Insulin Signaling and Heart Failure

Abstract

Heart failure is associated with generalized insulin resistance. Moreover, insulin resistant states such as type 2 diabetes and obesity increases the risk of heart failure even after adjusting for traditional risk factors. Insulin resistance or type 2 diabetes alters the systemic and neurohumoral milieu leading to changes in metabolism and signaling pathways in the heart that may contribute to myocardial dysfunction. In addition, changes in insulin signaling within cardiomyocytes develop in the failing heart. The changes range from activation of proximal insulin signaling pathways that may contribute to adverse left ventricular remodeling and mitochondrial dysfunction to repression of distal elements of insulin signaling pathways such as forkhead (FOXO) transcriptional signaling or glucose transport which may also impair cardiac metabolism, structure and function. This article will review the complexities of insulin signaling within the myocardium and ways in which these pathways are altered in heart failure or in conditions associated with generalized insulin resistance. The implications of these changes for therapeutic approaches to treating or preventing heart failure will be discussed.

Cardiovascular disease (CVD) remains the leading cause of death globally in developed and developing economies ¹. The morbidity from CVD is broad-based and includes consequences of atherosclerosis, hypertension leading to ischemic heart disease, stroke and heart failure. Heart failure is a major cause of death in industrialized nations and exceeds the mortality of most types of cancers. It is estimated that about 5.1 million patients are affected by heart failure in the United States. The annual incidence is approximately 670,000 cases, with about half of these patients dying within 5 years after the initial diagnosis ². The associated annual costs are about $34.4 billion, resulting in an enormous economic burden ³. Thus, heart failure once it develops has a dismal prognosis that has not significantly improved despite advances in pharmacological therapies. Much work has been done to examine diverse signaling pathways that contribute to heart failure, pathological cardiac hypertrophy and adverse left ventricular remodeling and have been reviewed extensively ^{4, 5}. A relatively understudied signaling pathway that could represent an important contributor to the pathophysiology of heart failure is insulin signaling, which will be the subject of this review.

The connection between abnormal insulin signaling and heart failure arises in part from the established epidemiological association between obesity, type 2 diabetes, insulin resistance and heart failure. Diabetes increases the risk for ischemic heart disease as evidenced by a two- to four-fold increase in myocardial infarction in diabetic patients compared to non-diabetics and is associated with a poorer prognosis ⁶. In addition, coexistence of ischemic cardiomyopathy and diabetes accelerate the progression to heart failure ⁷. Numerous epidemiologic studies identified insulin resistance and diabetes as risk factors for the development of heart failure independent of myocardial ischemia. The Framingham Heart Study revealed that diabetes independently increases the risk of heart failure by about 2-fold in men and about 5-fold in women compared with age-matched control subjects, independent of blood pressure and serum cholesterol levels ^{8, 9}. Furthermore, a longitudinal study in the same cohort identified obesity as a risk factor for the future incidence of heart failure in both men and women ¹⁰. These data are supported by the United Kingdom Prospective Diabetes Study (UKPDS), which reported that the risk for developing heart failure increases by 16% for every 1% increase in serum HbA_1c concentrations ¹¹. Similar results reporting an association between diabetes and the incidence of heart failure have been provided by the Cardiovascular Health Study ¹², the New Haven cohort of the Established Populations for Epidemiological Studies in the Elderly ¹³, post hoc analyses of clinical trials¹⁴ and have been reviewed comprehensively¹⁵.

In contrast, a number of studies reported a significant increase in survival in obese subjects in the context of heart failure ^{16, 17} and following myocardial infarction after adjusting for age, severity of illness, and gender ¹⁸. This resulted in the term “obesity paradox” which has been previously reviewed ^{19, 20}. Although obesity and type 2 diabetes are well-described and commonly accepted risk factors for the development of cardiovascular disease, these studies suggest a protective role of obesity when cardiac decompensation occurs. Given that cachexia is independently associated with a worse prognosis in heart failure, the possibility exists that the obesity paradox could in part reflect the adverse consequences of cachexia, which would lower body weight. These lower body weight subjects with heart failure were reported for example to have higher circulating concentrations of tumor necrosis factor α (TNFα) ²¹, implying an adverse consequence of systemic inflammation in these subjects. In another cohort, the potential benefit of the obesity paradox in heart failure was seen in the elderly and those with acute heart failure and was not present in subject in diabetes ²¹. Thus many of the interactions between insulin resistance obesity and heart failure might differ between acute and chronic heart failure. The neurohumoral changes in obesity that might confer a survival benefit in acute heart failure are currently poorly understood. Moreover, the absence of a benefit in subjects with type 2 diabetes suggests that long-standing insulin resistance, hyperinsulinemia or hyperglycemia might offset any survival advantages that can be attributed to the “obesity paradox”. Thus, the pathophysiology of the increased risk of heart failure in insulin resistant states is complex and multifactorial, and the potential impact of generalized insulin resistance and hyperinsulinemia per se remains incompletely understood.

Moreover, studies in humans and animal models have revealed that heart failure is associated with generalized insulin resistance ²². Metabolic abnormalities observed in patients with advanced heart failure resulting in catabolism, cardiac cachexia and insulin resistance might correlate with decreased survival²³. Although it is has not been definitively shown that insulin resistance in the context of heart failure directly contributes to a worse prognosis, a recent report suggested that ventricular unloading leads to improvement in whole body insulin sensitivity ²⁴. This review will therefore explore the current state of knowledge linking generalized insulin resistance with heart failure with a focus on the physiological and pathophysiological roles of insulin signaling within the heart, the cardiac adaptations to generalized insulin resistance and the interaction between systemic insulin resistance and changes in myocardial insulin signaling, which may inform the pathophysiology of heart failure.

Insulin Signaling Pathways

Insulin mediates its signaling upon ligand binding to the insulin receptor, which is a member of the tyrosine kinase family of receptors and is closely related to and partially homologous with the receptor for insulin-like growth factor −1 (IGF-1). Upon ligand binding, insulin receptors are autophosphorylated, which in turn increases its tyrosine kinase activity for other substrates such as insulin receptor substrates (IRS) proteins (Figure 1). The activated insulin receptor interacts with IRS proteins and other binding partners to activate a network of intracellular signaling pathways including PI3K/Akt and MAP Kinases such as ERK ^{25, 26}. As summarized in Figure 1, an essential initial step in IR/IGF-1R receptor signaling is activation of IRS proteins. IRS1 and IRS2 are the two most abundant IRS isoforms in the heart and are required for insulin-mediated activation of PI3K. PI3K consists of a p110 catalytic subunit and a p85 regulatory subunit, which catalyzes the generation of the lipid product phosphatidylinositol (3,4,5)-triphosphate (PIP₃). This results in phosphoinositide dependent kinase 1 (PDPK1) mediated activation of the kinase Akt (protein kinase B). Three members of the Akt family have been identified. Akt1 and Akt2 represent the most abundant Akt isoforms in the heart. Akt mediates a variety of cellular processes by phosphorylating proteins involved in autophagy and members of the forkhead transcription factor (FOXO) family which inhibits nuclear translocation and transcriptional activity²⁷ of promoters that encode pro-apoptotic signaling molecules, such as members of the BCL-2 family. The isoform-specific contribution of Akt on the regulation of these processes is incompletely understood. However, Akt1 but not Akt2 signaling has been identified to mediate somatic growth^{28, 29}. Furthermore, it has been shown that Akt2 is required for insulin-stimulated glucose uptake and metabolism in isolated cardiomyocytes and this effect is independent of Akt1³⁰. Together, this suggests a predominant role for Akt1 in the regulation of somatic growth and for Akt2 in the regulation of metabolism. Inhibition of glycogen synthase kinase 3β (GSK-3β) by Akt-mediated phosphorylation promotes cardiac hypertrophy ³¹ and glycogen synthesis. In addition, Akt mediates phosphorylation of mechanistic target of Rapamycin (mTOR), a protein complex regulating somatic growth and protein synthesis. mTOR consists of two distinct complexes, mTORC1 and mTORC2³². Among other components, mTOR Complex 1 (mTORC1) is composed of Raptor (regulatory-associated protein of mTOR)³³. Rictor (Rapamycin-insensitive companion of mTOR) is part of mTOR Complex 2 (mTORC2)³⁴. mTOR mediated phosphorylation of 70 kDa ribosomal protein S6 kinase (p70S6K) and 4E-binding protein 1 (4E-BP1)³⁵ results in its dissociation from eukaryotic translation initiation factor 4E (eIF4E) and activation of cap-dependent translation³⁶, which is attenuated by the mTOR inhibitor Rapamycin.

Figure 1. Schematic of insulin signaling pathway.

Simplified summary of key elements involved in insulin signal transduction that ultimately regulate multiple cellular processes. Upon binding to its ligand, insulin and IGF-1 receptors undergo autophosphorylation, which increases their tyrosine kinase activities. Tyrosine phosphorylation and activation of the docking proteins insulin receptor substrates 1 and 2 (IRS1/2), engages regulatory subunits of the phosphatidylinositol-3-kinase (PI3K) that generates phosphatidylinositol 3,4,5 tris phosphate (PIP3) from phosphatidylinositol 3,4, bis phosphate (PIP2). The serine threonine kinases phosphoinositide dependent protein kinase 1 (PDPK1) and Akt1 or Akt2 bind to PIP3 by their PH domains. PDPK1 phosphorylates Akt on Thr 308 and mTORC2 phosphorylates Akt1/2 on Ser 473. Once activated, Akt in turn phosphorylates multiple targets of which a subset is shown. Phosphorylation of these targets induces pleiotropic cellular responses: Bcl2 phosphorylation inhibits apoptosis, FOXO protein phosphorylation promotes nuclear exclusion, thereby repressing the expression of FOXO-regulated transcripts that mediate autophagy and apoptosis. TSC1/2 phosphorylation promotes mTOR activation, which increases mRNA translation, promotes protein synthesis, cell growth, mitochondrial fusion and also inhibits autophagy. Phosphorylation of glycogen synthase kinase (GSK) removes the repression of glycogen synthase, which in concert with increased availability of glucose promotes glycogen synthesis. Phosphorylation of endothelial nitric oxide synthase (NOSIII) or eNOS by Akt increases the generation of nitric oxide (NO) to promote vasodilation. Akt phosphorylation mediates the translocation of vesicles containing the GLUT4 glucose transporter in part by phosphorylating downstream targets such as AS160 (not shown), leading to increased glucose transport following insertion into the plasma membrane. Akt also mediates in part, translocation of the fatty acid translocase CD36. Insulin signaling also promotes activation of the mitogen activated protein kinases (ERK1/2) to increase the expression of various genes.

In addition to activating the Akt / mTOR signaling cascade and mediating GLUT4 translocation, insulin activates the mitogen-activated protein kinase (MAPK) / extracellular regulated kinase (ERK) pathway. This pathway involves activation of a cascade including Raf, MEK and ERK kinases, resulting in nuclear translocation of ERK and phosphorylation of transcription factors such as p62TCF, initiating a transcriptional program that mediates cellular differentiation and proliferation³⁷. A variety of genetically altered mouse models with perturbed insulin / IGF-1 signaling has been utilized to investigate insulin signaling in the heart. Table 1 summarizes the characteristics of previously described transgenic models under basal conditions and superimposed pathological stressors.

Table 1.

Rodent models for perturbed insulin / IGF-1 signaling

Basal phenotype						Phenotype following superimposed pathological stressor
Signaling molecule	Genetic Manipul ation / Treatme nt	Cardiac size	Contractile Function	Notes	Ref	Stressor	Cardiac size	Contractile Function	Notes	Ref
IR/IGF-1R
IR	CKO	↓	= / ↓	age-dependent mitochondrial dysfunction	42, 73, 78, 137	ISO	=	↓	catecholamine- mediated injury ↑	138
						AC	=	↓	susceptibility to heart failure ↑	137
						MI	=	↓	left ventricular dysfunction post MI ↑	139
						STZ	↓	=	mitochondrial dysfunction ↑ / cardiac efficiency ↓	162
						IPC			efficacy of IPC to reduce infarct size ↑	163
	CKO het	=	=		114	AC	↓	↑		114
IGF-1R	CKO	=	=		78
	MKO	↑	=		164
IR/IGF- 1R DKO	MKO	↓	↓↓		164
IRS
IRS1	CKO	↓	=		76
IRS2	CKO	=	=		76
IRS1/2 DKO	CKO	= / ↓	↓↓		43, 150
	MKO	=	↓↓		165
	GKO	ND	ND	embryonic lethal?	166
	CKOi	↓	↓↓		151
IRS3	GKO	=			167
IRS4	GKO	NR			168
PI3K
p85	p85α MKO p85β GKO	↓	=		169
p110α	dn	↓	=		170, 171	AC	=	↓		170, 172
						MI	=	↓		173
						DCM		↓	survival following cross ↓	172
	ca	↑	=		171	AC	=	↑		172
						MI	=	↑		173
						DCM		↑	survival following cross ↑	172
p110γ	GKO	=	↑		174	AC		↓		175
						ISO		↑		174
PDK1
PDK1	CKO		↓↓	β1-adrenergic receptor ↓	176, 177
	MKO	↓	↓↓	Dilated cardiomyopathy	178
Akt
Akt1	GKO	=	=		79, 136	AC	↑	↓		136
	GKO het		=		114	AC		↑		114
Akt2	GKO	= / ↑	= / ↓	age-dependent hypertrophy and contractile dysfunction	30, 79, 179	AC	=	=		30
						ISO	↑	=		179
						MI		↓	infarct size ↑	180
Akt3	COE	↑	↓	age-dependent contractile dysfunction	181
	GKO	=			182
kdAkt	kd	=	=		183
caAkt	ca	↑	↓	age-dependent contractile dysfunction	183
	myrAkt	↑	=		184	MI			infarct size ↓	184
GSK-3β
GSK-3β	ca	=			185	AC	↓			185
	dn	↑	=		186	AC	=	↑		186
mTOR
mTOR	Rapamyc in	=	=		187, 188	AC	↓	= / ↑		187, 188
						MI	↓	↑	LV remodelling and cell death ↓	189
	CKO (mTOR)	↓	↓ ↓	embryonic lethal	190
	CKOi (mTOR)		↓ ↓		191	AC	↓	↓↓	rapid development of heart failure post AC	191
	CKOi (raptor)		↓	autophagy ↑, Glucose oxidation ↑, Palmitate oxidation ↓	192	AC	↓	↓↓	rapid development of heart failure post AC	192
	CKOi (mTOR)	=	↓	Glucose oxidation ↑, Palmitate oxidation ↓	193
	ca	=	=		194
	kd	=	↓		194	ISO	=	=	hypertrophic response similar to WT	194
p70S6K
p70S6K1	TG	↑	=	mutant with higher basal activity	195	AC	=			195
	kd	=	=		195
	GKO	=			195	AC	=			195
p70S6K2	kd	=	=		195
	GKO	=			195	AC	=			195
p70S6K1 /2	GKO	=			195	AC	=			195
GLUT4
GLUT4	CKO	↑	=		123, 196
	MKO (α-actin- Cre)	↑	↓	age-dependent contractile dysfunction	197, 198

Changes are expressed compared to WT controls same treatment; ↑ increased; ↓ decreased; = no difference observed.

AC, aortic constriction; ca, constitutively active mutant; CKO het, cardiac-specific heterozygous knockout (αMHC-Cre); CKO, cardiac-specific knockout (αMHC-Cre); CKOi, inducible cardiac-specific knockout; COE, cardiac-specific overexpression; DCM, cross to transgenic mouse model of dilated cardiomyopathy; dn, dominant-negative mutant; GKO het, global heterozygous knockout; GKO, global knockout; IPC, ischemic preconditioning; ISO, Isoproterenol treatment; kd, kinase-deficient mutant; MI, Myocardial infarction; MKO, muscle-specific knockout (MCK-Cre, unless otherwise indicated); myrAkt, myristoylated Akt; STZ, streptozotocin treatment; TG, transgenic model; ND, not determined.

Insulin receptors are ubiquitously expressed and are highly abundant in tissues of the cardiovascular system such as the heart and endothelial cells ³⁸. Given its identification as a hormone that is intimately involved in fuel metabolic homeostasis, much of the work on insulin signaling has focused on its impact on tissues such as adipose, liver, skeletal muscle and brain and the mechanisms by which insulin regulates glucose uptake, lipogenesis and hepatic glucose utilization for example ^39–41 . Subsequent studies have revealed that insulin signaling by activating cell survival pathways may play an important role in the regulation of apoptosis and autophagy and cellular growth ^{42, 43}. Thus a brief review of the physiological role of insulin signaling in the heart and vasculature is warranted to set the stage for a deeper discussion of the pathophysiological consequences of insulin signaling in heart failure.

Insulin Signaling in the Vasculature

The major cell types in the vasculature in which insulin signaling has been evaluated are endothelial and vascular smooth muscle cells. In endothelial cells, insulin signaling via PI3K and Akt leads to activation of endothelial nitric oxide synthase. Nitric oxide release not only promotes vasorelaxation, but may also activate anti-inflammatory and anti-atherosclerotic pathways. As such, impaired endothelial insulin signaling has been suggested to contribute to obesity and diabetes-related vascular dysfunction that may contribute to hypertension and accelerated atherosclerosis ^{44, 45}. It is important to note that whereas there is broad consensus that impaired nitric oxide availability will contribute to vascular pathology in insulin resistant states, impairment of eNOS activity may occur independently of defects in insulin signaling per-se ^46–48. Another important consideration in understanding impaired endothelial insulin signaling in the context of insulin resistant states is the concept of imbalanced insulin signaling, whereby peripheral hyperinsulinemia, leads to activation of certain branches of the insulin signaling pathway, while others remain inhibited. For example, in certain circumstances, whereas insulin signaling to PI3K-Akt and eNOS might be impaired in endothelial cells, activation of MAPK signaling pathways leading to induction of endothelin1 expression might remain active and indeed be activated by ambient hyperinsulinemia, thereby leading to increased vasoconstriction ^49–51. Similarly, in vascular smooth muscle cells, the hyperinsulinemia that accompanies insulin resistant states has been suggested to be a contributor to vascular smooth muscle cell hyperplasia, which may promote intimal injury and vascular stenosis ⁵². The role of impaired insulin signaling in the coronary vasculature and the pathophysiology of heart failure in insulin resistant states is incompletely understood. However, subjects with insulin resistance and type 2 diabetes were shown to have significantly impaired endothelial-dependent coronary vasodilation ⁵³. Thus the possibility exists that impaired coronary vasodilation in insulin resistant states could contribute to impaired myocardial contractile reserve in heart failure, particularly in subjects with insulin resistance.

Insulin Signaling and the Regulation of Substrate Metabolism in Cardiomyocytes

Like skeletal muscle, activation of insulin signaling pathways in cardiomyocytes leads to an increase in glucose uptake by promoting the translocation of GLUT4 containing vesicles to the plasma membrane ⁵⁴. These vesicles ultimately fuse with the plasma membrane and expose GLUT4 proteins on the cellular surface. This process involves the modulation of filamentous actin networks, SNARE proteins, and small GTPases of the Rab family⁵⁵. Insulin signaling mediates GLUT4 translocation at multiple checkpoints. Recently identified mechanisms include Akt-mediated phosphorylation of Akt substrate of 160 kDa (AS160), a GTPase-activating protein for the Rab family of small G proteins. Phosphorylation of AS160 in response to insulin regulates its interaction with 14-3-3 and inhibits AS160 GTPase activity. This results in increased Rab activity and translocation of GLUT4 vesicles to the plasma membrane ^{56, 57}. It has also been shown that selective activation of Akt might be sufficient to stimulate GLUT4 translocation to an extent which is similar to insulin⁵⁸. However, Akt-independent mechanisms have also been identified ⁵⁷. In perfused hearts or in vivo, the increase in glucose utilization (uptake and oxidation) that is mediated by insulin signaling is associated with reduced fatty acid oxidation on the basis of the Randle phenomenon ^{42, 59}.

Transmembrane uptake of fatty acids (FA) in cardiomyocytes is mediated in part by fatty acyl translocases of the FA transporter family and CD36 ⁶⁰. FAs are then converted to FA-CoA by the enzyme Acyl-CoA synthase (ACS) and may be stored in the form of triglycerides. However, excess lipids may also be shunted into non-oxidative pathways which results in the generation of toxic lipid intermediates. This eventually results in mitochondrial dysfunction⁶¹, perturbed cellular signaling⁶², incomplete FA oxidation and increased apoptosis⁶³, a phenomenon called lipotoxicity (reviewed in ⁶⁴). Lipotoxicity has been extensively studied in animal models of obesity and type 2 diabetes, including ob/ob, db/db mice (models for obesity and insulin resistance based on leptin deficiency or resistance, respectively) and Zucker fatty rats⁶⁵. Importantly though, insulin also promotes the translocation of the fatty acid translocase CD36, which may increase FA uptake that may be partitioned towards lipid synthesis and storage ⁶⁰. This insulin-mediated increase and translocation of CD36 is transduced by PI3K/Akt signaling pathways⁶⁶ and may contribute to increased FA oxidation and increased myocardial oxygen consumption (mVO₂) in hyperinsulinemic states that is often associated with increased availability of triglycerides and FA to the heart. Increased FA utilization is associated with decreased cardiac efficiency (cardiac work / mVO₂). Studies in ob/ob and db/db mice suggest that mitochondrial uncoupling and increased reactive oxygen species (ROS) parallel the increase in FA oxidation^{67, 68}. Mitochondria-derived ROS can activate uncoupling proteins (UCPs)⁶⁹ allowing protons to bypass the ATP synthase of the electron transport chain and decrease coupling of mitochondrial ATP production to O₂ consumption. As a consequence, fatty acid oxidation further increases, which subsequently decreases cardiac efficiency. Similar to these studies performed in animal models, a recent study using human myocardial samples from type 2 diabetics showed mitochondrial dysfunction and increased oxidative stress⁷⁰. Furthermore, increased mVO₂ along with increased FA oxidation was reported for young women with obesity, insulin resistance and increased body mass index⁷¹. Increased ROS production and mitochondrial uncoupling was not observed in rodent models of type 1 diabetes⁷². This suggests that ROS mediated mitochondrial uncoupling might not be attributable to hyperglycemia per se, but is more likely a consequence of generalized insulin resistance and type 2 diabetes. Interestingly, hearts from cardiomyocyte-specific insulin receptor knockout mice show increased ROS generation and mitochondrial uncoupling, in the absence of hyperglycemia⁷³.

Insulin may also directly regulate mitochondrial metabolism by promoting mitochondrial fusion via a mechanism that involves the induction of OPA1 ⁷⁴. Short-term Insulin-mediated Akt activation may also activate pro-survival signaling pathways related to Akt inhibition of apoptotic signaling ⁷⁵. Studies in gene targeted mice with cardiomyocyte-restricted loss of insulin receptors or IRS1 proteins have also suggested that insulin signaling may play an important trophic role within the heart ⁴². Thus, loss of insulin receptors or IRS-1 is associated with a reduction in cardiomyocyte size with little hemodynamic consequences in non-stressed animals ^{42, 76}. In addition insulin/IGF-1 or IRS1 or IRS2 signaling pathways appear to be required for physiological cardiac hypertrophy in response to exercise ^76–78. Insulin signaling also regulates mitochondrial oxidative capacity. Thus mice with genetic deletion of the insulin receptor, or IRS proteins develop reduced mitochondrial oxidative capacity via partially understood mechanisms that may involve repressed expression of nuclear-encoded mitochondrial genes ^{43, 73, 76}. We have also shown that activation of PI3K signaling is responsible for increasing mitochondrial oxidative capacity in response to physiological cardiac hypertrophy ^{76, 79, 80}. However, this effect of PI3K appears to be independent of changes in Akt signaling and as will be discussed later, constitutive activation of Akt appears to repress mitochondrial oxidative capacity.

Insulin Resistance

Insulin resistance has been classically defined as the inability of insulin to promote its metabolic actions in organs such as adipose tissue, skeletal muscle and liver (Figure 2). Insulin resistance is a hallmark of obesity and type 2 diabetes and also is a characteristic of heart failure. Thus in insulin resistant states there is impaired insulin-mediated glucose uptake in muscle and adipocytes, impaired suppression of hepatic glucose production and impaired suppression of lipolysis ⁸¹. Paradoxically, triglyceride synthesis and hepatic secretion of VLDL is increased on the basis of increased insulin mediated activation of lipogenesis ⁸². This underscores the important concept of selective insulin resistance, indicating that whereas certain targets of insulin signaling might be repressed in insulin resistant states other targets may retain their “sensitivity” ⁸³. The metabolic milieu in insulin resistant individuals is therefore characterized by increased circulating concentrations of FFA, and triglycerides, and variable increases in glucose concentrations. Moreover, beta cells adapt to the insulin resistant environment by increasing insulin release, leading to hyperinsulinemia. In addition to these systemic metabolic abnormalities, insulin resistance is associated with additional changes to the systemic metabolic milieu including low-grade inflammation, increasing circulating concentrations of various cytokines and altered secretion of adipokines such as leptin, resistin and adiponectin ^{84, 85}. Mechanisms for insulin resistance include ectopic accumulation of lipid intermediates in skeletal muscle and liver. In the liver, a proposed mechanism for steatosis is hyperinsulinemia-driven de novo lipogenesis in the face of excess caloric intake. The molecular mechanisms for hepatic insulin resistance are incompletely understood; however, increased endoplasmic reticulum stress, ROS generation and mitochondrial dysfunction have been suggested to contribute to the development of insulin resistance⁸⁶. Ectopic lipid accumulation and oxidative stress have been proposed to contribute to skeletal muscle insulin resistance ⁸⁷. Additional mechanisms contributing to insulin resistance include adipose tissue dysfunction that impairs the release of “insulin sensitizing” adipokines such as adiponectin, mitochondrial dysfunction, incomplete FA oxidation, and activation of inflammatory responses leading to increased promulgation of inflammatory cytokines such as TNFα and interleukins. Mechanistically, TNFα, IL-6 and free fatty acids (FFAs) activate intracellular kinases that induce serine phosphorylation of IRS proteins, thereby attenuating insulin signaling and inducing insulin resistance⁸⁸. This complex pathophysiology of insulin resistance has been classically associated with obesity and a sedentary lifestyle. However, in heart failure generalized insulin resistance develops in patients with cardiac cachexia. Most studies seeking to explore the mechanisms for insulin resistance have been performed in subjects with obesity or type 2 diabetes. Fewer studies have been conducted in patients with heart failure.

Figure 2. Summary of mechanisms that lead to insulin resistance in heart failure or in the metabolic syndrome.

Signaling events in skeletal muscle, liver, adipose tissue and brain (not shown) impair insulin signaling in each respective organ leading to metabolic perturbations as illustrated, which alter the systemic milieu in ways that may adversely impact cardiac structure and function.

Insulin resistance in heart failure also correlates with increased serum concentrations of proinflammatory cytokines, catecholamines, catabolic steroids and growth hormone^89–93. In addition, decreased physical activity, loss of skeletal muscle mass and sympathetic over-activity, increases lipolysis thereby increasing the circulating concentrations of free fatty acids, which may further exacerbate insulin resistance in skeletal muscle and liver by promoting the accumulation of bioactive lipids. FFAs may further increase the activity of the sympathetic nervous system⁹⁴, which may independently attenuate peripheral insulin signaling and reduce glucose utilization in skeletal muscle⁹⁵. Furthermore, increased catecholamine concentrations stimulate hepatic glycogenolysis and gluconeogenesis leading to increased circulating glucose. Heart failure is also associated with hyperactivation of the renin-angiotensin-aldosterone system (RAAS), system, which has been implicated in the pathophysiology of insulin resistance ⁹⁶. In addition to increasing circulating concentrations of angiotensinogen (originating from adipose tissue)⁹⁷, chronic hyperinsulinemia may also increase the expression of angiotensin II receptors, which may further exacerbate adverse left ventricular remodeling⁹⁸. Recent studies have also indicated that heart failure patients with insulin resistance also exhibit evidence of ectopic accumulation of bioactive lipids such as ceramide in their skeletal muscles, adipose tissue inflammation and hypo-adiponectinemia, all of which may contribute to generalized insulin resistance ^{99, 100}. Moreover, skeletal muscle hypo-perfusion impairs skeletal muscle oxidative capacity. This induces growth hormone resistance and a state of low-grade inflammation which might also contribute to insulin resistance ^101–103. Growth hormone resistance described in patients with heart failure results in decreased hepatic IGF-1 production in response to growth hormone, but circulating growth hormone concentrations increase ⁹². Increased growth hormone receptor signaling has been shown to induce insulin resistance in multiple tissues including the heart^104–106, thereby exacerbating peripheral insulin resistance which might further impair cardiac function (Figure 2).

The changes in the metabolic milieu summarized above may contribute to the myocardial mal-adaptations that have been described in insulin resistant states, often referred to as diabetic cardiomyopathy and which have been extensively reviewed ^{107, 108}. In brief and as outlined above, multiple mechanisms have been proposed to contribute to the increased vulnerability of the heart in insulin resistant states such as increased FA utilization, impaired mitochondrial oxidative capacity, mitochondrial dysfunction, decreased cardiac efficiency, oxidative stress, accumulation of bio-active lipids, inflammation, increased apoptosis, altered calcium metabolism and signaling, increased apoptosis and myocardial fibrosis. In addition, expansion of local adipose tissue depots might compromise cardiac function under conditions of obesity and type 2 diabetes. Epicardial adipose tissue (EAT) is a visceral-like adipose depot that is located on the surface of the ventricles, along the coronary arteries, and at the apex of the heart. Perivascular adipose tissue (PVAT) surrounds the blood vessels. EAT mass and diameter are increased in obesity ^{109, 110}, and the volume of PVAT is associated with visceral obesity ¹¹¹. Neither, EAT nor PVAT, are separated by a fascial layer from the surrounding tissue. Therefore, it has been suggested that adipokines and possibly fatty acids secreted from EAT and PVAT might directly contribute to the development of cardiovascular dysfunction in type 2 diabetes¹¹². The induction of generalized insulin resistance and hyperinsulinemia in the context of pressure overload accelerates LV remodeling ^{113, 114}. Thus it is plausible that insulin resistance that accompanies heart failure could further exacerbate myocardial dysfunction by similar mechanisms. There is a paucity of studies that have directly addressed the possible beneficial effects of therapeutic manipulations that might increase insulin sensitivity in patients with heart failure to determine if such changes will meaningfully impact LV function or alter long-term prognosis. In a cohort of patients with advanced heart failure the Schulze group recently reported that mechanical unloading resulted in a significant reduction in indices of systemic insulin resistance that occurred in concert with improvement in hemodynamic indices ²⁴. Their study design did not address the question of whether or not the improvement in insulin sensitivity directly contributed to improved LV function. However, the authors reported that improvement in insulin sensitivity did correlate with reversal of certain metabolic and signaling abnormalities in the heart. An important question that these observations raise is: What are the contributions of altered insulin signaling and insulin resistance per se in cardiomyocytes and the potential contribution of these changes to heart failure and LV remodeling?

Cardiac Muscle Insulin Signaling in Insulin Resistant States or in Heart Failure

Prior to summarizing important studies in the field, it is important to discuss the various ways in which myocardial insulin resistance has been defined in the literature. In most instances, “insulin resistance” has been defined as a reduction in insulin-mediated myocardial glucose uptake. This definition is advantageous in the sense that it enables the inclusion of non-invasive studies that evaluate myocardial glucose utilization. The limitation however, is that this definition of myocardial insulin resistance does not account for changes in the activation of signaling intermediates that are downstream of the insulin receptor. For example, a reduction in GLUT4 protein could account for decreased insulin mediated glucose uptake in the face of an otherwise intact insulin signaling pathway, whereas impaired intracellular signaling could also reduce insulin-mediated glucose uptake even when GLUT4 protein levels are normal. As will be discussed, intact versus impaired intracellular signaling might influence LV remodeling in distinct ways.

Nearly all studies that have examined insulin-stimulated glucose uptake in humans or in animals with generalized insulin resistance, usually induced by diet-induced obesity, have demonstrated impaired insulin-mediated glucose uptake ¹¹⁵. Studies of myocardial glucose uptake in subjects with heart failure need to be interpreted in light of whether or not measurements of glucose uptake were performed under basal conditions or in response to physiological or pharmacological hyperinsulinemia as would occur during a euglycemic hyperinsulinemic glucose clamp. In many animal studies in which heart failure is induced by aortic banding (thoracic or abdominal) or by coronary artery ligation, basal glucose uptake has been reported to be increased, and this has been attributable to an increase in the GLUT1 transporter ^116–119. However, when the augmentation of glucose uptake in response to insulin was evaluated, the fold increase in insulin-mediated glucose uptake was impaired relative to non-failing control groups ¹²⁰. Importantly, in contrast to GLUT1, GLUT4 protein levels in the hearts of experimental models of heart failure or pathological cardiac hypertrophy are variable ^{117, 118, 121, 122} and the impact of GLUT4 expression and translocation in the context of heart failure is incompletely understood. Studies in transgenic mouse lines revealed that GLUT4 expression is dispensable for contractile function under basal conditions¹²³, while these mice exhibited increased fibrosis¹²⁴ and a poor response to ischemic injury¹²⁵. In contrast, deletion of GLUT1 in the hearts of mice did not exacerbate heart failure progression following TAC ¹¹⁷. Thus it is likely that GLUT1 and GLUT4 mediated glucose transport may play divergent as well as overlapping roles in the modulation of myocardial glucose uptake in heart failure. It is important to note that whereas myocardial glucose utilization might be increased in compensated stages of LV hypertrophy and early in the course of heart failure, glucose utilization may be reduced in the hearts of patients and animals with advanced heart failure ¹²⁶.

Insulin stimulation of the heart activates PI3K and Akt, which are requisite upstream signaling steps that promote GLUT4 translocation and increased glucose uptake. Studies of insulin-mediated activation of Akt in the heart in rodent models of generalized insulin resistance have yielded variable and potentially divergent results, which might be explicable by differences in the models used and differences in experimental conditions, such as dietary lipid composition. Studies based on 45% fat diets of varying durations have suggested that the ability of insulin to stimulate Akt might be augmented or preserved at a time when insulin-mediated glucose uptake is impaired ¹²⁷. In contrast, studies in more severe models of insulin resistance such as ob/ob mice or following a more aggressive high-fat protocol with 60% fat leads to impaired insulin-mediated activation of Akt and with FOXO1 dephosphorylation and nuclear localization ¹²⁸. It is important to put in context that the usual rodent diet is significantly lower in fat than human diets. Thus caution is warranted in extrapolating findings from rodent students in which a large increase in dietary fat is imposed for variable periods of time, given that species-specific differences may exist in the myocardial adaptations to a pathological lipid load, relative to those that might already be partially adapted to diets that are higher in fat content. In studies of failing hearts in experimental models, usually induced by transverse aortic constriction (on normal chow), Akt phosphorylation is increased in concert with increased IRS-1 phosphorylation ^{114, 129}. In one study, genetically reducing the expression of insulin receptors, Akt1 or inducing systemic hypoinsulinemia reduced Akt phosphorylation and interestingly attenuated cardiac hypertrophy, and prevented adverse LV remodeling and heart failure ¹¹⁴.

Limited availability of human heart tissue has hampered the ability of investigators to evaluate changes in insulin signaling pathways in heart failure or insulin resistant states. Early studies suggested dynamic regulation of uncoupling protein and GLUT4 content in human heart atrial appendages in response to increasing circulating FA concentrations indicating that increased myocardial lipid delivery could repress myocardial GLUT4 ¹³⁰. Subsequently in an elegant study combining LV biopsy with metabolic analyses, Cook and colleagues reported that a significant reduction in plasma membrane GLUT4 accounted for the reduction in myocardial glucose utilization in humans with heart failure (average EF 27%) or in subjects with type 2 diabetes without heart failure despite minimal changes in overall GLUT4 content ¹³¹. The depletion of sarcolemmal GLUT4 was corroborated by animal studies that revealed that only two weeks of high fat feeding significantly reduced myocardial GLUT4 content and sarcolemmal translocation even prior to any weight gain and at a time when myocardial insulin signaling was preserved ¹²⁷. Importantly, the human studies by Cook’s group also showed that PI3K and Akt signaling was substantially increased in the hearts of subjects with type 2 diabetes or heart failure and positively correlated with circulating insulin concentrations and with indices of insulin resistance. These studies suggest a paradigm whereby insulin-mediated glucose uptake in the heart in insulin resistant states could be dissociated from activation of insulin signaling to PI3K and Akt. The mechanisms contributing to reduced GLUT4 expression and translocation in the heart in obesity are incompletely understood, but could be due in part to altered regulation of SNARE proteins ¹³¹.

In contrast, in an analysis of left ventricular samples obtained from subjects with advanced heart failure at the time of left ventricular assist device (LVAD) implantation, Schulze and colleagues reported a significant repression in myocardial Akt phosphorylation that occurred in concert with evidence of accumulated bio-active lipids such as diacyl glycerol and ceramides leading to activation of protein kinase C isoforms that have been implicated as mediators of impaired insulin signaling ²⁴. In materials obtained after mechanical unloading and following evidence of recovery of LV function, these signaling changes were reversed, leading to the conclusion that in end-stage heart failure there is significant repression of insulin signaling pathways and that recovery correlated with improvement in insulin action. In a study from a different group, mechanical unloading was associated with a decrease in the phosphorylation of the Akt target FOXO3 in human hearts ¹³². This study did not measure Akt phosphorylation, but in light of other studies that have reported activation of Akt signaling in human failing hearts, the implications of this observation is that in some cases ventricular unloading could potentially reduce myocardial Akt signaling.

The complexity of the Akt-FOXO interactions is further underscored by studies showing that 6-months of treatment with a 60% high fat diet leads to heart failure, which is associated with reduced Akt phosphorylation and dephosphorylation and nuclear localization of FOXO1 ¹²⁸. Genetic deletion of FOXO1 prevented the development of heart failure and the associated metabolic abnormalities that was induced by high fat feeding. Interestingly genetic deletion of FOXO1 also restored insulin sensitivity by reducing serine phosphorylation of IRS1 proteins and restoring Akt phosphorylation in cardiomyocytes in response to insulin stimulation. Thus FOXO activation could represent an important mechanism leading to impaired insulin signaling in heart failure that is induced by lipotoxic cardiomyopathy. It is noteworthy that additional post-translational modifications of FOXO proteins could potentially drive nuclear localization ¹³³; therefore it will be of interest to determine if nuclear FOXO localization precedes impaired insulin-IRS1-Akt signaling in metabolic cardiomyopathy or occurs as a consequence of it.

Taken together, these studies underscore the complex responses of insulin signaling pathways in the myocardium to the changes in the systemic metabolic milieu that characterizes insulin resistant states or that accompany left ventricular remodeling in the failing heart. Current experimental data and studies in human samples could suggest a model whereby an early adaptation of the heart to systemic insulin resistance or to heart failure might be the repression of GLUT4 expression and GLUT4-mediated glucose transport. This occurs at a time when proximal insulin signaling to PI3K and Akt might remain intact and could potentially accelerate mal-adaptive ventricular hypertrophy. Over time, in response to persistent nutrient overload or to ongoing ventricular remodeling, FOXO activation (i.e. increased nuclear localization) might accelerate the desensitization of proximal insulin signaling leading to loss of Akt signaling and reduction in pro-survival signaling pathways that may further exacerbate cardiomyocyte loss. It has been proposed that myocardial insulin resistance might represent a mechanism by which the heart attempts to protect itself against nutrient overload in insulin resistant states¹³⁴. Early on, attempts to over-ride myocardial insulin resistance could exacerbate the early phases of adverse ventricular remodeling. However, it is possible that over time, a transition may take place in which there is subsequent loss of proximal insulin signaling to Akt, which may exacerbate LV dysfunction via additional mechanisms related to further loss of trophic effects of insulin signaling on myocyte mass and mitochondrial capacity and loss of its anti-apoptotic actions, all of which could accelerate or exacerbate LV dysfunction (Figure 3). We will next explore potential mechanisms linking changes in myocardial insulin signaling to left ventricular dysfunction and adverse LV remodeling. Given the potential bi-phasic changes in insulin signaling that occur in the evolution of heart failure or systemic insulin resistance namely an activation of Akt, followed later by loss of insulin-mediated Akt signaling, pathways by which Akt activation may promote heart failure versus mechanisms linking impaired insulin signaling with heart failure will be discussed separately.

Figure 3. Molecular mechanisms linking increased or decreased insulin signaling with adverse left ventricular (LV) remodeling and heart failure progression.

In early stages of heart failure mediated by pressure overload, or in the early stages of the metabolic syndrome, hyperinsulinemia leads to activation of insulin receptor substrate 1(IRS1) and Akt1 to promote pathological hypertrophy, mitochondrial dysfunction and decreased autophagy, which contribute to accelerated LV remodeling. As heart failure progresses, despite systemic hyperinsulinemia insulin signaling pathways may become desensitized leading to loss of cytoprotective consequences of Akt signaling and persistent nuclear localization of forkhead (FOXO) proteins that will accelerate heart failure by various mechanisms such as increased apoptosis and exacerbation of lipotoxicity.

Excessive Insulin Signaling and Heart Failure

As summarized above, there is clear evidence that hyperactivation of proximal insulin signaling pathways such as increased IRS-1 phosphorylation and Akt hyperactivation may occur in the context of pathological cardiac hypertrophy and subsequent heart failure, and that reduction of insulin or Akt signaling in this context might reverse LV remodeling and preserve cardiac function ¹¹⁴. How might persistent activation of Akt lead to pathological hypertrophy? Interesting insights have been obtained from various lines of mutant mice with constitutive activation of various downstream components of the insulin signaling pathway. These models mimic the effects of hyperinsulinemia in type 2 diabetes and obesity on myocardial insulin signaling. Although constitutive activation of PI3K leads to compensated cardiac hypertrophy, our group recently showed that this leads to desensitization of insulin mediated glucose uptake followed by impaired ability of insulin to activate Akt ⁵⁴. Of note this impairment in glucose uptake is due in part to a defect in activation of glucose transport following insertion GLUT4 in the plasma membrane.

In mice with inducible transgenic overexpression of Akt, cardiac hypertrophy is reversible and LV function preserved after short-term activation. However, if Akt expression is induced long-term, heart failure ensues, in part to a failure of angiogenesis. Importantly if Akt signaling is subsequently reduced in these failing hearts, the rate of progression to mortality is accelerated ¹³⁵. It is tempting to speculate that this temporal relationship between an adverse consequence of Akt activation leading to pathological remodeling that is then exacerbated by reduction of Akt expression might parallel the temporal relationship between an initial increase in Akt signaling in earlier stages of heart failure that subsequently transitions to reduced Akt signaling in end-stage cardiomyopathy. We recently reported an additional mechanism by which persistent Akt signaling impairs mitochondrial energy metabolism that likely contributes to the Akt-mediated LV dysfunction. The mechanism for mitochondrial dysfunction is due to Akt-dependent repression of the expression nuclear-encoded mitochondrial genes via FOXO dependent and independent signaling pathways. We also showed that short-term activation of Akt promotes a metabolic switch characterized by increased glycolysis while impairing mitochondrial FA oxidation and we speculated that this short term adaptation might represent a compensatory response to protect mitochondrial integrity and capacity. Importantly the mitochondrial repression occurs early and is independent of the Akt-induced cardiac hypertrophy ¹²⁹. However, in the long-term, a mismatch between impaired mitochondrial capacity and the increase in energetic requirements of Akt-mediated cardiac hypertrophy accelerates the onset of heart failure. Partial reduction of insulin or Akt signaling in murine models of TAC-induced cardiac hypertrophy prevents heart failure, which supports a model in which Akt-driven mechanisms might contribute to adverse LV remodeling ¹¹⁴. It is important to emphasize though, that in these studies Akt signaling was not abolished, but rather maintained at basal levels. In contrast, when Akt1 or insulin receptor signaling is completely abolished, then in response to stressors such as pressure overload, catecholamine toxicity or myocardial ischemia, all of these models exhibit a more rapid progression to heart failure ^136–139. Taken together, these observations suggest that whereas residual insulin-PI3K-Akt signaling is required to maintain cardiomyocyte viability, excessive signaling by these pathways drives adverse LV remodeling and ventricular dysfunction via multiple mechanisms.

We recently described that hyperinsulinemia may impair myocardial contractility via a mechanism by which the G-protein receptor kinase (GRK2) leads to G_i biased β₂AR signaling that inhibits adenylate cyclase, cAMP generation and cardiac contractility ¹⁴⁰. The inhibition of βAR signaling via GRK2 by hyperinsulinemia requires insulin receptor and IRS-1 mediated signaling pathways (Figure 3). Ob/ob mice or mice that have been treated for 12-weeks with a 45% HFD, exhibit increased myocardial GRK2 content ¹⁴¹. GRK2 overexpression has been shown to reduce myocardial insulin signaling, and in germline GRK2 heterozygous KO mice, GLUT4 content is increased, insulin signaling was enhanced in aging mice or when these animals were placed on a high-fat diet ^{141, 142}. Moreover, reduced GRK2 signaling promoted a “physiological cardiac hypertrophy” profile as animals aged. Thus GRK2 activation might also contribute to impairing myocardial insulin signaling with aging and in response to nutrient overload. GRK2 overactivation may also increase mitochondrial-mediated apoptosis by increasing mitochondrial calcium sensitivity, MPTP opening and cytochrome C release ¹⁴². Taken together, GRK2 activation could represent another important mechanism linking hyperinsulinemia with adverse LV remodeling. Although increased sympathetic activity is an established mechanism by which GRK2 is induced in the heart, whether or not hyperinsulinemia induces GRK2 by this mechanism remains to be definitively elucidated.

Suppression of PI3K signaling in the heart was reported to reduce signs of cardiac aging, in part by sustaining physiological levels of autophagy ¹⁴³. Kim and colleagues also recently reported that aging was associated with increasing expression of IGF-1 receptors in cardiomyocytes and in mice with deficiency of myocardial IGF1R signaling, cardiac senescence and age-dependent cardiac fibrosis was reduced ¹⁴⁴. Finally, hyperactivation of insulin signaling pathways will lead to autophagy suppression ¹⁴⁵. Autophagy plays a critical role in organellar quality control, thus as has been recently reviewed hyperinsulinemic activation of Akt/mTOR signaling could suppress autophagy thereby contributing to LV dysfunction ²⁷. In addition to Akt/mTOR-mediated suppression of autophagy, we also reported a novel mechanism that may occur in insulin resistant states and the metabolic syndrome by which bioactive lipids that accumulate in the heart will activate NOX2 via a PKC-dependent mechanism that ultimately inhibits lysosomal function and autophagic flux ¹⁴⁶. In this model autophagic flux was impaired despite reduced phosphorylation of mTOR. Recent reports of the time course of autophagy following TAC suggests that although autophagy might be transiently induced, a repression of autophagy coincides with the transition from compensated hypertrophy to heart failure ¹⁴⁷. Consistent with this model are many observations that mouse models with impaired expression of autophagy mediators develop LV dysfunction either at baseline or more rapid decompensation following pressure overload ^{148, 149}. Thus constitutive repression of myocardial autophagy could represent a plausible mechanism by which hyperinsulinemia and activation of Akt signaling could impair left ventricular function and accelerate adverse LV remodeling. Impaired autophagy might also impair mitochondrial quality control. Whether or not insulin signaling specifically regulates mitophagy remains incompletely understood, and whether or not they are linked in heart failure is currently unknown.

Decreased Insulin Signaling and LV Remodeling

Mechanisms that are distinct from those associated with hyperactivation of insulin signaling pathways might also contribute to promoting LV dysfunction when insulin-PI3K and Akt signaling ultimately become impaired. In this regard, valuable insights have been obtained from genetic knockouts of components of the insulin signaling pathway (Table 1). It is important to emphasize that there is significant redundancy in insulin and IGF-1 signal transduction pathways ⁷⁷. Thus, there is clear crosstalk between insulin and IGF-1R signaling pathways in the heart, which might reflect the relatively benign phenotypes of single knockouts of the IGF-1R or the IR respectively ^{42, 78}. This is also true for IRS isoforms in that loss of IRS1 or IRS2 individually do not precipitate LV dysfunction in unstressed hearts and indeed we observed no differences in the ability of insulin to activate Akt or promote glucose uptake ⁷⁶. However, combined deletion of IR and IGF-1receptors or IRS1 and IRS2 leads to catastrophic heart failure due in part to unrestrained autophagy, apoptosis and profound mitochondrial dysfunction ^{43, 150}. Similar phenotypes have been observed in animals with targeted loss of the phosphoinositide dependent protein kinase (pdpk1), which is required for IR or IGF-1R dependent activation of Akt ⁸⁰.

The second mechanism by which reduced insulin signaling might contribute to LV dysfunction is by persistent nuclear localization of FOXO1. Guo and colleagues in mice with cardiomyocyte-restricted deletion of IRS1 and IRS2 showed that they could ameliorate the heart failure in part by genetically deleting FOXO1 expression ^{150, 151}. Similarly heart failure associated with long-term high-fat feeding could also be prevented when FOXO1 was deleted ¹²⁸. It is likely that the mechanism by which persistent nuclear localization of FOXO proteins might lead to LV dysfunction is multifactorial. Mechanisms by which FOXO may mediate LV remodeling include myosin isoform switching, FOXO mediated activation of autophagic and atrophic pathways and FOXO-mediated activation of lipid metabolism pathways that would promote lipotoxicity ^{128, 132, 151}. The potential bi-phasic response in insulin signaling is also of particular importance for the design of future studies. Therefore, it is of great interest to perform time course studies, especially in the context heart failure and insulin resistance.

Implications for Therapy and Future Research

Taken together, this review of the field identifies an important need for uniformity in determining changes in the specific components of the insulin signaling cascade at varying stages in the evolution from compensated cardiac hypertrophy to decompensated cardiac hypertrophy and ultimately to heart failure. Furthermore, a careful analysis of the relationship between altered myocardial insulin signaling in models of generalized insulin resistance in concert with cardiac adaptations to hemodynamic stress will advance our understanding of the complex interactions between generalized insulin resistance, myocardial insulin signaling and heart failure. To more closely mimic human disease, studies in which heart failure is induced in animals that have developed generalized insulin resistance would be informative. Generalized insulin resistance is usually evident in high-fat fed rodent models after high-fat diets of >10-week duration. We would also suggest that measuring glucose uptake in response to insulin gives only a partial picture of changes in insulin signaling in the heart and as such it is important to evaluate the phosphorylation states of IRS1/2, Akt1/2 and FOXO proteins under conditions of ambient insulinemia and in response to an insulin stimulus. These analyses would be informative if determined in conjunction with assays of signaling defects in skeletal muscle, adipose tissue and liver, as they will not only identify differences in the adaptations of these tissues to insulin resistance but also the potential for crosstalk between these classical insulin signaling targets and the heart in models of heart failure. Moreover, they will test the hypothesis that changes in insulin signaling, either an increase in the early stages followed by impaired signaling in late-stage heart failure, may independently contribute to the pathophysiology of heart failure.

Modulation of myocardial insulin signaling is not currently a consideration in the management of heart failure. As discussed above, there are profound changes in myocardial insulin signaling that develop in the failing heart and are also present in subjects with generalized insulin resistance that may sensitize the heart to LV dysfunction. This raises important questions regarding whether or not modulation of generalized insulin resistance, normalization of hyperinsulinemia and other disturbances in the metabolic milieu will alter the prognosis of heart failure. Second, it remains to be determined if strategies that will directly correct the changes in myocardial insulin signaling described above will reverse heart failure. Oxidative stress is a widely accepted mechanism for cardiac dysfunction in insulin resistant states and it was recently reported that targeted anti-oxidant therapy may partially restore abnormal insulin signaling and improve cardiac structure and function ¹⁵². Combined treatment with polyphenols resveratrol and S17834 also significantly restored cardiac function in diet-induced heart failure in rodent models via multiple mechanisms including reducing oxidative stress, reversing oxidative modifications of proteins, reducing hyperinsulinemia and increasing circulating adiponectin concentrations ¹⁵³. As alluded to above, GRK2 could represent an attractive target that could ameliorate pathological LV remodeling via mechanisms that could involve direct modulation of insulin signaling pathways within cardiomyocytes as well as by effects on systemic metabolic homeostasis ^{141, 142, 154}. Whereas ventricular unloading improved myocardial insulin signaling and lipotoxicity, it is not clear if the myocardial changes were primary or secondary to improvement in peripheral insulin resistance ²⁴. It is probable that ventricular unloading by reversing some of the mechanisms that promote generalized insulin resistance in subjects with heart failure might have contributed to a feed-forward mechanism to increase myocardial recovery.

An important conundrum that has been increasingly recognized is the unique challenges in managing heart failure in the patient with type 2 diabetes and insulin resistance. Although it is clear that diabetic patients with heart failure will respond to conventional heart failure therapies, many recent clinical trials have revealed a disturbing association between multiple diabetes therapies and worsening of heart failure or increased heart failure hospitalization. This topic has been the subject of recent reviews ^{155, 156}. However, it is clear that many diabetes therapies including thiazolidinediones and even insulin might be associated with volume expansion, which may exacerbate heart failure. Similar effects have also been observed with certain DDPIV inhibitors. Specifically, treatment with the DDPIV inhibitor saxagliptin resulted in increased heart failure hospitalization¹⁵⁷. DPPIV degrades GLP-1 and other vasoactive peptides¹⁵⁸ including brain natriuretic peptide (BNP), which though increased in heart failure could represent an adaptive response¹⁵⁹. Moreover, the increase in circulating concentrations of insulin that occur as a result of incretin therapy could lead to ligand-dependent activation of insulin signaling pathways. Thus, if hyperactivation of insulin signaling indeed contributes to adverse ventricular remodeling in heart failure then a plausible mechanism for the relationship between these therapies and heart failure exacerbation could be the increase in myocardial insulin signaling and Akt hyperactivation. In contrast, metformin, which acts as an insulin sensitizer and reduces hyperinsulinemia might afford some degree of protection in terms of heart failure ¹⁶⁰. Even though not directly proven, decreased circulating insulin levels following metformin treatment might attenuate cardiac insulin signaling and provide a potential explanation for the cardioprotective effects observed. A recent trial using a novel class of diabetes therapeutic, an inhibitor of the renal sodium glucose co-transporter (SGLT2) had a dramatic impact on subsequent risk of hospitalization for heart failure ¹⁶¹. It is not known if this beneficial effect was secondary to the mild volume depletion or anti-hypertensive effects of this agent or to effects on altering the metabolic milieu such as reducing circulating levels of insulin. Thus additional studies are required to determine if therapeutic strategies that directly impact the abnormal systemic milieu that characterizes the insulin resistant state will impact LV structure and function in heart failure and impact prognosis and survival.

Concluding Remarks

The heart is an insulin responsive organ and insulin signaling plays an important role in cardiovascular homeostasis. Cardiac hypertrophy and heart failure are associated with profound changes in myocardial insulin signaling. Moreover, heart failure is an insulin resistant state that leads to an altered metabolic milieu that will influence myocardial structure and function via mechanisms that are dependent or independent of alterations in myocardial insulin signaling. It is possible that strategies that might correct the systemic metabolic disturbances that are associated with insulin resistance could have an impact on the outcome and prognosis of subjects with heart failure. An important imperative for the future will be to design trials that can rigorously test this hypothesis. Second, an important opportunity for the future will be to identify and evaluate therapeutic strategies that might directly influence changes in myocardial insulin signaling that accompany heart failure. Given the dynamic nature of these changes, it will be important that subjects are carefully stratified or phenotyped to determine which insulin signaling pathways are perturbed and the pathophysiological mechanisms that will most likely benefit from targeted therapies.

Nonstandard Abbreviations and Acronyms

FA

Fatty Acids

EAT

Epicardial adipose tissue