The dynamic interplay between cardiac mitochondrial health and myocardial structural remodeling in metabolic heart disease, aging, and heart failure

This review provides a holistic perspective on the bi-directional relationship between cardiac mitochondrial dysfunction and myocardial structural remodeling in the context of metabolic heart disease, natural cardiac aging, and heart failure. First, a review of the physiologic and molecular drivers of cardiac mitochondrial dysfunction across a range of increasingly prevalent conditions such as metabolic syndrome and cardiac aging is presented, followed by a general review of the mechanisms of mitochondrial quality control (QC) in the heart. Several important mechanisms by which cardiac mitochondrial dysfunction triggers or contributes to structural remodeling of the heart are discussed: accumulated metabolic byproducts, oxidative damage, impaired mitochondrial QC, and mitochondrial-mediated cell death identified as substantial mechanistic contributors to cardiac structural remodeling such as hypertrophy and myocardial fibrosis. Subsequently, the less studied but nevertheless important reverse relationship is explored: the mechanisms by which cardiac structural remodeling feeds back to further alter mitochondrial bioenergetic function. We then provide a condensed pathogenesis of several increasingly important clinical conditions in which these relationships are central: diabetic cardiomyopathy, age-associated declines in cardiac function, and the progression to heart failure, with or without preserved ejection fraction. Finally, we identify promising therapeutic opportunities targeting mitochondrial function in these conditions.

Cardiac aging is an intrinsic process that results in cellular and molecular changes that impair cardiac function. Due to the high energy demand of the heart, it is not surprising that age-related mitochondria defects are associated with diminished cardiac function [ 55 ] . Many factors contribute to the reduced energetic capacity of cardiac mitochondria during aging, including mutations and deletions in the mitochondria genome, increased ROS production, inflammation, altered mitophagy, and dysregulation in proteostasis and mitochondrial biogenesis [ 56 ] . It has been documented that the activity of mitochondrial respiratory chain complexes and proteins involved in mitochondrial metabolism, including those in FA metabolism, declines with age in the heart [ 57 ] . In contrast, extracellular structural proteins and glycolytic pathways increase substantially with aging [ 58 ] . Additionally, studies have consistently shown that age-related increases in mitochondrial ROS result in deleterious lipid and protein oxidation and accumulation of mtDNA that impairs the mitochondrial respiratory efficacy and further increases ROS production, forming a vicious cycle [ 59 – 61 ] . Aging also likely contributes to diminished replication fidelity and quantity of mtDNA, which promotes the accumulation of dysfunctional mitochondria, leading to adverse outcomes [ 62 , 63 ] . In humans and several model systems, evidence suggests that mitochondrial structure is disrupted by the aging process. A disrupted morphology of the mitochondria and loss of cristae in the aged inner mitochondrial membrane has been shown with electron microscopy [ 64 – 66 ] . Studies have also indicated reduction and remodeling of cardiolipin in aging mitochondria [ 67 , 68 ] . As cardiolipin is responsible for maintaining optimal mitochondrial function and structure through its role in maintaining the proton gradient, cristae curvature, and preventing apoptosis, its loss is detrimental to the cardiac bioenergetic milieu [ 69 , 70 ] . Insulin-like growth factor signaling, the mammalian target of rapamycin (mTOR), and regulation of histone acetylation by sirtuins are among the regulatory pathways that modulate cardiac health and aging. Modification of these pathways during the aging process can trigger mitochondrial dysfunction to accelerate cardiac impairment. In humans, an age-dependent decrease in serum insulin-like growth factor 1 was shown to be correlated with an enhanced risk of heart failure [ 71 ] . Autophagy, protein translation, lipid synthesis, and ribosome biogenesis are just a few of the crucial processes that mTOR controls. In various model organisms, the mTOR inhibitor rapamycin is known to increase lifespan [ 72 , 73 ] while remodeling the aged heart proteome to a more youthful composition associated with improved mitochondrial function and decreased abundance of glycolytic pathway proteins. These findings might point to proteomic and metabolic remodeling as a mechanism behind the cardiac functional benefits granted by mTOR inhibitors [ 58 ] . Increased mitochondrial protein hyperacetylation has been observed in myocardial tissues from the failing hearts of both humans and animals. It has been demonstrated that protein hyperacetylation reduces the activity of the TCA cycle enzymes succinate dehydrogenase, pyruvate dehydrogenase, as well as the malate-aspartate shuttle [ 74 ] . Acetylation-mediated impairment of the malate-aspartate shuttle limits the transfer of cytosolic NADH into the mitochondria and alters the cytosolic redox state [ 75 ] . Increased myocardial short-chain acyl-CoA content may contribute to increased protein acetylation; reduced protein deacetylation by sirtuins is another probable cause [ Figure 1 ] [ 76 , 77 ] . Given their regulation by intermediate metabolites, sirtuins have been suggested to act as sensors of metabolic flux, and they are known to play a role in metabolic heart disease, natural aging, and heart failure [ 78 , 79 ] . Three of the seven mammalian sirtuins reside in the mitochondria (SIRT3, SIRT4, and SIRT5), with SIRT3 acting as the main driver of deacetylation [ 80 ] . Downregulation of SIRT3 has been observed in aged and failing hearts and may be attributable to decreased NAD + levels or NAD + /NADH ratios, as low NAD + levels inhibit SIRT3, leading to mitochondrial protein hyperacetylation and dysfunction [ 75 , 81 ] . Sirtuins also help mediate apoptosis signaling and reduce ROS by regulating antioxidant enzymes such as manganese superoxide dismutase (MnSOD) and catalase [ 82 , 83 ] . Genetic models have shown that SIRT3 knockout results in a reduction in complex I and III of the electron transport chain and decreased FA oxidation leading to a more glycolytic state [ 83 ] .

Accumulated metabolic byproducts The heart consumes more ATP than any other organ in the body, and without a substantial local energy storage system, it relies on substrate uptake from the circulation to produce ATP primarily through OXPHOS. Thus, a mismatch between substrate uptake and oxidation can have devastating consequences on heart structure and function. This is evident when substrate uptake is compromised, such as for glucose in the setting of insulin resistance, wherein the heart is forced to almost exclusively utilize fatty acids [ 156 ] . Similarly, when FA oxidation is compromised, the heart switches to glucose oxidation or the oxidation of alternative substrates, such as in the failing heart [ 157 , 158 ] . It is accepted that a defect in mitochondrial function may contribute to altered substrate oxidation in the diseased heart [ 159 – 163 ] ; thus, progressive failure of mitochondrial respiratory machinery may result in the accumulation of backlogged metabolites and metabolic intermediates. Many of these compounds are unstable and prone to oxidation or other modifications that form toxic byproducts, negatively feeding back on mitochondrial function and further driving cardiac structural changes in a cyclic cascade degenerative remodeling. A defect in cardiac mitochondrial FA oxidation coupled with an increase in FA uptake causes the accumulation of toxic lipid species [ Figure 3 ] [ 164 ] . To directly test the implication of cardiac lipid overload on contractile function, several mouse models of cardiac lipotoxicity have been generated [ 164 ] . Cardiac hypertrophy, diastolic dysfunction, apoptotic cell death, cardiac fibrosis, and development of heart failure have all been observed after genetic manipulation of FA uptake or oxidation in the heart [ 165 – 171 ] . In addition, mice with leptin or leptin receptor deficiency have been used as models of lipotoxicity, with early studies establishing a link between cardiac TG accumulation and diastolic dysfunction in these mice [ 172 , 173 ] . Whether the early diastolic dysfunction associated with cardiac lipid accumulation is a consequence of cardiac fibrosis has not been directly tested. However, studies have shown that exposing cardiomyocytes to the saturated lipid palmitate in vitro induced cell death [ 174 , 175 ] and led to replacement fibrosis in vivo [ 166 , 167 , 174 ] . Similarly, defects in mitochondrial FA oxidation and increased FA uptake can lead to the formation of signaling lipid species that are known to cause cell death. This is the case for the sphingolipid intermediates ceramides, which are elevated in mouse models with enhanced cardiac FA uptake [ Figure 3 ] [ 167 , 174 ] . Ceramides have been shown to induce cell death in cardiomyocytes in vitro [ 176 ] , but whether they cause cell death and replacement fibrosis in the heart in vivo has yet to be demonstrated. As mentioned above, a common feature of altered FA oxidation in the heart is the accumulation of lipids and the development of cardiac hypertrophy. This has been recapitulated genetically by directly inhibiting the mitochondrial transport of FA through the deletion of Cpt1b in the heart [ Figure 3 ]. Heterozygous Cpt1b knockout mice develop cardiac hypertrophy, increased fibrosis, and die within two weeks following transverse aortic constriction [ 177 ] . Similarly, conditional deletion of Cpt1b in skeletal and cardiac muscle caused massive cardiac hypertrophy and reduced survival due to the development of congestive heart failure, but cardiac fibrosis was not assessed in these mice [ 178 ] . Similar findings were observed in mice with defective beta-oxidation or lacking the master regulator of FA metabolism Ppparα [ 179 – 181 ] , where the accumulation of cardiac fibrosis in Pparα null mice developed at a later stage than the other alterations observed in these animals [ 180 ] . Together, these studies associate cardiac lipid accumulation with cardiac fibrosis, but a causal relationship has not been demonstrated; in fact, another study detected cardiac lipid accumulation in the absence of cardiac fibrosis in the hearts of type 1 Akita mice [ 182 ] . There are multiple mechanisms underlying mitochondrial dysfunction-induced cardiac fibrosis (discussed subsequently). However, few studies have explored the pro-fibrotic signaling by cardiac lipids. As mentioned above, lipid overload in cardiac cells induced ER stress and caused cardiomyocyte apoptosis. Loss of cardiomyocytes has been consistently observed in mice with cardiac lipid overload and coincides with the appearance of replacement fibrosis [ 166 , 167 , 174 , 176 ] . The mechanisms implicated in lipid-induced cardiomyocyte apoptosis may involve an increase in lipid peroxide, enhanced iNOS expression and NF-κB activation [ 183 ] . Excess polyunsaturated fatty acids such as linoleic, linolenic, or arachidonic acids have been shown to increase collagen I/III ratio in the mouse myocardium, leading to stiffening chara

As discussed above, maintenance of a dynamic mitochondrial population, including fission and clearance, is considered a protective mechanism during myocardial stress. Alteration in mitochondrial fission and clearance leads to the accumulation of damaged organelles that produce excessive ROS, exacerbating cardiac injury. In the adult heart, mitochondrial dynamics and mitophagy are minimal and deletion of proteins involved in mitochondrial dynamics or mitophagy has minimal effect under basal conditions. However, the necessity of these processes becomes evident during aging or in response to stress. While the causal relationship between mitochondrial QC and cardiac remodeling has not yet been directly investigated, indirect evidence suggests that the accumulation of dysfunctional mitochondria may exacerbate cardiac fibrosis through mechanisms involving ROS and inflammasome activation. Consistent with this idea, Pink null mice exhibit enhanced oxidative stress and fibrosis that was exacerbated with aging [ 147 ] . In contrast, Pink1 overexpression in a rat model of HFpEF stimulated mitochondrial fission and prevented cardiac fibrosis [ 139 ] . Similarly, deletion of Parkin in the adult mouse heart had no effect but mitigated excessive mitophagy and prevented cardiac remodeling and replacement fibrosis in the context of Drp1 deletion [ 149 ] . Similarly, impaired mitophagy via Ulk1 deletion exacerbated fibrotic remodeling after transverse aortic constriction and high-fat feeding in mice [ 155 , 239 ] . Perturbation of mitochondrial dynamics also resulted in the accumulation of cardiac fibrosis. Lack of the mitochondrial fission factor Mff led to the development of cardiomyopathy associated with fibroblast proliferation and replacement fibrosis [ 240 ] . Although the authors found increased apoptotic cell death in Mff knockout mice, they proposed that cell death alone did not account for the extent of fibrosis, suggesting additional mechanisms. Since Mff is a receptor for Drp1, knockout of Drp1 in the adult heart produced a similar cardiomyopathy phenotype characterized by the accumulation of substantial fibrosis [ 122 ] . Surprisingly, fibrosis is absent in Mfn1/Mfn2 double mutant mice, suggesting that altered mitophagy, rather than dynamics, may be causal for the development of fibrosis. The mechanisms by which altered mitophagy leads to cardiac fibrosis are not well understood but likely involve apoptotic cell death, oxidative damage, and immune cell infiltration. Indeed, impairing lysosomal degradation of mtDNA led to the activation of toll-like receptor 9 (Tlr9), which in turn activated immune cell infiltration in the heart [ 152 ] . Treatment of cardiac-specific DNase 2a knockout mice with a Tlr9 inhibitor or deletion of Tlr9 attenuated cardiac fibrosis [ 152 ] . These results suggest that either excessive or stalled mitophagy leads to cardiomyocyte death, oxidative stress, mtDNA accumulation, and inflammation, all of which contribute to significant fibrotic remodeling of the heart [ Figure 3 ].

While the mechanisms by which disrupted mitochondrial function may potentiate cardiac structural changes have been widely studied, as described above, fewer studies have explored the effects of myocardial structural remodeling on mitochondrial milieu and bioenergetic demand. However, the bidirectionality of structural and energetic remodeling cascades may prove particularly important during the transition and decompensation stages of heart failure, when the pathological mechanical and biochemical environment overwhelms disrupted energy systems beyond their compensatory capacity. Thus, future studies of the mechanistic progression to heart failure should consider altered mitochondrial function as both a cause and consequence of structural cardiac remodeling, such as hypertrophy and fibrosis. Compensatory hypertrophy is believed to develop initially as an adaptive response to help maintain cardiac output and mitigate tissue-level stresses through thickening and stiffening of the ventricular walls [ 41 ] . However, along with increased energy demand due to chronic overload and decreased energy production capacity, hypertrophy progresses maladaptively and mechanistically contributes to the bioenergetic deficit precipitating heart failure. Additionally, it has been shown that the growth pathways involved in the development of cardiac hypertrophy directly regulate mitochondrial morphology and bioenergetic function, including regulation of TCA cycle and FAO enzymes [ 251 ] . Mitochondrial morphology is highly dynamic, and this functional plasticity endows the organelles with substantial adaptive capabilities. In healthy cardiomyocytes, mitochondria are abundant and contain intact membranes and clear cristae structures. In hypertrophic hearts subject to chronic overload, mitochondrial density is decreased, and the organelles may become swollen, elongated, and deformed, exhibiting ruptured membranes and irregular cristae structures [ 252 , 253 ] . These deformities reduce the biogenetic capacity of the mitochondria in the hypertrophic heart, which further exacerbates the detrimental effects of cardiac structural remodeling and may contribute significantly to the progression into the decompensated stage of PO-induced heart failure. While studies have shown that growth pathways involved in cardiac hypertrophy significantly interact with metabolic regulatory processes, the molecular mechanisms by which hypertrophic remodeling alters mitochondrial function are not yet fully elucidated. PKB, also known as Akt, is a serine-threonine kinase that has been widely studied across diverse physiologic and pathologic settings [ 254 ] . Activation of Akt is strongly associated with hypertrophic cardiac growth; while adaptive in the short-term, persistent stimulation of Akt signaling is deleterious due to feedback inhibition of insulin receptor substrate, PI3K signaling, and GLUT4-mediated glucose uptake, and may help precipitate heart failure due to a mismatch between cardiac hypertrophy and angiogenesis [ Figure 4 ] [ 255 – 257 ] . Furthermore, using mice with cardiomyocyte-specific constitutively activated Akt1 (caAkt) signaling, it has been shown that persistent activation of Akt directly alters cardiac mitochondrial bioenergetic function. Wende et al. observed selectively repressed expression levels of TCA cycle enzymes and proteins involved in OXPHOS, FAO, and mitochondrial biogenesis in caAkt hearts [ 251 ] . This was accompanied by a robust increase in left ventricular mass and contractile dysfunction at six weeks of age, as well as reduced functional measures of mitochondrial efficiency. Additionally, 18-week-old caAkt mouse hearts subject to ischemia-reperfusion showed decreased rates of glucose oxidation, palmitate oxidation, and myocardial oxygen consumption concomitant with increased glycolysis. A canonical pathways analysis also revealed that several mitochondrial metabolic and signaling pathways were differentially regulated in caAkt hearts, including Foxo1, Pparα, and Pgc-1α. Thus, hypertrophic factors such as Akt have a direct effect on mitochondrial bioenergetics and morphology. Numerous animal studies of pressure-induced hypertrophy have reported reduced FAO rates during the compensated stages of hypertrophy preceding the onset of overt heart failure [ 47 , 258 , 259 ] . In addition to decreased fatty acid oxidative capacity, carnitine deficiency and reduced CPT-1 activity in cardiac hypertrophy may limit mitochondrial uptake and utilization of fatty acids [ 260 , 261 ] . Meanwhile, glucose oxidation has been variously observed to increase, remain unchanged, or decrease during cardiac hypertrophy and heart failure [ 47 , 258 , 262 , 263 ] . While decreased fatty acid and glucose oxidation in hypertrophy may be partially compensated by anaerobic glycolysis and increased anaplerotic flux through the TCA cycle, this is less energetically efficient and may contribute to energy insufficiency during the transition to h

Approximately one in five individuals over 80 years old are at risk of cardiac dysfunction and heart failure, and among patients with congestive HF, arterial fibrillation, and coronary heart disease, over 70% are elderly [ 281 , 282 ] . Intrinsic cardiac aging is understood as the gradual progression of structural alterations and functional declines occurring with age independent of the prolonged exposure to canonical cardiovascular risk factors. The ‘free radical theory of aging’ suggests accumulated macromolecule oxidative damage as a hallmark of aging [ 283 ] . This paradigm is supported by observations that aged cardiomyocytes show increased markers of oxidative damage and associated abnormalities in mitochondrial structure, such as loss of cristae, enlarged organelles, and matrix derangement [ 284 ] . Additionally, age-dependent reductions in mitochondrial OXPHOS are correlated with a decline in the activity of ETC complexes I and IV, which may be driven by increased ROS generation and electron leakage. Furthermore, aged hearts exhibit four-fold increases in deletion frequency and point mutations in mtDNA, as well as increased mtDNA copy number [ 285 ] . Significantly downregulated autophagy is also widely observed in the aging heart; the inability to eliminate damaged cells and repair damaged organelles is particularly important in post-mitotic cells such as cardiomyocytes [ 286 ] . Therefore, therapeutic approaches that counteract age-related changes in mitochondrial QC, autophagy regulators, inflammation, and ROS generation would be potential targets to mediate the deleterious effects of cardiac aging.

Myriad therapeutics have been developed to aid in the prevention and treatment of heart disease over the last century. In this section, we will primarily focus on therapeutic targets involving mitochondria in the context of metabolic heart disease, cardiac aging, and heart failure. Human and animal data both suggest that reduced mitochondrial biogenesis and increased oxidative stress are hallmarks of several heart disease and heart failure phenotypes, including diabetic cardiomyopathy, cardiac aging, and heart failure. As such, therapeutics targeted at increasing mitochondrial biogenesis and reducing ROS are likely to prove beneficial in the treatment of various HF phenotypes [ 305 ] . One promising pharmacological avenue to stimulate mitochondrial biogenesis is through AMP-activated protein kinase (AMPK) [ 299 , 306 ] . AMPK is a highly conserved regulator of energy homeostasis and metabolism that is known to activate PGC-1α, the transcriptional coactivator considered to be the master regulator of mitochondrial biogenesis [ 277 ] . AMPK has been shown to increase mitochondrial biogenesis via both direct phosphorylation of PGC-1α and activation of NRF1/TFAM [ Figure 5 ] [ 307 – 309 ] . Several widely used cardioprotective therapies have been suggested to target AMPK activation indirectly, such as metformin, telmisartan, and thiazolidinediones [ 305 , 310 , 311 ] . For example, metformin and thiazolidinediones improve systemic and tissue insulin sensitivity, improving cardiomyocyte glucose uptake and cardiac function concomitant with the activation of PPARγ and AMPK [ 277 ] . AMPK can also enhance the expression and translocation of GLUT4 and glucose uptake, which is particularly beneficial in the context of diabetic cardiomyopathy [ 312 ] . Resveratrol is a naturally occurring polyphenolic stilbene that has been shown to increase mitochondrial biogenesis through both AMPK and NO-dependent mechanisms through activation of PGC-1α, NRFs, and TFAM. Importantly, several studies have reported improved mitochondrial biogenesis and cardiac functional parameters with resveratrol administration during hypertension-mediated HF in both humans and animals without a measurable reduction in blood pressure, suggesting direct effects on the heart [ 313 – 316 ] . Various direct activators of AMPK have also been developed and tested; however, the development of AMPK activators is complicated by variable expression levels and differential effects of various subunits and isoforms [ 317 ] . For example, several groups have reported that amino acid substitution within the C-terminal side of the γ2 subunit of AMPK leads to the development of aberrant conduction systems and severe cardiac hypertrophy [ 318 – 320 ] , although it has been suggested that this hypertrophy phenotype may be attributable to increased carbohydrate storage as opposed to myocyte cytoskeletal growth [ 320 , 321 ] . Interestingly, it has also been observed that systemic pan-activation of AMPK, for example, by MK-8722, induced as opposed to ameliorated cardiac hypertrophy [ 322 ] . Another promising avenue for increasing mitochondrial biogenesis is the soluble guanylyl cyclase/cyclic guanosine monophosphate (sGC/cGMP) pathway. Nitric oxide synthesized in the vasculature binds to sGC in vessel smooth muscle, catalyzing the conversion of guanosine triphosphate to cGMP. cGMP clearance is regulated by hydrolyzing phosphodiesterases, including PDE5A, which is known to be expressed in cardiomyocytes [ 323 ] . cGMP and its effector kinase, PKG, have been shown to regulate cardiac structure and function via regulation of calcium flux, phosphorylation of contractile proteins, and several other mechanisms [ Figure 5 ] [ 324 ] . It has also been suggested that this pathway stimulates mitochondrial biogenesis via PGC-1α and inhibits mitochondrially mediated cell death [ 325 – 327 ] . sGC stimulators have been previously used to treat pulmonary hypertension and have recently emerged as potential therapeutics for heart failure [ 305 , 306 ] . Recent clinical trials with the sGC stimulator vericiguat showed some promise by significantly reducing high-risk HFrEF patient hospitalization [ 328 ] and improving patient quality of life [ 329 ] , nevertheless, no recent trials have shown benefits of sGC/cGMP stimulation in patients with HFpEF [ 306 , 330 ] . As PDE5 expression is upregulated in hypertrophic and failing hearts, leading to decreased cGMP levels, inhibition of PDE5 represents an attractive therapeutic target, and it has been shown that the PDE5 inhibitor, sildenafil, ameliorates PO-induced cardiac hypertrophic remodeling in mice via deactivation of PKG and inhibition of the L-type Ca 2+ channel [ 331 ] . Increasing evidence from animal models suggests that targeted inhibition of ROS within mitochondria, rather than globally, may be cardioprotective, as it was shown in the GISSI-Prevenzione trial that chronic supplementation with the global antioxidant α-tocopherol resu