histone-modifiers should be further explored as part of therapeutic strategies against age-associated cognitive decline, although it is still unclear whether these interventions influence aging and longevity through purely epigenetic mechanisms, by impinging on DNA repair and genome stability or via transcriptional alterations affecting metabolic or signaling pathways.

Chromatin remodeling

      Besides DNA- and histone-modifiers, several chromosomal proteins and chromatin remodeling factors, such as the heterochromatin protein 1a (HP1a) and Polycomb group proteins which are implicated in genomic stability DNA repair, may modulate aging.128 Alterations in these epigenetic factors result in profound changes in chromatin architecture, including global heterochromatin loss and redistribution, which are common events in aged cells.

      The causal relevance of these chromatin alterations in aging has been largely studied in invertebrates in which loss-of-function mutations in HP1a decrease longevity, whereas its overexpression expands healthspan and lifespan129 (Table 1). Similar studies in mammals are still limited, but most studies indicate that heterochromatin relaxation contributes to aging and aging-related pathologies, whereas maintenance of heterochromatin promotes longevity. For example, loss of PIN1—a prolyl isomerase essential to preserve heterochromatin is associated with premature aging and neurodegeneration in different species from Drosophila to mammals130 (Table 1). Nevertheless, experiments aimed at extending vertebrate longevity by gain of function of chromatin remodeling factors are still missing.

Non-coding RNAs

      The large and growing universe of ncRNAs, including lncRNAs (such as telomeric RNAs or TERRA), microRNAs (miRNAs), and circular RNAs, has emerged as epigenetic factors with ability to influence aging. ncRNAs modulate healthspan and lifespan by post-transcriptional targeting of components of longevity networks or by regulating stem cell behavior.131 A circular RNA mediates the effect of the insulin/IGF-1 signaling pathway on Drosophila lifespan,132 but most studies have focused on miRNAs, and there is still debate on the extent to which other ncRNAs may derive from transcriptional noise, with their regulatory roles in human physiology and pathology only circumscribed to few specific cases.133

      Gain- and loss-of-function studies first confirmed the capacity of several miRNAs to modulate longevity in invertebrates. Subsequent studies in mice have provided causal evidence on the functional relevance of miRNAs in aging (Table 1). For example, miRNA-188-3p expression is upregulated in skeletal endothelium during aging and contributes to vascular problems associated with the passage of time. Depletion of miR-188 in mice alleviates the age-related decline in beneficial bone capillary subtypes, whereas endothelial-specific overexpression of this miRNA decreases bone mass and delays bone regeneration.30 Conversely, depletion of miR-455-3p in mice exhibits deleterious effects on mitochondrial dynamics, cognitive behavior, and lifespan, whereas its overexpression preserves these functions and extends lifespan.31 Overall, these findings suggest that miRNAs may causally contribute to aging and aging-related pathologies and represent potential therapeutic targets for delaying or ameliorating these conditions.

Derepression of retrotransposons

      Recent studies have unveiled the role of retrotransposons in aging of complex metazoans, including humans.134 These retrotransposable elements are mobile genetic units that can move from one genomic location to another, using a molecular mechanism that involves an RNA intermediate. Retrotransposons consist of long interspersed nuclear elements (LINEs), which encode the required proteins for retrotransposition, and SINEs, which are short, non-coding RNAs that hijack the LINE protein machinery. Retrotransposons are reactivated in senescent cells and during lifetime and generate deleterious effects through genetic and epigenetic changes or by activation of immune pathways triggered after identification of retrotransposon nucleic acids as foreign DNA.134 Mechanistically, epigenetic derepression of LINE-1 RNA inhibits the epigenetic reader Suv39H1,2 resulting in global reduction of H3K9me3 and heterochomatin,135 whereas reverse transcription of LINE-1 RNA results in double-stranded cDNA that activates the cGAS/STING/interferon pathway.136

      Treatments with nucleoside reverse-transcriptase inhibitors (NRTIs), which suppress or attenuate retrotransposition, extendm lifespan of Sirt6-null mice and improve healthspan, ameliorating bone and muscle phenotypes (Table 1). Likewise, treatment of aged wild-type mice with NRTIs reduces the levels of DNA damage markers.32 Moreover, in vivo targeting of retrotransposons with antisense oligonucleotides increases the lifespan of progeroid mice.135 Notably, a rare SIRT6 variant in centenarians is a stronger suppressor of LINE1 retrotransposons, enhances genome stability, and can more robustly kill cancer cells than wild-type SIRT6. 137 Collectively, these findings suggest that retrotransposons causally contribute to the aging process and that interventions that oppose retrotransposon activity might improve healthy longevity. Further clinical studies in aged populations with drugs targeting the different functions of retrotransposons may delineate novel intervention strategies on aging and aging-related pathologies.

Gene expression changes

      The mechanisms underlying the effects of all the above epigenetic factors converge at the modulation of gene expression levels. Aging causes an increase of the transcriptional noise and an aberrant production and maturation of many mRNAs.138,139 Microarraybased comparisons of young and old tissues from human and other species have identified age-related transcriptional signatures that result from epigenetic changes occurring during aging. Environmental exposures also cause alterations in gene regulation via DNA methylation alterations and histone modifications and promote aging-related epigenetic changes including the acceleration of epigenetic clocks.140 Single-cell transcriptomic and plasma proteomics of multiple cell types and organs at several ages across the entire mouse life span have unveiled remarkable gene expression shifts during aging. 138 These changes specially affect certain biological processes, such as inflammation, protein folding, extracellular matrix (ECM) regulation, and mitochondrial function, which are widely deregulated in aging.141 The common expression patterns observed during aging in different tissues may help to guide future interventions aimed at improving healthspan and lifespan (Table 1). Likewise, the observed decline in transcriptional and post-transcriptional efficiency and fidelity in the course of aging, and its negative consequences on the proteome health may also open new opportunities for prolongevity strategies.139

LOSS OF PROTEOSTASIS

      Aging and several age-related morbidities, such as amyotrophic lateral sclerosis (ALS), Alzheimer’s disease, Parkinson’s disease, and cataract, are associated with impaired protein homeostasis or proteostasis, leading to the accumulation of misfolded, oxidized, glycated, or ubiquitinylated proteins that often form aggregates as intracellular inclusion bodies or extracellular amyloid plaques. 142

Proteostasis collapse

      Intracellular proteostasis can be disrupted due to the enhanced production of erroneously translated, misfolded or incomplete proteins (Figure 3). Genetic manipulation of the ribosomal protein RPS23 to improve the accuracy of RNA-to-protein translation extends lifespan in Schizosaccharomyces pombe, Caenorhabditis elegans, and Drosophila melanogaster, 143 whereas a mutation in RPS9 that favors error-prone translation causes premature aging in mice.144 Another mechanism driving the collapse of the proteostasis network resides in slowed translation elongation and cumulative oxidative damage of proteins, increasingly distracting the chaperones from folding healthy proteins required for cellular fitness.145 In addition, numerous age-related neurodegenerative diseases including ALS and Alzheimer can be caused by mutations in proteins that render them intrinsically prone to misfolding and aggregation, hence saturating the mechanisms of protein repair, removal, and turnover that are required for maintenance of the healthy state.146

      The proteostasis network also collapses when mechanisms as suring quality control fail, for instance, due to reduced function of the unfolded protein response (UPR) in the endoplasmic reticulum (ER),147 when stabilization of correctly folded proteins is compromised or when mechanisms for the degradation of proteins by the proteasome or the lysosome become insufficient (Figure 3). Reduction of proteasome activity has been observed in aged organs including the brain of the short-lived fish Nothobranchius furzeri. 148 Moreover, some mono-ubiquitinylated proteins accumulate in aging tissues from flies, mice, monkeys, and humans, as documented for histone 2A.149

Figure 3. Loss of protein and organellar turnover

Loss of proteostasis and disabled macroautophagy are characterized by a deviation from the young equilibrium state in which an accumulation of waste products results from a variety of age-associated alterations and simultaneously waste removal is compromised through a variety of mechanisms. The functional consequences of these alterations are listed. Some strategies for reestablishing proteostasis and autophagy are exemplified on the left and on the right.

      The degradation of proteins by the lysosome can be achieved in a specific fashion, through chaperone-mediated autophagy (CMA), wherein proteins exposing a pentapeptide motif resembling KFERQ first bind to heat shock protein HSC70 and then to lysosome-associated membrane protein type 2A (LAMP2A), which facilitates the translocation of the client protein into the lumen of the lysosome.150 Hepatic LAMP2A expression declines with age in mice, and its transgenic re-expression reduces liver aging. 151 Protein aggregates can also be removed by macroautophagy upon their inclusion in two-membrane vesicles, the autophagosomes, for their later fusion with lysosomes.152 Since autophagosomes can envelop non-proteinaceous structures, this process will be discussed separately from proteostasis in the next hallmark section (disabled macroautophagy). Nonetheless, stimulation of autophagy constitutes a valid strategy for the elimination of intracellular protein aggregates.153

      Proteostasis, aging, and longevity Perturbation of general proteostasis accelerates aging. For example, feeding D. melanogaster with advanced glycation end products (AGEs) or lipofuscin (an aggregate of covalently cross-linked proteins, sugars, and lipids) causes the accumulation of AGE-modified and carbonylated proteins with a reduction of healthspan and lifespan that is further accentuated upon knockdown of the lysosomal protease cathepsin D.154 Loss of the protease ZMPSTE24 abolishes the normal proteolytic maturation of prelamin A and causes a progeroid syndrome in mice, phenocopying that observed in humans with loss-of-function mutations of ZMPSTE24.18 In mice, knockout of LAMP2A (essential for CMA) in neurons profoundly affects the proteome, yielding similar changes as found in Alzheimer patients. Indeed, inhibition of CMA in mice exacerbates experimental Alzheimer’s disease, whereas its stimulation by a pharmacological CMA activator attenuates the pathology. 155

      Experimental amelioration of proteostasis can retard the aging process (Table 1). Intranasal application of recombinant human HSP70 protein to mice enhances proteasome activity, reduces brain lipofuscin levels, enhances cognitive functions, and extends lifespan.35 Similarly, administration of the chemical chaperone 4-phenylbutyrate to aged mice reduces ER stress in the brain and improves cognition.36 In nematodes and flies, transfection-enforced overexpression of isolated proteasome subunits improves proteostasis and increases lifespan.156 In mice, stimulation of CMA by transgenic expression of LAMP2a HSCs improve the survival of the targeted cell populations,33 in line with the observation that pharmacological enhancement of CMA attenuates Alzheimer’s pathology and arteriosclerosis.155,34 Hence, activation of CMA may constitute a valid strategy for delaying the aging process.

       A phase 3 clinical trial has revealed that in patients with recent ALS diagnosis, administration of the antihypertensive guanabenz inhibits progression to the life-threatening bulbar stage.37 Guanabenz may act to stimulate the phosphorylation (or to inhibit the dephosphorylation) of eukaryotic translation initiation factor 2a (eIF2a), which occurs in the context of the ‘‘integrated stress response (ISR)’’ as part of the UPR,157 although it remains under debate to what extent the actions of guanabenz are mediated by the stimulation of the ISR.158 Importantly, eIF2a phosphorylation causes a switch from 50 cap-dependent to 50 cap-independent RNA translation, knowing that the latter is enhanced by several longevity-extending manipulations.159 Moreover, eIF2a phosphorylation is essential for the induction of stress granules, which are required for longevity extension by dietary restriction in worms.160 Finally, eIF2a phosphorylation is indispensable for the induction of autophagy,161 which is a major anti-aging mechanism (see below), suggesting a crosstalk between UPR and autophagy in prolongevity pathways. Future studies must determine whether the capacity of guanabenz to attenuate neurodegeneration is mediated via ISR stimulation or alternative mechanisms. Indeed, it has been proposed that inhibitors of ISR might be also used for the treatment of neurodegenerative diseases.162

DISABLED MACROAUTOPHAGY

      Macroautophagy (that we will refer to as ‘‘autophagy’’) involves the sequestration of cytoplasmic material in two-membrane vesicles, the autophagosomes, which later fuse with lysosomes for the digestion of luminal content.152 Thus, autophagy is not only involved in proteostasis but also affects non-proteinaceous macromolecules (such as ectopic cytosolic DNA, lipid vesicles, and glycogen) and entire organelles (including dysfunctional mitochondria targeted by ‘‘mitophagy,’’ and other organelles leading to ‘‘lysophagy,’’ ‘‘reticulophagy,’’ or ‘‘pexophagy’’), as well as invading pathogens (‘‘xenophagy’’).152 An age-related decline in autophagy constitutes one of the most important mechanisms of reduced organelle turnover, justifying its discussion as a new hallmark of aging. As a note of caution, genes and proteins that participate in the autophagic process are also involved in alternative degradation processes such as LC3-associated phagocytosis of extracellular material,163 and the extrusion of intracellular waste (e.g., dysfunctional mitochondria) in the form of exospheres for their subsequent removal by macrophages. 164 That said, there is strong evidence that the core process of autophagy is relevant to aging (Figure 3).

      Accelerated aging due to autophagy inhibition In humans, the expression of autophagy-related genes, such as ATG5, ATG7, and BECN1, declines with age.165 CD4+ T lymphocytes isolated from the offspring of parents with exceptional longevity show enhanced autophagic activity compared with age-matched controls.166 Decreased autophagy in circulating B and T lymphocytes from aging donors is accompanied by a reduction of the pro-autophagic metabolite spermidine.167,168 Similarly, in rodents, a progressive deterioration of autophagy has been described for some organs, pleading in favor of the idea that autophagic flux is compromised with age. Reduction of autophagic flux may participate in the accumulation of protein aggregates and dysfunctional organelles, reduced elimination of pathogens, and enhanced inflammation because autophagy eliminates proteins involved in inflammasome and their upstream triggers.169

      Genetic inhibition of autophagy accelerates the aging process in model organisms. This process is partially reversible, as illustrated in mice in which Atg5 is downregulated by a doxycyclineinducible shRNA. Atg5 knockdown causes the premature degeneration and senescence of multiple organ systems leading to premature death.170 Upon withdrawal of doxycycline, autophagy restoration is accompanied by attenuated systemic inflammation and segmental reduction of aging. Of note, in this model, the transient inhibition of autophagy is followed by a major increase in the incidence of malignancies. Hence, autophagy apparently acts as a tumor-suppressive mechanism, which may involve cell-autonomous processes and cancer immunosurveillance. 153 In patients, loss-of-function mutations of genes that regulate or execute autophagy have been causally linked to a broad spectrum of cardiovascular, infectious, neurodegenerative, metabolic, musculoskeletal, ocular, and pulmonary disorders, many of which resemble to premature aging at the histopathological and functional levels.152,153

Autophagy stimulation for decelerated aging

      There is ample evidence that stimulation of autophagic flux increases healthspan and lifespan in model organisms (Table 1). For example, increasing autophagy solely in the enterocytes of the intestine increases Drosophila lifespan.120 In mice, transgenic overexpression of Atg5 under the control of a ubiquitously expressed promoter is sufficient to extend lifespan and to improve metabolic health and motor function.38 Moreover, knockin mutation of beclin 1 (Becn1F121A/F121A) to reduce its inhibition by Bcl-2 causes an increase in autophagic flux, as well as an extension of lifespan. This effect is coupled to a reduction of age-associated pathologies and spontaneous tumorigenesis,39 as well as to increased neurogenesis.40

      Oral supplementation of spermidine to mice induces autophagy in multiple organs and extends longevity by up to 25%, accompanied by reduced cardiac aging. This latter effect is lost upon cardiomyocyte-specific knockout of Atg7, suggesting that it relies on autophagy.41 Mechanistically, the pro-autophagic effects of spermidine have been linked to an inhibition of the acetyl transferase EP300 (resulting in reduced acetylation of several core autophagy proteins)171 or to the hypusination of eIF5A, which is essential for the synthesis of the autophagy transcription factor TFEB.167 Among these factors, EP300 is the target of the longevity-enhancing drugs nordihydroguairaretic acid43 and salicylate.42 Pharmacological inhibition of EP300 with C646 mimics the stimulatory effects of spermidine on autophagy and cancer immunosurveillance.172 When circulating B lymphocytes or CD8+ T cells from aged human donors are cultured in the presence of spermidine, the cells recover juvenile levels of TFEB and eIF5A, coupled to a normalization of autophagic flux.167,168 Moreover, in Drosophila, hypusination deficiency due to a heterozygous mutation or knockdown of deoxyhypusine synthase abolished lifespan extension by spermidine supplementation.173 Deoxyhypusine synthase defi- ciency in murine T cells triggers severe intestinal inflammation coupled to epigenetic remodeling and rewiring of the tricarboxylic acid cycle,174 whereas spermidine treatment of wild-type mice protects against colitis and colon carcinogenesis.175 Hence, both EP300 inhibition and eIF5A hypusination appear plausible targets to explain the in vivo effects of spermidine.

      Pharmacological agents that induce mitophagy and have a positive impact on murine healthspan include NAD+ precursors (such as nicotinamide, nicotinamide mononucleotide, and nicotinamide riboside)176 and urolithin A.177 Clinical trials have demonstrated the efficacy of NAD+ precursors in the chemoprevention of non-melanoma skin cancer,46 in reversing insulin resistance in prediabetic women,44 and in reducing neuroinflammation in patients with Parkinson’s disease.45 Moreover, a phase 3 trial has revealed the capacity of urolithin A to improve muscle strength and to reduce C-reactive protein (CRP).47

To be continued