|
Targeting the molecular & cellular pillars of human aging with exercise
Jorming Goh,Esther Wong,Janjira Soh,Andrea Britta Maier,Brian Keith Kennedy
First published: 30 December 2021
https://doi.org/10.1111/febs.16337
Abstract
Biological aging is the main driver of age-associated chronic diseases. In 2014, the United States National Institute of Aging (NIA) sponsored a meeting between several investigators in the field of aging biology, who identified seven biological pillars of aging and a consensus review, “Geroscience: Linking Aging to Chronic Disease,” was published. The pillars of aging demonstrated the conservation of aging pathways in diverse model organisms and thus represent a useful framework with which to study human aging.
In this present review, we revisit the seven pillars of aging from the perspective of exercise and discuss how regular physical exercise can modulate these pillars to stave off age-related chronic diseases and maintain functional capacity.
생물학적 노화는 노화 관련 만성 질환의 주요 원인. 2014년에 미국 국립노화연구소(NIA)는 노화의 7가지 생물학적 기둥을 식별한 노화 생물학 분야의 여러 연구자 간의 회의를 후원하고 "Geroscience: Linking Aging to Chronic Disease"라는 합의 검토를 발표. . 노화의 기둥은 다양한 모델 유기체에서 노화 경로의 보존을 보여주므로 인간 노화를 연구하는 데 유용한 프레임워크를 나타냄. 이 현재 리뷰에서 우리는 운동의 관점에서 노화의 7가지 기둥을 다시 살펴보고 규칙적인 신체 운동이 노화 관련 만성 질환을 예방하고 기능적 능력을 유지하기 위해 이러한 기둥을 어떻게 조절할 수 있는지 논의.
Introduction: exercise as a polypill to enhance healthspan
A single session of aerobic exercise changes the expression of ~ 9800 molecular analytes in systemic circulation that span transcripts, proteins, metabolites, and lipid classes [[1]]. Such wide-spanning effects of exercise have routinely led to its credence as a polypill, as it confers pleiotropic benefits in multiple organ systems [[2]]. In fact, exercise is as effective as drug interventions in the secondary prevention of coronary heart disease (CHD) [[3]], and even outperforms standard metformin treatment in preventing type 2 diabetes [[4]]. In overweight/obese middle-aged adults with impaired fasting glucose adopting either the World Health Organization (WHO)'s guidelines on regular physical activity (walking at least 150 min·week−1) or taking metformin, the incidence of type 2 diabetes was 58% lower in the participants undergoing physical activity intervention, and 31% percent lower in the metformin group, compared with the placebo group. Furthermore, participants in the physical activity intervention arm had a 39% reduced incidence of diabetes relative to the participants in the metformin arm [[4]].
When extrapolated to the population level, epidemiological studies support the role of regular exercise in improving human longevity, as indicated by several outcomes [[5]] such as frailty [[6]], physical function [[7]], falls [[8]], muscle strength and power [[9]], cognition [[10]], and mortality [[11]]. On one extreme, training regimens of Olympians represent the pinnacle of physical activity volume. A recent study of ~ 8000 former U.S. Olympians reported a ~ 5 year higher mean lifespan compared with the general U.S. population [[12]], with the greatest benefits derived from averting premature deaths (~ 2.2 and 1.5 years of lives saved from cardiovascular diseases and cancer, respectively). Within the general population, participation in moderate-intensity physical activity staves off mortality from, and incidence of, CHD for both men and women. A seminal epidemiological study conducted by Jeremy Morris on ~ 18 000 British male civil servants, demonstrated that, after an average of 8.5 years of follow-up and 150 000 person-years of exposure, the rate of death from CHD was two-fold higher in subjects who reported not participating in at least 5 min of vigorous-intensity exercise than those that did [[13]]. Similarly, in the Nurses' Health Study [[14]], ~ 72 488 female nurses between the ages of 40 and 65 years were followed for 8 years and their physical activity patterns were matched to incident coronary events. Compared with women in the lowest quintile group for energy expenditure (expressed as metabolic equivalents -METs), women in higher quintile groups showed relative risks of 0.77, 0.65, 0.54, and 0.46 for coronary events, after adjustment for chronological age, suggesting a dose–response relationship between physical activity and incidental CHD. More recently, a meta-analysis of prospective observational studies comparing the relationship between physical activity “doses” with all-cause mortality was performed in patients diagnosed with chronic, noncommunicable diseases [[15]]. In this study, every 10 METs increase in physical activity per week was associated with a 22% lower mortality rate in women with breast cancer and a 12% decrease in patients with ischemic heart disease.
These epidemiological studies support the reverse J-shaped curve, where the risks of CVD and premature mortality decrease in a dose-dependent fashion, until a threshold of physical activity is exceeded, at which point many health benefits are attenuated compared to less intensive doses of physical activity. It has been speculated that high volumes of moderate and high-intensity physical activity can trigger ventricular arrhythmias or sudden cardiac death in individuals with underlying cardiac conditions during or after their activities [[16]].
Physical activity is a broad term that encompasses exercise; the latter including additional measurement criteria, including the type (e.g., aerobic versus resistance exercise), frequency, duration, as well as intensity of exercise, and is targeted at improving different aspects of fitness (e.g., cardiopulmonary fitness, muscular endurance, and strength), thus conferring even more robust physiological adaptations than walking. In this review, we revisit the pillars of aging [[17]] and discuss how exercise, both from a single session (acute) and habitually (chronic), targets and modulates each of these pillars, thus staving off chronic disease and maintaining a longer healthspan.
Molecular pillars of aging
The imprints of aging are ubiquitous from molecules to cells, and in all tissues and organs. Advancement in geroscience indicates that multiple processes are responsible for the progressive accumulation of cellular damage and loss of functions with age [[18]].
Broadly, scientists have identified seven interconnected processes that drive the aging process in different organisms, including macromolecular damage, dysregulated stress response, disruption in proteostasis, metabolic regulation, epigenetic drift, inflammaging, and stem-cell exhaustion (Fig. 1) [[17]]. Scientists can rely on these conserved drivers of aging to modify the underlying biological mechanisms of aging and hence derive novel interventions that may potentially slow the rate of aging. Challenges remain, however, in dissecting the hierarchical connections between the pillars of aging and their relative contributions to aging. Fundamentally, the rate of biological aging is coupled to the age-dependent disruption of these homeostatic networks.
Interventions such as exercise training, which can positively influence multiple pillars of aging, will offer broad-range benefits to delay aging, maintain functional capacity, and delay the onset of chronic diseases of old age. The significance of each hallmark is briefly discussed in the subsequent sections, with the potential role of long-term exercise training in modifying each hallmark also discussed. It is beyond the scope of the present review to delve deeply into the pillars that are modulated by exercise. Instead, some of the pertinent features of the pillars that respond to exercise training will be summarized (Fig. 1).
Macromolecular damage
Aged cells accrue substantial damage to DNA, proteins, and lipids from the exposure to exogenous and endogenous stressors through the passage of time [[19]]. The external stressors include environmental toxins and UV radiation, while reactive oxygen species (ROS), a byproduct of mitochondrial respiration, is a major internal cause of macromolecular damage [[20]]. The accumulation of macromolecular damage impedes cell function and collectively increases organismal functional deficits, eventually leading to age-associated diseases. The consequences of protein damage in aging are covered in “Proteostasis”; herein, the effects of genetic damage accrued throughout life are discussed. Somatic mutations, translocations, deletions, chromosomal aneuploidies, and gene disruptions by insertion of viruses and transposons have been associated with increased genomic instability seen in aging [[18]]. Increased somatic mutations in mitochondrial DNA (mtDNA) are also highly represented in aged cells [[21]]. These DNA alterations impact transcription, leading to dysfunctional mitochondria and cells. DNA damage elicits cellular responses that may lead to apoptosis or cell senescence, resulting in stem-cell depletion and deregulation in tissue and organismal homeostasis and renewal [[18]]. Besides genomic damage, malfunction or inappropriate stimulation of the mitochondrial and nuclear DNA repair systems cause accelerated aging in mice and in human progeroid syndromes [[22-24]]. Telomere attrition during somatic cell division constitutes another type of age-related chromosomal deterioration and has been linked to chronological aging [[25, 26]]. Telomere shortening increases susceptibility to rampant chromosomal end fusions and activates a prolonged DNA damage response, leading to the induction of cell senescence [[18]]. Many current aging research focus on strategies that reinforce macromolecular stability, such as enhancing repair mechanisms, to delay aging rather than to target the source of damage, which by far has yielded inconsistent results.
Exercise training upregulates endogenous antioxidant concentration and enzyme activities that modulates DNA repair. A study in red skeletal muscle of 21-month-old male rats demonstrated that 48 h after cessation of daily treadmill running (10 weeks), the content of 8-hydroxy-2′-deoxyguanosine (8-OHdg), a DNA lesion, was decreased, whereas the activities of Oxoguanine DNA glycosylase (OGG1) and Uracil DNA glycosylase (UDG) were increased, by 31% and 43%, respectively [[27]]. As well, Safdar et al. [[28]] showed that treadmill running rescued a number of aging phenotypes in a progeroid mouse model with concomitant upregulation of DNA repair enzymes. In this study, experimental male and female mice were heterozygous for the mitochondrial polymerase gamma knock-in (C57BL/6J, PolgA+/D257A), a mutation resulting in increased point mutation in mtDNA and leading to systemic mitochondrial dysfunction, accelerated aging and a short lifespan. At 3 months of age, mice were subjected to treadmill running (3 days·week−1; 45 min·day−1) for 5 months. Remarkably, early mortality was completely attenuated in mice that ran, bringing them to equivalent survivorship to that of wild-type mice. In addition, whereas mutator mice in the control (nontreadmill exercised) group displayed increased mtDNA depletion in multiple organs and tissues (skeletal muscle, heart, liver) relative to wild-type mice, this deficiency was completely abrogated in treadmill-ran mutator mice, with concomitant improvements in skeletal muscle mitochondrial biogenesis. Whether exercise would elicit benefits in other progeria models remains largely untested.
The capacity to initiate DNA repair mechanisms after exercise in humans is similar to preclinical animal models. In healthy, recreationally active young men (n = 14), an acute session of stationary cycling to achieve VO2max followed by 30 min of cycling at 85% of VO2max, increased double-stranded DNA breaks (quantified as Ɣ-H2AX foci) in PBMCs, with a peak in numbers of Ɣ-H2AX positive foci immediately after exercise, and subsequent decreases at 2 and 4 h postexercise [[29]]. These increased DNA lesions after exercise coincided with a parallel trend in 53-Binding Protein 1 (53BP1, a DNA repair protein) foci expression per cell, with the highest number of foci detected immediately after exercise and gradual reductions at 2 and 4 h. The results suggest DNA repair proteins are activated by exercise and may be involved in repairing double-stranded DNA breaks. It is probable that DNA damage repair capacity is influenced by aerobic fitness. In another study [[30]], healthy and sedentary young men (n = 6) with low aerobic capacity (VO2max < 45 mL·kg−1·min−1) demonstrated an attenuated DNA repair capacity in ex vivo-irradiated peripheral blood mononuclear cells (PBMCs) after an acute bout of cycling to exhaustion, compared with endurance-trained male athletes (n = 6) with higher aerobic capacity (VO2max > 55 mL·kg−1·min−1).
While the evidence presented above suggests that exercise training in preclinical animal models and young healthy individuals is associated with increased DNA damage repair mechanisms, no studies have yet demonstrated this in older humans and should be investigated in future work.
Epigenetic drift
Renewed interest in the application of epigenetics in geroscience research emerged in 2005, with the finding that young (~ 20 years old) monozygotic twins displayed few epigenetic differences, but such differences were more pronounced in older (~ 50 years old) monozygotic twins, a phenomenon termed the “epigenetic drift” [[31]]. Thus, epigenetic changes during aging may be a useful biomarker, given that changes therein may reflect the rate of biological aging, consequent to environmentally induced or spontaneous changes in the methylome. Genome-wide and CpG site-specific methylation experiments have been performed in recent years to identify the epigenetic regulators of aging and have been considered to be the most robust in predicting biological age [[32]]. The earliest epigenetic “clock” algorithms were developed by Bocklandt et al. in 2011, wherein salivary DNA was obtained from 34 male twins aged between 21 and 55 years, and a prediction model was then trained on a group of men and women (n = 60; 18–70 years) [[33]]. The study reported three methylation sites that correlated with advanced age and explained ~ 70% of the variation in chronological age. More refinement in epigenetic age testing continued in the next 2 years. In ~ 650 Caucasian and Hispanic volunteers (19–101 years), genome-wide methylation status was assessed across ~ 450 000 CpG markers from whole blood, and the rate of aging was predicted from 71 methylation markers using a penalized multivariate regression model [[34]]. The age prediction algorithm developed from this study was subsequently referred to as the “Hannum Clock,” soon to be followed by the “Horvath Clock”—developed a year later, also using whole-genome sequencing data and penalized regression statistics, which identified 353 CpG sites from 82 public methylation array datasets from multiple human tissues [[35]]. These first-generation epigenetic clocks were trained to predict chronological age, with newer methylation clocks emerging in the last 3 years that were trained on physiological correlates that were indicative of risk of morbidity or mortality [[36, 37]].
These population cohort studies demonstrate promise for DNA methylation as a promising aging biomarker, and potential target for exercise intervention. In this regard, however, recent evidence is less convincing, as neither the Finnish twin cohort study nor a subcohort of the Lothian Birth cohort study, both of which involved whole epigenome sequencing as well as incorporating the Horvath clock algorithm, have shown significant effects of lifetime exercise on differences in DNA methylation status [[38, 39]]. Given that exercise epigenetics is a nascent field, more studies in different populations and exercise types etc. should be evaluated.
Disruption in proteostasis
The regulation of de novo protein synthesis, folding, assembly, as well as its export, breakdown and degradation, sums up the proteostatic network (PN) in mammalian systems. A disruption in proteostasis is intimately linked to aging and age-related diseases, such as Alzheimer's and Parkinson's disease [[40]], sarcopenia [[41]], and atherosclerotic cardiovascular disease [[42]]. During organismal aging, the ability to preserve proteome solubility and functionality is compromised in many cells and organs due to a failure of the PN [[43]]. Age-related impairments observed in the different quality control components of PN can involve molecular chaperones, the proteasome, as well as the process of autophagy. In neurodegenerative diseases, the deterioration of the PN is often exacerbated by the presence of mutated proteins. Cellular stress such as oxidative damage can lead to proteostasis imbalance and result in the persistent presence of damaged proteins that accumulate as toxic protein aggregates, leading to proteotoxicity in the aged and diseased cells. To restore protein homeostasis, cellular defense programs are upregulated in the mitochondria, the endoplasmic reticulum (ER) and the cytosol, and are respectively, the mitochondrial unfolded protein response UPRmt; activated by the mitonuclear balance ratio [[44]], the unfolded protein response in ER (UPRer) and finally, the heat shock response (HSR) [[42]]. These defense mechanisms are evolutionarily conserved across diverse eukaryotes [[44-46]] and are relevant molecular targets in aging biology.
Many preclinical studies support a close relationship between proteostasis and healthy aging [[43]], where proteomes of many long-lived animals exhibit exceptional stability and enhanced resistance to stress. Moreover, lifestyle interventions that enhance the integrity of the PN were demonstrated to successfully extend healthspan and lifespan in various experimental models [[47]]. Furthermore, the various protein quality control mechanisms can communicate and compensate for each other, both intra- and extracellularly [[48]]. As an illustration, the UPRer, UPRmt, and HSR responses interact with one another. Proteotoxic stress activates the HSR, and in mammals, a widely observed phenomenon is the upregulation of molecular chaperones, such as heat shock protein (Hsp)70, which helps to maintain protein folding in the cytosol [[49]]. Hsp70 comprises a group of molecular chaperones that mediates HSR, the UPRmt and UPRer. During homeostasis, Hsp70 binds to the transactivating domain of HSF1 in the cytosol, repressing its transcriptional activity [[49]]. Mis-folded cytosolic proteins get bound by Hsp70, whereupon HSF1 is released, trimerizes and transcriptionally activates nuclear genes by binding to promoters with specific HSF-1 binding sites, including that of Hsp70. In addition, Hsp70 can also bind newly synthesized proteins to be transported into the mitochondria, where translocase of outer membrane (Tom)40, a mitochondrial membrane channel protein, and its associated receptor proteins, Tom20 and Tom70, would interact with, and form a complex with Hsp70 and its protein cargo [[50]]. Tan et al. also demonstrated that during heat stress, single-stranded DNA-binding protein (SSBP)-1, which is involved in mtDNA replication and maintenance, would translocate from the mitochondria to the nucleus to form a complex with cytosolic HSF-1 [[51]]. This SSBP-1/HSF complex was shown by the authors to directly induce transcription of Hsp70, as well as other molecular chaperones during heat shock via the recruitment of brahma-related-gene (BRG)1, a chromatin-modifying enzyme complex. Critically, when SSBP-1 was knocked down with adenovirus expressing short hairpin RNA (shRNA), the induction of the chaperone genes was reduced several folds, suggesting a critical role for SSBP-1 in protein transport/chaperoning into the mitochondria. The role of UPRmt and UPRer serves to dampen the protein load and increase the activity of molecular chaperones and proteases to re-establish homeostasis, failing which, would trigger the autophagic pathway [[52]].
Exercise and the UPRmt
Exercise training targets and activates some of the defense mechanisms implicated in the PN, including the UPRmt, UPRer, and HSR. In two preclinical animal studies, 4 weeks of daily treadmill running (60% maximal speed to exhaustion; 60 min per session) [[53]] or 4 weeks of high intensity interval treadmill running [[54]] for 5 days·week−1 (10 × 2 min bouts at 85–90% VO2max) induced mitonuclear imbalance in skeletal muscle of 24–25-month-old, male C57BL/6 mice, relative to sedentary controls (gastrocnemius muscle samples were harvested 24 h after the last exercise session), as evidenced by the increase in the mitochondrially encoded cytochrome c oxidase I (MTCO1)/Succinate dehydrogenase complex subunit A (SDHA) ratio. These two studies also demonstrated that exercise induced upregulation of proteins involved in the UPRmt, such as the ATP-dependent Zn metalloprotease-1 (Yme1L1) and the Lon Protease homologue-1 (Lonp1) in skeletal muscles of the aged mice [[53, 54]]. In addition, the investigators reported increased protein expression of citrate synthase, NRF-1, and mitochondrial transcription factor A (Tfam)—all factors implicated in enhanced mitochondrial biogenesis. Hence, aerobic exercise induced UPRmt response in skeletal muscle, which may favorably upregulate mitochondrial oxidative capacity.
Exercise and the UPRer
The UPRer responds favorably to intense skeletal muscle contraction similar to the UPRmt, in that a 7-day low frequency (10-Hz) electrical stimulation of rat skeletal muscle led to the upregulation of ER stress response genes, such as increased transcription of activating transcription factor (ATF)4 and sliced X-box binding protein (XBP)1 by 1.5- and 3.3-folds, respectively, concomitant with transcriptional and translational upregulation of CCAAT-enhancer binding protein (C/EBP) homologous protein (CHOP) and binding immunoglobulin protein (BiP) [[55]]. A key finding in this study was that the UPRer responses occurred prior to mitochondrial adaptations and induction of autophagy, suggesting that the exercise-induced stress response triggered through the UPRer is an early signaling event. In addition, treatment with tauroursodeoxycholic acid (TUDCA, a naturally occurring bile acid that blocks the UPRer response) attenuated CHOP and Hsp72 protein expression, despite continued electrical stimulation of the muscles, demonstrating that modulation of the UPRer response to skeletal muscle contraction can be partially inhibited pharmacologically. Most importantly, however, was the finding that despite the blocking UPRer with TUDCA, there were no significant changes in peroxisome proliferator-activated receptor gamma coactivator (PGC)-1α mRNA nor mitochondrial cytochrome oxidase (COX) enzymatic activity, suggesting the presence of parallel pathways responsible for induced mitochondrial biogenesis.
Few studies have explored the effects of exercise on the UPRer responses in older human participants. At present, the effects of aerobic exercise on the UPRer are unknown, with only a handful of studies examining the effects of resistance exercise. A single session of resistance exercise (unilateral leg extension exercise at 70% of 1-RM; 8 sets of 10 repetitions) upregulated UPR-related genes in skeletal muscle tissue from young (n = 12; 18–35 years) and older (n = 12; 65–85 years) men and women (vastus lateralis muscles were biopsied at 18 h after the exercise session), although the increase did not reach statistical significance in the older participants [[56]]. Furthermore, the acute exercise resulted in a greater enrichment of genes associated with the Protein Kinase R-like ER kinase (PERK) pathway, which represents a signal activator of the UPRer [[57]], notably in the young participants, but with a reduced response in the older participants. Ogborn et al. [[58]] and Hentila et al. [[59]] reported contrasting observations in similar studies. First, a single session of resistance exercise comprising knee extension (4 sets of 10 repetitions at 75% 1-RM) in younger (n = 9; ~ 21 years) and older (n = 9; ~ 70 years) untrained men resulted in comparable upregulation of skeletal muscle PERK protein expression 48 h after exercise, although PERK mRNA remained unchanged [[58]]. Furthermore, mRNA expressions of ATF4 and CHOP were unaffected by the exercise intervention, whereas ATF6 and inositol-requiring enzyme (IRE)1α mRNA were upregulated at 24- and 48-h postexercise, suggesting that specific UPR pathways were modulated by a single session of resistance exercise [[58]]. Hentila et al. [[59]]. showed that a single session of bi-lateral leg press (5 sets of 10 repetitions) upregulated the protein expression of ATF4 and BiP in skeletal muscle tissue from younger (n = 12; ~ 27 years) and only ATF4 in older (n = 8; ~ 62 years) men, 48 h postexercise. Curiously, in this study, 7 months of resistance exercise training (2 sessions per week; progressive whole-body approach utilizing multiple repetition and sets) did not result in changes in the UPRer protein markers in skeletal muscle tissue.
Aside from skeletal muscle, the UPRer response to exercise also extends to other tissues in human participants. An 8-week resistance training program (2 sessions per week of whole-body program) in older healthy men and women (n = 30; ~ 72 years) resulted in increased PERK phosphorylation and ATF4 and XBP1 protein expression, in PBMCs (obtained 5–6 days before, and after exercise training program) concomitant with increased PGC-1α protein levels [[60]]. The fact that the UPR response to exercise is conserved in PBMCs suggests that resistance exercise can exert modulatory effects on proteostasis in a systems manner. Future studies should investigate whether myokines released from skeletal muscle are upstream effectors these responses.
Exercise and the HSR
Aging is associated with decreased cellular and induced response of Hsp72 to physiological stress in animal models [[61]] and human volunteers [[62]]. In contrast, a higher concentration of circulating Hsp72 was associated with lower hand grip strength, lower muscle mass, and slower gait in older community dwelling adults [[63]], and also reported in older adults with impaired insulin sensitivity [[64]] and type 2 diabetes [[65]]. Curiously, cellular mRNA expression of Hsp72 is attenuated in skeletal muscle of adults with type 2 diabetes [[66]], compared with healthy, age-matched controls, while both mRNA and protein expression of Hsp70 were depressed in adipose tissue and liver of adults with obesity-associated nonalcoholic fatty liver disease (NAFLD) [[67]]. The apparent dichotomy between decreased cellular Hsp72 content and expression, and higher systemic Hsp72 in aging can be resolved by the fact that cellular Hsp72 (intracellular) maintains proteostatic homeostasis by chaperoning proteins into either the mitochondria or ER as described previously, whereas Hsp72 in the systemic circulation (extracellular) is a “danger” signal, as it modulates stress-induced inflammatory responses [[68, 69]], including the acute exercise-induced inflammatory response, which is considered a physiological stressor [[70]].
A single session of physical exercise induces the expression of Hsp72 in different cells, including hepatocytes, myocytes, and immune cells, although the release of extracellular Hsp72 seems to originate primarily from the liver [[71]]. A pre-clinical study demonstrated an increase in the transcription of, and protein synthesis of Hsp72 in liver cells from 4-month old rats, after 60 min of treadmill running at speeds between 15–27 m·min−1 [[72]]. Human studies also corroborated the role of the liver in the exercise-induced release of Hsp72 release. Hepato-splanchnic release of Hsp72 was elevated in young healthy men (22 years) after a 120-min bout of cycling at 65% VO2max, as evidenced by increased net concentration of Hsp72 in the venous–arterial blood flow through the median cubital vein [[71]]. In addition to the liver, Hsp72 mRNA and protein expression were increased in contracting skeletal muscle after two-legged, knee extension exercise (4–5 h, at 40% peak power output until exhaustion), and this induction was only evident when muscle glycogen was depleted during the exercise, and not in glycogen abundant muscle fibers in the rested state [[73]].
Exercise-induced Hsp72 elevation in systemic circulation has also been observed in older adults. A single session of treadmill exercise until exhaustion increased Hsp72 protein expression by 1.4-fold in lymphocytes from physically active 85-year-old men and women, although this was not observed in either 25- or 65-year-old adults [[74]]. The same investigators [[75]] reported that 8 weeks of brisk walking at ventilatory threshold (40–48 min, 3 days·week−1) did not significantly change intracellular Hsp72 protein expression in leukocytes obtained from a small sample of septuagenarians (n = 8) supplemented daily with vitamin C (500 mg) and E (100 mg). In a different study, Konopka et al. [[76]] reported that 12 weeks of cycle ergometry improved VO2max by 30%, and downregulated Hsp70 mRNA, but not protein content, in skeletal muscle of older women (n = 9, ~ 70 years).
Resistance exercise has also been demonstrated to induce Hsp72 expression in aged animals. For instance, electrical stimulation designed to recapitulate the stretch-shortening contraction of resistance exercise training was applied to the tibialis anterior of anesthetized rats for 4.5 weeks [[77]]. While no changes in Hsp72 mRNA were observed in skeletal muscle of young (3 months) and middle-aged (10 months) rats, protein expression was increased by 9- and 4-folds, respectively. Few studies have been conducted to investigate the effects of resistance exercise on Hsp72 response in older adults. Yoon et al. [[78]] reported that light- or moderate-intensity resistance exercise training comprising of leg extension exercise training (light: 3 sets of 25 repetitions at 40% 1 repetition maximum (RM) combined with localized heating; moderate: 3 sets of 15–18 repetitions at 60% of 1 RM 3 times per week; 12 weeks) did not result in differences in systemic concentrations of = Hsp72 in healthy elderly women (n = 13; 65–75 years). Conversely, Perreault et al. [[79]] reported that 16 weeks of resistance exercise training (3 sets of 8 repetitions at 80% 1-RM; 3 times per week) in older (60–75 years) sarcopenic men (n = 26) decreased plasma concentrations of Hsp72 protein.
Of clinical relevance is that exercise in general has a beneficial effect on the heat shock response. Heat treatment of leukocytes harvested from healthy middle-aged (~ 49 years) or healthy older (~ 64 years) adults resulted in a 1.5-fold increase in Hsp72 protein concentrations in cell media, whereas there were no concentration changes in media when replicated with leukocytes from older diabetics (~ 69 years) [[80]]. Remarkably, resistance exercise training (12 weeks, 3 times per week, whole body exercise, 2–3 sets at 12–15 repetitions) increased supernatant protein concentrations of Hsp72 from these older type 2 diabetics by nearly 5-fold from baseline concentrations in vitro. Similarly, 8 weeks of aerobic exercise training (cycle ergometer, 3 times per week, 20–50-min progression, ~ 70–80% of peak heart rate) increased Hsp72 protein expression in skeletal muscle in diabetic men (~ 63 years; n = 7) [[81]].
Although the reviewed papers comprise diverse subject characteristics, tissue types, and exercise regimens being studied, the available evidence indicates a modulatory effect of aerobic or resistance exercise training on the heat shock response in older individuals. Future studies need to delineate the molecular significance of the systemic and tissue responses of Hsp70/72 to aerobic and resistance exercise, given that their intracellular and extracellular roles as molecular chaperones, or inflammation modulators, respectively.
Exercise and protein turnover: gold standard for measuring protein kinetics
While the previous sections describe the signaling responses of molecular chaperones of proteostasis to exercise, the translational implications need further elucidation, since the gold standard for proteostasis quantification is to measure fractional protein synthesis and degradation in vivo. This is further complicated by the wide variation in the half-lives of intracellular proteins within and between different mammalian cells. For instance, the median half-life of ~ 800 intracellular proteins from hepatocytes is ~ 8.7 h, with the majority ranging between 4 and 14 h [[82]]. Comparatively, in skeletal muscle, protein turnover is different between cytosolic and mitochondrial proteins, as well as between muscle fiber types. Kruse et al. [[83]] demonstrated that the median half-life of mitochondrial proteins in fast-twitch, young female C67BL/6 skeletal muscle fiber, in this case, extensor digitorum longus (EDL) was 24.9 days, whereas the median half-life of other cytosolic proteins in EDL was 22.2 days. This was also observed in the EDL harvested from aged C57BL/6 female mice, where the median half-life for the same comparison was 25.9 and 22.8 days, respectively. These observations were in contrast to different muscle fiber types, where the median half-lives of cytoplasmic proteins from soleus muscle fibers (slow-twitch) obtained from young and old C57BL/6 female mice were 16.4 and 16.8 days, respectively, representing a ~ 27% difference in turnover rate. In another preclinical study involving rats [[84]], the investigators found that fractional synthesis rate of skeletal muscle proteins on a per-protein basis differed significantly—proteins associated with fast-twitch fibers had faster turnover rates, compared with proteins associated with slow-twitch fibers.
In human studies, intra-individual variability in protein turnover and even between different proteins obtained from the same subject has also been reported. The protein synthesis rate in skeletal muscle in response to a 4-week sprint interval training (9 sessions; 4–8 bouts of cycling at maximal effort for 30 s) was measured using deuterium oxide, and varied in a gender- and protein-specific manner, among a group of young, recreationally active adults (n = 21; 21–24 years) [[85]]. Specifically, there was greater synthesis of total and cytosolic protein in male skeletal muscle after the exercise intervention than in female skeletal muscle. In addition, while mitochondrial biogenesis was increased, as evidenced by greater post-exercise protein concentration of PGC-1α in muscle from both genders, there were no significant increases in COX IV protein content for either male or female participants. Thus, while many investigators have traditionally used exercise-induced changes in cellular protein content as a marker of improved proteostasis, those findings need to be considered in a protein-, and even organelle- and tissue-specific manner. Future studies that intend to quantify proteostasis in humans should be evaluated from a combined in vivo approach using deuterium oxide as well as localized sampling of protein markers from skeletal muscle.
Metabolic dysregulation
Metabolic dysregulation, for example, hyperglycemia and insulin resistance, increase with age, with numerous studies supporting the contribution of defective mitochondria [[86-89]]. Skeletal muscle fibers of old rats contain lower copy numbers of mtDNA than those of young rats [[90]], which was also accompanied by reduced COX transcripts in highly oxidative muscles (soleus), but not in glycolytic muscles (gastrocnemius). In humans, advanced age is also associated with increased mtDNA deletions in human skeletal muscle [[91]]. These mitochondrial deficits are associated with lower substrate oxidation, particularly from lipid sources, which is characteristic of diabetic skeletal muscle [[92]]. Reduced fat oxidation causes accumulation of intracellular fatty acid intermediates and activates ROS. Consequently, ROS induce lipid peroxidation and damage to mitochondrial proteins, thus increasing oxidative stress and reducing the [ATP] available for normal cellular functions [[93]]. Depletion of cellular [ATP] precedes cell apoptosis, especially of type II muscle fibers [[94]], which in older adults, also coincides with the onset of sarcopenia, an age-associated loss in skeletal muscle mass and function.
Exercise mitigates metabolic dysregulation and sarcopenia, with aerobic or strength/resistance exercise targeting disparate, as well as overlapping molecular pathways. In a preclinical study, 3 months of treadmill running (17.5 m·min−1 at 10% incline, 45 min·day−1, and 5 days·week−1) increased gene expression and nuclear content of PGC-1α in skeletal muscle of 22-month-old Fischer 344/BNF1 rats by two-fold, compared with either young or old rats in the sedentary group [[88]]. Exercise-induced upregulation of PGC-α was concomitant with upregulation of Tfam, Cyt c and mtDNA content in these animals. Exercise training in human volunteers has also been demonstrated to corroborate these preclinical findings. In previously sedentary, but healthy older adults (~ 67 years; 3 women, 5 men), 12 weeks of aerobic exercise training (treadmill or stationary bicycle exercises; 30–40 min at 50–70% VO2max and 4–6 sessions weekly) increased skeletal muscle mtDNA content, and activities of complex I and II in the electron transport chain (ETC) [[95]]. In parallel, Short et al. reported improved VO2peak, increased enzymatic activities of mitochondrial proteins in skeletal muscle, as well as upregulation of genes involved in mitochondrial biogenesis, including PGC-1α, NRF-1, TfAM [[96]]. Healthy men and women (n = 65; 21–87 years) in this study underwent 16 weeks of aerobic exercise training (stationary cycle ergometry; 3–4 sessions per week; 20–40 min per session; 70–80% of maximal heart rate). Importantly, the exercise program also improved skeletal muscle glucose transporter (GLUT)4 mRNA and protein by 30–52%, while reducing abdominal fat mass (5%) and circulating triglycerides (25%). These adaptations were observed in the older participants, suggesting that such exercise-mediated parameters were well preserved even in advanced age.
Resistance exercise training has also been shown to be efficacious in improving skeletal muscle metabolism, particularly from the mitochondrial aspect. A 14-week whole-body training program (3 sessions per week; 3 sets × 10–12 repetitions; 50–80% of 1-RM) increased mitochondrial creatine kinase concentrations as well as complex IV enzymatic activity in skeletal muscle tissues of older men and women (n = 28, ~ 69 years) [[97]], suggesting an improvement in muscle oxidative capacity. Curiously, skeletal muscle from untrained older adults (n = 25; ~ 70 years, 12 women) showed ~ 600 genes expressed differentially, compared with young adults (n = 26; ~ 21 years, 14 women) [[98]]. When the older adults underwent whole-body resistance exercise (2 times per week over 26 weeks; 3 sets of 10 repetitions each, progressing from 50% to 80% of 1-RM), their muscle transcription signatures changed to resemble those of the younger participants [[98]].
Dysregulated stress response
Enhanced stress resistance is characteristic of long-lived animals and humans [[99]]. For example, a more efficient DNA double-strand break repair is key to the extreme longevity in long-lived species [[100]]. Stress response is an evolutionarily conserved, adaptive mechanism that mobilizes a vast network of cellular processes to counteract a wide variety of stressors and boost cell survival. Age-dependent decline in stress response is seen in aging. Conversely, augmenting stress adaptation robustly correlates with lifespan extension across species [[18]]. It may be easier to modulate the ability of the organism to combat stress than to eradicate all sources of stressors. This is particularly so as different levels of stressors exhibit varied effects on the rate of aging. While strong acute or mild chronic stressors results in exhaustion or dysregulation of an efficient stress response, which accelerates aging, some mild stressors appear to fortify resilience through hormesis. Hormesis describes a bi-phasic response to biological insult or stress, where low doses of the stressor can result in tolerance and adaptation of the organism, whereas high doses can have deleterious consequences. In this section, we begin with a discussion of antioxidant supplementation during exercise training and how the consumption of antioxidants blunts some aspects of exercise training responses in human participants. We then focus on mitohormesis as a classic example of hormesis, where moderate mitochondrial oxidative stress primes cells to upregulate stress responses (e.g., mitochondrial unfolded protein response; UPRmt) that confers cells with a cytoprotective edge against future mitochondrial insults [[101]]. These protective adaptations had been discussed earlier in the Proteostasis section; mitochondrial quality control and biogenesis are discussed in the following subsections.
Oxidative stress and antioxidants
The definitive nature of exercise training involves repeated bouts of skeletal muscle contraction, during which both skeletal muscle tissue and other visceral organs and tissues are exposed to changes in mechanical load, oxygen tension, metabolites, temperature, and oxidative stress at multiple levels—cellular, tissue, and systemic. Exercise-induced oxidative stress is a curious paradox as it fits the hormesis criterion. First, regular moderate-intensity exercise training upregulates cellular concentrations of antioxidants (e.g., super oxide dismutase) as well as their enzymatic activity in skeletal muscle [[102]] and cardiac muscle [[103]], which mitigates exercise-induced elevation in oxidative stress [[27]]. However, prolonged intense exercise without sufficient postexercise recovery can increase oxidative stress that damage skeletal muscle fibers as well as their cellular components. This redox imbalance stemming from excess ROS or reactive nitrogen species (RNS) negate positive adaptations to exercise training and is a purported mechanism that explains the overtraining syndrome (OTS)—a state of perpetual decline in physical performance, with lower force production and disturbed mood states typically lasting months, usually afflicting elite athletes [[104]]. It is unlikely that regular exercise in the general population would perturb the redox state negatively in most adults, unless there is underlying chronic disease or they engage in ultra-endurance exercises, which in the past decade have become extremely popular with the general public. These types of exercises can result in pathological cardiac wall remodeling [[105]], likely a consequence of extensive ROS production and impaired clearance [[106]], as well as inflammation, both of which can drive cardiac edema.
Although the oxidative-damage hypothesis suggests that increased ROS can result in redox imbalance, tissue damage, and reduced exercise adaption, some studies have shown the contrary, where supplementing with oral antioxidants paradoxically prevented some beneficial effects of exercise, at least partially. In one study, 8 weeks of vitamin C supplementation (1000 mg) in 14 young men (27–36 years) blunted the aerobic training adaptation [[107]]. Specifically, VO2max increased by 22% in the participants not taking vitamin C, and by 10.8% in the supplemented group. Given the small sample size, it was possible that the power was not high enough to detect significant differences. Ristow et al. [[108]] conducted a different study that corroborated the findings from Gomez-Cabrera. In this study, healthy young men (n = 39) underwent vitamin C (1000 mg daily) and vitamin E (400 IU daily) supplementation in combination with 4 weeks of combined aerobic and circuit resistance training. Glucose infusion rates during a hyperinsulinemic-euglycemic clamps, together with markers of insulin sensitivity, such as plasma adiponectin, and skeletal muscle expression of endogenous antioxidants (super oxide dismutase- SOD1 and 2; glutathione peroxidase), and markers of mitochondrial biogenesis, for example, peroxisome proliferator-activated receptor gamma co-activator 1-alpha (PGC-1α), were upregulated only in the non-supplemented group, demonstrating that antioxidants blunted the molecular changes with combined aerobic and circuit resistance exercise training.
Not all exercise studies show a blunted effect of antioxidant supplementation. In one study [[109]], 54 young men and women were stratified into a vitamin C (1000 mg) and E (235 mg), daily supplementation group or a placebo-control group for 11 weeks, while undergoing aerobic training three to four times per week. Skeletal muscle mRNA expression of cytochrome c oxidase subunit IV (COX 4) and PGC-1α increased by ~ 60% and 20%, respectively, in the placebo group, whereas the gene expression for the two markers decreased by ~ 13% in the supplementation group. Curiously, both groups of participants demonstrated equivalent increase in their VO2max, suggesting that nonmitochondrial-related mechanisms may be responsible for such training adaptations. For instance, vitamin C consumption (0.24 mg·cm−2 of body surface area (BSA) daily for 6 weeks, about 4-fold higher than in the human dose) downregulated PGC-1α, nuclear respiratory factor (NRF)-1, and mitochondrial transcription factor A (TFAM) in rat skeletal muscle, obtained 72 h after the last bout of treadmill running [[107]]. However, 500 mg of daily vitamin C supplementation in healthy men (18–40 years) did not change the lactate threshold, muscle glycogen storage, nor enzymatic activities or protein concentrations of citrate synthase (CS) and MnSOD [[110]] (muscle biopsies were performed between 72 and 96 h after the final training session) after 12 weeks of cycle training (5 days·week−1), respectively. Finally, Wyckelsma et al. [[111]] reported that 3 weeks of daily vitamin C (1 g·day−1) and E (235 mg·day−1) supplementation in older men (~ 65 years) did not impair power at exhaustion during incremental cycling. Given the different study designs and exercise modes and training intensities/durations, etc., the jury is still out with respect to the efficacy of consuming antioxidants over the long-term, given their potential interference with beneficial adaptations to exercise. A caveat is that many exercise training studies in humans typically recruit young healthy participants where the consumed antioxidants may not be beneficial, compared with middle-aged or older adults who may have decreased capacity to buffer ROS. Hence, exercise training studies involving older adults, from a broad spectrum of health states should be conducted to determine how oral antioxidants interact with exercise training to influence tissue and systemic redox balance.
Exercise and hormetic responses: ROS as intermediary signals in mitochondrial biogenesis and remodeling
As presented in the earlier subsection, antioxidant supplementation appears to abrogate many of the physiological benefits conferred by exercise training. In the past decade, emerging data from exercise studies have supported a role for ROS in cellular signaling, where they potentially mediate exercise-induced, beneficial physiological effects [[112]] and hence support the hormetic theory of a threshold effect of biological stressors (e.g., toxins and chemicals). In the exercise epidemiology/immunology literature, it has been consistently reported that the optimal protection for cardiometabolic and immune health appears to coincide at the moderate-intensity level [[113]], which interestingly, also corresponds to the stimulatory range of 30–60% above basal levels that were observed to drive hormetic adaptations in other biological phenomena [[114]].
Radak et al. [[115]] were the first to propose in 2005 that exercise-induced ROS may serve as important signaling mediators to drive physiological tolerance and adaptation. They specifically proposed that the intermittent, albeit transient, elevated concentrations of ROS mediate signaling pathways (e.g., NF-κB-mediated transcription), or induce mild DNA damage in various cells that stimulate DNA repair and endogenous antioxidant responses. Since then, the role of exercise-induced ROS in the control of mitochondrial quality and quantity (mitostasis) has piqued the interests of exercise and aging biologists.
Mitostasis is tightly regulated in mammalian cells. A typical lifecycle of this energy producing organelle involves the obligatory quality control of biogenesis, fusion, fission, and mitophagy. These events shape the physical structure of the organelle according to cellular metabolic demands [[116]]. Mitochondrial biogenesis is the de novo increase in populations of new mitochondria within cells and involves the co-ordinating efforts of the master transcriptional factor, PGC-1α. Numerous studies since the mid-2000s [[117-119]] have shown aerobic exercise to upregulate PGC-1α phosphorylation and activation in skeletal muscle. The cascade of events following PGC-1α activation includes activation of other transcriptional factors, such as TFAM, NRF-1, and Nuclear factor erythroid 2-related factor 2 (NRF-2). These transcription factors are key mediators of mitochondrial gene transcription, such as those genes encoding for proteins in the ETC, as well as antioxidant genes. Mitochondrial biogenesis regulates the coordinated expression of mitochondrial proteins encoded by both nuclear and mitochondrial genes; hence, the structural and functional quality of mitochondria depends on efficient import of intact proteins from the cytosol, as covered in the Proteostasis section. Age-related increases in ROS production, and poor scavenging of the free radicals due to diminished antioxidative capacity can result in mitochondrial damage, wherein damaged mitochondria can fuse with healthy counterparts to alleviate stress and continue to supply the requisite ATP needed by the cell. Fusion thus compensates for faulty organelle components in mitochondria by sharing healthy resources such as RNA or protein [[116]]. Fission, on the other hand, is driven by irreparable mitochondria and leads to the segregation of such damaged mitochondria for degradation through the autophagosome-lysosomal pathway (mitophagy), which leaves only healthy mitochondria within the mitochondria pool [[116]].
The benefits of exercise-induced, ROS-mediated mitochondrial adaptations (mitohormesis) can be observed from the reciprocal crosstalk between the redox-sensitive transcription factor, NRF-2, and the master regulator of mitochondrial biogenesis, PGC-1α [[120]]. A preclinical study by Merry and Ristow illustrates this symbiotic relationship [[121]], where NRF-2−/− mice demonstrated an abrogation of mtDNA content and citrate synthase activity, as well as decreased TFAM and COX4 gene expression in skeletal muscle after 6 weeks of treadmill running, compared with wild-type littermates. These results suggest that NRF-2 is essential for exercise-induced mitochondrial biogenesis. Next, the authors also reported that C2C12 murine myoblasts with NRF-2 knocked-down with shRNA, did not demonstrate increased TFAM and COX4 mRNA expression after H2O2 treatment, unlike control C2C12 myoblasts treated with scrambled shRNA. Others [[122]] have also demonstrated that acute exercise (treadmill running for 60 min,10% incline, 14 m·min−1 over 2 consecutive days) maintained protein concentrations of a number of endogenous antioxidants (e.g., catalase and glutathione) in heart muscle of wild-type mice, whereas the antioxidant concentrations were attenuated in heart muscle from NRF-2−/− mice. In short, disrupting NRF-2 lowers the endogenous antioxidative capacity in heart muscle during exercise and illustrates that redox and mitochondrial adaptive responses to exercise is significantly influenced by NRF-2 signaling.
Other elements of mitochondrial network dynamics, such as the regulatory proteins, involved in mitochondrial fusion are also amenable to exercise training. Aerobic exercise training (daily treadmill running at 13 m·min−1, 10% incline for 60 min, 2 weeks in duration) in 26-month-old Wistar rats upregulated mitofusin-1 (Mfn-1) [[123]]. As well, in the exercise-trained animals, differences in age-related decrease in protein content of TFAM, AMPK, phosphorylated AMPK, and PGC-1α were either minimized or restored to levels close to that of 3-month-old, exercise-trained rats. Such adaptive responses in mitochondrial network remodeling have been reported in exercise-trained human subjects. Konopka et al. [[124]] were the first to document changes with mitochondrial fusion after 12 weeks of progressive aerobic exercise training (cycle ergometry for 20–45 min at 60–80% of heart rate reserve; 3–4 sessions per week). Mfn-1 and Mfn-2 protein concentrations in biopsied vastus lateralis (obtained 48 h after the last exercise session) increased by ~ 93% and 36%, respectively, in older men (n = 6; ~ 74 years), compared with ~ 55% and 41%, respectively, in young men (n = 7; ~ 20 years). Both groups also increased VO2max as well as cross-sectional area of skeletal muscle. In another study [[125]], 12 weeks of aerobic exercise training (treadmill walking and cycle ergometry, 5 days·week−1, 85% of maximal heart rate) improved the ratio of fusion to fission proteins in vastus lateralis of previously sedentary, older obese subjects (n = 10; ~ 68 years), suggesting increased mitochondrial fusion activity.
While there is accumulating knowledge on the mechanisms involved in mitochondrial remodeling with exercise, such research studies are currently limited to measuring changes in gene and protein expression of key mitochondrial fusion regulators, such as Mfn-1 and Mfn-2, as discussed in the preceding paragraphs. The effects of exercise on mitophagy are even less well-understood because measuring mitophagy flux is complicated and is beyond the scope of this review. Interested readers are encouraged to refer to an excellent review written by Philp et al. [[126]].
Inflammaging
Immunosenescence is the deterioration of the immune system as one ages and is typically characterized by poorer vaccine response and an increase in chronic, systemic inflammation—a condition termed “inflammaging” [[127]]. A connection between age-related sarcopenia and inflammaging has been reported, wherein low-grade, systemic chronic inflammation is often present in the elderly with sarcopenia and frailty [[127, 128]]. Interleukin (IL)-6 and tumor-necrosis factor (TNF)-α are two well-defined, pro-inflammatory cytokines, and high serum concentrations were shown to predict sarcopenia in the elderly [[129]]. Although not all studies report this finding consistently, it is now appreciated that the level and chronicity of elevated pro-inflammatory cytokines may be the major determinant in the loss of muscle mass and strength [[127]]. These inflammatory cytokines contribute to sarcopenia by targeting the ubiquitin proteosome pathway, which increases muscle atrophy and via the inhibition of the mammalian target of rapamycin (mTOR)-Akt signaling pathway, which controls protein synthesis and stress response pathways [[130]]. Conversely, circulating concentrations of anti-inflammatory cytokines such as IL-10 decrease with age [[131]]. This suggests that the aging circulation may be skewed toward a pro- rather than an anti-inflammatory state.
A physically active lifestyle is effective against a wide gamut of ageing-associated chronic diseases and conditions attributable to low-grade systemic inflammation [[132, 133]]. Chronic low- and moderate-intensity physical exercise has been linked to lowered circulating inflammatory markers as well as enhanced immune function [[134]]. Physical exercise is theorized to exert beneficial effects via various mechanisms that lower pro-inflammatory activation including the involvement of toll-like receptors (TLR)s, wherein a growing evidence has propounded their association with a sedentary lifestyle, systemic inflammation and aging-related diseases [[135]]. Of these TLRs, TLR2, and TLR4 have come to the fore among exercise scientists [[136, 137]]. Their anti-inflammatory effects induced by exercise modulate phosphoinositide 3-kinase (PI3K)/protein kinase B (AKT)/mTOR signaling following the activation of adaptor proteins, causing the release of the anti-inflammatory cytokines including IL-10. Elevation in IL-10 concentrations is likely attributed to the transient rise in the supposedly pro-inflammatory cytokine IL-6 during and after exercise due in part to skeletal muscle engagement [[138, 139]], of which the latter cytokine positively correlates with exercise intensity and duration [[140]]. A systematic review conducted by Cavalcante et al. [[141]] investigated the roles of TLR2 and TLR4 in both acute and chronic conditions of aerobic, resistance, and concurrent exercises in rodents and humans; and suggested a greater propensity for aerobic exercise in inducing an inflammatory response via TLR2 and TLR4 compared with resistance exercise. Judging from alterations in TLR4 and/or TLR2 expression that correlated with inflammatory cytokines like TNF-α and IL-6, they concluded that overall, in both acute and chronic instances, aerobic exercises aggravated the inflammatory response when compared with resistance exercises [[141]]. Therein, the lowered expression of these TLRs and inflammatory cytokines generally exhibited in resistance exercise were deemed to have reduced potentially detrimental pro-inflammatory responses, and hence were more favorable over aerobic exercise.
Early-onset, lifelong aerobic exercise through voluntary wheel-running was shown to have ameliorated both inflammaging and all-cause mortality in naturally aging C57BL/J6 mice, while concomitantly improving both health and lifespan when compared with age-matched sedentary controls [[142]]. In humans, older adults who spent a considerable part of their lives maintaining a highly active lifestyle through cycling regularly, displayed an ameliorated immunosenescence [[63, 64]]. Compared with their sedentary age-matched counterparts, physically active older adults showed similarities in the frequency of naïve T-cells with recent thymic emigrants (RTE) as younger adults [[63, 64]]. These examples serve to illustrate the potential of maintaining habitual aerobic exercise from youth to later life to combat against inflammaging.
Stem cell exhaustion
Senescent stem cells (defined by increased p16INK4A expression) in human tissue at various age levels show decreased rejuvenation abilities. In addition, intrinsic characteristics of aged stem cells include changes in cell polarity and ability for self-renewal, poor capacity for differentiation, and increased propensity for apoptosis and senescence [[143]]. In addition, age-associated decreased hematopoiesis is likely consequent to increased activity of the cell cycle inhibitor, p16INK4A, increased burden of oxidative damage, and attenuated repair mechanisms in the stem cell genome etc [[144, 145]]. Changes in these intrinsic factors can impair normal functions of stem cells; human mesenchymal stem cells (MSCs) harvested from adipose tissue of older individuals (50–70 years) showed decreased ability to differentiate into osteocytes and chondrocytes in vitro, but increased propensity to undergo adipogenesis [[146]]. These cell-autonomous factors contribute to the age-related dysfunction of bone marrow-derived hematopoietic stem cells (HSCs) and MSCs, as well as tissue resident progenitor cells, such as in the heart and blood vessels (endothelial progenitors), skeletal bone and muscle (osteoblasts, satellite cells), and the nervous system [[147]].
In recent years, there has been greater appreciation for the role of extrinsic factors in stem cell aging. The microenvironment, in this case, the stem cell niche, becomes less favorable for normal stem cell growth and differentiation. Within the bone marrow, the extracellular matrix is home to numerous stromal cells and makes up the HSC niche [[148]]. Aging is associated with a remodeling of this niche; (a) MSCs have attenuated ability to form colonies, (b) deterioration of the bone marrow arterioles, (c) loss of α-smooth muscle actin-positive (α-SMA+) cell density, and iv) loss of adrenergic innervation of bone marrow [[149]]. These functional changes in the HSC niche often accompany the increase in senescent mesenchymal stromal cells, as well as the shift in differentiation of HSCs towards myeloid lineages, at the expense of lymphoid lineages [[150]]. The prevailing concept has been that the skewing of myeloid-biased HSCs could explain the decline in adaptive immunity, such as the decrease in lymphocyte numbers in older adults [[150]].
Effects of exercise training on circulating progenitor cells
While the research field of exercise and stem cell biology is in its infancy, there is substantial interest in the clinical applications of exercise to mobilize HSCs and other progenitor cells (e.g., endothelial progenitor cells—EPCs) as this adjuvant approach could serve to potentiate tissue repair and remodeling, which would be particularly beneficial for patients recovering from chronic heart failure [[151, 152]]. Van Craenenbroeck et al. studied the acute [[151]] and chronic effects of aerobic exercise [[152]] on EPCs as well as monocyte progenitor cells with pro-angiogenic capacity (circulating angiogenic cells—CACs). In the former case, patients with chronic heart failure (CHF) (n = 41; 22 mild, 19 severe) underwent a cardiopulmonary exercise test (CPX; ramp protocol on bicycle ergometer at an intensity starting with 20 or 40 W, and incrementally by 10 or 20 W·min−1 until symptom-limited effort within 8–10 min of testing). CD34+ EPC numbers and CAC migration capacity were lower in CHF patients compared with healthy controls at rest, while CPX restored EPC cell numbers in CHF patients to that of healthy controls [[151]].
In the latter study [[152]], twenty-one previously sedentary CHF patients underwent 6 months of exercise training (3 times per week; 60 min each session; intensity at 90% of heart rate corresponding to each patient's anaerobic threshold) and results were compared to age-matched, nontrained healthy controls (n = 17). Flow-mediated dilation (FMD) and CPX were performed at baseline and after 6 months of training. Baseline CAC migratory capacity was significantly impaired in sedentary CHF patients but was normalized to that of healthy controls after training. Further, circulating numbers of CD34+/KDR+ EPCs as assessed with flow cytometry increased significantly after the 6-month exercise training, which was also associated with improvement in FMD and endothelial function. These results demonstrated that exercise training enhances EPCs and CACs mobilization, with exercise-associated improvement in CAC function also a positive contributor to the repair of vascular endothelia in patients with CHF.
In healthy older individuals (n = 12) with more than 30 years of participation in moderate-to-high intensity aerobic exercise training, 10 days of detraining (complete cessation of regular exercise) induced a 44% decrease in circulating CD34+ cells, while percentage changes in CD34+/VEGFR2+ cells with detraining were positively associated with a concomitant decrease in forearm blood flow response to reactive hyperemia [[153]]. This study suggests that vascular adaptations as a consequence of exercise cessation may be related to a decrease in CD34+ cells.
Exercise-induced HSC mobilization can also benefit patients that receive autologous stem cell transplants. This procedure requires large numbers of HSCs which can be challenging to harvest from healthy donors; exercise has been suggested and demonstrated to be a useful nonpharmacological approach to enhance HSCs recovery [[154]], given that exercise-induced increase in vascular shear stress and circulating epinephrine quickly mobilize HSCs from the bone marrow into systemic blood.
Although exercise represents a promising clinical approach to enhance hematopoietic stem and progenitor cells (HSPCs) mobilization, some investigators do not observe upregulation of HSPCs with aerobic exercise training in healthy [[153]] or clinical populations [[155]]. Differences in exercise regimens, cell surface markers, and flow cytometry gating strategies may have contributed to these equivocal findings [[156]]. This means the specific progenitor cell populations responding to either a single session of exercise, or longer term training still await further clarification.
Future directions and summary
We have revisited the seven pillars of aging (Fig. 1) [[17]], and have discussed how exercise in preclinical animal models and relevant clinical diseases have been demonstrated to be efficacious in modulating each of these pillars. A relevant question to interrogate in future studies is how these pillars of aging interact as a network, and how they are modulated by exercise. The latter point highlights an obvious challenge for the study design when all seven pillars of aging are simultaneously investigated after exercise training. Further, whether the upregulation of stress response proteins to mitigate DNA damage and proteostasis is beneficial for longevity is unknown. Finally, the cellular and molecular interactions between exercise, diet, and other nutritional supplements are of immense interest. With increasing use of artificial intelligence in assessing exercise training responses, there is now the exciting prospect of personalized dosing of exercise to enhance human healthspan.
|