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

There is evidence for a progressive decline of protein quality control mechanisms during the process of cardiac aging. This enables the accumulation of protein aggregates and damaged organelles that contribute to age-associated cardiac dysfunction. Macroautophagy (referred to as autophagy) 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 : (i) repressed autophagosome formation and protein aggregate accumulation in the heart; (ii) systolic and diastolic dysfunction; and (iii) reduced exercise capacity, vs. 8-month old (adult) mice (all p< .05). Separate cohorts of 21 month old mice completed a 3-month progressive resistance treadmill-running program (old-ETR) that improved (all < .05) : (i) body composition; (ii) exercise capacity; and (iii) soleus muscle citrate synthase activity, vs. age-matched mice that did not train (old-SED). Importantly, (iv) protein expression of autophagy markers indicated trafficking of the autophagosome to the lysosome increased, (v) protein aggregate clearance improved, and (vi) overall function was enhanced (all p<0.05), in hearts from old-ETR vs. old- SED mice. Dietary maneuvers and pharmacological interventions shown to elevate basal autophagy are reported to mitigate / reverse age-associated cardiac dysfunction. Here we show the first evidence that a physiological intervention initiated late-in-life improves autophagic flux, protein aggregate clearance, and overall function in mouse hearts.


INTRODUCTION
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 myocardial autophagy can be influenced positively by a well-accepted lifestyle intervention strategy (i.e., regular physical activity) in a preclinical model of primary aging.
Protein aggregates accumulate and organelles become damaged and / or dysfunctional during the process of aging. A progressive loss of the cellular quality control mechanism autophagy contributes importantly in many organs to this age-associated proteotoxicity and the subsequent decline in cellular function (Cuervo & Dice, 2000;Donati, Recchia, Cavallini, & Bergamini, 2008;Koga, Kaushik, & Cuervo, 2011;Vittorini et al., 1999). Post-mitotic cells with limited proliferative capacity such as cardiac myocytes are particularly reliant upon autophagy to maintain proteostasis and thereby preserve myocardial function during aging Terman & Brunk, 2005). In support of this, agerelated 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 cardiac-selective Atg7 overexpression (Bhuiyan et al., 2013).
Most literature indicates that primary aging precipitates myocardial dysfunction in C57BL/6J mice (Dai & Rabinovitch, 2009;Dai, Rabinovitch, & Ungvari, 2012;Dai et al., 2009), but comparisons of cardiac autophagy between older and adult mice have not yielded consistent findings (Boyle et al., 2011;Li et al., 2017;Taneike et al., 2010;Wang et al., 2018;Wu et al., 2016;. Inconsistencies likely arise from conclusions being based solely upon steady-state measures of autophagy markers including MAP1LC3/LC3 (microtubule associated protein 1 light chain 3) and SQSTM1/p62 (Klionsky et al., 2016;. 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 i.e., autophagic flux. A review of the literature reveals this methodological approach has been implemented once in aged mice (Wu et al., 2016), and once in cardiomyocytes isolated from older rats (L. . Both studies provided support for an age-associated repression of cardiac autophagic flux. Here we substantiated these observations and additionally showed accrual of ubiquitinated proteins and protein aggregates in the myocardium, cardiac dysfunction, and reduced exercise capacity, in 24mo vs. 8-mo old animals. These findings allowed us to test our primary hypothesis.

Myocardium from older vs. adult mice displays repressed autophagic flux.
The impact of aging on cardiac autophagy in pre-clinical murine models is not uniform (Boyle et al., 2011;Li et al., 2017;Taneike et al., 2010;Wang et al., 2018;Wu et al., 2016;. Few studies have assessed the influence of aging on different steps involved in the process of myocardial autophagy (L. Wu et al., 2016). To address both issues, adult and older male C57BL/6J mice completed TD-NMR analyses to determine body composition. This was required so that the lysosomal acidification inhibitor chloroquine (CQ) could be administered using a dose based on lean body mass (Glick, Barth, & Macleod, 2010;Gottlieb, Andres, Sin, & Taylor, 2015;Pires et al., 2017). A schematic of our procedures is shown in Figure 1a. Twenty-four h after TD-NMR, CQ (75 mg IP / g lean body mass) or vehicle-control (phosphate-buffered saline; VEH) (Gottlieb et al., 2015) was administered to both groups. After 4 h hearts were collected from isoflurane anesthetized mice. Based on available literature, we hypothesized that autophagic flux would be quantifying protein expression of LC3-II and / or p62. The membrane-bound lipidated form of cytosolic LC3-I i.e., 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 . 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;Wang et al., 2018;Wu et al., 2016;, 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 (Taneike et al., 2010) and LC3II / LC3I  declined in hearts from ~ 26 mo vs. ~ 4 mo animals; (ii) LC3II / LC3I increased in 18 mo vs. 2 mo mice; (Boyle et al., 2011) and (iii) LC3II / LC3I was not different between ~ 23 mo and ~ 4 mo animals (Li et al., 2020;Wu et al., 2016). We observed increased LC3I:GAPDH, LC3II:GAPDH, and p62:GAPDH in older vs. adult mice from two independent cohorts treated identically (Figure 1, Figure 3). Because Atg3 mRNA and protein expression was lower in both groups of older vs. adult mice ( Figure S2, S6), elevated LC3I:GAPDH observed in older animals could have resulted from an inability to perform the lipid conjugation step whereby cytosolic LC3I is converted to membrane-bound LC3II. With regard to LC3II 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 lysosomal acidification inhibitor CQ to assess autophagic flux (Glick et al., 2010;Gottlieb et al., 2015;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 x 7 days) which impairs lysosomal acidification, blocks autophagosome-lysosome fusion, and thereby prevents degradation of autophagolysosome (Kawai, Uchiyama, Takano, Nakamura, & Ohkuma, 2007;. Compared to mice that were administered a vehicle-control, bafilomycin increased LC3II : LC3I and p62 protein expression in myocardial 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 mo rats displayed greater LC3 puncta and p62 expression after treatment with bafilomycin (100 nM x ~ 4h) compared to results obtained from ~24 mo rats (L. . Both studies concluded that constitutive autophagosome formation is robust in myocardium from adult but not older mice and our results are supportive. Specifically, CQ-treatment increased LC3I:GAPDH, LC3II:GAPDH, and p62:GAPDH in hearts from adult but not old mice, underscoring the statement that autophagosome clearance capacity is compromised in aged myocardium ( Figure   1).
Repressed myocardial autophagic flux associates with cardiac proteotoxicity, oxidant stress, 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 i.e., protein aggregate accrual, interstitial fibrosis, LV hypertrophy, oxidant stress, and / or cardiac dysfunction. In addition to compromised autophagic flux, old vs. adult mice in the present study displayed each one of these features of cardiac aging. Highlighting an association between repressed cardiac autophagy and myocardial dysfunction, elevated cardiac p62 protein expression correlated significantly with a well-accepted estimate of overall LV dysfunction i.e., the MPI (Goroshi & Chand, 2016;Tei et al., 1995) (Figure 2l). 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 . As would be predicted, autophagic flux was greater, and accrual of cytotoxic proteins, impaired cardiac performance, and early mortality was less severe, in CryAB R120G x 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 myocardium, it is not unreasonable to suggest by older vs. adult hearts in our study. While precise mechanisms responsible for the ageassociated reduction in Atg3 in cardiomyocytes have not been reported, evidence exists that oxidative stress inhibits Atg3 enzyme function in HEK 293 cells (Frudd, Burgoyne, & Burgoyne, 2018) and Atg3 protein expression in mouse brain endothelial cells (Kamat, Kalani, Tyagi, & Tyagi, 2015).

Late-in-life interventions attenuate cardiac dysfunction by activating autophagy.
Although genetic manipulations (e.g., Atg7 overexpression) cannot be used clinically to upregulate autophagy in conditions associated with cardiac proteotoxicity at present, 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). For example, cardiac hypertrophy was attenuated and diastolic function was preserved in C57BL/6J mice that consumed spermidine-supplemented vs. vehicle-treated water from 18-24 months of age. The autophagyboosting effect of spermidine shown originally in flies and yeast (Eisenberg et al., 2009;Morselli et al., 2011) was confirmed in the myocardium, and cardiac-selective Atg5 deficient animals were refractory to this intervention, substantiating that autophagy is necessary for the benefits of this polyamine to be observed (Eisenberg et al., 2016). Sheng et al. fed one group of C57BL/6J mice standard chow ad libitum, whereas another cohort consumed 40% fewer calories from the same diet, from 19-22 months of age (Sheng et al., 2017). AMPK is phosphorylated and activated by nutrient insufficiency to an extent that increases expression of autophagy-associated genes (Kubli & Gustafsson, 2014; and this was confirmed in myocardium from calorie-restricted vs. ad libitum fed mice. Age-associated cardiac fibrosis, LV hypertrophy, and compromised ejection fraction were less severe in mice that ingested calorie-restricted vs. standard diet from 19-22 months of age i.e., late-in-life. While inhibiting TOR (target of rapamycin) signaling extends lifespan in organisms from worms to mice Kapahi et al., 2004;Vellai et al., 2003), Flynn et al. first reported the cardiovascular benefits of this intervention in the context of aging (Flynn et al., 2013). Mice treated with rapamycin from 24-27 months of age had repressed pro-inflammatory signaling in the myocardium, less LV hypertrophy, and preserved systolic function vs. age-matched controls. Although rapamycin repressed mTORC1 signaling in the myocardium as demonstrated by reduced pS6K : total S6, evidence for upregulated cardiac autophagy per se was not presented. While each of these approaches to boost autophagy in the context of aging attenuated indexes of cardiac dysfunction (Eisenberg et al., 2016;Flynn et al., 2013;Sheng et al., 2017), it is unknown if this benefit associated positively with improved autophagic flux and protein clearance in the myocardium because neither of these endpoints was assessed.
A lifestyle intervention with potential to improve autophagy, clear damaged proteins, and beneficially influence the aging-associated decline in cardiac function is dynamic exercise. He et al. first showed in mice that 30-80 -min treadmill-running increases LC3-GFP puncta, LC3II:LC3I, and p62 degradation in the heart . Beta cell lymphoma/leukemia 2 (Bcl-2) is an anti-apoptotic 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 e.g., Alzheimers disease (Lazarov et al., 2005) the Robbins laboratory group sought to determine whether this form of "environmental enrichment" might initiate myocardial 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 . The same investigative team later demonstrated that VWR increased mRNA expression of Atg4, Atg5, and Wipi1 in myocardium 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, EF) appear identical between CryAB R120G x Atg7 transgenic mice and CryAB R120G mice that completed VWR, suggesting that exercise-training conferred benefits similar to genetic autophagy activation and vice versa.
In light of the interesting findings from interventions involving nutraceuticals, pharmaceuticals, and lifestyle alterations (e.g., caloric restriction, regular physical activity) in older mice, we tested whether late-in-life exercise training induces cardiac autophagy to an extent that improves proteostasis and lessens myocardial dysfunction. As anticipated, the intensity, frequency, and duration of "forced" exercise training produced functional and biochemical evidence indicating our protocol was efficacious ( Figure S5). In support of our hypothesis, aging-associated increases in LC3I:GAPDH and p62:GAPDH were lowered in hearts from old-ETR vs. old-SED mice (Figure 3c,f; Figure 4c,f). Importantly, CQ-induced p62:GAPDH accumulation occurred in hearts from old-ETR but not old-SED mice ( Figure 4f).
Collectively, these results indicate that 3-mo treadmill-running improves autophagic flux in hearts from older mice, and these observations are concurrent with reduced myocardial protein aggregates (Figure 5a, b, c), ubiquitinated proteins ( Figure 5d, e), and lipid peroxides (i.e., 4-HNE; Figure 5f, g). It is not unreasonable to suggest that elevated Atg3 in hearts from old-ETR vs. old-SED mice ( Figure S7b) facilitated the conversion of cytosolic LC3I to LC3II to thus lower LC3I:GAPDH. While the mechanism(s) responsible for elevated mRNA and protein expression of Atg3 in old-ETR vs. old-SED mice is unclear and has not been reported earlier, improved redox balance ( Figure 5) displayed by hearts from old-ETR mice might be responsible, based on results from studies using other cell types (Frudd et al., 2018;Kamat et al., 2015). The ability of CQ to promote accrual of p62 in hearts from old-ETR but not old-SED mice ( Figure   4f) indicates improved autophagic flux and likely explains the training-induced reduction of p62:GAPDH we observed in two separate cohorts of old-ETR vs. old-SED that were untreated (Figure 3f) or treated with vehicle ( Figure 4f). Three mo of treadmill-training did not prevent aging-associated cardiac fibrosis (Figure S11a-d). However, multiple indexes of systolic performance, together with a doppler-derived measure of overall LV function that incorporates the time intervals of mitral valve inflow and aortic valve outflow (i.e., MPI) (Tei, Nishimura, Seward, & Tajik, 1997), improved in trained vs. untrained older mice ( Figure 6). Highlighting the association between training-induced benefits concerning cardiac autophagy (i.e., reduced cardiac p62 protein expression) and LV function (i.e., lower MPI), a significant correlation existed between these two endpoints in older trained mice ( Figure 6k). On balance, our data indicate that rejuvenated autophagic flux contributes importantly to greater protein clearance and attenuated cardiac dysfunction in the context of aging.
We observed a strong association between the: (i) age-related accrual of p62:GAPDH (repressed autophagy) and the increase (i.e., worsening) of MPI (Figure 2l); and the (ii) training-induced reduction in p62:GAPDH (improved autophagy) and the decrease (i.e., improvement) of MPI ( Figure 6k). While these findings indicate training-induced elevations in autophagic flux associate positively with preserved myocardial 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 e.g., 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 myocardial 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.

EXPERIMENTAL PROCEDURES
Animals and housing. Male C57BL/6J mice were obtained from the Jackson Laboratories at 4 months of age, and from the National Institute on Aging rodent colony at 18 months of age. All mice were handled according to Institutional approved procedures documented in protocol number 19-07010.
Histology, morphology, mRNA gene expression. Twenty-four -48 h after assessing myocardial function, mice were anesthetized and hearts were segmented to assess protein indexes of autophagy, histology, morphology, and mRNA expression of pro-fibrotic and antioxidant genes (Bharath et al., 2017;Bharath et al., 2015;S.-Y. Park et al., 2016;Pires et al., 2017;Symons et al., 2011;. Exercise training. Body composition was assessed using TD-NMR in 5-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) x total running time (min) x final running speed (m/min) x treadmill grade (25%)] (Symons et al., 2011;. 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 mo progressive resistance treadmill-running program. After 3 mo, body composition, exercise tolerance (maximal workload capacity), and myocardial function were assessed. Each evaluation was separated by 24 h. Twenty-four h after measuring myocardial function, all mice were anesthetized as described, a blood sample was obtained via cardiac puncture, and the heart was excised and segmented to assess protein indexes of autophagy, histology, morphology, and mRNA expression of pro-fibrotic genes (described earlier), together with protein aggregate accumulation. Soleus muscle also was dissected free from both hindlimbs to assess CS enzyme activity (Sigma-Aldrich) .
Cardiac protein aggregation. After determining myocardial 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;Watanabe et al., 2012). Electron microscopy was used as a second approach to estimate cardiac protein aggregates (DiMemmo et al., 2017;. Details concerning each of these methods is provided in online "Experimental procedures." Statistical analyses. Data are presented as mean ± standard of error of the mean. Significance was accepted when p < .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 adult, adult-CQ, old, old-chloroquine). In cases when a significant main effect was obtained, a Tukey post-hoc test was used to determine the location of the differences.

CONFLICT OF INTEREST STATEMENT
None of the authors has any conflicts of interest to disclose.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.  The correlation between protein expression of p62/GAPDH and MPI was strong in hearts from A and O mice. For (b-i, k) n=12-13; for (l) n=16-17. *p<0.05 vs A.    Myocardial autophagy. First we tested the hypotheses that static (i.e., basal) myocardial autophagy and / or autophagosome formation (i.e., autophagic flux) are repressed in hearts from 23-24-month vs. 7-8-month old mice. A schematic is shown in Figure 1a. Lean mass, fat mass, and fluid mass were assessed in all mice using time-domain-nuclear magnetic resonance (TD-NMR; Bruker minispec, Bruker Biospin Corporation) (Bharath et al., 2017;Bharath et al., 2015;Pires et al., 2017). Twenty-four h later, the lysosomal acidification inhibitor chloroquine (CQ; 75 mg IP/g lean body mass) or vehicle-control (phosphate-buffered saline; PBS) (Gottlieb et al., 2015) was administered to separate cohorts of older and adult mice. Four h later, mice were anesthetized using 2-5% inhaled isoflurane combined with 100% oxygen. When a stable plane of anesthesia was attained, the chest was opened using aseptic procedures, the heart was exposed, excised, and segmented to assess protein indexes of autophagy (Bharath et al., 2017), ubiquitinated proteins, and 4-hydroxy-2-nonenal (4-HNE), an α, β-unsaturated hydroxyalkenal that is an estimate of lipid peroxidation. Protein isolation and immunoblotting analyses were performed as described by us (Bharath et al., 2017;Bharath et al., 2015;S.-Y. Park et al., 2016;Pires et al., 2017;Symons et al., 2011;. In brief, total protein from each heart was separated by SDS-PAGE (4-20%), transferred onto polyvinylidene difluoride (ThermoFisher), and probed with LC3II, LC3I, p62/SQSTM1, total-ubiquitin, 4-HNE, and GAPDH primary antibodies. Alexa Fluor anti-rabbit 680 (Invitrogen) and anti-mouse 800 (VWR International) served as secondary antibodies.

Experimental
Fluorescence was quantified using the Odyssey imager (LI-COR Biosciences).
Myocardial function. It was necessary for us to substantiate previous reports that myocardial function observed in adult mice is compromised in older animals (Dai & Rabinovitch, 2009;Dai et al., 2009;Flynn et al., 2013). A schematic is shown in Figure 2a. Separate cohorts of adult and older mice were anesthetized lightly with 1-3% isoflurane anesthesia combined with 100% oxygen while myocardial function was assessed using transthoracic echocardiography (Pires et al., 2017;Symons et al., 2011). Acquisitions were made using a Vevo 2100 high-resolution unit equipped with a 22-55 MHz transducer (Visual Sonic) (Pires et al., 2017;Symons et al., 2011).
An investigator blinded to mouse age performed the analyses and reduced the data using the Vevo-Strain software / Vevo 2100 imaging system. Parasternal long axis images acquired in Bmode and parasternal short axis images at the level of the papillary muscles acquired in M-mode were captured using an MS 550D transducer. Indices of systolic function [stroke volume (SV), ejection fraction (EF), fractional shortening (FS), cardiac output (CO)], together with enddiastolic volume, end-systolic volume, and end-systolic left ventricular (LV) mass were assessed.
Parasternal short-axis measures in M-mode were used to estimate LV anterior wall thickness in diastole and systole, LV internal diameter in diastole and systole, and end-diastolic and endsystolic volumes. Using pulse wave doppler, diastolic function was estimated by measuring passive (i.e., early due to pressure gradient; E) and active (i.e, later due to atrial contraction; A) velocity of blood flow through the mitral valve (MV). These results were used to calculate the E/A ratio. The myocardial performance index (MPI), an indicator of global left ventricular function, was calculated after measuring isovolumetric contraction time (ICT), isovolumetric relaxation time (IRT), and ejection time (ET), as (ICT + IRT) / ET (Goroshi & Chand, 2016;Tei et al., 1995). VevoLab analysis software 3.1.0 was used to quantify all measures.
Histology, morphology, mRNA gene expression. Twenty-four -48 h after assessing myocardial function, mice were anesthetized as described, and excised hearts were segmented to assess protein indexes of autophagy (described earlier), histology, morphology, and mRNA expression of pro-fibrotic and antioxidant genes. Segments of frozen tissue were cut and embedded in optimal cutting temperature (OCT) compound (Fisher Scientific). Five sections (5 µm thick) per sample were placed on glass slides, and stored at -80°C. To assess type 1 collagen, sections were air dried for 60-min, fixed in ice cold acetone (Fisher Scientific) for 10 min, blocked for 60-min in 3% BSA blocking buffer, and stained with anti-Collagen I antibody (1:200, Abcam) at 4°C overnight. Slides were washed three times with 1X PBS, and incubated with Alexa Fluor Plus 647 goat anti-Rabbit IgG secondary antibody (1:500, Invitrogen) for 60-min at room temperature. Negative control sections were treated with secondary antibody only. Next, sections were mounted with ProLong™ Gold Antifade Mountant with DAPI (Invitrogen). Ten 20X images were randomly acquired per section with an XM10 Olympus fluorescence camera (Dai & Rabinovitch, 2009;Tei et al., 1995). Image quantification was performed using the CellSens Dimension software (Olympus). Percentage of the collagen area in total heart tissue was quantified by ImageJ software (NIH). To assess cardiac myocyte area, sections were stained with wheat germ agglutinin (WGA, Alexa 488, Thermo Fisher) for 60-min at room temperature, followed by DAPI (Alexa 405) for 5-min at room temperature. Ten 20X images were randomly acquired per section with an XM10 Olympus fluorescence camera, and cardiomyocyte cross sectional area (µm 2 ) was quantified using CellSens Dimension software (Olympus) (Pires et al., 2017). To assess mRNA expression of fibrosis, total RNA was extracted from segments of myocardium from adult and old mice using the RNeasy Mini Kit (Qiagen). (Bharath et al., 2017;Bharath et al., 2015;S.-Y. Park et al., 2016;Pires et al., 2017;Symons et al., 2011;. Total RNA was reverse transcribed to cDNA using the QuantiTect Reverse Transcription Kit (Qiagen). The quantitative gene expression assay was performed with specific primers for fibrillin (Fbn) 1, Fbn2, transforming growth factor-b (Tgfb)1, Tgfb2, connective tissue growth factor (Ctgf), superoxide dismutase (SOD) 1 and SOD2, and catalase. Relative gene expression was normalized to 18S. Primer sequences are shown in Table S2.
Exercise training. Our main focus was to determine whether late-in-life exercise-training improves basal cardiac autophagy and trafficking of the autophagosome to the lysosome to an extent that rejuvenates protein aggregate clearance and recovers myocardial function. A schematic is shown in Figure 3a. Body composition was assessed using TD-NMR in 5-month and 21-month old mice. Twenty-four to 48 h later, mice were familiarized with walking/running on a motorized treadmill (Columbus Instruments) for 3 consecutive days x 5-10 min/day. On day 4, a workload capacity evaluation test was completed on each mouse i.e., 1 min x 5 m/min x 25% grade, followed by 1 m/min increases in speed each min until maximal exercise capacity was achieved. Electrical shocks were not used. Mice were encouraged to run by tapping their rear using test tube cleaning brushes. Total workload was calculated as [body weight (kg) x total running time (min) x final running speed (m/min) x treadmill grade (25%)] (Symons et al., 2011;. 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 mo progressive resistance treadmill-running program. Adult-SED and old-SED mice ran on the treadmill 1 day/week x 5 m/min x 10% grade for 5-min to maintain familiarization. This was required so that SED mice would be able to complete a second workload capacity evaluation test after 3 mo. Adult-ETR and old-ETR mice trained 6 days/week x 3 mo. The initial intensity corresponded to 70% of their workload capacity e.g., 30 min x 11.4 m/min x 5% grade. Over the next 3 mo, running duration, treadmill speed, and/or treadmill grade were increased every four days to 60 min x 17.4 m/min x 15% grade. After 3 mo, body composition, exercise tolerance (maximal workload capacity), and myocardial function were assessed. Each evaluation was separated by 24 h. Twenty-four h after measuring myocardial function, all mice were anesthetized as described, a blood sample was obtained via cardiac puncture, and the heart was excised and segmented to assess protein indexes of autophagy, histology, morphology, and mRNA expression of pro-fibrotic genes (described earlier), together with protein aggregate accumulation. Soleus muscle also was dissected free from both hindlimbs to assess CS enzyme activity (Sigma-Aldrich) .
Cardiac protein aggregation. After determining myocardial 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;Watanabe et al., 2012). In brief, samples (3 mg/ml) were added to a 96-well plate together with 2 μl of detection buffer, and incubated for 5-min in the dark. Fluorescence output was measured at an excitation setting of 550 nm and emission filter of 600 nm (Varioskan Lux, ThermoFisher). Protein aggregation (%) was calculated based on an 8 point standard curve ranging from 0-12% aggregated IgG. Electron microscopy was used as a second approach to estimate cardiac protein aggregates (DiMemmo et al., 2017;. Cardiac segments placed in 2.5 % glutaraldehyde at the time of collection were embedded in plastic, sectioned at 0.5 um with glass knives, and further trimmed with a diamond knife. Sections were placed onto 200 mesh copper grids, and subsequently stained with saturated uranyl acetate, which extends preservation, improves contrast of the extracellular matrix, membranes, cytoplasm, and DNA, and facilitates the quantification of protein aggregation (Erickson, Anderson, & Fisher, 1987). Three fields of view were chosen based on the quality of the images. Images were photographed at 1100x, 2700x, 4400x, and 11000x magnification (FEI Tecnai T-12, ThermoFisher). Results were calculated by quantifying the number of protein aggregate clusters at 2700x magnification due to the clear and detailed images obtained using NIH Image J software.