Late‐in‐life treadmill training rejuvenates autophagy, protein aggregate clearance, and function in mouse hearts

Abstract Protein quality control mechanisms decline during the process of cardiac aging. This enables the accumulation of protein aggregates and damaged organelles that contribute to age‐associated cardiac dysfunction. Macroautophagy is the process by which post‐mitotic cells such as cardiomyocytes clear defective proteins and organelles. We hypothesized that late‐in‐life exercise training improves autophagy, protein aggregate clearance, and function that is otherwise dysregulated in hearts from old vs. adult mice. As expected, 24‐month‐old male C57BL/6J mice (old) exhibited repressed autophagosome formation and protein aggregate accumulation in the heart, systolic and diastolic dysfunction, and reduced exercise capacity vs. 8‐month‐old (adult) mice (all p < 0.05). To investigate the influence of late‐in‐life exercise training, additional cohorts of 21‐month‐old mice did (old‐ETR) or did not (old‐SED) complete a 3‐month progressive resistance treadmill running program. Body composition, exercise capacity, and soleus muscle citrate synthase activity improved in old‐ETR vs. old‐SED mice at 24 months (all p < 0.05). Importantly, protein expression of autophagy markers indicate trafficking of the autophagosome to the lysosome increased, protein aggregate clearance improved, and overall function was enhanced (all p < 0.05) in hearts from old‐ETR vs. old‐SED mice. These data provide the first evidence that a physiological intervention initiated late‐in‐life improves autophagic flux, protein aggregate clearance, and contractile performance in mouse hearts.


| INTRODUC TI ON
The incidence of cardiovascular disease (CVD) is ~20%, ~50%, ~80%, and ~90% in individuals 18-44, 45-64, 65-79, and 80+ years of age, respectively (Benjamin et al., 2019). Treating cardiovascular complications associated with the aging demographic creates an enormous economic challenge to the healthcare system in particular, and society in general. New therapeutic targets need to be identified so that practical intervention strategies can be designed, optimized, and implemented. We sought to determine whether cardiac autophagy can be influenced positively by a well-accepted lifestyle intervention strategy (i.e., regular physical activity) in a pre-clinical model of primary aging.
Protein aggregates accumulate and organelles become damaged and/or dysfunctional during the process of aging. A progressive loss of protein quality control and autophagy contributes importantly in many organs to this age-associated proteotoxicity and the subsequent decline in cellular function (Cuervo & Dice, 2000;Koga et al., 2011;Rubinsztein et al., 2011). Post-mitotic cells with limited proliferative capacity such as cardiac myocytes are particularly reliant upon autophagy to maintain proteostasis and thereby preserve cardiac function during aging (Rubinsztein et al., 2011;Terman & Brunk, 2005). In support of this, age-related cardiomyopathy is recapitulated in adult mice by cardiac-specific Atg5 deletion (Taneike et al., 2010) and mTORC1 activation Taneike et al., 2016), whereas desmin-related cardiomyopathy, characterized by the accumulation of cytotoxic misfolded proteins, is prevented by cardiacselective Atg7 overexpression (Bhuiyan et al., 2013).
Most literature indicates that primary aging precipitates myocardial dysfunction in C57BL/6J mice Dai et al., , 2012, but comparisons of cardiac autophagy between older and adult mice have not yielded consistent findings.
Inconsistencies likely arise from conclusions being based solely upon steady-state measures of autophagy including the quantification of MAP1LC3/LC3 (microtubule-associated protein 1 light chain 3) and SQSTM1/p62 (Klionsky et al., 2021). However, because autophagy is a highly dynamic process, it is best practice to pharmacologically block the turnover of these proteins to most accurately quantify the scale of autophagosome formation, that is, autophagic flux. This methodological approach has been implemented once in aged mice (Wu et al., 2016) and once in cardiomyocytes isolated from older rats . Both studies provided support for an age-associated repression of cardiac autophagic flux. Here, we substantiated these observations using chloroquine, and additionally showed accrual of ubiquitinated proteins and protein aggregates in the myocardium, oxidative stress, impaired mitochondrial quality, cardiac dysfunction, and reduced exercise capacity, in 24-month-vs.
8-month-old animals. These findings allowed us to test our primary hypothesis that late-in-life exercise training re-establishes autophagic flux to an extent that improves protein clearance, redox status, and cardiac function.
A growing area of research inquiry is whether upregulating the process of autophagy has therapeutic benefit. For example, late-in-life interventions that increase autophagy such as supplementation with the natural polyamine spermidine (Eisenberg et al., 2016), caloric restriction (Sheng et al., 2017), and mTORC1 inhibition using rapamycin (Flynn et al., 2013) lessen age-associated cardiac dysfunction in C57BL/6J mice. An alternative or complementary approach with potential to improve cardiac autophagy and attenuate the aging-associated decline in cardiac function is dynamic exercise.
In this regard, He et al. observed that an acute bout of treadmill running elevates protein indexes of autophagy in murine hearts (He et al., 2012), and Bhuiyan et al. reported that autophagy, protein clearance, and function improved in hearts from mice with desminrelated cardiomyopathy that did vs. did not have long-term access (i.e., 6-month) to wheel running (Bhuiyan et al., 2013). The potential for a physiological maneuver, that is, late-in-life exercise training, to rejuvenate cardiac autophagy has not been investigated. Here, we present what we believe to be the first report that exercise capacity, autophagic flux, protein aggregate clearance, redox balance, mitochondrial quality, and cardiac function improve in hearts from older mice that complete a progressive, resistance treadmill running program vs. animals that do not train.

| Hearts from older mice display repressed autophagic flux, accumulation of ubiquitinated proteins, and oxidative stress
The impact of aging on cardiac autophagy in pre-clinical murine models is not uniform. Further, few studies have assessed the influence of aging on different steps involved in the process of myocardial autophagy Wu et al., 2016). We sought clarity concerning the influence of aging on steady-state autophagy and autophagic flux in the murine heart. Because lean mass is likely to change in an age-associated and/or exercise-training-related manner, adult and older male C57BL/6J mice completed TD-NMR analyses to determine body composition. This allowed chloroquine (CQ) to be administered at a dose based on lean body mass (Figure 1a) . Twenty-four h after TD-NMR, CQ (75 mg IP/g lean body mass) or vehicle-control (phosphate-buffered saline; VEH) was administered to both groups. After 4 h, hearts were collected from isoflurane-anesthetized mice (Gottlieb et al., 2015;Klionsky et al., 2021;Pires et al., 2017). This CQ regimen is referred to in Figure 1a as "4 h." We hypothesized that autophagic flux would be impaired in hearts from older vs. adult mice, and that this would associate with accrual of ubiquitinated proteins and heightened oxidative stress. was higher (p < 0.05), and Atg3:GAPDH was lower (Figure S1c, d; p < 0.05), in hearts from old-VEH vs. adult-VEH mice (histogram 3 vs. 1), whereas LC3-II:LC3-I, Atg5:GAPDH, and Atg7:GAPDH ( Figure   S1b, e, f) were similar between groups. Regarding mRNA expression, Atg3 was lower and p62 was higher ( Figure S2, both p < 0.05) in hearts from old vs. adult mice, whereas LC3B mRNA expression was similar between groups ( Figure S2). These findings were concurrent with accrual of ubiquitinated proteins (Figure 1e, f) and elevated 4-hydroxy-2-nonenal (4-HNE; Figure 1g, h; both p < 0.05). mRNA expression of superoxide dismutase (SOD) 2 trended higher (p = 0.06) and catalase was lower (p < 0.05) in hearts from old vs. adult mice, whereas SOD1 was similar between groups ( Figure S2).
Collectively, these findings strongly suggest that steady-state autophagy is attenuated in hearts from old vs. adult mice.
It is not conclusive whether aging impairs autophagosome formation, trafficking of the autophagosome to the lysosome, and/or lysosomal degradation of the autophagosome in the heart. We ad-

| Hearts from older mice display impaired systolic and diastolic function
While some discrepancies exist, the balance of available literature indicates that primary aging precipitates cardiac dysfunction in C57BL/6J mice Dai et al., , 2012.
We hypothesized that aging-associated cardiac dysfunction ex-  Figure S3a, c, d; all p < 0.05) indicating LV hypertrophy and systolic dysfunction exist in hearts from aged animals. Further, older F I G U R E 1 Hearts from old mice display impaired autophagic flux, accumulation of ubiquitinated proteins, and oxidative stress. (a) Body composition was assessed using TD-NMR in adult (A; 8-mo) and older (O; 24-mo) mice. Vehicle (VEH; phosphate-buffered saline, PBS) or chloroquine (CQ; 75 mg/kg lean muscle mass) was administered to A and O mice and tissues were obtained 4 h later. Representative images (b, e, g) and mean data ± standard error are shown for LC3-II (c), p62 (d), poly-ubiquitin (total ubiquitin; f), and 4-hydroxy-2-nonenal (4-HNE; h). LC3-II and p62 were greater in hearts from O vs. A mice treated with VEH. In A mice, LC3-II and p62 increased further in CQ vs. VEHtreated cohorts, whereas these endpoints were similar in O mice treated with CQ. These findings demonstrate that autophagic flux is robust in A but not O mice. Total ubiquitin and 4-HNE were elevated in hearts from O vs. A mice, but CQ did not alter responses in either group. For panels (c, d), n = 9-24, *p < 0.05 vs. A-VEH mice. For Although type 1 collagen staining indicated increased fibrosis in hearts from older vs. adult mice, cardiomyocyte area was not affected ( Figure S4a-c). Of the profibrotic genes that were assessed, mRNA expression of fibrillin (Fbn) 1 and transforming growth factorβ (Tgfb) 2 were elevated (p < 0.05), and a trend existed for increased connective tissue growth factor (Ctgf; p = 0.07) in hearts from older vs. adult mice, whereas Fbn2 and Tgfb1 were not different between groups ( Figure S4d). adult mice were anticipated. Substantiating these observations was necessary to test the hypothesis that exercise training lessens the age-associated disruptions. Adult mice did (adult-ETR) or did not (adult-SED) complete progressive resistance treadmill training from 5 to 8 months of age. Likewise, older mice did (old-ETR) or did not (old-SED) train from 21 to 24 months of age ( Figure 3a). As expected, fat mass and body mass were greater ( Figure S5a Figure S7) of Atg3 was higher and p62 was lower (both p < 0.05) in hearts from old-ETR vs. old-SED mice, whereas LC3B was similar between groups. Collectively, these findings indicate that late-in-life exercise training improves steady-state autophagy in mouse hearts.

| Exercise training improves clearance of ubiquitinated proteins and oxidative stress in hearts from older mice
Bolstering our findings shown in Figure 1e mRNA expression of SOD1, SOD2, and catalase were elevated by late-in-life exercise training ( Figure S7).

| Exercise training improves autophagic flux in hearts from older mice
Here, we tested whether exercise training improves the ageassociated repression of autophagic flux observed in Figure 1.
Because exercise training did not influence steady-state autophagy F I G U R E 3 Late-in-life exercise training improves steady-state autophagy, clearance of ubiquitinated proteins, and oxidative stress in mouse hearts. (a) Adult (A, 5 mo) and older (O, 21 mo) mice did (ETR) or did not (SED) complete 12 weeks of treadmill running. At least 24 h following the last exercise bout, A (8 mo) and O (24 mo) hearts were excised and prepared for immunoblotting. Representative images (b, e, g) and mean data ± standard error are shown for LC3-II (c), p62 (d), total ubiquitin (f), and 4-HNE (h). LC3-II (c) and p62 (d) were greater in hearts from O-SED vs. A-SED mice. While no differences existed between A-SED and A-ETR mice, LC3-II increased (c) and p62 decreased

F I G U R E 4
Late-in-life exercise training improves steady-state autophagy and autophagic flux in mouse hearts. (a) Adult (A, 5 mo) and older (O, 21 mo) mice did (ETR) or did not (SED) complete 12 wk of treadmill running. Mice were treated with VEH or CQ 48, 24, and 4 h prior to tissue collection. Representative images (b, e) and mean data ± standard error are shown for LC3-II (c, f) and p62 (d, g). In A hearts treated with VEH (b-d), ETR did not influence steady-state LC3-II or p62. Robust CQ-evoked increases in LC3-II and p62 observed in A hearts were similar regardless of exercise training (b-d). In O hearts treated with VEH, (e-g), LC3-II trended upwards (p = 0.08) and p62 decreased in response to ETR. These data indicate that late-in-life exercise training improves steady-state autophagy. While CQ did not influence LC3-II or p62 in hearts from O-SED mice (e-g), significant accumulation of LC3-II (f) and p62 (g)  adult-SED animals ( Figure S12, S13). Collagen type 1 area ( Figure   S14a, b), cardiomyocyte area ( Figure S14a, c), and mRNA expression of profibrotic genes ( Figure S14d) were similar between old-SED and old-ETR mice.

| Exercise training improves protein aggregate removal and mitochondrial quality in hearts from older mice
Electron microscopy images indicate aging-associated cardiac protein aggregation is normalized by exercise training (Figure 6a

| DISCUSS ION
Our primary aim was to test the hypothesis that late-in-life exercise training rejuvenates indexes of cardiac autophagy, improves clearance of ubiquitinated proteins, and re-establishes cardiac function.
First, we substantiated earlier findings that repressed autophagic flux in the heart of older vs. adult mice exists, and that this is associated with protein aggregate accrual, oxidative stress, and cardiac dysfunction. Next, we demonstrated for the first time that a physiological intervention, that is, progressive resistance treadmill running, improves autophagic flux, protein clearance, redox balance, mitochondrial quality, and cardiac function, in hearts from older mice. These data indicate positive crosstalk exists between regular physical activity and cardiac autophagy in the context of primary aging.

| Autophagic flux is depressed in hearts from older mice
Repressed autophagy is observed in a wide variety of conditions associated with aging, including neurodegenerative diseases, normal brain aging, osteoarthritis, insulin resistance, atherosclerosis, macular degeneration, suppressed hepatic proteolysis, heart failure, and endothelial cell dysfunction (Bharath et al., 2017;Campos et al., 2017;Park et al., 2019;Pires et al., 2017;Rubinsztein et al., 2011). A close examination of the literature reveals that the impact of aging on cardiac autophagy in pre-clinical murine models is not congruent. Most studies investigating this issue have assessed steady-state autophagy by quantifying protein expression of LC3-II and/or p62. The membrane-bound lipidated form of cytosolic LC3-I, that is, LC3-II accumulates as the phagophore membrane is formed and extended during the process of autophagy. Atg3 is the phosphatidylethanolamine-transferase that performs the final lipid conjugation modification of LC3 required for completing the conversion of cytosolic LC3-I to membrane-bound LC3-II (Mizushima, 2007). The adaptor protein p62, which tethers targeted cargo destined to become engulfed in the autophagosome, is degraded as autophagy proceeds.
A variety of studies indicate p62 accumulates in hearts from aged vs. adult mice (Li et al., 2020;Liang et al., 2020;Ren et al., 2017;Wang et al., 2018;Wu et al., 2016;Zhang et al., 2017), and translational relevance of these findings to older humans was recently reported (Li et al., 2020). Results concerning LC3-II are less clear. With regard to C57BL/6J mice: (i) LC3II:GAPDH (Liang et al., 2020;Taneike et al., 2010) and LC3-II:LC3-I  declined in hearts from ~26 months vs. ~4 months animals; (ii) LC3-II/ LC3-I increased in 18 months vs. 2 months mice; (Boyle et al., 2011) and (iii) LC3-II/LC3-I was not different between ~23 months and ~4 months animals (Li et al., 2020;Wu et al., 2016). We observed increased LC3-I:GAPDH, LC3-II:GAPDH, and p62:GAPDH in older vs. adult mice from two independent cohorts treated identically (Figures 1, 3). Because Atg3 mRNA and protein expression was lower in older vs. adult mice ( Figures S1, S2, S6), elevated LC3-I:GAPDH observed in older animals might result from an inability to perform the lipid conjugation step whereby cytosolic LC3-I is converted to membrane-bound LC3-II. With regard to LC3-II and p62 accrual observed in hearts from older vs. adult mice, this might be secondary to a defect that exists later in the process of autophagy and we tested this. Separate cohorts of adult and old mice were treated with the autophagosome-lysosome fusion inhibitor CQ to assess autophagic flux (Gottlieb et al., 2015;Klionsky et al., 2021;Pires et al., 2017). This approach has been used in the context of cardiac aging on two occasions. Wu et al. treated C57BL/6J mice with the vacuolar H + -ATPase inhibitor bafilomycin (0.3 mg/ kg IP × 7 days), which impairs lysosomal acidification, blocks autophagosome-lysosome fusion, and thereby prevents degradation of autophagolysosomes. Compared to mice that were administered a vehicle-control, bafilomycin increased LC3-II: LC3-I and p62 protein expression in cardiac lysates from 4 but not 22-mo-old mice (Wu et al., 2016). Using a different species and experimental setting, Ma et al. reported that cardiomyocytes isolated from hearts of 4 months rats displayed greater LC3 puncta and p62 expression after treatment with bafilomycin (100 nM × ~4 h) compared to results obtained from ~24 months rats .
Both studies concluded that constitutive autophagosome formation is robust in hearts from adult but not older mice and our results after 4 h ( Figure 1) and 48 h (Figure 4) CQ administration are supportive. Specifically, CQ increased LC3-II:GAPDH and p62:GAPDH in hearts from adult but not older mice (Figures 1, 4). These data substantiate that autophagosome clearance capacity is compromised in aged mouse hearts.

| Repressed cardiac autophagic flux associates with proteotoxicity, oxidative stress, impaired mitochondrial quality, and contractile dysfunction
Strong rationale exists that repressed autophagosome formation contributes importantly to accelerated cardiac aging. In a loss of autophagy approach, adult mice with cardiac-selective Atg5 deletion (Taneike et al., 2010), and cardiac-specific mTOR activation via miR-199a overexpression  or tuberous sclerosis complex 1 and 2 depletion (Taneike et al., 2016), exhibit important characteristics of cardiac aging, that is, protein aggregate accrual, interstitial fibrosis, LV hypertrophy, oxidative stress, and/or cardiac dysfunction.
In addition to compromised autophagic flux, old vs. adult mice in the present study displayed each of these features of cardiac aging.
Highlighting an association between repressed cardiac autophagy and cardiac dysfunction, elevated cardiac p62 protein expression correlated significantly with a well-accepted estimate of overall LV dysfunction, that is, the MPI (Figure 2m) (Goroshi & Chand, 2016).
Using a gain of autophagy procedure, mice with cardiac-selective Atg7 overexpression (Atg7 transgenic mice) were crossed with CryAB R120G mice, a model of desmin-related cardiomyopathy that exhibits impaired autophagic flux together with the accumulation of preamyloid oligomer (PAO), a toxic component in many of the protein misfolding based neurodegenerative diseases (Maloyan et al., 2007;Pattison et al., 2011). As anticipated, autophagic flux was greater, and accrual of cytotoxic proteins, impaired cardiac performance, and early mortality was less severe, in CryAB R120G × Atg7 transgenic mice vs. CryAB R120G animals (Bhuiyan et al., 2013). Based on previous results using loss of autophagy and gain of autophagy approaches involving the heart, it is not unreasonable to suggest that impaired autophagic flux (Figures 1, 4) contributed importantly to the accrual of ubiquitinated proteins and elevated lipid peroxidation (Figures 1, 3), protein aggregate accumulation (Figure 6 . Mitochondrial number did not differ among groups (d). The ageassociated reduction in Pink1 mRNA was restored by late-in-life exercise training (e), whereas training improved Park2 mRNA in older mice (f). The age-associated reduction in UQCRC2 (j, complex III) and MTCO1 (k, complex IV) was restored by exercise training. Aging did not impact NDUFB8 (h; complex I), SDHB (i, complex II), or ATP5A (l, complex V). For (b), n = 4 mice per group, n = 8-16 fields of view, data are expressed as area of protein aggregation (µm 2 ); (c), n = 6-11, data are expressed as % of protein aggregation; (d), n = 9-16, data expressed as number of mitochondria per area (µm 2 ); (e, f), n = 6-8 mice; (g-l), n = 6 mice; (b-k), *p < 0.05 vs. A-SED; #p < 0.05 vs. O-SED in HEK 293 cells (Frudd et al., 2018) and Atg3 protein expression in mouse brain endothelial cells (Kamat et al., 2015). Because autophagy is an important driver of mitochondrial clearance, stalled mitophagy in aged hearts could precipitate reactive oxygen species generation from dysregulated mitochondria, and our results concerning repressed Pink1, Complex III, and Complex IV in hearts from older vs. adult mice support this notion (Figure 6).

| Late-in-life interventions that activate autophagy associate with attenuated cardiac dysfunction
Genetic manipulations (e.g., Atg7 overexpression) cannot be used clinically to upregulate autophagy in conditions associated with cardiac proteotoxicity at present. However, benefits from inducing this protein degradation pathway late-in-life via nutraceutical (e.g., spermidine), lifestyle (e.g., caloric restriction), and pharmacological (e.g., rapamycin) maneuvers have been demonstrated (Eisenberg et al., 2016;Flynn et al., 2013;Sheng et al., 2017). While each of these autophagy-boosting approaches attenuated age-associated cardiac dysfunction (Eisenberg et al., 2016;Flynn et al., 2013;Sheng et al., 2017), it is unknown if functional benefits associated positively with improved autophagic flux and protein clearance in the heart because neither of these endpoints was assessed.
A lifestyle intervention with potential to improve autophagy, clear damaged proteins, and beneficially influence the agingassociated decline in cardiac function is dynamic exercise. He et al.
first showed in mice that 30-80-min treadmill running increases LC3-GFP puncta, LC3-II:LC3-I, and p62 degradation in the heart (He et al., 2012). Beta cell lymphoma/leukemia 2 (Bcl-2) is an antiapoptotic and anti-autophagy protein that inhibits autophagy by directly interacting with beclin 1 at the endoplasmic reticulum. The authors reported that the Bcl-2-beclin-1 complex dissociates in response to treadmill running, and this finding was confirmed by Bhuiyan et al. in mice that completed an acute bout of voluntary wheel running (VWR) (Bhuiyan et al., 2013). Because long-term VWR decreased the amyloid load in mice with neurodegenerative disorders, for example, Alzheimer's disease (Lazarov et al., 2005) the Robbins laboratory group sought to determine whether this form of "environmental enrichment" initiates cardiac autophagy to an extent that improves cardiac proteostasis. Providing strong proof of concept, the authors reported a 47% reduction in cardiac PAO accumulation in CryAB R120G mice that completed 6 months of VWR vs. CryAB R120G animals that did not train, but indexes of autophagy were not assessed (Maloyan et al., 2007). The same investigative team later demonstrated that VWR increased mRNA expression of Atg4, Atg5, and Wipi1 in hearts from CryAB R120G mice vs. untrained mice, but neither autophagic flux nor protein aggregate accrual were assessed (Bhuiyan et al., 2013). In the latter study, it is extremely interesting to note that functional endpoints assessed via echocardiography (LVIDs, LVIDd, and EF) appear identical between CryAB R120G × Atg7 transgenic mice and CryAB R120G mice that completed VWR, suggesting that exercise training conferred benefits similar to genetic autophagy activation and vice versa.

| Late-in-life exercise training improves cardiac autophagic flux that associates positively with cardiac function
The interesting findings from interventions involving nutraceuticals, pharmaceuticals, and lifestyle alterations in older mice inspired us to test whether late-in-life exercise training induces cardiac autophagy to an extent that improves proteostasis and lessens cardiac dysfunction. As anticipated, the intensity, frequency, and duration of "forced" exercise training produced functional and biochemical evidence of efficacy ( Figure S5). In support of our hypothesis, age-associated dysregulation of LC3-II:GAPDH and p62:GAPDH improved in hearts from old-ETR vs. old-SED mice (Figures 3, 4).  (Frudd et al., 2018;Kamat et al., 2015). In our investigation, elevated Pink1 and Park2 mRNA in hearts from old-ETR vs.
old-SED mice represent a greater ability to clear ROS generating mitochondria to an extent that normalizes mitochondrial quality ( Figure 6). We realize that associations are presented and cause and effect relationships remain to be identified.
Three months of treadmill training did not prevent agingassociated cardiac fibrosis ( Figure S11a We observed a strong association between the age-related accrual of p62:GAPDH (repressed autophagy) and the increase (i.e., worsening) of MPI ( Figure 2m); and the training-induced reduction in p62:GAPDH (improved autophagy) and the decrease (i.e., improvement) of MPI (Figure 5l). While these findings indicate training-induced elevations in autophagic flux associate positively with preserved cardiac function in the context of primary aging, evidence exists that an 8-week exercise program preserves autophagic flux in adult rats in the setting of a common age-related pathology, for example, heart failure (Campos et al., 2017). Specifically, 12 weeks following myocardial infarction-induced heart failure via left anterior descending coronary artery ligation, indexes of cardiac autophagic flux and cardiac function were improved in rats that completed treadmill running from weeks 4-12 vs. those that did not train. At present, it is unknown whether late-in-life exercise training rejuvenates autophagic flux to an extent that improves tolerance to infarction-induced heart failure, but these studies are ongoing in our laboratory. Two substrains of C57BL/6 mice were used in the present study, that is, C57BL/6J for adult mice (Jackson Laboratories) and C57BL/6N for older mice (NIA rodent colony). While both substrains share the same core genetic background, nuances do exist which could influence the phenotypes we describe, and this limitation should be considered when integrating our findings into what is known currently.

| Animals and housing
Male C57BL/6J mice were obtained from the Jackson Laboratories at 4 months of age, and C57BL/6N animals were obtained from the National Institute on Aging rodent colony at 18 months of age. All mice were housed 4-per cage under controlled temperature (22°C) and light (12:12-h light-dark cycle) conditions, and were provided with food (Diet #2920, Teklad Diets, Madison, WI) and water ad libitum. Animals were handled according to Institutional approved procedures documented in protocol number 19-07010 JDS.

| Cardiac function
Separate cohorts of adult and older mice were anesthetized lightly with 1%-3% inhaled isoflurane combined with 100% oxygen, and transthoracic echocardiography was completed to assess indexes of systolic and diastolic function Symons et al., 2011).
Mice were recovered from anesthesia following the echocardiography measurements. These procedures were initiated at the same time of day.

| Exercise training
Body composition was assessed using TD-NMR in sym5-month-and 21-month-old mice. Twenty-four to 48 h later, mice were familiarized with walking/running on a motorized treadmill (Columbus Instruments). On day 4, a workload capacity evaluation test was completed on each mouse. Total workload was calculated as [body weight (kg) × total running time (min) × final running speed (m/ min) × treadmill grade (25%)] (Symons et al., 2000). After all mice finished the workload capacity evaluation, they were separated randomly into groups that did not (adult-SED and old-SED) or did (adult-ETR and old-ETR) complete a 3-month progressive resistance treadmill running program. Exercise-training bouts were completed between 0600 and 1000 each day. After 3 months, body composition, exercise tolerance (maximal workload capacity), and cardiac function were assessed. Each evaluation was separated by 24 h and completed at the same time of day. Twenty-four hours after measuring cardiac function, all mice were anesthetized as described, and the heart was excised to assess protein indexes of autophagy, histology, morphology, mRNA expression of profibrotic (described earlier) and mitophagy-related genes, protein aggregate accumulation, and protein expression of Complex I-V of the electron transport chain.
Soleus muscle was dissected free from both hindlimbs to assess CS enzyme activity (Sigma-Aldrich) (Symons et al., 2000).

| Mitochondrial number and quality
Mitochondrial number was assessed using electron microscopy, mitophagy-related gene expression was assessed using qPCR, and Complex I-V protein expression of the electron transport chain was assessed by immunoblotting. Details are provided in online "Experimental procedures."

| Cardiac protein aggregation
After determining cardiac protein concentrations (Pierce BCA Protein Assay; ThermoFisher), protein aggregate accrual in the heart was measured using a commercially available kit (Proteostat; Enzo Life Sciences) (Laor et al., 2019). Electron microscopy was used as an additional approach to estimate cardiac protein aggregates (DiMemmo et al., 2017). Details are provided in online "Experimental procedures."

| Statistical analyses
Data are presented as mean ± standard of error of the mean.
Significance was accepted when p < 0.05. To determine normality of the distribution for each data set, GraphPad Prism software was used.
An unpaired t test (e.g., EF in adult vs. old mice) was used, as appropriate, to compare two mean values. Comparison among four means was completed using a one-way ANOVA (e.g., cardiac p62:GAPDH among old-SED-VEH, old-SED-CQ , old-ETR-VEH, and old-ETR-CQ ).
In cases when a significant main effect was obtained, a Tukey post hoc test was used to determine the location of the differences.

CO N FLI C T O F I NTE R E S T
None of the authors has any conflicts of interest to disclose.

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.