Molecular mediators of the cardiac benefits of exercise

Abstract

Exercise exerts myriad cardiovascular benefits and protects against most forms of cardiovascular disease. While extensive epidemiological evidence supports the clinical benefits of exercise, our understanding of the molecular underpinnings of its benefits in the heart remains incomplete. Pinpointing these mechanisms is essential to identifying molecular targets modulated by exercise for therapeutic gain. In this Review, we discuss the cellular and molecular mechanisms through which exercise benefits the heart, with a focus on the cardiomyocyte. We highlight cardiomyocyte secreted mediators of inter-cell communication, intracellular signaling pathways, transcriptional regulation and post-transcriptional mechanisms through noncoding RNAs that have been implicated in the cardiac exercise response. We emphasize pathways and mediators regulated by exercise training that may provide therapeutic targets in heart failure.

Introduction

Exercise is a form of physical activity involving intentional physical effort to sustain or improve health and fitness. Exercise confers numerous benefits at the systemic and organ-level to reduce the incidence and morbidity of multiple chronic diseases^1–3. This includes almost all forms of cardiovascular disease. Initial epidemiological support from observational studies is further strengthened by randomized clinical trials of exercise^2,4,5. Regular moderate physical activity at least 150 minutes a week is recommended by the American Heart Association and American College of Cardiology for all adults. This has been shown to delay disease onset, reduce severity, and improve quality of life for those with many forms of cardiac dysfunction⁶. Importantly, routine exercise also improves functional capacity and outcomes in age-related heart diseases. Both endurance and resistance forms of exercise have been demonstrated to confer cardiovascular benefits⁷. Despite this, <25% of US adults meet these recommendations⁸. Many patients are unable to exercise adequately to derive these benefits.

Despite strong evidence supporting the benefits of exercise, we currently have a limited understanding of the underlying molecular mechanisms. There are several reasons why it is imperative that we explore these beneficial mechanisms in greater detail. First, studying these pathways will reveal new candidate mediators with therapeutic potential to pharmacologically mimic exercise. Notably, much work has shown that the pathways that remodel the heart in response to exercise training are distinct from those activated in models of cardiac injury. Second, pathways functionally important in the heart’s response to exercise are highly enriched for targets that protect against pathological stress, preventing or mitigating disease^9,10. Defining the candidates from mechanistic studies of exercise adaptations may thus provide actionable pathways and targets for therapeutic development.

In addition, exercise remains the only known physiological stimulus that induces cardiomyogenesis in the adult mammalian heart^11,12. Loss of cardiomyocytes is a critical feature of many forms of heart disease. Thus, identifying the mediators of cardiomyocyte proliferation induced by exercise could guide new therapies to mitigate cardiomyocyte loss.

Given limited tissue access, many studies seeking to understand the cardiovascular benefits of exercise in humans have focused on circulating biomarkers such as plasma proteins and metabolites, correlating individual components or networks with exercise performance parameters and cardio-metabolic risk factors and longitudinal disease outcomes¹³. These studies have identified numerous biomarkers suggestive of an exercise response. However, this approach alone cannot establish causal mechanisms. In this context, animal models of exercise have been useful tools for elucidating the intrinsic cardiac pathways contributing to cardiovascular benefits.

In this review, we highlight some of the predominant molecular mechanisms underlying the heart’s response to exercise. Numerous other cell types likely contribute to cardiac function in both exercise and pathological injury. Indeed, recent single-nucleus and single-cell RNA transcriptomics in human and rodent hearts have demonstrated marked cellular diversity with >20 cell types and subtypes^14–16. There is likely dynamic regulation of these cell populations and transcriptional profiles due to exercise. These may underlie some of the molecular changes in the heart due to exercise. We refer the reader to prior studies highlighting cellular variation in the composition of the heart and cell-type focused responses to exercise training^{10,12,14,17–23}. However, less is known about these adaptations and the cardiac benefits of exercise.

Here, we focus on the molecular mechanisms responsible for principle phenotypic adaptations that cardiomyocytes undergo in response to exercise training (Figure). We highlight those that we believe hold promise as therapeutic targets, particularly for heart failure and age-related cardiac dysfunction as two growing unmet clinical needs.

Figure. Molecular mechanisms of cardiomyocyte adaptations to exercise.

Diagram of mononucleated cardiomyocyte depicting the four predominant molecular mechanisms in response to exercise training that are discussed here (yellow rectangles): transcriptional regulation, post-transcriptional regulation, intracellular signaling, and cell-cell communication through secreted factors. Letters ‘A’ to ‘G’ arbitrarily signify components of an intracellular kinase signaling cascade; TF – transcription factor; ncRNA – noncoding RNA; lncRNA – long noncoding RNA; miRNA – microRNA; circRNA – circular RNA. Image created in BioRender.

Efforts to characterize these adaptations rely heavily on different animal models of exercise training. Studies often involved voluntary wheel running or forced swimming of rodents which have produced reliable physiological and molecular changes in cardiomyocytes. We previously reviewed both methods for endurance exercise training in rodent and other models²⁴. Where applicable, we also summarize evidence for specific molecular adaptations to exercise from studies in humans. We place a special emphasis on those mechanisms that are induced by exercise training that consistently and potently protect against cardiac injury in disease models. This paradigm paves the way for consideration of exercise-induced molecular mechanisms as beacons for potential novel therapies for cardiac injury.

Cardiomyocyte adaptations to exercise

There are several key phenotypic adaptations of cardiomyocytes in response to exercise, that encompass bioenergetics, contractility, physiological hypertrophy, cardiomyogenesis, and ischemic cardioprotection.

Nutrient metabolism in cardiomyocytes undergoes substantial rewiring with exercise training^25–28. Exercise induces expression of multiple programs important for oxidative metabolism. This includes nutrient transporters, tricarboxylic acid cycle machinery, ATP synthesis enzymes and the signals to generate new mitochondria. Increased workload from exercise creates a 3- to 10-fold increase in myocardial oxygen consumption above resting rates^29–31. Correspondingly, there is an increase in ADP concentrations which promotes oxidative phosphorylation to allow ATP production. This need for increased energy production is met with the systemic increase in substrates that promote oxidative metabolism in exercise. These include free fatty acids from catecholamine-stimulated lipolysis in adipose and lactate from glycolysis in skeletal muscle^32–34. These substrates contribute substantially to ATP production during exercise. With exercise training, the heart adapts by increasing expression and activity of pathways to support fatty acid oxidation and by decreasing glycolysis and glucose oxidation pathways. This includes tight regulation of phosphofructokinase 1, the rate-limiting enzyme of glycolysis. Exercise training-induced regulation of PFK1 allows cardiomyocytes to efficiently rely on fatty acid oxidation primarily but shift to glycolysis during acute demands, as reviewed previously³⁵. Metabolite signaling by increased adenine monophosphate (AMP) and glucose-6-phosphate (G6P) supports these fuel adaptations by transcriptional activity to promote mitochondrial biogenesis and quality control. In pathological cardiac hypertrophy, cardiomyocytes switch from an adult to fetal program for fuel utilization^35–37. This includes decreases in fatty acid oxidation and branched chain amino acid catabolism and increases in glycolytic flux and ketone body oxidation^35,36,38. This is due in part to suppression of the transcriptional programs supporting mitochondrial biogenesis and branched chain amino acid transport and oxidation³⁸. In addition, there is an increase in both glucose import through the glucose transporter 1 (GLUT1) and activation of PFK activity³⁵.

Exercise training also optimizes contractility of cardiomyocytes¹. This occurs through increased myocyte shortening, heightened calcium sensitivity, improved calcium recycling through increased sarco/endoplasmic reticulum Ca2+ ATPase (SERCA) 2a activity, improved mitochondrial calcium buffering, and other mechanisms. Exercise training also modulates the activity of calcium and potassium channels to increase efficiency, prolonging the action potential during rest and shortening it during peak activity. Exercise training after cardiac injury such as myocardial infarction has been shown to promote several of these adaptations.

Cardiac and cardiomyocyte hypertrophy (“physiological hypertrophy”) is perhaps the most recognized hallmark of cardiomyocyte adaptations to exercise^10,41,42. This is characterized by an increase in both cardiomyocyte length and width. Cardiomyocytes also increase in size in response to pathological stress such as pressure-overload. However importantly the physical changes in cardiomyocytes undergoing pathological hypertrophy are distinct with cardiomyocyte length often increasing disproportionately to width. At a cellular level, physiological and pathological hypertrophy are distinct⁴¹. A critical aspect of this is the difference in contractility of cardiomyocytes exposed to these two stresses. Cardiomyocytes express both α- and β-myosin heavy chains to optimize the force of contractility⁴³. β-myosin is a slower, low-force contractile protein with lower ATPase activity compared to α-myosin. Physiological cardiomyocyte hypertrophy increases the expression of α-myosin relative to β-myosin. In contrast both pathological cardiomyocyte hypertrophy in response to pressure-overload and other forms of cardiomyocyte injury increase β- relative to α-myosin. This results in relatively reduced metabolic activity and resultant decrease in contractile speed and adverse cardiac remodeling. Thus, changes in the sarcomere during physiological hypertrophy also contribute to the enhanced cardiomyocyte contractility noted above and the reverse appears true for pathological hypertrophy. As discussed above, metabolic fuel utilization is also an important distinction in physiological vs pathological cardiomyocyte hypertrophy. Metabolic flexibility in hearts with physiological hypertrophy mitigates oxidative stress through efficient utilization of fatty acids and other fuel sources (BCAAs, glycolysis) during acute stress in part through mitochondrial augmentation and adaptation³⁵. In contrast, pathologically remodeled cardiomyocytes relatively inefficient reliance on glycolysis for ATP production and growth, which results in lipid and glucose accumulation leading to toxicity from under-utilized substrates and increased oxidative stress^44–46. This stress promotes inflammation and fibrosis associated with pathological hypertrophy, which in part undermines the reversibility of this form of remodeling.

The generation of new cardiomyocytes (cardiomyogenesis) through exercise training offers great therapeutic promise¹⁰. Various types of endurance training can induce cardiomyocyte proliferation markers in young mice as well as in aged mice and in states of cardiac injury. Of note, endurance exercise is the only physiological stimulus known to induce cardiomyogenesis in adult mice. However, we want to emphasize that while the relative increase in cardiomyogenesis is large (~4.6-fold in young adult mice), this reflects the very low basal rate of cardiomyogenesis and the absolute number of new cells formed is undoubtedly modest. The change in heart size seen in exercise in attributable primarily to cardiomyocyte hypertrophy noted above.

One potential therapeutically important adaptation derived from exercise training is protection of cardiomyocytes from future ischemic injury (“ischemic cardio-protection”)⁴⁷. This molecularly involves multiple mechanisms including signaling through pre-conditioning kinase cascades, low-dose generation of protective reactive oxygen species, stress mediated increases in nitric oxide production, and expression of stress-induced proteins that may protect from proteostatic and other cellular stress^10,47–49. Both local and remote ischemic pre-conditioning through exercise training has been suggested to protect cardiomyocytes from severe ischemic injury in experimental models. This has provided motivation for both recommending exercise to reduce incidence and complications of myocardial ischemia in humans and for better understanding the molecular mechanisms.

Mechanisms underlying cardiac adaptations to exercise

Multiple coordinated cellular mechanisms contribute to the phenotypic adaptations to exercise described above. As a framework for discussing these mechanisms, we will consider them from ‘outside-to-inside’ with respect to the cell (Figure). An emerging area regulating exercise adaptations in other tissues that we will explore in cardiomyocytes is cell-cell communication via secreted factors and the role that exercise may play in modulating this in the heart⁵⁰. The discovery of novel low-abundance muscle-derived secreted proteins and metabolites in response to exercise (“exerkines”) that may act as paracrine and endocrine factors to impart the systemic benefits seen in exercise training has increased interest in discovery of such mechanisms in other tissues⁵¹. Exercise activates several signaling pathways in cardiomyocytes including kinase-mediated growth pathways, G-protein coupled receptors and others discussed below. Broad transcriptional programs contribute to the effects of exercise^28,37,52, regulated by ligand-driven nuclear receptors and other transcription factors, and transcriptional coactivators, notably including peroxisome proliferator activated receptor gamma coactivator 1α (PGC-1α)⁵³. We will discuss the gene programs regulated by PGC-1α and DNA-binding transcription factors that rewire metabolism and regulate cardiomyocyte growth, contractility, and protection from ischemic injury in cardiomyocytes after exercise. Additionally, there are important nodes of post-transcriptional regulation, including long noncoding RNAs (lncRNAs), microRNAs (miRNAs) and circular RNAs^54,55, discussed below.

Exercise-induced secreted factors and cell-cell communication

An emerging concept in studying exercise adaptations is that stressed cell types may communicate with each other or with other tissues locally or distally as compensatory measures. Numerous low-abundance secreted proteins and metabolites have been identified from tissues after exercise such as muscle (myokines) and adipose tissue (adipokines)⁵¹. Several of these appear to be regulated by transcriptional programs already essential to exercise adaptations such as PGC-1α in muscle⁵⁶. Relative to the muscle and adipose tissue, the autocrine, paracrine or endocrine secretory capacity of the heart has been less extensively studied in both exercise and disease contexts. We summarize some known exercise-induced heart-secreted proteins (Table 1) and discuss key ones here.

Table 1:

Heart-derived myokines related to exercise

Name	Sources	Effect of Exercise	Effects on the Heart	Key References
FSTL1	Cardiomyocytes, skeletal muscle	↑ with endurance training and pathological cardiac remodeling	Protection from cardiac injury, ↑ angiogenesis	217
GDF15	Many cell/tissue types	↑ with acute exercise and pathological cardiac remodeling	Suppression of pathological cardiac hypertrophy, exercise-induced adaptations unclear	87,218
CTGF	Cardiomyocytes, fibroblasts	↑ with pathological cardiac remodeling	↑ Fibrosis and hypertrophy	219
Natriuretic peptides (ANP, BNP)	Cardiomyocytes	↑ with acute exercise and pathological cardiac remodeling	Vasodilation, natriuresis, favorable anti-fibrotic remodeling	58–61,64
MCP-1	Cardiomyocytes, immunocytes	↑ in prolonged exercise and heart failure	↑ Monocyte chemoattraction, inflammation, fibrosis	220
SPARC	Cardiomyocytes, skeletal muscle	↑ with endurance training	extracellular matrix remodeling, anti-fibrotic remodeling	221
Apelin	cardiac endothelial cells	↑ with endurance and resistance training	↑ cardiac contractility, vasodilation, anti-apoptotic effects	222
IGF-1	Many cell/tissue types	↑ with endurance and resistance training	↑ Cardiac hypertrophy, anti-apoptosis	68
FGF21	Many cell/tissue types	↑ with acute exercise and pathological cardiac remodeling	↑ favorable glucose and lipid utilization, anti-apoptotic effects	93–95,97
NRG-1	cardiac endothelial cells	↑ with acute exercise and endurance training	↑ physiological cardiomyocyte hypertrophy and regeneration, anti-apoptotic effects	76,77,81

Cardiac secreted factors modulate cellular metabolism in response to exercise training. An early example of this is the natriuretic peptide family. Atrial, brain and C-type natriuretic peptides are released from different types of cardiomyocytes in response to pressure or volume overload. These peptides function both locally and systemically through a GPCR family of natriuretic peptide receptors (NPRs) and through cyclic guanidine monophosphate-protein kinase G (cGMP-PKG) signaling to exert many potentially beneficial actions⁵⁷. Local actions of natriuretic peptides include suppression of pathological cardiomyocyte hypertrophy^58–60, reduced ischemia-mediated cardiomyocyte apoptosis, and increased contractility⁶¹. Natriuretic peptides also act systemically to exert several other potential benefits. In the adrenal glands and kidneys, they decrease renin and aldosterone secretion, promoting endothelial vasodilation to reduce blood pressure, and reducing salt intake through central activity^62–64. NPR signaling promotes mitochondrial oxidation in white adipose tissue and skeletal muscle, possibly through activation of PGC-1α in those tissues^65,66. Natriuretic peptide secretion is transiently elevated after acute exercise in untrained and endurance trained individuals. Direct contributions of natriuretic peptide signaling on cardiomyocytes in the context of exercise training are unclear.

Insulin-like growth factor 1 (IGF1)⁶⁷ is a peptide hormone secreted from many tissues in response to exercise and other stimuli that promotes tissue growth^10,68–70. Plasma levels of IGF1 increase in humans and experimental models after acute or chronic exercise training⁷¹. Exercise-induced cardiomyocyte IGF1 secretion promotes physiological cardiomyocyte hypertrophy through binding the cellular surface IGF1 receptor. This induces a signaling cascade through phosphoinositide-3-kinase and protein kinase B (PI3K/AKT) which activates mammalian target of rapamycin (mTOR) to promote protein synthesis, permitting cardiomyocyte growth⁷² and contractility^70,73. PI3K/AKT also promote cardiomyocyte survival and functional recovery after ischemic injury in vitro⁷⁴ and in vivo⁷⁵. IGF1 also activates gene expression programs regulating mitochondrial biogenesis and ATP synthesis, as further detailed below. Exercise induced IGF1 may also reduce proinflammatory cytokines, which supports a role in mitigating cardiac inflammation and fibrosis⁶⁸.

Another exercise-induced secreted growth factor is neuregulin 1 (NRG-1). NRG-1 is a member of the epidermal growth factor family that acts through the tyrosine kinase receptor ErbB4 expressed on many cell types including cardiomyocytes⁷⁶. Studies in rodents and humans have demonstrated that acute exercise and exercise training increase circulating NRG-1^76–78. Studies in rodents and endothelial cell models have shown that shear stress, hypoxia or catecholamine activity increases endothelial expression of NRG-1, providing a possible mechanism for the exercise-induced increase of secreted NRG-1. Multiple studies suggested that NRG-1-ErbB receptor signaling was important to the renewal of adult cardiomyocytes^79–81, sparking interest in the notion that NRG-1 may promote the generation of new cardiomyocytes after injury^76,77,82. Additional work has suggested that NRG-1 activity on other cell types may reduce fibrosis and promote cardiac repair. NRG-1 appears to work through the transcriptional co-activator, CITED4 (CBP/p300-interacting transactivator with ED-rich carboxy-terminal domain 4), to promote the cardiomyocyte hypertrophy proportionate in length and width⁸³ that is characteristic of physiological hypertrophy as noted above.

Two stress-induced secreted proteins implicated in cardiac responses to exercise are growth differentiation factor 15 (GDF15)⁵¹ and fibroblast growth factor 21 (FGF21). GDF15 is released from many tissues in response to mitochondrial stress. Systemically, it promotes energy expenditure, decreased food intake, and weight loss in mice and humans in various settings⁸⁴. This includes cachexia due to cancer and heart failure as well as hyperemesis gravidarum, severe nausea and vomiting during pregnancy^85–88. GDF15 expression has also been suggested to increase after acute exercise in untrained and trained humans⁸⁹. GDF15 expression in the heart increases immediately after exhaustive treadmill exercise in mice. The role of GDF15 in cardiomyocytes or other cardiac cells during exercise is currently unclear. Some murine cardiac injury models including transverse aortic constriction (TAC) and myocardial infarction have suggested that cardiac-derived GDF15 may play protective anti-inflammatory and anti-fibrotic roles during acute cardiac injury^90,91. However, other studies of GDF15 elevation during prolonged cardiac stress have suggested that tonic elevation of cardiac derived GDF15 promotes cardiac dysfunction and systemic wasting^87,92. Additional studies will be needed to determine whether exercise-induced GDF15 secretion is adaptive in cardiomyocytes.

FGF21 plays roles in systemic lipid metabolism in response to physiological and pathological stress⁹³. While mostly derived from the liver, FGF21 can be secreted by cardiomyocytes in response to stress, though it remains unclear if exercise induces cardiomyocyte FGF21 secretion specifically⁹³. FGF21 signaling through obligate receptor β-Klotho on cardiomyocytes promotes lipolysis, fatty acid oxidation and mitochondrial ATP production and reduces lipotoxic stress^94,95. Exercise training-induced FGF21-mediated reduction in mitochondrial oxidative stress in a diabetic cardiomyopathy model was abolished with deletion of cardiomyocyte β-Klotho⁹⁴. Other studies have suggested that FGF21 reduces pathological cardiomyocyte hypertrophy in cells and mice⁹⁶. However, other work found that in a diabetic model of diastolic dysfunction, FGF21-FGF receptor 4 signaling promotes concentric hypertrophy⁹⁷. Like GDF15, the specific actions of cardiomyocyte-derived FGF21 in response to exercise on cardiomyocytes and non-cardiomyocytes remain important unanswered areas for study.

Some secreted proteins released or inhibited during exercise training may act locally in the heart. Several cytokines and extracellular matrix remodeling proteins such as interleukin-15, transforming growth factor β, periostin, and others have been suggested to increase differentially with exercise^50,51. Ongoing studies will be needed to identify and functionally investigate possible heart-secreted proteins during acute and chronic exercise in experimental models.

Many investigations of the heart-secreted proteins regulated by exercise have suggested a potential promise as therapeutic targets. For example, augmentation of natriuretic peptide activity through pharmacological inhibition of neprilysin, a natural enzyme that cleaves circulating natriuretic peptides, is one of the mainstays of current therapy for heart failure with preserved or reduced ejection fraction^98,99. It is currently unknown whether phasic elevation of natriuretic peptides in exercise affects cardiomyocyte adaptation in a manner differently to tonic elevation seen in heart failure. Recently a small molecule activating the NRG-1 receptor ErbB4 was identified through a chemical screen for molecules stabilizing the ErbB4 dimer¹⁰⁰. The lead compound EF-1 was shown to reduce cardiomyocyte apoptosis, pathological hypertrophy, and collagen fibrosis¹⁰⁰. This molecule also showed efficacy in reducing cardiac dysfunction in female myocardial infarction (MI) and doxorubicin cardiac injury mouse models. Conversely, one recent study of a model of cardiac cachexia due to genetically mediated ER stress and radiation injury evaluated inhibition of systemic GDF15 signaling⁸⁷. In this model, monoclonal antibody-mediated antagonism of the hypothalamic GDF15 receptor GFRAL improved cardiac function⁸⁷. Additional studies will be required to ascertain the molecular mechanisms underlying this protective effect and role for GDF15 inhibition in other models of cardiac injury.

Intracellular signaling pathways

Exercise imparts characteristic changes to intracellular signaling that enable many of the phenotypic adaptations in cardiomyocytes. Cardiomyocyte growth and physiological hypertrophy is, for example, induced by several parallel and coordinated signaling pathways that promote protein synthesis. Here we summarize many of the known intracellular signaling and growth pathways in cardiomyocytes that are influenced by exercise.

IGF1, PI3K and AKT signaling

As mentioned briefly above, a critical signaling pathway driving physiological hypertrophy is the PI3K/AKT pathway^56–58. This pathway is driven by multiple circulating factors including IGF1 and NRG-1. Plasma levels of IGF1 increase in humans and experimental models after acute or chronic exercise training⁷¹. Exercise-induced IGF1 secretion promotes cardiomyocyte hypertrophy through binding the cell surface IGF1 receptor and downstream PI3K and AKT phosphorylation. These promote growth without fibrosis through increasing protein synthesis through the mTOR pathway, among other adaptations^68,73.

Several of the signaling pathways above are important for the metabolic adaptations that cardiomyocytes make in response to exercise training. IGF1-PI3K-AKT signaling promotes glucose uptake in cardiomyocytes through enhanced GLUT4 expression¹⁰¹. It also promotes mitochondrial biogenesis by AKT-dependent activation of PGC-1α^68,70. This can occur through increasing activation of forkhead box O (FOXO) transcription factors which directly increase PGC-1α gene expression. IGF1-mediated AKT activity through mTOR promotes protein synthesis which may increase ATP consumption. This indirectly raises levels of adenine monophosphate (AMP) and the activity of AMP kinase (AMPK), which promotes PGC-1α activity through phosphorylation. Studies in mice deleted for specific AKT isoforms demonstrate that AKT1 is the isoform required for physiological growth¹⁰², while AKT2 is responsible for most of its metabolic effects in the heart¹⁰³. Exercise-induced calcium cycling in cardiomyocytes can also activate PGC-1α through CaMKII/calcineurin dependent PGC-1α phosphorylation¹⁰⁴. Exercise also promotes anti-oxidant defense in the setting of future ischemic stress through activation of the nuclear deacetylase proteins sirtuins (SIRTs) 1-3^105–109. SIRT1 can activate PGC-1α while SIRT3 increases mitochondrial proteins superoxide dismutase 2 and uncoupling protein 2. These act to further reduce oxidative stress.

The IGF1-PI3K-AKT-MTOR pathway is critical for promoting cardiomyocyte contractility. It does so through increasing the activity of the sarcoendoplasmic reticulum (SR) calcium transport ATPase 2a (SERCA2a), a critical mediator of calcium transport from the cytosol to the sarcoplasmic reticulum that enables efficient calcium cycling and contraction-relaxation cycles^69,110. IGF1-mediated increase in SERCA2a activity also promotes AMPK activity in response to increased cytosolic calcium. The increase in AMPK activity is critical to the efficient uptake and utilization of fatty acids as nutrient sources to support oxidative metabolism in cardiomyocytes.

MAP kinase signaling

Another key signaling mediator of physiological cardiomyocyte hypertrophy is the mitogen-activated protein kinase (MAPK) pathway. This pathway integrates neurohormonal inputs such as β-adrenergic signaling, nitric oxide (NO), and NPR signaling to activate extracellular signal regulated kinases (ERKs), c-Jun N-terminal kinases (JNKs), and p38 MAPKs to promote cell growth without activating stress pathways¹¹¹. ERK1/2 also promote physiological hypertrophy through activation of transcriptional mediators GATA4 and MEF2^112,113. Another important growth pathway is mediated by intracellular calcium signaling. Exercise-induced increases in intracellular calcium activate calmodulin. This promotes the activity of Ca2+/calmodulin dependent kinase II (CaMKII) and calcineurin^39,114,115. Calcineurin dephosphorylates nuclear factor of activated T cells (NFAT) which permits its nuclear entry and action as a transcriptional activator of physiological hypertrophy related genes. Notably chronic activation of this pathway can promote pathological cardiomyocyte hypertrophy^114,116,117.

Hippo/YAP/TAZ pathway

A negative regulator of cardiomyocyte growth suppressed by exercise is the Hippo yes-associated protein (YAP) pathway^118–122. Exercise training in a model of doxorubicin mediated cardiac injury suppressed Hippo kinases (MST1/2 and LATS1/2)¹²². This reduced phosphorylation of YAP and transcriptional coactivator with PDZ-binding motif (TAZ)¹²². De-repression of YAP and TAZ coactivators allow their activation of genes important in cytoskeletal remodeling, protein synthesis and metabolic adaptation¹²³. Hippo-YAP/TAZ signaling may also impact contractility through increasing SERCA2a expression to promote calcium reuptake^124,125. This pathway promotes actin and myosin gene expression to increase sarcomeric proteins for force generation.

Calcineurin signaling

Calcineurin signaling seems to play an important role in cardiomyogenesis. A study in aged mice demonstrated that exercise reverses the global downregulation of cell cycle pathway-associated genes in aged hearts¹²⁶. The authors showed that cardiomyogenesis can be stimulated by exercise also in the aged heart¹¹. A transcriptional comparison between young and aged hearts with or without exercise revealed exercise was associated with genes that cluster in the circadian rhythm biological pathway. The regulator of calcineurin, RCAN1.4, was specifically induced with exercise in aged hearts and was accompanied by reduced calcineurin activity¹¹. RCAN1.4 had previously been reported to play a role in cardiomyogenesis¹²⁷, and further molecular studies revealed that induction of RCAN1.4 in aged cardiomyocytes led to an increase in cell cycle genes¹¹. Interestingly, while the treatment with the thymidylate synthase inhibitor5-FU before swim training did not influence hypertrophy, the hearts were still more vulnerable to injury, suggesting that cell cycle activity in cardiac cells is necessary for the protective effect of exercise¹²⁸. This study also indicated that proliferation in non-CM cells was important in the heart’s response to exercise, such as endothelial cells and fibroblasts. Additional work will be needed to ascertain the signaling mechanisms responsible for those effects.

Other signaling pathways

Aside from these pathways, other key signaling pathways have been implicated in regulation of physiological cardiomyocyte growth. These include some G-protein coupled receptors. One study demonstrated that the inositol 1,4,5-triphosphate (IP₃) receptor type 2 (IP₃R2) was an important mediator of intracellular calcium release during both physiological and pathological hypertrophy¹²⁹. Genetic overexpression of IP₃R2 promoted increased hypertrophy in response to forced swim training or hypertension induced by 2-week infusion of isoproterenol. Protein kinase A (PKA) is a kinase that may activate mTOR under cAMP or sympathetic adrenergic stimulation. PKA activity was shown to be increased in both exercise-induced and pressure-overload hypertrophy¹³⁰. Cardiomyocyte-specific inducible inhibition of PKA activity in mice was shown to prevent both forms of hypertrophy in vivo. Whether context-specific or transient activation of PKA may be beneficial to recapitulate physiological cardiomyocyte hypertrophy in specific settings is not known. PKA activity also may regulate cardiomyocyte contractility. For example, PKA phosphorylation of phospholamban (PLN) prevents inhibition of SERCA2a to enhance calcium re-uptake¹³¹. Additionally it has been shown that endurance exercise induces β-adrenergic activity and promotes PKA dependent phosphorylation of L-type calcium channels to increase intracellular calcium¹³². Other pathways important to optimizing contractility with exercise include the calcium/calmodulin/CaMKII and and PKG signaling pathways. Exercise-mediated increase in calcium flux activates CaMKII which activates targets similarly to PKA¹³³. Activation of PKG through nitric oxide (NO) and natriuretic peptides can improve contractility as well through PLN inhibition to promote SERCA2a activity and improvement calcium sensitivity to promote efficient relaxation¹³⁴.

Signaling pathways and protection from cardiac injury

Many of the same pathways important to growth, metabolic and contractile rewiring in response to exercise also protect remodeled cardiomyocytes from ischemic damage and fibrosis. For example, IGF1 activity suppresses apoptosis through inhibition of proapoptotic proteins Bad and Bax and activation of the anti-apoptotic protein Bcl-2^135–138. Consistent with this, AKT activation mitigated cardiomyocyte death in vitro⁷⁴, and myocardial injury and dysfunction after ischemia-reperfusion in vivo⁷⁵ in mice. All these actions support the physiological expansion of cardiomyocyte size and contractility without pathological cardiac stress under increased energy needs^68,69,137. The AMPK dependent activation of PGC-1α protects against oxidative stress through upregulation of multiple antioxidant genes. Exercise-induced activation of SIRT1 reduces oxidative stress further through PGC-1α activation and the deacetylation of p53 and FOXO proteins which enable cell survival under stress^105,106,139. This pathway also mitigates pro-fibrotic changes by decreasing fibroblast deposition of extracellular matrix through TGF-β/SMAD protein activation^106,139,140. Similar anti-fibrotic and pro-survival pathways are induced by Hippo-YAP/TAZ and cGMP/PKG signaling^57,118,122.

The many ways in which growth programs integrate to promote favorable exercise-induced adaptations in cardiomyocytes offer insights into therapeutic targets based on these pathways. Pharmacological agonism or gene transfer studies of the IGF1-PI3K-AKT pathway suggest that activating this pathway can reduce ischemia-related cardiac dysfunction⁷⁵. This is similarly true for NRG1-related agonism which also acts through PI3K/AKT activation^79,141. Therapeutically targeting the AMPK/PGC-1α pathway in experimental models has also been demonstrated. Metformin, an AMPK activator, has been suggested to reduce damage from ischemia-reperfusion injury in experimental models¹⁴². Similar results have been seen with another AMPK agonist AICAR¹⁴³. Nitric oxide and cGMP/PKG mediated therapies such as phosphodiesterase inhibitor sildenafil^144,145 and the soluble guanylate cyclase activator, vericiguat¹⁴⁶, are used as treatments to promote pulmonary vasodilation in heart failure¹⁴⁷. These also have anti-fibrotic effects. Of course, a critical challenge inherent in all these therapies is attaining tissue specificity and minimizing off-target effects given the widespread expression and pleiotropic function of these molecules.

Transcriptional responses to exercise

Numerous transcription factors are induced by endurance training in animal models. As touched upon above, many of these converge on the regulation of nutrient utilization and oxidative metabolism. A central transcriptional pathway is that coordinated by transcriptional coactivator peroxisome proliferator-activated receptor PGC-1α^53,148,149. PGC-1α activates many metabolic gene programs across tissues. PGC-1α was initially identified as a cold-inducible protein that regulates mitochondrial biogenesis in brown adipose tissue¹⁵⁰. Subsequent studies demonstrated that PGC-1α in skeletal muscle was induced by exercise acutely and with training¹⁵¹. This work also showed that PGC-1α promotes multiple endurance-related adaptations in muscle including mitochondrial biogenesis, fiber type switching, and neuromuscular junction remodeling. The heart expresses abundant PGC-1α whose expression increases with acute exercise. PGC-1α governs many diverse gene programs by recruiting various DNA-binding transcription factors such as PPARα. PPARβ, PPARδ, estrogen-related receptors (ERRs), nuclear respiratory factors (NRFs), and others to activate specific genes^53,148,149. PGC-1α recruitment of PPARα is important in upregulating mitochondrial fatty acid oxidation genes in the heart^28,152–154. Notably, expression of both PGC-1α and PPARα is upregulated by endurance training and downregulated in heart failure murine models and human failing hearts^28,155. PGC-1α regulates other components of mitochondrial metabolism through other DNA-binding factors. For example, it recruits ERRs (e.g. ERRα) and yin-yang 1 (YY1) to promote expression of most genes involved in oxidative phosphorylation including the tricarboxylic acid cycle, electron transport chain proteins and mitochondrial ATP synthesis enzymes^28,156–158. PGC-1α recruitment of NRFs promotes expression of mitochondrial transcription factor A (TFAM) which increases mitochondrial DNA replication, a critical requirement for mitochondrial biogenesis^28,159. PGC-1α also cooperates with myocyte enhancer factor 2c (MEF2c) to promote fatty acid and glucose oxidation after exercise¹⁶⁰.

PGC-1α loss-of-function either globally or specifically in cardiomyocytes reduces cardiac metabolic gene expression, mitochondrial biogenesis, and bioenergetic capacity in mouse models of cardiac dysfunction^53,161–165. These effects have been recapitulated in mouse knockout models of DNA-binding partners of PGC-1α^28,152,158. This includes models of pressure-overload cardiac hypertrophy, peripartum cardiomyopathy, and ischemia/reperfusion injury. However, the requirement of PGC-1α or key DNA-binding partner transcription factors such as ERRα for the known cardiomyocyte adaptations to exercise training, a physiological stressor distinct from the above pathological insults, has not yet been demonstrated.

PGC-1α and other transcription factors modulate exercise-induced cardiomyocyte growth. A quantitative PCR-based transcriptomic (Quanttrx) screen for transcriptional pathways differentially induced by exercise vs. pressure-overload cardiac hypertrophy in mice found several factors differentially regulated by these two stimuli²¹. This included the known regulators of cardiomyocyte growth GATA4 and MEF2c. This approach also first implicated CITED4 as a contributor to physiological cardiomyocyte hypertrophy and cardiomyocyte proliferation. CITED4 expression increased in hearts of mice after 14 days of swim training. In cultured cardiomyocytes, CITED4 overexpression promoted cardiomyocyte expansion and hypertrophy. In vivo studies corroborated these findings, as CITED4 overexpression in the mouse heart promoted physiological cardiomyocyte hypertrophy through mammalian target of rapamycin (mTOR) induced growth¹⁶⁶. This also protected mice from ischemia-reperfusion injury. Subsequent work showed that deficiency of cardiomyocyte CITED4 in cardiomyocytes caused dilated cardiomyopathy and impaired tolerance in mice after exercise training²³. These mice also showed increased fibrosis, mitochondrial dysfunction and cardiomyocyte apoptosis after transverse aortic constriction (TAC), suggesting that exercise-induced CITED4 may block this pathology. CITED4 was shown to regulate the expression of multiple microRNAs important for cardiomyocyte-fibroblast crosstalk. The loss of microRNA 30d expression due to CITED4 deficiency promoted cardiac fibrosis. Subsequent work coupling diffusion tensor magnetic resonance imaging (DT-MRI) of the mouse heart to RNA fluorescence in-situ hybridization (RNA-FISH) showed that CITED4 expression contributed significantly to the increased microstructural helicity of cardiomyocytes after exercise¹⁶⁷. This suggests a unique and region-specific contribution of CITED4 to exercise-induced adaptation of cardiomyocytes.

Exercise training may also suppress certain transcriptional programs such as those that normally prevent cell growth and proliferation. An instructive example of this is the CCAAT/enhancer-binding protein beta (CEBPβ) transcription factor. CEBPβ was identified in the Quanttrx screen of exercise training vs. TAC-induced hypertrophy as downregulated with acute exercise and endurance training²¹. Experiments in cultured cardiomyocytes showed that inhibiting CEBPβ promoted an increase in cardiomyocyte proliferation, cardiomyocyte numbers and size. Further experiments showed that CEBPβ exerted its actions through inhibiting the transcription factor serum-response factor (SRF). SRF normally upregulates GATA4 and α-MHC expression. Heterozygous loss-of-function of CEBPβ promoted physiological cardiomyocyte hypertrophy, cardiomyocyte proliferation and protection from TAC-induced dysfunction. Importantly, downregulation of CEBPβ promoted expression of CITED4, leading to its discovery as a positive regulator of cardiomyocyte proliferation as described above. Another key example is exercise-induced suppression of Hippo, which permits the activity of YAP and TAZ transactivators to promote expression of genes related to cardiomyocyte hypertrophy, mitochondrial oxidative metabolism, and reduction of apoptosis.

Several of these transcriptional pathways also regulate cardiac contractility in response to exercise. For example, MEF2c promotes expression of sarcomeric proteins troponin I and α-MHC and promotes SERCA2a expression, which improves efficiency of contractile machinery and calcium recycling^168–170. GATA4 induced by exercise similarly promotes α-MHC expression and increases connexin 43 expression^169,171,172. The latter promotes electrical signal transmission to enhance contractility.

Activation of other transcription factors may also protect cardiomyocytes from ischemic or fibrotic injury in addition to the ones above. For example, another transcriptional profiling study of exercise vs. TAC induced hypertrophy in mice demonstrated an upregulation of heat shock factor 1 (HSF1) with exercise training¹⁷³. Deficiency of HSF1 promoted cardiac dysfunction and exercise tolerance while constitutive activation of HSF1 reduced cardiomyocyte apoptosis in response to prolonged TAC. One mechanism suggested for these effects was that upregulation of heat shock proteins such as HSP70 by HSF1 transcriptional activity may allow proper protein folding during the repeated stress of endurance training. Another example studied activation of the PPAR transcription factor PPARβ¹⁷⁴. Pharmacological ligand activation of PPARβ conferred cardiomyocyte length enlargement, cardiac mass and angiogenic markers in just 96 hours without inducing expression of stress markers. These adaptations were comparable to those seen in mice undergoing voluntary wheel running training for 5 weeks. Cyclic adenine mononucleotide (cAMP) response element binding protein (CREB) is a transcription factor activated with endurance exercise training in rats and mice¹⁷⁵. CREB activation through phosphorylation was shown to be induced by treadmill-based exercise training in rats for 10 days. This correlated with increased heart size and increased expression of PGC-1α, some PGC-1α target genes and the anti-apoptotic protein Bcl2. These studies suggest a correlation of these transcription factors with endurance training in mice. For many of the transcription factors, additional work will be needed to demonstrate that these pathways are required for exercise-induced cardiomyocyte adaptations.

Attempts have been made to therapeutically augment the above transcription factors in experimental models. Activation of exercise-induced transcription factors exerts cardioprotection, in part by promoting mitochondrial function and cardiomyocyte proliferation, in different injury models. However, targeting transcriptional pathways is challenging for several reasons. Pharmacological activation or inactivation of transcription factor proteins requires delivering molecules through both cell and nuclear membranes. This is difficult to do efficiently. Additionally, transcription factors are frequently widely expressed, making it problematic to deliver pharmacological agents such as small molecules in a tissue-specific manner.

One early study suggested the potential efficacy of viral gene delivery to promote cardiac repair near infarcts in murine models. A more recent study of adeno-associated virus (AAV) vector as a gene delivery tool for the candidate transcription factor CITED4 suggested the potential promise of gene therapy for delivering beneficial transcription factors in cardiomyocytes to correct dysfunction¹⁷⁶. Recombinant AAV has been a tool to somatically overexpress transgenes for gene therapy in experimental models and some human diseases. AAV serotype 9 has a relatively high tropism for cardiomyocytes. When coupled with a cardiomyocyte-selective promoter to drive transgene expression (for example cardiac troponin), this allows for cardiomyocyte-specific gene overexpression. CITED4 delivery to cardiomyocytes via AAV serotype 9 mediated expression promoted physiological hypertrophy without adverse remodeling in wild-type mice. This approach also demonstrated efficacy in reducing ischemia-reperfusion injury in mice overexpressing CITED4. This study provides proof-of-concept for gene delivery of transcription factors that mediate the adaptive benefits of exercise in cardiomyocytes.

Several exercise-induced transcription factors such as PPARα and PGC-1α are downregulated in aged cardiomyocytes. This motivates efforts to activate these pathways to prevent some of the age-related cardiac dysfunction which exercise appears to mitigate. ERR transcription factors are among the key DNA-binding factors enabling the PGC-1α-induced adaptations to exercise training. Recently two pan-ERR transcription factor agonists were shown to prevent pathological remodeling and promote mitochondrial fatty acid oxidation in mouse hearts after TAC induced pressure overload¹⁷⁷. Future studies are needed to fully evaluate the potential of these pharmacological agents in stimulating physiological cardiomyocyte hypertrophy and other exercise adaptations in cardiomyocytes in aging and other disease models.

Post-transcriptional programs in cardiomyocytes

MicroRNAs

MicroRNAs (miRNAs) are a class of noncoding RNAs, typically 18-22 nucleotides in length, that exhibit high conservation. miRNAs are generally transcribed as primary transcripts by RNA polymerase II and processed into mature miRNAs by the endonucleases Dicer and Drosha¹⁷⁸. Drosha cleaves the primary transcript in the nucleus to produce a precursor miRNA (pri-miRNA), which is further processed by Dicer in the cytoplasm to generate mature miRNAs¹⁷⁸. miRNAs typically bind to the 3’-untranslated regions of target mRNAs, leading to transcript degradation or translational repression^179,180. However, miRNAs can also target the 5’-untranslated regions, coding regions, or promoters of target mRNAs to activate gene expression^179,180. miRNAs are the most widely studied noncoding RNAs in the regulation of cardiac function^181,182. including the cardiac response to exercise^181,182.

miRNAs in the heart and circulation are dynamically regulated in response to exercise and have been demonstrated to play functional roles in exercise-induced cardiac adaptations. Liu and colleagues profiled cardiac miRNAs in two distinct mouse exercise models: voluntary wheel running and ramp swimming exercise¹⁸³. They identified miRNA-222, whose expression was upregulated, predominantly in cardiomyocytes, in the heart in both mouse exercise models¹⁸³. Cardiomyocyte-specific miR-222 overexpression mimicked exercise-induced physiological cardiac growth, including physiological cardiac hypertrophy and cardiomyogenesis¹⁸³. Inhibition of miR-222 using locked nucleic acid (LNA)-enhanced antisense oligonucleotides not only canceled exercise-induced cardiac hypertrophy but also completely blocked exercise-induced cardiomyogenesis¹² in adult mouse hearts, which was measured by multi-isotope imaging mass spectrometry (MIMS)¹². These suggest that miR-222 is necessary for exercise-mediated cardiac growth. Importantly, further studies showed that cardiomyocyte-specific miR-222 overexpression protected the heart against stresses such as myocardial ischemia-reperfusion injury^12,183 and pressure overload-induced pathological hypertrophy and heart failure¹⁸⁴. These findings highlight the notion that virtually every pathway identified as functionally important in the heart’s response to exercise also protects the heart against pathological stress when mimicking changes observed during exercise^21,166,176. Interestingly, it appears that the molecular mechanisms governing miR-222’s function in the heart are context-dependent. In exercise-induced physiological cardiac growth and myocardial ischemia-reperfusion, miR-222 conferred its effects mainly through p27, HMBOX1, and HIP1/2¹⁸³. However, in pressure overload-induced heart failure, p27 and HIP1/2 were not affected by miR-222, which exerted its effects primarily through NFATc3, PUMA, and HMBOX1¹⁸⁴. In addition to miR-222, other miRNAs have been associated with exercise-induced cardiac growth, such as miR-1^185,186 and miR-133a¹⁸⁵. However, further functional studies with either gain-of-function or loss-of-function approaches are needed to examine their roles in the heart’s response to exercise.

Interestingly, miR-222 has been reported to increase in the peripheral blood of healthy young athletes after 90 days of organized, team-based rowing training¹⁸⁷. Similarly, circulating miR-222 was elevated in heart failure patients who underwent a symptom-limited cardiopulmonary exercise test on a bicycle ergometer using a standardized exercise protocol¹⁸³. This highlights the potential clinical relevance of miR-222 as a biomarker and therapeutic target for heart disease and suggests that exercise may stimulate the secretion of miRNAs from organs, other than the heart, to exert effects on the heart. In mice, four weeks of swimming training protected the heart against myocardial ischemia-reperfusion injury, which was associated with the increased release of extracellular vesicles from brown adipose tissue enriched with miR-125b-5p, miR-128-3p, and miR-30d-5p¹⁸⁸. Suppression of extracellular vesicle release by injecting AAV9 encoding shRNA targeting Rab27a (a GTPase required for small extracellular vesicle secretion) or specific inhibition of these miRNAs in brown adipose tissues attenuated the cardioprotection conferred by exercise¹⁸⁸. This indicates that miRNAs mediate inter-organ communication in exercise-induced cardioprotection. In healthy young adults performing organized, team-based rowing training for over one year and in rats subjected to six weeks of swimming exercise, miR-342-5p was elevated in exosomes isolated from plasma after training¹⁸⁹. Injection of these exosomes into mice subjected to myocardial ischemia-reperfusion improved post-ischemic cardiac function and protected against cardiomyocyte death¹⁸⁹. In contrast, the cardioprotective effects of exercise were abolished by inhibition of miR-342-5p¹⁸⁹. indicating that exercise confers cardioprotection through circulating miR-342-5p. However, it remains unclear which organ is the primary source of miR-342-5p in response to exercise in this study. Likely, multiple organs contribute, as pri- and mature-miR-342-5p were increased in the heart, aorta, liver, skeletal muscle, and kidney after exercise¹⁸⁹. These findings underscore the critical roles of miRNAs in mediating inter-organ communication during exercise-induced cardiac adaptation and their high potential as biomarkers and therapeutic targets for heart disease. Indeed, inhibition of miRNAs using antisense oligonucleotides has proven effective in preclinical large animal models¹⁹⁰ and clinical trial^191–193.

Long Noncoding RNAs

Long noncoding RNAs (lncRNAs) are RNA molecules exceeding 200 nucleotides in length. Transcribed by RNA polymerase II, lncRNAs constitute the majority of the noncoding transcriptome¹⁹⁴. lncRNAs are expressed at lower levels relative to protein-coding genes and exhibit limited sequence conservation across species¹⁹⁵. lncRNAs exert their functions through various mechanisms, including sequestration of miRNAs (acting as miRNA sponges), direct binding to proteins and DNA, or serving as scaffolds to recruit transcriptional repressors or activators^194,195. The roles of lncRNAs in the heart have been documented in cardiac development and disease. lncRNAs have also proven functionally important in the development and progression of cardiac diseases^196–200.

Despite the well-documented roles of lncRNAs in regulating heart diseases, their roles in cardiac adaptation in response to exercise are just emerging. Gao et al. profiled lncRNAs in the hearts of mice subjected to eight weeks of swimming exercise²⁰¹. They identified cardiac physiological hypertrophy associated regulator (CPhar), whose expression in cardiomyocytes, but not fibroblasts, in the heart was upregulated by exercise²⁰¹. Inhibition of CPhar using AAV9-shRNA inhibited exercise-induced cardiac hypertrophy and markers of cardiomyocyte proliferation, suggesting its necessity for exercise-induced physiological cardiac growth. Cardiomyocyte-specific CPhar overexpression protected against, while its inhibition exacerbated, myocardial ischemia-induced injury²⁰¹. Mechanistically, CPhar mediated exercise-induced physiological cardiac growth and protected against injury by binding to DDX17 to downregulate ATF7 through sequestering C/EBPβ, the negative regulator of exercise-induced physiological cardiac growth discussed above²¹. Mice undergoing exercise hypertrophic preconditioning (three weeks of swimming training, terminated one week before inducing heart failure) exhibited improved cardiac function and reduced fibrosis after transverse aortic constriction-induced heart failure²⁰². These cardioprotective effects were attenuated by silencing lncRNA Mhrt779 (an antisense RNA of Myh7) while enhanced by Mhrt779 overexpression²⁰², indicating an essential role for Mhrt779 in exercise cardioprotection. In mice subjected to eight weeks of voluntary wheel running, Li et al identified a series of lncRNAs dynamically regulated by exercise in the heart, called long noncoding exercise-associated cardiac transcripts (lncExACTs)²². These lncExACTs were also altered in pathological cardiac hypertrophy and heart failure²². Interestingly, in every case, their changes during exercise were opposite to those observed in the disease models²², highlighting the distinct nature of these responses despite superficial similarities between cardiac growth induced by exercise and pathological stress. Among these lncExACTs, lncExACT1 was highly conserved across species¹⁰. Like observations in animal models, lncExACT1 expression was elevated in the heart and plasma of patients with heart failure²², underscoring its clinical translational potential. Inhibition of lncExACT1 using antisense oligonucleotides, mimicked its changes seen in exercise, and recapitulated many exercise-induced cardiac phenotypes, including physiological cardiac hypertrophy and markers of cardiomyogenesis²². Mechanistically, lncExACT1 exerted its regulatory effects on the heart by promoting DCHS2 transcription and binding to miR-222, a miRNA previously shown to be necessary for exercise-induced physiological cardiac growth¹⁸³. Importantly, inhibition of lncExACT1 improved cardiac function, increased markers of cardiomyocyte proliferation, and reduced fibrosis in two distinct heart failure models²². This included the delivery of lncExACT1 inhibition at the time of reperfusion, a clinically relevant timeframe, in a mouse model of myocardial ischemia-reperfusion²². These findings reinforce the idea that virtually every pathway identified as functionally important in the heart’s response to exercise also protects the heart against pathological stress when mimicking exercise-induced changes.

Circular RNAs

Circular RNAs (circRNAs) are a subclass of lncRNAs that form closed loops²⁰³, which have also been implicated in cardiac exercise adaptation. Sequencing of circRNAs in cardiomyocytes isolated from mice after four weeks of swimming exercise identified circUtrn²⁰⁴, a circRNA derived from exons 60-64 of the Utrn gene. In the heart, circUtrn was upregulated by exercise but downregulated under various stresses, including myocardial infarction, myocardial ischemia-reperfusion, transverse aortic constriction-induced pressure overload, and doxorubicin-induced cardiotoxicity²⁰⁴. Inhibition of circUtrn using AAV9-shRNA attenuated exercise-induced cardiac hypertrophy and markers of cardiomyocyte proliferation²⁰⁴, suggesting that circUtrn is required for exercise-induced physiological cardiac growth. Cardiomyocyte-specific overexpression of circUtrn improved cardiac function, reduced pathological hypertrophy, and decreased cardiac fibrosis in mice subjected to myocardial ischemia-reperfusion²⁰⁴. In contrast, circUtrn inhibition abolished exercise’s cardioprotective effects against myocardial ischemia-reperfusion injury²⁰⁴.

Circ-Ddx60, another circRNA derived from the Ddx60 gene, has also been shown to contribute to the cardiac benefits of exercise²⁰⁵. Circ-Ddx60 was increased in the heart one week after mice completed three weeks of swimming (exercise hypertrophic preconditioning)²⁰⁵. Inhibition of circ-Ddx60 attenuated exercise’s hypertrophic preconditioning-mediated anti-pathological cardiac hypertrophy and heart failure via eEF2²⁰⁵. These findings suggest a role for circRNAs in the cardiac benefits of exercise. As newly discovered noncoding RNAs, the functions of circRNAs in the heart and their roles in the cardiac response to exercise remain largely unexplored and deserve further investigation.

It is important to note that targeting noncoding RNAs presents unique challenges: (1) noncoding RNAs are typically expressed at relatively low levels, which may limit the efficacy of antisense oligonucleotide-mediated RNA inhibition. This could be partially addressed by conjugating antisense oligonucleotides with cell-type-specific sensors; (2) noncoding RNAs other than miRNAs exhibit low species conservation, necessitating careful selection of candidate noncoding RNAs for therapeutic targeting; (3) noncoding RNAs, particularly lncRNAs, confer their effects through a wide range of mechanisms. Targeting a specific activity of a noncoding RNA may lead to unanticipated off-target effects, requiring further investigation. Despite these challenges, research has shown promise in targeting some noncoding RNAs to treat cardiac diseases, particularly through antisense oligonucleotide-mediated noncoding RNA degradation and cleavage.

RNA Modifications

Eukaryotic RNA chemical modifications, initially identified on tRNAs and rRNAs, are now increasingly observed on other RNA species, including mRNAs, miRNAs, and lncRNAs²⁰⁶. More than 100 internal modifications have been identified on mRNAs and noncoding RNAs²⁰⁶. Among these, N6-methyladenosine (m6A) is the most abundant and extensively studied internal RNA modification, regulating RNA stability, protein-binding capability, and interactions with other RNAs²⁰⁶. The role of m6A in the cardiac response to exercise is just beginning to emerge. Wang et al. found that after four weeks of swimming exercise, the total cardiac m6A level decreased that was associated with downregulation of cardiac METTL14²⁰⁷, an m6A methyltransferase (i.e., m6A writer that deposits m6A modifications on target RNAs). Cardiomyocyte-specific overexpression of METTL14 increased cardiac total m6A and canceled exercise-induced cardiac hypertrophy and markers of proliferation²⁰⁷. These changes were blunted by inactivation of METTL14’s methyltransferase activity²⁰⁷. Inhibition of cardiomyocyte METTL14 reduced m6A deposition on Phlpp2 mRNA, thereby activating Akt phosphorylation, decreasing cardiomyocyte death, and reducing post-ischemic infarction, ultimately improving post-ischemic cardiac function and remodeling²⁰⁷. Using the same exercise model, the same group identified YTHDF2, an m6A-binding protein (i.e., m6A reader that recognizes m6A modifications on RNAs), which was downregulated in exercised hearts but upregulated in hearts after myocardial ischemia-reperfusion²⁰⁸. YTHDF2 overexpression abolished exercise-induced cardiac growth, while its inhibition attenuated post-ischemic myocardial injury and fibrosis²⁰⁸. These findings suggest that YTHDF2 suppression is required for exercise-induced cardiac growth. Interestingly, YTHDF2 conferred these effects through G3BP1 and appeared independent of its m6A regulatory property²⁰⁸. These studies provide early evidence of the involvement of RNA modifications, such as m6A, in cardiac adaptation to exercise. Given that RNA modifications are dynamic processes, including modification deposition and removal, future studies are needed to uncover the full picture of their roles in the heart, particularly their involvement in the cardiac response to exercise.

Outstanding questions and future directions

The pathways described above underscore the incredible mechanistic diversity by which cardiomyocytes respond to the adaptive stress of exercise training. Exercise-induced pathways show remarkable consistency in protecting the heart from pathological injury across different model systems. Therapeutically targeting the pathways offers great promise in treating cardiac disease and dysfunction (Table 2). However, several important questions remain in how to strategically manipulate these molecular adaptations.

Table 2:

Potential therapeutic targets inspired from exercise-induced molecular adaptations in cardiomyocytes

Molecular Mechanism	Approach	Phenotype in preclinical model(s)	Advantages	Disadvantages	References
Heart-secreted mediators of cell-cell communication	Natriuretic peptide agonism (Sacubitril-Valsartan: neprilysin treatment in conjunction with angiotensin receptor blocker)	Favorable remodeling, improved contractility, survival benefit in humans with HF of multiple etiologies	Oral medication, FDA approved for HF in humans as first-line therapy	Potentially limiting anti-hypertensive actions in some patients; cardiomyocyte specific effects not completely elucidated	61,64–66,223
	GDF15 antagonism (GFRAL receptor neutralizing monoclonal antibody)	Improved cardiac function, reduced adipose and muscle wasting in mouse model of HF due to ER stress and radiation exposure	Human monoclonal antibody with demonstrated efficacy for cancer cachexia	Unclear effects on cardiomyocyte/cardiac adaptations to exercise; repeat dosing required	87
	NRG-1 mimetic (ErB4 activating small molecule)	Improved cardiac function and reduced fibrosis in mouse models of MI and doxorubicin cardiotoxicity	Increased tissue specificity due to receptor specificity	Limited effect in cardiac injury models using male mice; need for frequent dosing	100
Intracellular signaling	IGF-1/PI3K/Akt Pathway Activation (IGF1 tethered nanofiber)	Improved cardiac function, cardiomyogenesis, increased capillary density, and reduced apoptosis in rat model of MI	Local activity at site of cardiac injury; cardiomyogenic potential	Administration only studied at infarct zone; Unclear long-term effects	224
Transcriptional regulation	CITED4 gene transfer (intravenous delivery of AAV9-cTnT-CITED4 in mice)	Physiological cardiomyocyte hypertrophy; Reduced cardiomyocyte apoptosis, inflammation, fibrosis, and dysfunction in IRI model	Cardiomyocyte specific delivery; long-term gene expression; Low immunogenicity	Dose-dependent off-target effects; Possible time-dependent reduction in transgene expression; Cost	176
	ERR activation (pan-ERR agonists SLU-PP-32 and SLU-PP-915)	Improved mitochondrial function, FA oxidation, cardiac function in mice with TAC induced HF; No improvement in pathological hypertrophy	Potential effects in other ERR-responsive organs (skeletal muscle); Systemic delivery without genetic modifications	Limited efficacy in preclinical models for adaptations beyond mitochondrial function; Potential off-target effects across tissues; Unclear dose scaling and frequency for efficacy in larger animal models	177,225
	PGC-1a activation through AMPK or Sirtuin activation (AICAR, Resveratrol, SRT2104)	Improved cardiac contractility, mitochondrial function, cardiac function, and decreased oxidative stress in aged mice (26-28 months); improved cardiac function and mitochondrial function in diabetic cardiomyopathy mouse model	Reduced ischemia-reperfusion and improved anti-oxidant activity	Risk of chronic activation-induced catabolism, cardiac and systemic dysfunction; limited tissue specificity; unclear long-term safety; need for continuous dosing	226–229
	YAP/TAZ activation (Hippo inhibitor XMU-MP-1)	Improved cardiac function and pathological hypertrophy, reduced cardiomyocyte apoptosis and fibrosis in mice 3 weeks after TAC	Non-genetic manipulation	Potential off-target increase in extra-cardiac cell proliferation (liver, small intestinal epithelium), lack of long-term safety evaluation, need for continuous dosing	230
Post-transcriptional regulation	lncExACT1 inhibitor (locked nucleic acid [LNA]-GapmeR)	Increased physiological hypertrophy, cardiomyogenesis, improved fibrosis, improved cardiac function in mice	High specificity of target, enhanced stability through LNA modifications including nuclease resistance, potential to target otherwise difficult-to-drug pathway (ncRNAs)	Unclear efficiency of cardiac vs. extra-cardiac uptake, potential off-target tissue effects, need for continuous dosing	22

Cardioprotective mechanisms of exercise were often demonstrated in ischemia and pressure-overload models of cardiac injury in rodents. There is a rising prevalence and unmet need for treatment of other forms of heart failure, such as peripartum cardiomyopathy, age-related cardiac dysfunction, and other forms of heart failure with preserved ejection fraction. It will thus be important to evaluate any benefit of these mechanisms in a broader range of cardiac disease.

Cardiomyocyte adaptations to exercise may convey different phenotypes depending on specific stages disease. Currently, the recommendation for exercise is a mainstay for reduction of both incidence and complications of most cardiovascular illnesses. However, it is not well understood where the benefits of exercise precisely lie on the spectrum of health: as preventative before any signs of cardiac disease versus restorative after disease onset. It is even conceivable, that some effects such as cardiomyocyte growth and augmentation of energy utilization could be detrimental in states of severe cardiac dysfunction. This is especially relevant to end-stage heart failure which is frequently associated with cardiac cachexia. Exercise-induced mechanisms that promote extra-cardiac energy utilization in those settings may complement the decreased energetic needs of a failing and wasting myocardium.

Aside from this, some studies have shed light on some potentially non-adaptive responses to chronic exercise training in experimental models and humans. For example, one study in rats subjected to 16 weeks of forced wheel running suggested that extended endurance training resulted in maladaptive right ventricular and left atrial remodeling including fibrosis and arrhythmias²⁰⁹. The impact of extended exercise on right ventricular remodeling has also been observed in highly trained human athletes²¹⁰. The longitudinal impact of these changes and relevance to moderate endurance training in humans is unclear. Other studies have suggested that longitudinal exercise training is associated with increased coronary artery calcification (CAC), a surrogate measure of atherosclerosis^211,212. Importantly, these studies did not identify any increase in mortality or cardiovascular events in highly trained subjects despite increased CAC, suggesting favorable atherosclerotic plaque characteristics despite the increased presence. These and other findings highlight the importance of understanding the specific molecular, cell, and tissue changes with exercise along with the type of exercise training and underlying cardiac substrate to fully appreciate the potential adaptive and non-adaptive changes associated with endurance training in the heart.

Emerging studies incorporating complementary approaches to further discover and detail exercise-induced pathways will undoubtedly help delineate those that have the most potential for therapy. These include integrated genomic, transcriptomic, proteomic, and epigenomic interrogations of endurance exercise effects in model systems such as rats by groups like the Molecular Transducers of Physical Activity Consortium (MoTrPAC)²¹³ and others. Such efforts will identify novel molecular pathways specific to cardiomyocytes or other cardiac cells that mediate the benefits of exercise. These studies will also help elucidate the complex mechanistic orchestration of exercise responses temporally, spatially and systemically across cell and tissues types. Nomination of specific analytes or regulatory programs from these and other integrative studies will help provide initial biomarkers for better understanding the cardiac exercise response. These will then require rigorous individual mechanistic interrogation through gain- and loss-of-function approaches in experimental models (e.g. rodents and cultured cells) related to exercise and cardiac injury to test individual importance and causality to the exercise-induced responses. Further support for new biomarkers and causal contributors to specific cardiomyocyte exercise responses may then come from consistent observations from human studies of exercise or heart failure that implicate the same or related analytes or pathways. This will provide further basis for considering new therapeutic efforts focused on a new molecular pathway.

Importantly, ongoing efforts may help resolve differences in exercise adaptations and therapeutic implications between biological sexes. A recent, impactful prospective study of >400,000 US adults assessed leisure-time physical activity and all-cause cardiovascular mortality over approximately 22 years. Strikingly, this study found that whereas men achieved their maximum survival benefit from exercising ~300 minutes/week, women achieved a similar benefit from only ~140 minutes/week²¹⁴. This underscores the important sex-specific differences in cardiovascular adaptations to both physiological and pathological stress. Integrated tissue and molecular phenotyping of men and women in both healthy and diseased states will hopefully aid in the identification of sex-specific biomarkers and pathways related to exercise. This may further sub-phenotype cardiovascular health and increase the specificity of recommendations across sexes. Such specificity may create phenotype-specific exercise recommendations or molecular pathway-inspired treatment targets to further optimize cardiovascular health across a wide range of conditions.

It will also be important to contextualize studies in humans with those from experimental models. The MoTrPAC Study Group multi-omics interrogation of exercise responses across tissues recently suggested a sex-consistent response to exercise training in increasing analytes related to mitochondrial biogenesis, translation, and cellular response to heat stress by 8 weeks of endurance training in rats²¹⁵. Similar findings were observed for the cardiac mitochondrial acetylome for both sexes²¹⁶. Additional similar investigations of exercise-responsive adaptations in the heart across sexes and in different disease states or forms of exercise may help further identify sex-specific adaptations in the heart.

In summary, multiple complementary molecular mechanisms transmit the behavioral events of acute or chronic exercise to physiological adaptations in the heart. Several of these mechanisms may have pleiotropic effects and provide benefits across cardiac cell types and in extra-cardiac tissues. Ongoing and future efforts to characterize these molecular adaptations and establish those most relevant to human exercise-observed benefits will hopefully lead to new therapeutic targets and treatments for cardio-metabolic diseases.

Nonstandard abbreviations and acronyms

SERCA

Sarco/endoplasmic reticulum Ca2+ ATPase

PGC-1α

peroxisome proliferator activated receptor gamma coactivator 1α

lncRNA

long noncoding RNA

miRNA

microRNA

GPCR

G-protein coupled receptor

NPR

natriuretic peptide receptor

cGMP-PKG

cyclic guanidine monophosphate-protein kinase G

IGF1

insulin-like growth factor 1

PI3K

Phosphoinositide-3-kinase

AKT

protein kinase B

CITED4

CBP/p300-interacting transactivator with ED-rich carboxy-terminal domain 4

Erb-B2

Receptor Tyrosine Kinase 4

GDF15

growth differentiation factor 15

FGF21

fibroblast growth factor 21

TAC

transverse aortic constriction

MAPK

mitogen-activated protein kinase

NO

nitric oxide

ERK

extracellular signal regulated kinase

JNK

c-Jun N-terminal kinase

GFRAL

GDNF receptor-alpha-like

MEF2

myocyte enhancer factor-2

GATA4

GATA binding protein 4

MST1/2

STE20-like serine/threonine-protein kinase 1/2

LATS1/2

large tumor suppressor 1/2

YAP

Hippo yes-associated protein

TAZ

transcriptional coactivator with PDZ-binding motif

IP₃

inositol 1,4,5-triphosphate

IP₃R2

receptor type 2

PKA

Protein kinase A

mTOR

mammalian target of rapamycin

FOXO1

forkhead box O

SIRT

sirtuin

PLN

phospholamban

PPAR

peroxisome proliferator activated receptor

ERR

estrogen-related receptor

NRF

nuclear respiratory factor

TFAM

mitochondrial transcription factor A

CEBPβ

CCAAT/enhancer-binding protein beta

SRF

serum-response factor

HSF1

heat shock factor 1

cAMP

cyclic adenine mononucleotide

CREB

response element binding protein

AAV

adeno-associated virus

MIMS

multi-isotope imaging mass spectrometry

CPhar

cardiac physiological hypertrophy associated regulator

lncExACT

long noncoding exercise-associated cardiac transcripts

circRNA

Circular RNA

MoTrPAC

Molecular Transducers of Physical Activity Consortium

BCAA

branched chain amino acid