Under physiological conditions, the myocardium is able to switch between metabolic substrates, mainly fatty acids and carbohydrates, in response to changes in nutrient availability (71). However, systemic low-grade inflammation goes along with cardiac insulin resistance which is accompanied by a shift in substrate utilization toward fatty acid metabolism favoring the accumulation of toxic lipid metabolites (29).

TNF-α causes cardiac insulin resistance by activation of both NF-κB- and the JNK-dependent signaling pathways converging toward serine phosphorylation and proteasomal degradation of the insulin response substrate (IRS)1 (29). Moreover, IL-6 interferes with insulin signal transduction through signal transducer and activator of transcription (STAT)3-dependent suppressor of cytokine signaling (SOCS)3 upregulation, which impairs the coupling of IRS1 with the insulin receptor (72). In line with this notion, genetic knockout of IL-6 attenuates cardiac insulin resistance and inflammation in obese mice (49). At the myocardial level, insulin resistance is further promoted by inflammation-induced oxidative stress (73).

Contrasting reduced cardiac glucose uptake via the insulin-dependent glucose transporter 4 (GLUT4) in insulin resistant states, metabolic stress promotes increased cardiac fatty acid uptake through upregulation of cluster of differentiation (CD)36, the main fatty acid transporter in cardiomyocytes. CD36 is regulated by the peroxisome proliferator-activated receptor (PPAR)-γ/retinoid X receptor (RXR) complex allowing for its enhanced expression in response to nutrient excess (Figure 2) (29, 74). Paired with high levels of circulating FFAs, increased abundance of CD36 on the sarcolemmal membrane raises intracellular fatty acid availability and turnover (75).

FFA overload leads to mitochondrial dysfunction and uncoupling of fatty acid oxidation from ADP phosphorylation in cardiomyocytes (17). As a result of deranged cardiac lipid metabolism, cardiac triacylglycerols and toxic intermediate products such as diacylglycerols and ceramides are formed (17) and accumulate in the heart of obese and diabetic patients (76–78). Cardiac lipotoxicity has been implicated in the generation of reactive oxygen species (ROS), cell apoptosis, defective insulin signaling, and impaired calcium handling (79–83). While the exact mechanisms underlying cardiac lipotoxicity remain elusive and are subject of ongoing investigations, the combination of myocardial inflammation, insulin resistance and excess supply of FFA emerges as a decisive factor (17).

Nutrient overload activates different inflammatory signaling cascades in cardiomyocytes which contribute to cell hypertrophy, apoptosis, and mechanical dysfunction (38). The regulation of inflammatory programs in cardiomyocytes is closely linked to intracellular ROS accumulation resulting from deranged cardiac substrate utilization in diabetes and obesity (17, 38). Excess availability of lipids and glucose favors the production of ROS (17, 84) which in turn enhances the transcription and functional activity of NF-κB (85–89). Cardiomyocyte-specific inhibition of NF-kB signaling through overexpression of inhibitor of NF-κB (IκB)-α mitigates cardiac alterations in hyperglycemic mice—highlighting the importance of this axis (90).

In addition, high glucose concentrations directly activate a number of pro-inflammatory pathways in cardiomyocytes converging toward NF-κB. Exposure to high glucose levels enhances the expression of high-mobility group box 1 (HMGB1) protein in cardiomyocytes thereby activating mitogen-activated protein kinase (MAPK) and NF-κB which leads to TNF-α and IL-6 secretion (41). High glucose also induces upregulation of TNF-α, IL-1β, IL-6, and IL-12 through activation of JNK and NF-κB (45). Another mechanism linking glucose metabolism to inflammation is histone 3 lysine 9 trimethylation (H3K9me3) at the IL-6 promoter under high glucose conditions favoring its upregulation (91). Moreover, posttranslational modification of the NF-κB p65 subunit by O-linked N-acetylglucosamine (O-GlcNAc) enhancing its transcriptional activity under hyperglycemic conditions may also apply to cardiomyocytes (92). Likewise, hyperglycemia-induced epigenetic changes that increase p65 expression may be of relevance in cardiomyocytes (38, 44, 93).

Excess availability of FFAs contributes to deranged substrate utilization of the heart in high metabolic states leading to lipotoxicity and ROS formation (17). Exposure of human cardiomyocytes to saturated fatty acids enhances NF-κB binding activity and raises nuclear p65 protein levels leading to enhanced expression of TNF-α, IL-6, and CCL-2 (94). Similar findings were reported in hearts from mice fed a high-fat diet (94). Direct activation of the NOD-, LRR- and pyrin domain-containing protein (NLRP) 3 inflammasome by accumulating ceramides has been demonstrated in other cell types including adipocytes and may also be of importance in cardiomyocytes.

Endothelial cells are a central component of the cardiac vasculature forming a barrier between blood and myocardial tissue. Aside from their regulatory function in substrate exchange, endothelial cells control myocardial blood flow, and immune cell recruitment (95–97). Endothelial nitric oxide (NO) production regulates the vascular tone and hinges on functional insulin signaling in endothelial cells (98). In metabolic disorders, such as obesity and T2D, coronary endothelial cell function is markedly impaired by high levels of circulating inflammatory mediators (e.g., TNF-α, IL-1β, and IL-6) contributing to insulin resistance (99). In addition, excess metabolic substrates, namely glucose and FFA, exert a rage of detrimental effects on endothelial cell function linked to ROS formation and inflammatory pathway activation (98).

Exposure of endothelial cells to high glucose levels activates IκB kinase (IKK)β and NF-κB signaling which leads to upregulation of inflammatory cytokine expression, reduced insulin sensitivity and diminished NO production (100, 101). Excess glucose also leads to tight junction disruption—a hallmark of endothelial barrier dysfunction—through activation of the NLRP3 inflammasome (102). In line, high glucose levels associate with increased inflammatory markers in the circulation and in endothelial cells in the setting of acute coronary syndrome (103, 104).

High levels of circulating FFAs disrupt endothelial cell function via induction inflammatory signaling cascades and increased ROS formation (105). FFAs induce vascular inflammation via TLR4-dependent activation of IKKβ and NF-κB which has been linked to endothelial insulin resistance and decreased NO availability (105–108). It has also been reported that FFAs selectively stimulate NF-κB and activator protein (AP)1 transcriptional activation leading to enhanced expression of inflammatory mediators such as TNF-α, CCL-2, and ICAM-1 (105, 109). Conversely, genetic inhibition of NF-κB in endothelial cells blocks ROS formation, improves insulin sensitivity, downregulates vascular adhesion molecules and increases the expression of endothelial NO synthetase (eNOS) in obesity (110). Moreover, palmitic acid, a long-chain saturated fatty acid, activates the NLRP3 inflammasome and increases the expression of IL-1β in endothelial cells thereby contributing to endothelial dysfunction (111).

Fibroblasts are one of the largest non-cardiomyocyte cell populations in the heart and regulate the ECM composition, structure, and turnover (112). Expansion of the cardiac interstitium through accumulation of ECM proteins (i.e., interstitial and perivascular fibrosis) in patients with obesity and diabetes reflects a maladaptive response to glucometabolic disturbances (112, 113). Exposure to high glucose increased the expression of transforming growth factor (TGF)-β, the main fibrogenic cytokine in the heart, and promotes fibroblast proliferation and ECM protein synthesis in vitro (113–118). High glucose levels also activate the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (Akt)/MAPK signaling pathway and leads to upregulation of pro-inflammatory IL-17 synthesis and IL-17 receptor (IL-17R) expression, thus stimulating increased collagen synthesis (119). Apart from the MAPK pathway, these effects may be partially favored by a pro-inflammatory state in cardiac fibroblasts under high glucose conditions manifest from activation of NF-κB and enhanced expression of TNF-α, IL-6, and IL-1β (41, 114, 117, 120). A key event in myocardial remodeling is the conversion of fibroblasts to activated myofibroblasts, the main ECM producing cells (112). Due to differences in study conditions, conflicting data have been reported on the effect of high glucose on myofibroblast transition with the majority of studies pointing toward increased myofibroblast conversion under high glucose conditions (102, 115, 118, 121, 122). Interestingly, obese diabetic (db/db) mice, characterized by increased body weight, hyperglycemia and hyperlipidemia, display cardiac fibrosis in the absence of myofibroblast conversion, suggesting the activation of alternative matrix-synthetic programs in fibroblasts (112, 123).

Macrophages are the predominant immune cell type in the resting heart and have an important role in the regulation of tissue homeostasis (124). In response to chronic nutrient overload, resident macrophages expand and interact with other cardiac cell types via paracrine mechanisms (49–51, 124, 125). In fact, myocardial remodeling observed in patients with obesity or diabetes is largely mediated and amplified by cardiac macrophages (69). Increased levels of circulating nutrients (i.e., glucose and FFAs) alter macrophage function favoring their polarization from a regulatory (M2) toward a pro-inflammatory (M1) phenotype through different mechanisms (60, 126, 127).

First, overnutrition leads to global insulin resistance accompanied by chronic hyperglycemia, which promotes increased glucose uptake by macrophages via the insulin-independent GLUT1. In contrast to cardiomyocytes, macrophages are not sensitive to insulin and maintain glucose uptake in insulin resistant states (69, 128). Elevated intracellular glucose availability shifts the macrophage metabolism toward glycolysis and away from oxidative phosphorylation leading to increased pro-inflammatory gene expression (129). In parallel, the pentose phosphate pathway is activated and generates nicotinamide adenine dinucleotide phosphate (NADPH) which supports the synthesis of inflammatory prostaglandins and leukotrienes, thus activating NF-κB (69, 129).

Second, obesity and diabetes are both associated with elevated circulating and cardiac lipid levels which act as extra- and intracellular pro-inflammatory signaling molecules (17, 69, 130). Saturated fatty acids drive inflammatory responses in macrophages mediated by TLR4 on the cellular surface (14, 59). Multiple studies have demonstrated that long-chain saturated fatty acids (e.g., palmitic acid), but not short-chain saturated fatty acids or long-chain unsaturated fatty acids, induce the expression of inflammatory cytokines in macrophages (e.g., TNF-α) via the JNK signaling pathway in a TLR4-dependent manner (59–61, 126). Mechanistically, it is uncertain whether this effect is mediated by direct binding of FFAs to TLR4 or by an indirect TLR4-dependent mechanism—with a recent systematic study indicating the latter (61). Within the cell, saturated fatty acids also activate the NLRP3 inflammasome via an AMP-activated kinase (AMPK)-dependent pathway hinging on mitochondrial ROS production and cause IL-1β and IL-18 synthesis (131). In addition, excess intracellular fatty acid availability promotes anabolic pathways in macrophages including triacylglycerol, phospholipid, and ceramide synthesis (69, 132). Fatty acid-derived ceramide production activates the NLRP3 inflammasome thereby promoting lipotoxicity and M1 polarization (132, 133). Moreover, oxidized low-density lipoprotein (LDL) induces CD36-dependent mitochondrial ROS production in macrophages which facilitates NF-κB activation and inflammatory cytokine generation (69, 134). In aggregate, inflammatory processes in macrophages are tightly coupled to nutrient metabolism and therefore dictated by the availability of energetic substrates.

Cardiomyocytes are exposed to a broad range of cytokines originating from other cardiomyocytes, non-cardiomyocyte cardiac cells, and extracardiac tissues (135). Inflammatory cytokines such as TNF-α and IL-6, highly abundant in obesity and T2D, bind to receptors on the cardiomyocyte surface which triggers downstream activation of NF-κB and other central regulators of cell metabolism with differential impact on cardiomyocyte function (7–9, 86, 136–138). A number of deleterious effects of inflammatory cytokines on cardiomyocytes have been documented, namely cardiomyocyte hypertrophy, progressive cardiomyocyte loss through apoptosis, activation of fetal gene programs, impaired contractility, and increased passive tension (Table 1) (35, 36, 137, 139–141).

TNF-α exerts intracellular effects via binding to two different cell surface receptors, TNF receptor (TNFR)1 and TNFR2, both of which are expressed in cardiomyocytes (204). Exposure to TNF-α stimulates protein synthesis and blunts protein degradation in cardiomyocytes leading to cell hypertrophy (35, 142) via Akt/NF-κB and JNK activation (205). IL-1β induces cardiomyocyte hypertrophy through (1) direct interaction with cardiomyocytes (154), and (2) signal transducer and activator of transcription (STAT)3-dependent induction of insulin-like growth factor (IGF)1 by cardiac fibroblasts (155). IL-6 induces cardiomyocyte hypertrophy through Ca²⁺/calmodulin-dependent protein kinase (CaMK)II-dependent activation of STAT3 (160). IL-18, another upregulated pro-inflammatory cytokine in patients with obesity and T2D, induces cardiomyocyte hypertrophy via PI3K/Akt/GATA binding protein (GATA)4 signaling (167). In line with in vitro results, a number of studies have confirmed the role of pro-inflammatory cytokines in cardiac hypertrophy in vivo. Administration of TNF-α (35, 206) and IL-1β (141, 207) leads to left ventricular (LV) hypertrophy and dysfunction in rodents. Conversely, genetic deletion of TNF-α (208) and IL-1β (155) reduces LV hypertrophy and dysfunction in response to pressure overload. Marked LV hypertrophy and impaired diastolic relaxation and can be induced by infusion of IL-6 (209). Conclusively, IL-6 knockout attenuates myocardial hypertrophy and improves diastolic function in response to pressure overload (160).

Inflammatory cytokines modulate a range of processes controlling cardiomyocyte apoptosis. Sustained TNF signaling induces apoptosis via activation of intrinsic and extrinsic cell death pathways leading to activation of caspases-9 and−3 via cytosolic upregulation of cytochrome c, second mitochondria-derived activator of caspase (Smac), and HtrA serine peptidase (HtrA)2, and to cleavage of BH3 interacting domain death agonist (Bid) to truncated (t-)Bid, respectively (144). IL-1β promotes cardiomyocyte apoptosis (1) by induction of inducible nitric oxide synthase (iNOS) and subsequent generation of oxygen free radicals that alter the cellular balance of BCL2-antagonist/killer (Bak) and BCL-extra-large (Bcl-xL), and (2) by increasing endoplasmatic reticulum stress which promotes interleukin 1 receptor associated kinase (IRAK)2/C/EBP homologous protein (CHOP) signaling (156, 157). In addition, activation of the NLRP3 inflammasome induces cardiomyocyte cell death via caspase-1 (210).

Activation of the NLRP3 inflammasome/caspase-1 has also been liked to LV diastolic dysfunction. In diabetic cardiomyopathy, inhibition of caspase-1 leading to diminished IL-1β and IL-18 synthesis improves diastolic function and reduces myocardial fibrosis (54). Likewise, inhibition of IL-1β and IL-18 synthesis by knockdown of NLRP3 improves diastolic LV function in diabetic rats (211).

Alongside LV diastolic dysfunction, impaired systolic LV function is common in diabetic cardiomyopathy and obese patients with HFpEF (212, 213). In parallel, negative inotropic effects in vivo and in vitro have been reported for TNF-α, IL-1β, IL-6, and IL-18 (37). Cytokines mediate a rapid and reversible reduction of cardiomyocyte contractility by activating myocardial iNOS (143). Moreover, IL-1β and IL-6 decrease the expression of sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA)2a, which in turn may impair cardiomyocyte contractility through altered calcium handling (38).

Structural and functional abnormalities of the coronary microvasculature are propagated by chronic metabolic inflammation and can occur in the absence of macrovascular coronary artery disease (21, 213–217). The combination of systemic inflammation, hyperglycemia, and hyperlipidemia alters the release of vasoactive substances, such as NO, from the vascular endothelium leading to impaired smooth muscle relaxation and decreased myocardial perfusion (21, 147, 213–215, 217). Attenuated vasodilator response irrespective of macrovascular alterations is observed in subjects with diabetes (214, 218, 219), obese subjects with or without diabetes (220–223), subjects at increased risk to develop HFpEF (224) and patients diagnosed with HFpEF (213, 225, 226). Over the past two decades, the strong link between inflammation and microvascular dysfunction has been substantiated by a number of clinical studies (227–230). Notably, reduction in coronary flow reserve correlates with the degree of systemic inflammation assessed by CRP, IL-6, and white blood count (231). Likewise, in obese patients without coronary artery disease high circulating inflammatory markers (e.g., high sensitivity [hs]CRP, TNF-α, IL-6, and Leptin) associate with reduced coronary flow reserve (232, 233).

On a cellular level, TNF-α, IL-6, and leptin activate the NADPH-oxidase in the vessel wall leading to enhanced production of hyperoxide anion which in turn decreases NO availability and impairs vasodilation (147, 148). Moreover, obesity and diabetes-related microangiopathy is accompanied by microvascular rarefaction (222, 234–238). Reduction of coronary capillary density relative to cardiomyocyte surface area in turn promotes cardiomyocyte hypertrophy by decreasing NO-dependent protein kinase (PK)G activity (238). Depressed endothelial NO generation due to systemic inflammation has also been proposed as a leading cause of reduced cGMP-dependent PKG signaling in adjacent cardiomyocytes and impaired diastolic relaxation (7, 215).

Inflammatory processes in the myocardium of patients with the metabolic syndrome are amplified by endothelial activation and recruitment of circulating immune cells (38). Diet-induced obesity and diabetes alike enhance the expression of endothelial transmembrane proteins in the heart, such as VCAM-1 and ICAM-1, which facilitate leucocyte adhesion to the vascular wall and endothelial transmigration (50, 54, 239). Accordingly, both conditions are accompanied by increased abundance of cardiac macrophages and greater propensity of macrophages to assume a pro-inflammatory (M1) phenotype (47, 49, 50, 52, 53, 240). Mechanistically, circulating inflammatory cytokines (e.g., TNF-α, IL-1β) and, in advanced stages of myocardial functional impairment, elevated levels of Ang II induce upregulation of vascular adhesion molecules on the endothelial surface (215, 241–243).

During inflammatory states, cardiac cells secrete inflammatory and profibrotic cytokines which stimulate maladaptive remodeling through direct activation of fibroblasts and indirect effects (38, 112). TNF-α exerts multiple profibrotic effects including fibroblast activation and increased expression of TGF-β, the main profibrotic cytokine in the heart (38, 119, 151). Direct upregulation of collagen production by TNF-α via activation of WNT1 inducible signaling pathway protein (WISP) 1 has been reported (119, 151). Accordingly, pharmacological inhibition of TNF-α by monoclonal antibodies markedly reduces myocardial collagen I and III content and attenuates cardiac fibrosis in diabetic rats (54). Of note, TNF-α also activates matrix-degenerating programs in fibroblasts, such as the expression of matrix metalloproteinases (MMPs), suggesting that TNF-α mediated fibrosis may partly represent a response to ECM degradation (152). Another fibrogenic inflammatory mediator, upregulated under glucometabolic challenge by NF-κB activation, is IL-6 (14, 50). Abundant evidence indicates profibrotic effects of IL-6, mainly attributed to STAT3-dependent induction of collagen synthesis by cardiac fibroblasts and to enhancement of TGF-β expression (112, 164, 244). Genetic deletion of IL-6 mitigates cardiac fibrosis and dysfunction in diabetic mice (164). The inflammatory cytokine IL-1β is released upon activation of NLRP3/caspase 1 and is present in increased abundance in diabetic hearts (55). Il-1β has been implicated in cardiac fibrosis by exerting indirect profibrotic effects on fibroblasts via generation of ECM fragments and by induction of TGF-ß (112). Inhibition of caspase 1 reduces the biologically active form of IL-1β thereby improving cardiac fibrosis and LV function in diabetic rats (54).

Another process linking metabolic inflammation to myocardial fibrosis is endothelial-to-mesenchymal-transition. When exposed to inflammatory cytokines (e.g., TNF-α, IL-1β, IL-6, IL-13) or oxidized LDL endothelial cells adopt a fibroblast-like phenotype displaying mesenchymal cell morphology and function (245, 246). It has been suggested that endothelial-to-mesenchymal transition represents a general response to intracellular inflammation and may have a major role in cardiac ECM remodeling (245, 246).

AGEs are heterogenous molecules formed in a non-enzymatic reaction between the carbonyl group of a reducing sugar and the amino group of proteins, lipids, and nucleic acids. Chronic hyperglycemia leads to enhanced endogenous production and accumulation of AGEs in the cardiac ECM. Binding of AGEs to their cell surface receptor RAGE that is expressed in cardiomyocytes, fibroblasts, endothelial cells, and cardiac immune cells triggers the activation NF-κB via PI3K/Akt/MAPK (247, 248). The resulting pro-inflammatory state associates with enhanced intracellular ROS generation and alteration of cellular protein function (249). Overall, NF-κB activation by AGE-RAGE interaction leads to enhanced transcription and secretion of TNF-α, IL-1β, IL-2, and IL-6 contributing to the inflammatory milieu in the myocardium of hyperglycemic patients (250).

Several lines of evidence suggests that endocrine actions of pro- and anti-inflammatory adipokines in the systemic circulation along with paracrine effects of the epicardial adipose tissue contribute to myocardial inflammation (251, 252). Adipocyte hypertrophy promotes the secretion of leptin which has been linked to inflammatory effects in the myocardium and cardiomyocyte hypertrophy (14, 30, 253–255). In contrast, plasma levels of anti-inflammatory adiponectin are inversely correlated with body fat mass leading to reduced antagonism of inflammatory pathways in cardiometabolic patients (256). Adiponectin blocks TNF-α mediated activation of NF-κB through a protein kinase (PK)A-dependent mechanism (257). In addition, adiponectin potently stimulates ceramidase activity in cardiomyocytes and enhances ceramide catabolism thereby protecting from lipotoxic damage (258).

Given its anatomical intimacy with the underlying heart muscle and a shared microcirculation, the epicardial adipose tissue (EAT) is a pivotal regulator of myocardial inflammation (255). Unhindered passage of pro- and anti-inflammatory cytokines secreted by the EAT to the neighboring myocardium allows for paracrine interactions (251). At baseline, the EAT protects the myocardium form pro-inflammatory and hypertrophic stimuli through secretion of adiponectin (257–259). Along with growing body fat mass, macrophages are recruited to EAT where they foster local adipose tissue inflammation through upregulation of TNF-α, IL-6, IL-1β, and leptin while blunting the secretion of adiponectin (251, 260). Owed to its close proximity to the heart, EAT amplifies the effects of systemic metabolic disturbances on the myocardium (261). Clinically, EAT expansion correlates with elevated systemic inflammatory markers, increased LV mass index, abnormal coronary microcirculation (262), worsened parameters of diastolic function, and left atrial dilation—all features of the metabolic HFpEF phenotype (263–265). Besides, inflammation-induced invasion of pluripotent stem cells from the EAT to the outer myocardial layer and subsequent conversion to fibroblasts has also been proposed as mechanism of maladaptive myocardial remodeling (251).