Exercise-induced Cardiac Remodeling in the Young and Aged
Exercise capacity is among the strongest predictors of risk of death in men and women. Although exercise capacity is complex and driven by both genetics and environment, the overall benefits of exercise to improve cardiac structure and function have been established for decades. Exercise training improves cardiorespiratory fitness in subjects and models of all ages, and undoubtedly lowers cardiovascular risk mortality. Regular exercise induces beneficial structural and functional changes in the heart, collectively referred to as exercise-induced cardiac remodelling. However, despite clear and robust evidence for exercise as cardiovascular medicine, the mechanisms by which exercise-induced cardiac remodeling occurs in aged versus young hearts are not yet clear. In the following section, we discuss exercise-induced cardiac remodeling and how these mechanisms differ by age. Our discussion describes age-specific adaptation to exercise with a special focus on the contribution of hypertrophy and fibrosis to cardiac function. Although these processes were chosen based on our group's expertise,[14,15] as well as available literature in both young and aged models, we would be remiss if we did not mention that other age-specific differences underlie the response to exercise including mitochondrial function and dynamics, antioxidant defenses, and angiogenesis. Furthermore, it is likely that noncardiac adaptations to exercise differ by age such as vascular and skeletal muscle adaptations.
Preclinical Rodent Models of Exercise
Limited access to human cardiac tissue is a major obstacle to understanding the mechanisms of aging and of exercise-induced cardiac remodeling. To circumvent this issue, preclinical models, specifically rodents, have been useful because of similar cardiac aging phenotypes to humans as well as similar exercise physiology. The focus of this review is on preclinical rodent models of exercise and cardiac function, although specific human reports are mentioned where warranted. From a preclinical standpoint, most literature utilizes mice and rats that undergo forced treadmill training or voluntary wheel running, with smaller contributions from swim exercise, high-intensity interval training (HIIT), and resistance training. The advantages and disadvantage of preclinical exercise models have been well reviewed elsewhere, but in most instances, regardless of the modality, mice and rats undergo predictable cardiac remodeling that recapitulates human adaptations. However, subtle but critical differences exist, especially when considering age. Generally, rodents sprint for short durations in multiple installments over a 24-h period. Therefore, cumulative exercise for a rodent does not directly equate to voluntary physical activity levels in humans who likely do not engage in as many repeated installments. However, given that wheel running is voluntary, it is still generally accepted as mirroring the natural pattern of activity in humans, with some suggestion that it may more accurately represent human voluntary activity with advanced age than other forms of training. Although activity is highly strain dependent, running wheel distance and speed peak around 12–15 wk of age, then steadily decline because of reductions in both running velocity and duration. Therefore, the low engagement of voluntary wheel running can be a barrier to implementation of this modality in aged animals. Forced exercise offers the circumvention of low running distances in aged animals. However, in addition to the psychological and physiological stress of treadmill running, it is often accompanied by disadvantages in circadian timing of exercise with training occurring during the animal's inactive period. Furthermore, it can be difficult to age-match for intensity to compare young and aged animals. Awareness of these barriers and selection of appropriate models to answer age-specific questions will enhance the translatability of preclinical findings to the human heart.
Cardiac Function in the Young and Aged
Principal adaptations of cardiovascular function in response to exercise in the young include both improved systolic and diastolic functions. Endurance training increases stroke volume (SV) and cardiac output (CO) to improve delivery of blood to the periphery. Training also induces diastolic remodeling with increased end-diastolic volume (EDV) due to higher LV volumes and improved LV compliance. Together, higher EDV also contributes to higher SV due to Frank-Starling mechanisms (reviewed in).
Exercise-induced changes in cardiac function in the aged heart are more variable and typically diminished in comparison to the young. Preclinically, a few reports demonstrate modest improvements in ejection fraction (EF) and fractional shortening[24,25] after 8–12 wk of regular endurance training in aged mice and rats. However, in older humans, improvements in systolic function with endurance training are not commonly reported. Unpublished data from our group demonstrate that 2 wk of voluntary wheel-running improved systolic function in young mice as measured by higher fractional area change (FAC) and EF. However, these changes were blunted in aged mice (Figures 1A, B) as evidenced both by the significant interaction between age and running as well as within-age comparisons showing more robust adaptations in the adult heart. We note that in both our unpublished work and the preclinical studies previously mentioned, assessments of systolic function were made at rest, not during exercise. Declines in resting systolic function generally do not occur with aging, but rather declines in systolic reserve characterize the aging heart. Modest evidence exists for improvements of systolic reserve with exercise training in aged rats and humans (reviewed in), although the mechanisms are not yet fully described. Diastolic function, however, does markedly decline with aging. In most preclinical reports, diastolic function improves with exercise training, at least to the extent that it attenuates age-associated impairments in diastolic function.[25,27] The mechanism of improved relaxation differs in the young and aged hearts, with improved early diastolic filling in aged but not in adult hearts. Distinct mechanisms of exercise-induced diastolic function have also been reported in humans. Six months of progressive aerobic exercise training in men improved peak atrial filling rate in older exercisers while remaining unchanged in young. These findings make sense, given that diastolic filling in the aged heart has a larger contribution from atrial filling rather than passive relaxation. However, we do note that exercise-induced changes in diastolic function in humans remain controversial, with many reports suggesting less robust or absent changes in relaxation. Elegant exercise interventions of progressive and vigorous training protocols generally have not demonstrated significant improvements in LV stiffness in men or women,[30,31] despite favorable changes in aerobic capacity. Together, although exercise training may improve systolic or diastolic function at rest, these functional changes are modest in comparison to the well-established functional outcomes of training in young hearts.
Exercise-induced cardiac remodeling by echocardiography and cardiac morphometrics in young and aged male and female mice. Adult (4- to 6-month-old) or aged (18-month-old) C57Bl6 mice underwent 2 wk of voluntary wheel running. A. Ejection fraction (EF) was higher in adult female mice in response to exercise training but not different in aged mice. B. Fractional area change (FAC) was higher in adult mice while unchanged in aged mice. C. Adult male mice had larger heart weight normalized to body weight (Ht/BW) in response to wheel running, whereas aged mice of both sexes demonstrated a regression of cardiac mass after wheel running. D. Similarly, adult male mice had larger left ventricle weight normalized to body weight (LV/BW) in response to wheel running, whereas aged mice of both sexes had smaller LV/BW. Average nightly running distances were as follows: adult male, 3 ± 0.6 km; adult female, 5 ± 0.7 km; aged male, 1 ± 0.5 km; and aged female, 3 ± 0.9 km. *P < 0.05 sedentary versus wheel within age and sex. Data were analyzed by two-way analysis of variance, followed by Student's t-test within age and sex. Data are expressed as mean ± SEM. Sed, sedentary; Wheel, voluntary wheel running. Bruns DR, unpublished data, 2022.
Hypertrophy in the Young and Aged
Chronic exercise training stimulates LV hypertrophy. This type of hypertrophy, referred to as physiological hypertrophy, is distinct from age-related hypertrophy. Although differences exist with respect to the type of hypertrophy (eccentric or concentric, due to endurance training or strength training, respectively), exercise-induced hypertrophy is characterized by both higher LV size and volumes. Physiological hypertrophy is accompanied by an increase in LV chamber dimensions, increased LV EDV, and decreased ESV, as well as enhanced left atrial cavity size. Targeted genetic investigations in mice have elucidated the pathways responsible for physiological hypertrophy including insulin-like growth factor 1 (IGF1), phosphoinositide 3-kinase (PI3K), and protein kinase Akt (reviewed in). Higher expression of these mediators has been reported in professional athletes, further mechanistically linking these molecules to hypertrophic outcomes. Although most of the beneficial reports of exercise on cardiac remodeling have been gleaned from endurance exercise, over the last several years, growing recognition has emerged regarding other forms of exercise, particularly with respect to HIIT and reports of greater changes in LV mass compared with moderate-intensity exercise.
In contrast to the wealth of data demonstrating exercise-induced cardiac hypertrophy in young models, extensive variability exists in the aging heart. Although a few studies have reported that moderate-intensity treadmill running increases cardiac mass and myocyte cross-sectional area in 24-month old male mice, the bulk of the work to date suggests that hypertrophy is either unchanged[37,38] or paradoxically reversed with exercise training in the aged.[16,39] We recently compared the hypertrophic response of young (3-month) and aged (18-month) mice in response to 4 wk of voluntary wheel running (Bruns DR, unpublished data, 2022). We found consistent regression of LV and heart mass in 18-month-old mice — both when expressed as gross morphometric assessment and when normalized to body weight (Figures 1C, D). This finding that exercise regresses cardiac mass in the aged heart is significant and interesting in light of the observations that antiaging drugs, which improve cardiac function also promote the regression of LV mass. Furthermore, the notion that exercise can reverse or attenuate cardiac mass is not without precedence, as young animals engaging in physical activity after or preceding myocardial infarction also demonstrate attenuation of LV hypertrophy.
Unlike the molecular mediators for exercise-induced hypertrophy in the young heart, the mechanisms by which the aged heart changes cardiac mass are not well described. Although some work suggests that Akt is activated in response to 8 wk of swim exercise training in rats, whether this is similar in magnitude to the activation elicited by chronic training in young animals is not known because this study did not include a young exercise group. IGF1 and PI3K were upregulated in the hearts of middle-aged mice and rats by 12 wk of wheel running or swim training, respectively. However, statistical analyses did not assess whether the magnitude of these changes was the same as in younger animals, nor were older animals included. Exercise likely regulates hypertrophy in the aging heart through myocyte survival. Age-related cardiac hypertrophy is due at least in part to myocyte apoptosis and concomitant reactive hypertrophy to replace lost myocytes. Activation of apoptotic signals in the aged heart is reversed or at least attenuated by exercise training,[39,43] suggesting attenuation of cell death may be a mechanism by which exercise attenuates age-associated hypertrophic remodeling. However, it is not yet clear whether reduced apoptotic signals result in attenuated cell death and diminished hypertrophic response. Identification of these mechanisms is important for understanding exercise-mediated changes in cardiac mass and how this response differs in young versus aged hearts.
Clinically, exercise training in older adults also does not seem to be as robustly hypertrophic as in younger subjects. Cross-sectionally, no differences were noted in LV mass between apparently healthy sedentary seniors and master athletes. Older women who underwent 12 wk of aerobic, strength, or combined aerobic and strength training saw no change in cardiac mass by echocardiography, nor were changes in cardiac structure noted in octogenerians who underwent 9 months of exercise training at 85% of heart rate peak. With longer duration and higher-intensity interventions, modest increases in LV mass were noted in previously healthy sedentary seniors, perhaps suggesting that higher intensity and volumes of exercise are needed for the hypertrophic response. Although future work is warranted with respect to the intensity, duration, and timing of the exercise intervention and the molecular signals behind exercise-induced remodeling, it does not seem as though the aged heart undergoes similar hypertrophic adaptation as the young.
Fibrosis in the Young and Aged
The aging heart is characterized by excess deposition of ECM proteins, resulting in a fibrotic heart with elevated arrhythmia potential, diastolic deficits, and subtle contractile abnormalities during exercise. Myocardial ECM accumulation is the end result of the balance between synthesis and degradation of ECM proteins. A major component of the ECM is collagen, with collagen content including the summation of all types of collagens that ultimately reflects ECM quality. Excess accumulation of ECM in the aging heart has been suggested to be due to age-associated increases in profibrotic mediators such as transforming growth factor β1 (TGF-β1), as well as declines of mechanisms, which remove the matrix such as metalloproteinases (MMP). Although fibrosis and diastolic function are not always connected, as discussed in greater depth hereinafter, collagen accumulation is undoubtedly significant in the aged heart, given the strong correlation between collagen content and diastolic stiffness. Therefore, not surprisingly, exercise emerged as an antifibrotic intervention in the aged heart given its impact on diastolic function.[17,48–50] We were unable to find reports of the impact of exercise on fibrosis in the human heart. Given the robust reports of age-associated diastolic dysfunction and ECM deposition in human hearts, it is reasonable to hypothesize that some degree of antifibrotic remodeling occurs in response to exercise; however, the degree to which the ECM reverse remodels is not yet known.
Exercise may promote degradation of collagen and other ECM proteins by MMP, downregulate profibrotic factors, or both. Downregulation of TGF-β1 and other profibrotic molecules was suggested to be responsible for attenuated fibrosis in aged rats that underwent 12 wk of endurance training. However, exercise also stimulated the expression of MMP. The strongest evidence for mechanisms of reduced ECM deposition with exercise is due to posttranslational modification of collagen proteins — that is, reducing age-related cross-linking. Several studies report unchanged expression of MMP and profibrotic gene expression, but do report attenuation of collagen cross-linking in the aged exercised LV.[37,48,49] Additional support for the hypothesis that the quality of the ECM supports beneficial cardiac remodeling comes from studies demonstrating that exercise training can alter collagen characteristics even if total collagen remains unchanged.[37,48] Age-related changes in oxidative stress and inflammation also likely contribute to modifications of ECM proteins via collagen cross-linking. Exercise has well-established antioxidant and anti-inflammatory mechanisms in the young heart, suggesting that similar mechanisms may contribute to a beneficial ECM phenotype in the aged heart as well.
Reports of exercise as an antifibrotic intervention are incredibly sparse in young models free from cardiac disease, presumably due to the lack of ECM deposition in the young heart. However, exercise has well-reported antifibrotic properties in young models of heart disease that are characterized by elevated fibrotic deposition such as after myocardial infarction. Although the mechanisms by which exercise attenuates fibrosis after cardiac injury are not fully clear, the authors posited that an attenuated inflammatory response with exercise likely reduced fibrotic deposition. Indeed, attenuation of the inflammatory response is linked to antifibrotic outcomes of exercise in other instances of cardiac injury and suggests that the mechanisms for reversal or attenuation of fibrosis are intact in the young heart, albeit not activated until administration of a profibrotic stress. Whether these molecular signals for attenuation of fibrosis in the young heart overlap with those that regulate fibrosis in the aged heart is not yet clear. Delineating these pathways by age remains an important focus for future research efforts.
Integration of Hypertrophy, Fibrosis, and Cardiac Function
Exercise-induced cardiac remodeling is the ultimate outcome of both beneficial structural and functional changes in the heart in response to regular exercise training. For example, LV hypertrophy is strongly linked to and likely largely responsible for improved systolic function with exercise training in the young heart. To this end, the lack of hypertrophic remodeling with exercise training in the aged heart may be responsible for diminished improvements in SV and systolic function. However, as discussed previously, interventions that promote cardiovascular health and slow aging also seem to result in regression of LV mass, perhaps suggesting that hypertrophy is not a requirement for improvements in cardiac function or overall cardiac health. The mechanisms of cardiac mass regression are of interest, not only for the aging heart but also for models of cardiac disease that are characterized by pathological hypertrophy. The mechanisms of loss of cardiac mass are not yet clear but could be due to regression of myocyte size or attenuation of ECM deposition. As discussed previously, loss of myocytes with age results in both reactive hypertrophy and replacement fibrosis. Therefore, if exercise attenuates myocyte loss, it may both attenuate/reverse hypertrophy and fibrotic remodeling, thus regressing cardiac mass. Blood pressure also directly regulates cardiac mass, with elevated afterload such as with age-associated hypertension causing hypertrophy. To date, few studies have adequately addressed the blood pressure–lowering effects of exercise; thus, it remains unclear if attenuation of age-associated LV afterload is sufficient to stimulate cardiac mass regression. To begin to address some of these challenges in dissecting cardiac morphological from physiological remodeling, multifaceted approaches that take into consideration physiological, molecular, and morphological outcomes are warranted.
In the work discussed in the scope of this review, changes in fibrosis were often linked to improved relaxation. However, in several cases, changes in fibrosis were independent from diastolic function and vice versa. For example, aged rats demonstrated improved collagen accumulation likely because of enhanced breakdown of ECM in response to HIIT; however, diastolic function was not improved. ECM accumulation is strongly linked to LV stiffness. However, stiffness is also regulated by the myofilament. Relaxation is composed of a passive and active process and is dependent on the properties of the myofilament. Diastolic function therefore integrates sarcomeric function, extracellular coupling, along with noncardiac contributions from the vascular system. Myofilament-specific adaptations occur during aging and in response to exercise. Furthermore, the impact of exercise on myofilament adaptations differs in aged versus young hearts. For example, age-associated deficits in relaxation are in part due to attenuated sarcoplasmic reticulum Ca2+ATPase activity and prolongation of the calcium transient. These age-dependent alterations of calcium signaling are reversible by chronic exercise training. This finding is interesting given that the sarcoplasmic reticulum does not seem to respond to exercise training in young hearts (reviewed in). Thus, it is clear that the mechanisms by which exercise is cardioprotective in the young and aged hearts are not similar. Identification of the underpinnings by which these outcomes differ is of great interest.
Exerc Sport Sci Rev. 2022;50(3):137-144. © 2022 American College of Sports Medicine