Exercise training reverses cardiac aging phenotypes associated with heart failure with preserved ejection fraction in male mice

Abstract Heart failure with preserved ejection fraction (HFpEF) is the most common type of HF in older adults. Although no pharmacological therapy has yet improved survival in HFpEF, exercise training (ExT) has emerged as the most effective intervention to improving functional outcomes in this age‐related disease. The molecular mechanisms by which ExT induces its beneficial effects in HFpEF, however, remain largely unknown. Given the strong association between aging and HFpEF, we hypothesized that ExT might reverse cardiac aging phenotypes that contribute to HFpEF pathophysiology and additionally provide a platform for novel mechanistic and therapeutic discovery. Here, we show that aged (24–30 months) C57BL/6 male mice recapitulate many of the hallmark features of HFpEF, including preserved left ventricular ejection fraction, subclinical systolic dysfunction, diastolic dysfunction, impaired cardiac reserves, exercise intolerance, and pathologic cardiac hypertrophy. Similar to older humans, ExT in old mice improved exercise capacity, diastolic function, and contractile reserves, while reducing pulmonary congestion. Interestingly, RNAseq of explanted hearts showed that ExT did not significantly modulate biological pathways targeted by conventional HF medications. However, it reversed multiple age‐related pathways, including the global downregulation of cell cycle pathways seen in aged hearts, which was associated with increased capillary density, but no effects on cardiac mass or fibrosis. Taken together, these data demonstrate that the aged C57BL/6 male mouse is a valuable model for studying the role of aging biology in HFpEF pathophysiology, and provide a molecular framework for how ExT potentially reverses cardiac aging phenotypes in HFpEF.

Parasternal short-axis 2D images were acquired at ~400-600 frames per second for Speckle tracking strain analysis. Radial strain analysis was performed offline using GE EchoPACS software (version 201) on optimal 2D cine images in which endocardial borders could be well delineated in three consecutive cardiac cycles. Using the software's semi-automated tracing system, the region of interest was generated by tracing the endocardial and epicardial borders at end-diastole. LV radial strain and strain rate curves were then generated for six segments at the papillary muscle level for a full cardiac cycle set from end-diastole to end-diastole. Systolic strain was averaged from the peak value of segmental strain curves in systole. Early diastolic strain rate was averaged from the six segmental strain rates at 33% into diastole. Diastole was defined from aortic valve closure (AVC) to maximum LV dimension at end-diastole, and segmental early diastolic strain rates were then measured from the frame corresponding to 33% between AVC and enddiastole and then averaged.
Cardiac systolic function was assessed by fractional shortening (FS=[(LVEDD-LVESD)/LVEDD] x 100) and radial systolic strain. Echocardiographic assessment of diastolic function was done by early diastolic strain rate, which measures the rate of myocardial deformation during the active myocardial relaxation phase. This method was selected because it is less angle-and load-dependent than other methods, and enabled us to assess diastolic function in animals without sedation. A limitation of this study is that conventional diastolic function techniques, such as mitral inflow (E/A) ratios and E/e' ratios, could not be obtained due to physiological heart rates in the 500-700 bpm range and inability to acquire apical views without the use of anesthesia.

Stress Echocardiography Protocol
A modified stress echocardiography protocol was designed to measure exercise capacity and cardiac reserves in mice. The protocol was completed over two days with day 1 dedicated to resting baseline echocardiography and day 2 dedicated to exercise capacity testing and echocardiographic assessments at peak exercise.

Baseline Echocardiography:
Baseline resting echocardiography, as described in the echocardiography methods section above, was performed on a separate day to avoid any potential confounding effects on exercise testing.

Exercise capacity testing:
Prior to formal testing, mice were acclimated to the automated treadmill (Columbus Instruments). Animals were acclimated for three consecutive days by walking at a pace of 5-10 m/min for 5 min (day 1), 10 min (day 2), and 15 min (day 3). The treadmill protocol for exercise capacity consisted of a warmup phase and a run-to-exhaustion phase with the treadmill incline set at 10°. Warmup lasted for 5 min, during which the treadmill speed increased from 5 m/min to 15 m/min. The run phase started at a speed of 15 m/min and increased at a pace of 2 m/min/min until exhaustion was reached. Animals were motivated to run with a combination of tail tapping and puffs of compressed air directed at the hindlimbs. An animal was judged to be exhausted if it could not keep pace with the treadmill for a full three seconds without falling back on to the resting pad, a pattern that had to be repeated three times in a row. Once exhaustion was reached, a small amount of blood was immediately acquired by nicking the tip of the tail with a sharp razor blade. Maximum exercise effort was then corroborated by measuring point-of-care blood lactate with a StatStrip Xpress Lactate Meter (Nova Biomedical).

Echocardiography at Peak Exercise
Immediately after peak lactate measurement, animals were rapidly transferred to an echocardiographer who acquired parasternal short-axis M-mode images within 30 seconds of peak exercise. Chronotropic reserves were assessed by measuring HR at peak exercise, which was an average of nine beats (three consecutive beats from three independent images). Contractile reserves were assessed by comparing FS at peak exercise to baseline FS at rest (from day 1) to determine the degree of FS augmentation in response to exercise Blood Pressure Measurement: Systemic blood pressure was measured using a tail cuff approach with the CODA noninvasive blood pressure system (Kent Scientific). Mice were acclimated to the system and final measurements were done without the use of anesthesia. At least nine blood pressure measurements were obtained for each mouse, and then averaged, to provide a single averaged blood pressure per mouse. Final stress echocardiography testing was performed on 4 sedentary and 5 exercise trained (ExT) mice due to deaths prior to study completion. All animals were 30 months old at study completion. Peripheral blood lactate measurements were obtained at peak exercise to confirm exhaustion and adequate effort during exercise testing. Data shown as mean  SEM, with all individual data points plotted. Unpaired Student's t-test used for analyses. * p<0.05, ** p<0.01, *** p<0.001.

Supplemental Figure S6
Supplemental Figure S6: QPCR validation of ExT RNAseq candidates. Validation was done in cardiac samples from an independent cohort of 18-month-old C57BL/6 males subjected to eight weeks of voluntary wheel running (EXT) versus normal sedentary lifestyle (SED). n=5/group. A) Top five genes driving cell cycle-related pathway enrichment. B) Genes significantly upregulated by ExT in discovery cohort. C) Genes significantly downregulated by ExT in discovery cohort. Data shown as mean  SEM, with all individual data points plotted. Unpaired Student's t-test used for analyses. * p<0.05, ** p<0.01, *** p<0.001.