Endothelial cells form the lining of all blood and lymphatic vessels within the vascular network. The adult human body contains more than one trillion ECs, which weigh more than 100 g and cover a surface area of more than 1,000 square meters (16). As such, ECs can be regarded as an organ with endothelium forming a dynamic interface with all other organs in the body (17). Normal ECs line capillaries and establish specialized vascular niches or tissue-specific endothelium. On receiving cues from a wide variety of cells and tissues, this “organ” maintains vascular homeostasis by regulating each component of the endothelial function (18, 19). The most important endothelial function is to control vasomotor tone. The healthy ECs constitutively generate eNOS-derived NO to modulate vascular smooth muscle relaxation, which properly regulates both nutrient trafficking and blood fluidity. A well-controlled EC-derived NO is also essential to other aspects of the endothelial function, such as the capability of maintaining an anticoagulant, antiplatelet, and fibrinolytic phenotype of the endothelium (5). It has been well established that loss of endothelial function facilitates the development of atherosclerosis (20). Therefore, endothelium operates as a complex organ to maintain whole-body homeostasis by fulfilling proper endothelial function, which can be impaired under disease conditions, such as diabetes and obesity.

Obesity contributes significantly to the development and progression of cardiovascular disease (21), which is believed to be mediated by oxidative stress, inflammation and endothelial dysfunction (22). Recent studies have highlighted the importance of metabolic pathways in ECs and suggested the therapeutic potential of targeting EC metabolism, since microcirculation is a key player in obesity-associated cardiovascular disease. Obesity elevates circulating free fatty acids and alters adipokines and inflammatory cytokines that are released from visceral and perivascular fat, which leads to structural and functional changes in microvasculature (23). Given the recognized differences in the metabolism of healthy ECs vs. diabetes-associated dysfunctional ECs, EC metabolism could be targeted for therapeutic benefit (24). For instance, by restoring diabetes-impaired glycolytic-influx and angiogenic response of the ECs, diabetic vasculopathies can be improved (24). Endothelial insulin resistance is associated with diabetic cardiovascular complications, including atherosclerosis (25). The role of endothelial insulin resistance in the development of obesity and systemic insulin resistance is not as straight-forward. Common treatments for obesity involving strategies to increase insulin sensitivity (26) have often been associated with improvement of endothelial function and/or endothelial insulin sensitivity (5). For instance, administration of metformin, which is known to enhance whole-body insulin sensitivity, improved endothelium-dependent vessel relaxation (27). Another excellent example is that physical activity has been found to modulate EC phenotype by overcoming obesity-induced endothelial dysfunction (28). However, emerging evidence support the concept of endothelial regulation of metabolism, e.g., through endothelial interactions with the insulin targets in obesity. Advancement in this research area may provide new strategies to prevent and treat obesity and obesity-related diseases.

Adipose tissue angiogenesis and vascular functions have long been associated with obesity, adipose metabolism, and insulin sensitivity (40–42). Angiogenesis and adipogenesis are tightly coupled during development (43) and postnatal fat expansion (44). In vitro studies have shown that vascular-derived signals, e.g., endothelial-VEGF, directly affect the proliferation and differentiation of the surrounding pre-adipocytes, suggesting that angiogenesis can be a therapeutic target for obesity and metabolic diseases (45). Pharmacological inhibition of angiogenesis has been reported in vivo for the first time to reduce fat mass in distinct obesity models, likely through different vasculature-dependent mechanisms (46). However, the physiological consequences of modulating angiogenic activity seem to be context-dependent: gain of function studies show that overexpression of VEGF-A improves metabolic profiles, which maintains tissue function during the early phase of diet-induced obesity (44). In contrast, reducing vascularization promotes dysfunctional adipose cell death which may reverse obesity by improving metabolic outcomes in established obesity (47). Similar beneficial results were reported in nanoparticles-targeting adipose tissue transformation and angiogenesis, which prevented obesity in HFD-fed mice (48). It is unclear whether any EC-derived factors are involved in these studies.

Adipose stromal cells (ASCs) are therapeutically potent progenitor cells because they possess properties of pericytes. ASCs in combination with EC establish functional multilayer vessels. Transforming growth factor (TGF)-β is secreted by EC (49) and is able to induce αSMA expression in ASC, a marker of smooth muscle cell differentiation. A recent study found that ECs initiate this differentiation program in co-cultured ASC and propagate the program in distant ASC by induction of Activin A but not TGF-β (50, 51). Activin A is a secreted protein and a member of the TGF-β family that has pleiotropic effects on regulating apoptosis, proliferation, and cell differentiation (52). How ECs induce Activin A in ASC is yet to be determined (53).

Endocrine fibroblast growth factors (FGFs), such as FGF21, FGF15/19, and FGF23, are critical for maintaining whole-body homeostasis, because they operates in inter-organ endocrine signaling pathways that govern glucose, bile acid, and lipid metabolism (54, 55). FGF21 has gained much attention because of its favorable pharmacological properties in glucose and lipid metabolism (56). FGF21 is expressed in the liver, pancreas, adipose, and skeletal muscle, which can be altered by different stimuli, such as starvation or overfeeding (57). The role of endogenous FGF21 has just emerged (58). A recent study in HFD-fed mice (59) implicated that endothelial-derived FGF21 may contribute to improved metabolic profiles in mice lacking the autophagic protein LC3, although confirmative evidence is needed (e.g., demonstration of FGF21-dependency using FGF21-blocking antibodies (60) or endothelial FGF21 knockout mice). The role of autophagic proteins in the regulation of adipose tissue is not fully understood (61). If the role of LC3 is supported, it would be interesting and important to determine whether this feature also applies to other EC-derived FGFs.

MicroRNAs recently link ECs to adipocyte function in diabetes and CVD (62, 63). One example is miR-181b, which was downregulated in adipose ECs from mice fed an HFD (64). Downregulation of miR-181b induced insulin resistance and low-grade inflammation in adipose tissue, which can be reversed by genetic delivery of miR-181b. Mechanistically, miR-181b targets the PH domain and leucine-rich repeat protein phosphatase-2 in ECs and, thus, improves eNOS-NO signaling. As such, adipose tissue ECs promote glucose uptake in adipocytes in a paracrine manner (64). miR-181b also decreased protein levels of vascular cell adhesion molecule and intercellular adhesion molecule, consistent with previous studies on systemic delivery of miR-181b (65, 66). Other miRNAs, such as miR-126, have been shown to preserve normal endothelial function, likely via blocking unwanted endothelial activation (67). These findings highlight the role of adipose ECs in the development of obesity-induced insulin resistance and the potential of using miRNAs as a tool to modulate EC function (68).

Endothelial nitric oxide synthase (eNOS)-derived NO, the major endothelial-derived vasodilator, is central to endothelial regulation of insulin sensitivity in skeletal muscle, by stimulating blood vessel relaxation (in vascular smooth muscle cells) and perfusion (in skeletal muscles) (75). In addition, eNOS-derived NO regulates Akt signaling in skeletal muscle, through a cGMP-PI3K-dependent pathway (76). Interestingly, rapid formation of capillary ECs occurred in rat skeletal muscle after exposure to insulin (77), supporting the notion that insulin is a vascular hormone (78). However, the normal response of skeletal muscle capillary to NO is impaired in obesity-induced insulin resistance (79). Physical training appears to be able to improve it, likely because both endurance training and high intensity training with intervals increased eNOS protein contents specifically in the endothelium of capillaries and arterioles of skeletal muscle in previously sedentary lean and obese young men (80).

There is evidence indicating that insulin delivery to skeletal muscle interstitium through the ECs is rate-limiting in insulin-stimulated glucose uptake in mice (74, 81) and humans (82). This process seems to be impaired by insulin resistance in type 2 diabetes and obesity (83). Mechanistically, damage to the endothelial glycocalyx barrier in skeletal muscle is believed to be an early event, as shown in mice fed an HFD (84). A recent study showed that endothelial insulin resistance also plays a pivotal role in the regulation of glucose uptake by the skeletal muscle (83). Mechanistically, HFD-impaired insulin signaling in EC downregulated insulin-induced eNOS phosphorylation and attenuated capillary recruitment and insulin delivery, which, in turn, reduced glucose uptake by the skeletal muscle (83). Conversely, restoration of the insulin-induced eNOS phosphorylation in ECs normalized capillary recruitment and insulin delivery, which restored glucose uptake by the skeletal muscle (81).

Fatty acids represent a key energy source that is used by a number of tissues, which must be tightly controlled to avoid lipotoxicity induced by an excess of unoxidized fatty acids (85). Obesity often causes over-accumulation of lipids in non-adipose tissues (e.g., skeletal muscle), which can trigger a toxic reaction and impair the function of the tissues, contributing to insulin resistance (86). Previous studies have identified endothelium as a key regulator of fatty acid transport, mediated by a complex signaling pathway including VEGF-B, PPARγ, and Apelin (87). In a recent study, endothelial Fcγ receptor IIB activation was shown to blunt insulin delivery to skeletal muscle, causing insulin resistance in mice (88). A study found that the capillary endothelial fatty acid binding proteins 4 and 5 (FABP4/5) are required for fatty acid uptake in heart and skeletal muscle, organs that possess muscle-type continuous capillary (89). In contrast, the function of an EC-derive mitogen activated protein kinase kinase kinase kinase 4 (MAP4K4) has opposing functions in blood and lymphatic ECs: while MAP4K4 is required for lymphatic vascular integrity, blood EC-MAP4K4 induces insulin resistance by impairing vascular function in obesity (90).

Obesity and insulin resistance can cause NAFLD, which is characterized by liver metabolic dysregulation, such as hyperlipidemia (95). Recent epidemiology of NAFLD and its connection with CVD suggest that blocking endothelial dysfunction would be beneficial (96), since endothelial function is impaired in patients with NAFLD (97).

Liver sinusoidal endothelial cells are highly specialized ECs, which form the wall of liver sinusoids and represent up to 20% of liver cells (98). LSECs are well known for their crucial roles in liver regeneration (99–101) and chronic liver diseases associated with obesity (e.g., fibrosis) (102). LSECs maintain hepatic stellate cells and hepatocytes in quiescence at basal condition but induce hepatocyte proliferation and angiogenesis during regeneration and undergo morphological and functional changes during fibrosis (103). It is believed that LSECs accomplish these by interacting with other hepatic cell types to maintain liver homeostasis. Since healthy LSECs facilitate the rapid transfer of fatty acids formed from cholesteryl esters to parenchymal cells, LSECs injury contributes to NAFLD (104). LSECs injury has been associated with NAFLD progression, where NAFLD may alter LSECs-mediated transport of nutrients, lipids, and lipoproteins. A recent study in choline-deficient, l-amino acid-defined, and high-fat diet models that mimic human NAFLD indicated that LSEC injury functions as “gatekeeper” in the progression from simple steatosis to the early non-alcoholic steatohepatitis stage, and LSEC injury may activate Kupffer cells and hepatic stellate cells, which in turn results in chronic liver injuries (105). In this regard, LSECs operate as a traffic-director at the crossroad of regeneration and fibrosis (106). It is yet to be determined whether LSEC-derived factors are required for this function.

Iron overload often accompanies NAFLD. However, the key molecules involved in NAFLD-associated iron dysregulation have not been fully elucidated. A recent study found that LSEC secreted a bone morphogenetic protein-binding endothelial regulator that inhibited BMP-SMAD signaling in hepatocytes and reduced hepcidin protein expression (107). Fatty liver is the hepatic component of insulin resistance before developing NAFLD (108). In an attempt to identify physiological consequences of mice missing mucin-type O-glycans that are highly expressed in vascular ECs, a recent study found that the pups developed fatty liver disease, due to disconnected portal vein and intestinal lymphatic systems, leading to chylomicron deposition (109).