Adiponectin Synthesis, Secretion and Extravasation from Circulation to Interstitial Space

Abstract

Adiponectin, an adipokine that circulates as multiple multimeric complexes at high levels in serum, has antidiabetic, anti-inflammatory, antiatherogenic, and cardioprotective properties. Understanding the mechanisms regulating adiponectin’s physiological effects is likely to provide critical insight into the development of adiponectin-based therapeutics to treat various metabolic-related diseases. In this review, we summarize our current understanding on adiponectin action in its various target tissues and in cellular models. We also focus on recent advances in two particular regulatory aspects; namely, the regulation of adiponectin gene expression, multimerization, and secretion, as well as extravasation of circulating adiponectin to the interstitial space and its degradation. Finally, we discuss some potential therapeutic approaches using adiponectin as a target and the current challenges facing adiponectin-based therapeutic interventions.

Adiponectin Overview

Circulating Adiponectin Correlation with Disease States

 

The adipose tissue secretes a variety of peptide hormones, collectively called adipokines. In 1995 leptin was the first adipokine discovered, which was shown to be a critical regulator of appetite and, thus, energy intake (1). Following this, many other adipokines, and indeed organokines, were discovered, and in this review, we will focus specifically on adiponectin (Ad) (2, 3). Adiponectin was discovered between 1995 and 1996 by four independent groups using different approaches (4–7). Over the last 20 years, an impressive amount of literature has highlighted multiple intriguing aspects of the molecule’s structure and function (3, 8–10).

Adiponectin circulates in high levels in both rodents and humans (2–20µg/ml) and is found in multiple multimeric complexes, including hexameric and trimeric forms (11, 12). Various studies have demonstrated that adiponectin has a range of physiological properties, such as insulin-sensitizing, anti-atherogenic, and anti-inflammatory (13). Furthermore, chronic inflammation of metabolic disorders such as Type 2 diabetes, obesity, and cardiovascular disease, are correlated with lower levels of circulating adiponectin (14–17). When it comes to inflammation, adiponectin acts as a double-edged sword, with both anti-inflammatory and proinflammatory properties. On one side, adiponectin plays an anti-inflammatory role in diseases, such as obesity, Type 2 diabetes, and atherosclerosis (13). On the other hand, adiponectin acts as a proinflammatory agent for the development of rheumatoid arthritis, chronic kidney disease, and bowel disease (18). Recently, adiponectin has also been shown to play an important role in promoting longevity and prolonging lifespan (19–21).

Therefore, the establishment of these strong correlations with diseases indicates that understanding adiponectin physiological functions in various metabolic tissues will allow further comprehension of the disease pathogenesis and allow opportunities for therapeutic intervention in their progression and prevention.

Insight on Adiponectin Action Derived from Animal Models

To further understand adiponectin metabolic effects, various animal models have been established investigating the downstream effects of adiponectin signaling and function in different tissues. For instance, Ad-knockout (KO) mice present a fairly normal phenotype until challenged with a high-fat diet, and later develop metabolic perturbations, including insulin resistance (22). Additionally, adiponectin receptors (AdipoR) AdipoR1 and AdipoR2 knockout mice present increased lipid accumulation in various tissues, indicating the important relevance of the main mediators of adiponectin action in the body (23, 24).

Adiponectin has been extensively documented to mediate various cardioprotective effects (25), many mediated via AMP-activated protein kinase (AMPK) (26–28). For instance, cardiac remodeling and dysfunction in obesity and diabetes are associated with altered levels of adiponectin (25). Botta et al. (29) have demonstrated that adiponectin protects from high-fat diet (HFD)-induced cardiac insulin resistance. In Ad-KO mice fed with an HFD, it was observed that they presented insulin resistance and cardiomyocyte lipotoxicity, which was reversed by replenishing normal circulating adiponectin levels, which led to activation of AMPK signaling (29).

Among the various important remodeling events that characterize heart failure, autophagy is of growing interest. Early studies in genetic mouse models of autophagy deficiency demonstrated that decreased autophagic flux induction was positively correlated with the development of heart failure (30, 31). Interestingly, in an Ad-KO mouse subjected to aortic banding to induce pressure overload (PO) and heart failure, it was observed that Ad-KO animals exhibited less PO-induced cardiac autophagy compared to wild type (32). Additionally, it was evident the mitochondrial dysfunction was exacerbated in Ad-KO mice. This phenomenon may result in a lack of proper mitophagy in Ad-KO mice, which would usually clear out damaged mitochondria. Therefore, adiponectin may directly stimulate autophagy flux. Without adiponectin and in response to PO in aged Ad-KO mice, the cellular events that contribute to the development of cardiac dysfunction may be exacerbated (32).

Adiponectin is also expressed in muscle, and its intramyocellular localization, especially in the glycolytic type II fibers, is positively correlated with increased intramyocellular accumulation (33). In an Ad-KO mouse model, it was observed that lack of adiponectin resulted in contractile dysfunction and phenotypical changes in the skeletal muscle (34). Furthermore, male Ad-KO mice exposed to a HFD and treated with and without adiponectin injections presented muscle insulin resistance (35). However, adiponectin treatment reversed HFD-induced insulin resistance and defects in insulin signaling. Additionally, adiponectin replenishing also improved pathological mitochondrial morphology changes induced by HFD in the muscle (35). In a muscle-specific AdipoR1-knockout (muscle-R1KO) mouse model, Iwabu et al. (36) also demonstrated that muscle-R1KO animals presented lower AMPK phosphorylation concurrent with significant high plasma glucose and insulin levels, suggestive of insulin resistance.

Similar to the effects observed in the heart, one of the molecular mechanisms proposed by which adiponectin promotes antidiabetic properties is through promotion of skeletal muscle autophagy and antioxidant potential to reduce insulin resistance caused by HFD. Muscle samples were analyzed from an Ad-KO mouse model, where administration of recombinant adiponectin alleviated the HFD-induced insulin resistance effect. Moreover, HFD-induced autophagy flux was observed in wild-type mice, but not in knockout mice, suggesting that the autophagic flux activation in muscle is also dependent on adiponectin (37).

In the liver, adiponectin increases insulin sensitivity and promotes antidiabetic effects. In an Ad-KO mouse model fed with HFD, animals presented hepatic lipid accumulation and insulin resistance. These effects were attenuated by adiponectin supplementation, which also contributed to enhancing fatty acid oxidation (38). Finally, it has also been observed that lack of adiponectin contributes to endothelial activation and exacerbation of sepsis-related mortality. Sepsis has multiple contributing factors and often results in death (39). Therefore, adiponectin deficiency may predispose patients to sepsis-related complications in late stages of obesity, diabetes, and insulin resistance.

Direct Effects of Adiponectin on Cellular Models

Adiponectin mediates several tissue-specific signaling pathways. Some of the major direct effects of adiponectin on cells are targeted to the adipose tissue, skeletal muscle, vascular endothelium, macrophages, and insulin-sensitizing actions (13). The primary mechanism of action of adiponectin is via membrane receptors binding, such as AdipoR1, AdipoR2, and T-cadherin (40, 41). To induce a response in any target tissue, adiponectin must first exit circulation and cross the endothelial barrier (42). The two central adiponectin receptors are the AdipoR1 isoform, which is mostly found in skeletal muscle, and AdipoR2 isoform, found in the liver (23, 43).

Alterations in adiponectin expression have been associated with a positive correlation between receptor expression and insulin resistance (44). A crucial downstream signaling effector of adiponectin is AMPK (41). For instance, AdipoR1 KO mice presented glucose tolerance and impaired AMPK activation (24), while AdipoR2 KO mice showed better glucose tolerance and lipid metabolization (45). The main mechanism of adiponectin-induced glucose uptake in skeletal muscle has been highly elucidated (46). Briefly, adiponectin binds to cell surface AdipoRs and, subsequently, binds to the adaptor protein (APPL1), resulting in activation of AMPK and translocation of glucose transporter-four (GLUT4) to the plasma membrane (47, 48). Adiponectin can also activate insulin receptor substrate-one (IRS1) phosphorylation, and downstream PI3K-AKT axis, to promote cell survival in muscle and hepatocytes following stress (49). It also mediates activation of various pathways that contribute to increased fatty acid oxidation, including PPAR-α, AMPK, and p38 MAPK (50, 51).

On adipose tissue, adiponectin function is mostly related to the control of the adipose endocrine system. Adiponectin represses secretion of leptin and proinflammatory cytokines, such as IL-6 and TNF-α via AMPK and NF-κB signalling, therefore reducing its own expression level (52, 53). Similarly, on macrophages, adiponectin has anti-inflammatory effects by repressing M1 macrophage (proinflammatory) activation, and favors activation of M2 macrophage (anti-inflammatory) (54). Furthermore, in human macrophages, adiponectin also activates anti-inflammatory factors (IL-10) and decreases proinflammatory ones (IFN-γ, IL-6, and TNF-α) (55). Finally, vascular protection and endothelium relaxation are promoted by nitric oxide (NO) production activated by AMPK-endothelial nitric oxide synthase (eNOS) (56). In addition to that, adiponectin contributes to NO production and suppression of ROS and NF-κB (57). Therefore, adiponectin regulates vascular homeostasis on endothelial cells via activation of cAMP-PKA and AMPK signalling.

As noted, there is a controversy as to whether adiponectin acts as an anti- or proinflammatory mediator (54, 58–67). Some studies report that adiponectin functions as an anti-inflammatory mediator during the progression of metabolic diseases (62, 67). In agreement with these reports, extensive evidence shows that adiponectin acts as anti-inflammatory mediator through the regulation of M1 and M2 macrophage proliferation, plasticity and polarization (55, 63, 68–71). In contrast, other studies found that adiponectin promotes an inflammatory response by inhibiting group two innate lymphoid cell (ILC2) function, activating NF-κB and inducing inflammatory cytokines IL-1 and IL-6 under certain circumstances (64, 72, 73). The regulatory effect of adiponectin on innate immunity under physiological conditions may emerge as an important determinant of metabolic adaption. During the development of obesity, the decreased levels of adiponectin by obesity in circulation favor its anti-inflammatory action and subsequent insulin-sensitizing effect. Whereas adiponectin behaves as an initial proinflammatory factor in response to LPS and helps to desensitize macrophages to further proinflammatory stimuli (64, 72, 74). In agreement with this, adiponectin suppresses cold stress-induced ILC2 activation and type 2 inflammatory response through an AMPK-mediated feedback inhibition of IL-33 signaling, acting as a molecular brake on thermogenesis (73). A better understanding of the physiological role of adiponectin in the regulation of innate immunity may provide a basis for the development of adiponectin-based therapeutic strategies (54).

Summary of Adiponectin’s Principal Physiological Effects

Adiponectin has been demonstrated by various studies to promote antidiabetic effects. The majority of adiponectin in circulation is derived from adipocytes. However, other tissues, such as skeletal muscle, also produce adiponectin. Most adipokines present a proinflammatory action, such as tumor necrosis factor-α, adipocyte fatty acid-binding protein and lipocalin-2, but contrary to all other adipokines, adiponectin promotes anti-inflammatory effects, is antidiabetic and cardioprotective, and plays an important role in cancer (75–77).

The metabolic effects of adiponectin have been well characterized in skeletal muscle and liver, and other observations also indicate similar effects in the heart, where it enhances glucose uptake and fatty acid metabolism (78). The majority of adiponectin action in all target tissues has been shown to be regulated by AMPK, which plays an important role in regulating a wide range of physiological effects of adiponectin on glucose metabolism (79). Furthermore, exercise has been shown to be highly effective as an insulin sensitizer to promote glucose uptake in skeletal muscle, increase fatty acid oxidation, and reduce obesity (80). Exercise may also facilitate adiponectin action by upregulating AdipoRs expression, therefore, increasing the glucose-sensitizing effects (81, 82). A recent human study in obese and prediabetic patients demonstrated that intensive interval training could elevate circulating adiponectin levels, which may be clinically relevant in aiding to reduce the cardiometabolic risk (83).

Besides its physiological relevance in the muscle, heart, and liver, adiponectin also plays a role in respiratory complications (75). Decreased serum adiponectin levels have been correlated with deficient lung function and chronic obstructive pulmonary disease (COPD) (84). Additionally, recent studies have also shown a positive correlation between obesity and pulmonary hypertension (85). However, the mechanisms by which obesity drives pulmonary hypertension are still very complex and not well understood. It is well known that adiponectin plays a pleiotropic effect on inflammation and cell proliferation (86), representing a potential protective function on the pulmonary vasculature. Therefore, adiponectin could potentially become a therapeutic target for pulmonary hypertension treatment.

On the basis of all the physiological effects of adiponectin in various tissues, it is evident that it would make an attractive therapy option to treat various metabolic diseases. Currently, AdipoRon, a synthetic analog of endogenous adiponectin, presents great pharmacological properties mimicking adiponectin effects, which can bind and activate adipoR1 and adipoR2 receptors, contributing to the treatment of various metabolic diseases, including obesity, Type 2 diabetes, and cancer (87). Other therapies are also under current investigation that may be effective in increasing the adiponectin effect in the body; it has been extensively reviewed by Liu et al. (9).

Mechanisms Regulating Adiponectin’s Physiological Effects

Gene Expression and Secretion

Adiponectin is mainly produced from mature adipocytes in white adipose tissue. Originally thought to be exclusively expressed by adipose tissue; currently it is well established that adiponectin is also produced and secreted by numerous cell types, including skeletal and cardiac muscles (34, 88–91). It circulates in the plasma at very high concentrations (2–20 μg/mL), constituting approximately 0.01% to 0.05% of total serum proteins (92, 93). Recent studies have shown that adiponectin is secreted predominantly during the day, following a circadian-rhythm pattern. Even though the impact of the circadian rhythm on adiponectin expression is not yet well understood, some metabolic diseases, including obesity and insulin resistance, have been linked to disrupted circadian rhythm (33). Mice with high-fat diet-induced insulin resistance had significantly lower expression of adiponectin (94).

The regulation of adiponectin gene expression, specifically, is highly regulated by different transcription factors (97). Some of the transcription factors with active binding sites on both human and mouse adiponectin promoter include the peroxisome proliferator‐activated receptor (PPAR)‐response elements, C/EBP sites, FOXO sterol regulatory elements (SREs), and E‐box (98). The major positive regulator of adiponectin expression is PPARγ, mainly expressed in adipose tissue, in abundance (99, 100). Activation of PPARγ by antidiabetic drugs thiazolidinedione can stimulate adiponectin expression and secretion in adipose tissue of both humans and rodents (101). In contrast, the negative regulators of adiponectin expression levels are the inflammatory and obesity mediators known to induce insulin resistance, such as reactive oxygen species (ROS), TNF-α, and IL-6 (102, 103). Moreover, in vitro studies have demonstrated that adiponectin secretion in 3T3-L1 adipocytes is regulated by endothelin-1, an important vasoconstrictive agent that is increased in diabetes and obesity disease states (104). Therefore, various factors may be involved in the regulation of adiponectin gene expression and secretion, including vascularly derived factors, which represents an area to be further explored.

Recent studies have brought to attention the role of adiponectin gene DNA methylation in gestational diabetes (GDM). Children exposed to GDM are at a higher risk of becoming glucose intolerant later in life, among other metabolic complications (105–107). Early-life alterations of DNA methylation leads to disrupted adiponectin gene (ADIPOQ) expression, which was associated with alterations in maternal glucose levels, leading to hyperglycemia (108, 109). Although the precise molecular mechanism of how this happens is not really well known, epigenetics alterations around ADIPOQ are highly suggestive of being one possible mechanism involved in the metabolic perturbations observed in the offspring of GDM women (110). Furthermore, recent evidence shows that other chronic conditions during pregnancy, such as obstructive sleep apnoea, was associated with endothelial dysfunction, reduced perivascular adiponectin, and epigenetic modifications on the adiponectin gene promotor of adult male offspring (111).

Adiponectin exists as a full-length protein of 30 kDa, circulating in serum as three major isoforms: trimer, hexamer, and high-molecular weight (HMW) multimer (92). Different adiponectin multimers have been shown to exert distinct biological properties in their various target tissues (112). The HMW adiponectin is the major active form mediating its insulin-sensitizing effects, whereas the central action of adiponectin is attributed primarily to the hexameric and trimeric multimers (51, 113–118). A distinction in circulating levels of adiponectin has been observed in both humans and rodents, where total and HMW adiponectin is higher in females than in males. This is possibly because HMW adiponectin production is inhibited by testosterone in males (92, 119, 120). Also, the globular domain of adiponectin (∼18–25 kDa), which can be produced via the action of neutrophil elastase (121), has been shown to similarly present significant metabolic effects in various tissues (121). Many factors that regulate adiponectin gene expression and secretion can directly influence plasma levels of adiponectin, via the regulatory mechanisms at the levels of transcriptional, translational, and posttranslational modifications.

Posttranslational Modification and Multimerization

The selective reduction of the HMW adiponectin concentrations (or impaired adiponectin multimerization) was found to be associated with various metabolic diseases, such as obesity, insulin resistance, Type 2 diabetes, and arteriosclerosis (11, 113, 122–124). An increase in the HMW form, rather than the total level of adiponectin, was found to be associated with insulin-sensitizing effect of thiazolidinediones (TZDs) in both mice and human diabetic patients (124). Moreover, increasing the HMW form of adiponectin by fat-specific overexpression of DsbA-L-protected mice from diet-induced insulin resistance (125). In fact, it has been suggested that reduced levels of the HMW form rather than total levels of adiponectin could be a superior biomarker for insulin resistance, the metabolic syndrome, and Type 2 diabetes (113, 126).

The biosynthesis of the adiponectin multimers is a complex process involving extensive posttranslational modifications, as shown in FIGURE 1 (129). There is a conserved cysteine residue at the NH₂ terminus (Cys36 in human and Cys39 in mouse) and the formation of an intermolecular disulfide bond via this residue is essential for adiponectin multimerization and secretion (127). Cys39 was also identified as a key site of succination of adiponectin, which blocks adiponectin multimerization and may contribute to the decrease in plasma adiponectin in diabetes (128). In addition to disulfide bond formation, hydroxylation and glycosylation of several conserved lysine residues in the collagenous domain of adiponectin are necessary for the intracellular assembly and stabilization of its high-order multimeric structures (12, 129, 130). There are four conserved lysine residues in the collagenous domain of adiponectin (Lys65, Lys68, Lys77, and Lys101 for human adiponectin) and hydroxylation, and subsequent glycosylation at these sites are required for intracellular assembly of the trimer of adiponectin into the HMW multimer (12, 129). Whitehead and colleagues (131) identified three additional hydroxylations on Pro71, Pro76, and Pro95 in human adiponectin, including Pro71, Pro76, and Pro95, and demonstrate that inhibition of proline hydroxylation results in a more severe impairment of adiponectin multimerization .

FIGURE 1.

Adiponectin synthesis and multimerizationAdiponectin exists as a full-length protein of 30 kDa, circulating in serum as three major isoforms: trimer, hexamer, and high-molecular weight (HMW) multimer. Adiponectin is mostly secreted by the adipose tissue, highly regulated by the PPARγ signaling pathway. Hydroxylation and subsequent glycosylation of lysine residues in the collagenous domain of adiponectin are required for intracellular assembly of the trimer of adiponectin into the HMW multimer. Additionally, the multimerization and secretion of adiponectin are controlled by molecular chaperones in the ER (endoplasmic reticulum), including ERp44 (ER protein of 44 kDa), Ero1-Lα (ER oxidoreductase 1-Lα) and DsbA-L. The ERp44 forms a mixed disulfide bond (S-S) with adiponectin at the NH₂-terminal region and inhibits the secretion of adiponectin multimers. In contrast, Ero1-Lα enhances the secretion of HMW adiponectin through releasing adiponectin trapped by ERp44 and by transferring its own disulfide bond to protein disulfide-isomerase, allowing disulfide bond formation and adiponectin multimerization. Lastly, DsbA-L also play a role in specifically recognizing the LMW and MMW, and enhancing HMW assembly. This figure was created using Servier Medical Art (available at https://smart.servier.com/) and PAGES software.

The multimerization and secretion of adiponectin are also tightly controlled by a pair of molecular chaperones in the endoplasmic reticulum (ER), including ERp44 (ER protein of 44 kDa), Ero1-Lα (ER oxidoreductase 1-Lα), and DsbA-L, all of which are induced during adipogenesis, as shown in FIGURE 1 (125, 132, 138). ERp44 forms a mixed-disulfide bond with adiponectin through the cysteine residue (Cys36 in human and Cys39 in mouse) at the NH₂-terminal region and inhibits the secretion of adiponectin multimers through a thiol-mediated retention (132). In contrast, Ero1-Lα enhances the secretion of HMW adiponectin through releasing adiponectin trapped by ERp44 (133, 132, 138). Ero1-Lα also participates in adiponectin maturation by transferring its own disulfide bond to protein disulfide-isomerase (PDI) that is essential for the intermolecular disulfide bond formation and adiponectin multimerization (133). On the other hand, ERp44 was found to specifically recognizes the LMW and MMW in the cis-Golgi and enhance HMW assembly by regulating the population of adiponectin intermediates with appropriate oxidative state (134). Unlike Ero1-Lα and ERp44, DsbA-L does not contain CXXX motif, whereas it serves as an adiponectin interactive protein that selectively promotes adiponectin multimerization in adipocytes (125). In addition, DsbA-L alleviates ER stress-induced adiponectin downregulation via an autophagy-dependent mechanism and protects against obesity and ER stress-mediated suppression of adiponectin multimerization (135, 136). In support of this, the spliced form of XBP1s (X-box-binding protein 1), a key transcription factor of the ER stress response, promotes adiponectin multimerization through a direct regulation of the expression of several ER chaperones, including PDIA6 (protein disulfide isomerase family A, member 6) ERp44, and DsbA-L (137). The PPARγ (peroxisome-proliferator-activated receptor γ) agonists TZDs selectively enhance the secretion of HMW adiponectin through upregulation of Ero1-Lα and DsbA-L (125, 138). These data provide evidence for a link between impaired adiponectin multimerization and the causes of a diabetic phenotype in humans and suggest that not only total concentrations, but also multimer distribution, should always be a consideration in the interpretation of plasma adiponectin levels in health, as well as various disease states (139).

From the ER, adiponectin is engaged in the secretion pathway and undergoes multimerization into trimer, hexamer, and HMW adiponectin. Two pathways, multimerization and secretion, are coupled through shared components such as ER chaperones Ero1-Lα, ERp44, and DsbA-L (125, 132, 138). ERp44 is mainly localized in the ER-Golgi intermediate compartment/cis-Golgi region and retains adiponectin in the early secretory compartment (132, 133). Whereas Ero1-Lα is able to facilitate the secretion of HMW adiponectin through releasing adiponectin trapped by ERp44 (132, 138). Another chaperone DsbA-L promotes the secretion of HMW adiponectin through induction of adiponectin multimerization in adipocytes (125). However, what other factors are involved in adiponectin secretion is poorly understood. Future studies are also urgently needed to elucidate the precise mechanisms underlying the secretion of various multimeric forms of adiponectin given its promising therapeutic potential.

Adiponectin receptors, AdipoR1 and AdipoR2, have been found to bind to the globular form, as well as the full-length adiponectin, rapidly inducing intracellular signaling cascades (23, 48, 49). Despite the most active form of adiponectin, how HMW adiponectin functions in its target cells remains largely unknown. T-cadherin was identified as a selective receptor for the HMW adiponectin that potentiates adiponectin signaling (140). However, T-cadherin lacks the intracellular structural domain and appears to serve as a binding protein of adiponectin (141–143). The intracellular signaling events upon the binding of HMW adiponectin to T-cadherin remain to be defined.

Extravasation of Adiponectin from Vasculature to Interstitial Space

Structure and function of endothelium.

The endothelium consists of a monolayer of endothelial cells that act as an important barrier regulating various signaling molecules between the vascular lumen and the interstitial space (144). It also plays a range of roles, including maintenance of extracellular matrix, regulation of immune response, and regulation of cellular growth (144). The endothelial permeability can be regulated via two distinct mechanisms, transcellular and paracellular (FIGURE 2A). Transcellular pathway allows solutes to be transported actively across the endothelium. It is driven by active processes involving transport proteins, pumps, and receptors. Alternatively, the paracellular pathway allows solutes to move between adjacent endothelial cells (42). The regulatory mechanism of the paracellular movement requires the involvement of tight and adherent junctions, which have been shown to be essential in regulating hormone transport (145). An interesting distinction between both transport methods is that in the transcellular movement, larger macromolecules are carried across the endothelium via vesicle trafficking, while the paracellular route restricts the passage of large solutes, FIGURE 2B (146).

FIGURE 2.

Adiponectin extravasation into the circulationA: paracellular and transcellular movement. B: adiponectin extravasation from circulation into interstitial space on target tissue. Process mediated through paracellular movement across the endothelium. Decreased tightness of junctions allows higher amounts of full-length adiponectin extravasation into the interstitial space. This figure was created using Servier Medical Art (available at https://smart.servier.com/) and PAGES software.

Role of transcellular/paracellular transport.

An extensive literature currently indicates that endothelium not only plays a massive role in cardiovascular disease but also diabetes (144, 147–149). Interestingly, exocytosis of adiponectin from circulation has been shown to promote protective effects (3). Adiponectin must leave circulation into the interstitial space on target tissues in order to act, where vascular permeability may highly contribute to the regulation of adiponectin action, as illustrated in FIGURE 2B (150). Recent studies from Sweeney’s group demonstrated an underappreciated and novel link between vascular permeability and adiponectin flux during hyperglycemia or diabetes. In both in vivo and in vitro approaches, a correlation was observed between the tightness of an endothelial barrier upon hyperglycemic treatment. Reduction in tight junction proteins, such as claudin-7, significantly contributed to increased permeability in diabetic mice, allowing an increased flux of both low- and high-molecular weight proteins (150). Studies also suggest that another key player in adiponectin translocation may be due to AdipoR-mediated endocytosis (151). However, Yoon et al. (150) demonstrated that there were no significant changes in the levels of both AdipoR1 and AdipoR2 in the heart, following hyperglycemia.

Furthermore, altered transendothelial glucocorticoids are also great mediators of endothelium permeability (152). Recently, Sweeney’s group demonstrated that glucocorticoids may impair adiponectin action in target tissues by reducing the transendothelial flux of adiponectin across endothelial monolayers in vivo and in vitro (153). For instance, the adiponectin flux across the endothelial monolayers was associated with reduced claudin-10 mRNA levels in skeletal muscle of glucocorticoid-induced diabetic rats (153). Therefore, these all indicate that paracellular transport is a major mechanism in which glucocorticoid-mediated tightening can reduce adiponectin flux across endothelial monolayer cells. As a consequence, decreased adiponectin action due to reduced extravasation from circulation to interstitial space can contribute to a diabetic phenotype followed by glucocorticoid exposure.

Thus, adiponectin plays an essential role in vascular physiology by mediating cross-talk among endothelial cells, smooth muscle cells, leukocytes, and platelets (93). The endothelium-derived nitric oxide (NO) is an important factor promoter of vasodilation and inhibition of inflammation (144). More importantly, reduced NO is an early event in the atherosclerosis process; however, adiponectin acts via AdipoR1 and AdipoR2 to potentiate NO production via the AMPK signaling pathway (154). Endothelial dysfunction, characterized by impaired NO production, has been highly investigated and associated with not only cardiovascular disease, but also insulin resistance (155) and diabetes (148, 151). However, adiponectin can attenuate these metabolic defects via the cAMP/PKA pathway and inhibit proinflammatory actions via endothelial ROS inhibition (156). Currently, some antidiabetic drugs have been shown to protect the endothelium. For instance, there is strong evidence that metformin exerts a beneficial effect on endothelial function, as well as glucagon-like peptide 1 (GLP-1) agonists, and SGLT2 inhibitors (144, 157, 158).

Improving adiponectin access to interstitial space.

One possible mechanism of adiponectin action is primarily via binding of the membrane receptors, such as AdipoR1, AdipoR2, and T-cadherin. Alternatively, vascular permeability is another critical factor in determining adiponectin action (42). To promote an effect in a target tissue, adiponectin must first exit circulation into the interstitial space by crossing the endothelial barrier. Paracellular transport is one of the major determinants of adiponectin flux across the endothelium (FIGURE 2). However, little is known about its mechanisms, and alterations in the paracellular transport have been associated with endothelial dysfunction in diabetes.

The endothelial barrier that controls the adiponectin movement is cell and tissue specific. Previously, Dang et al. (153) demonstrated that glucocorticoids may impair adiponectin action in target tissues by reducing the transendothelial flux of adiponectin across endothelial monolayers in vivo and in vitro. However, in a more recent study, Yoon et al. (150) for the first time tracked adiponectin biodistribution via fluorescence rhodamine-labeled full-length adiponectin in streptozotocin-induced diabetic mice, where hyperglycemia increased vascular flux of adiponectin. Surprisingly, only a few studies have investigated adiponectin interstitial space levels, with two of them suggesting that exercise may contribute to increased interstitial adiponectin (159, 160).

Receptor Expression and Localization

AdipoR1 and AdipoR2 are widely expressed in various metabolic tissues, such as adipose tissue, liver, skeletal muscle, and the cardiovascular, as well as in the reproductive tract and brain. The AdipoR1 isoform is mostly found in skeletal muscle and AdipoR2 isoform in the liver (23, 43). The difference between the two is that AdipoR1 is a receptor for globular adiponectin, while AdipoR2 is a receptor for full-length adiponectin (23). Both receptors possess very similar homology at the protein level and contain seven transmembrane domains; however, structurally and functionally distinct from G protein-coupled receptors (8, 23). AdipoR1 and R2 contain a zinc-binding catalytic site and are closely located to the inner surface of the plasma membrane (161). The interaction between the ligand and receptor takes place between the globular domain of adiponectin and the extracellular surface of the receptors (23). Once adiponectin binds to the zinc‐binding motif of AdipoR1 and R2, a series of downstream signaling events are initiated in many target tissues, including skeletal muscle, liver, heart, kidney, and pancreas (162). Some of the major signals activated are AMPK, p38 mitogen-activated protein kinase (p38 MAPK), and higher activity of peroxisome proliferator-activated receptor alpha (PPARα) ligand, playing an important role in glucose and fatty acid metabolisms (23). These signals are regulated by the leucine zipper motif (APPL1), an adaptor protein discovered in 2006, which directly binds to intracellular regions of AdipoR1 and AdipoR2 (48). Another key regulator of adiponectin function is the T-cadherin receptors, discovered in 2004 (140). Even though T-cadherin receptor does not have a membrane or intracellular domain, it plays an important role in binding adiponectin and promoting adiponectin signaling in both smooth muscle and endothelial cells (140).

Degradation of Adiponectin

Adiponectin biosynthesis is well understood; however, its fate, once in circulation, is less well established. A decrease in plasma adiponectin levels is often associated with obesity; however, the underlying mechanisms remain unclear. In 2009, Shrerer’s group showed the clearance rate of adiponectin for the first time. The rate varies among different isoforms of adiponectin. For instance, HMW adiponectin is cleared the slowest and trimer the fastest (163). Also, the main site of clearance is the liver, while the degradation of final products is excreted by the kidney (163). Various studies have shown that reduced endogenous adiponectin levels might not be only a result of upstream defects at the exocytotic machinery. Defective adiponectin synthesis or a posttranslational modification could also sequester adiponectin to the endoplasmic reticulum (ER) to be degraded (7, 132, 134, 164).

In 3T3-L1 adipocytes, it was found that ER stress promotes autophagy-dependent adiponectin degradation, providing a mechanism underlying obesity-induced adiponectin downregulation (135). Even though inhibition of autophagy prevented adiponectin degradation, it did not rescue insulin sensitivity. Perhaps that is because of an adaptive role of autophagy during ER stress-induced insulin resistance (135). Conversely, our group extensively reported that autophagy is an important cellular mechanism induced by adiponectin (32, 37, 165). Recently, we demonstrated that adiponectin directly stimulated autophagy in skeletal muscle cells via an AMPK-dependent signaling mechanism and alleviated insulin resistance (37, 166). Gu and colleagues proposed an alternative signaling pathway that, once blunted, leads to adiponectin degradation. Rodents exposed to HFD developed obesity and a decline in plasma adiponectin, which was associated with suppression of ERK1/1 activation in adipocytes. Moreover, inhibition of the MEK/ERK1/2 pathway in adipocytes significantly decreased intracellular and secretory adiponectin levels, suggesting that it may promote adiponectin degradation (167).

Capitalizing on Knowledge of Adiponectin Action for Therapeutics

Recombinant Adiponectin

Adiponectin is primarily produced in adipocytes as a monomer and posttranslationally modified into diverse multimers. The recombinant adiponectin produced by Escherichia coli, on the other hand, is composed of only monomeric adiponectin, meaning that multimer formation requires posttranslational processing by mammalian adipocytes (8). The multimeric complexes highly contribute to the various biological effects of adiponectin (92, 127). In healthy individuals, the plasma concentration of adiponectin is usually high (∼5–30 μg/mL); however, in the case of obesity, the level of blood adiponectin is often lower, contributing to increase insulin resistance and Type 2 diabetes (168). Supplementation with recombinant adiponectin has been shown to improve insulin resistance, protect diabetic phenotype or metabolic challenges in both in vivo and in vitro studies (35, 169–171). However, there is controversy over whether recombinant adiponectin may be really effective in lowering blood glucose in diabetic animal models (172). Even though adiponectin has been suggested as a potent target for developing a therapeutic agent for treating metabolic diseases, such as diabetes, the preparation of recombinant adiponectin has been a challenge due to posttranslational modifications. As a result, currently, no drug targeting AdipoRs has yet been approved.

Adiponectin Peptides

AdipoR agonists, for a long time, have been the focus of research and development programs in cardiovascular and metabolic diseases (10, 165). Designing adiponectin agonists that activate adiponectin receptor-mediated downstream signaling is highly desired; however, it poses great challenges in the production of biologically active adiponectin and optimization of the right dose and via of administration. Since the discovery of AdipoR1 and AdipoR2 by Kadowaki’s group (23), various small molecules or peptides have been discovered (10). Recently, an adiponectin-mimetic peptide, ADP355 (174) and a small molecule, AdipoRon (173), have been developed as agonists to AdipoRs.

To date, ADP355 is the first synthetic adiponectin-mimetic peptide composed of non-natural amino acids (174, 175). Initially intended for development for cancer treatment, ADP355 presents the ability to activate the AdipoRs-mediated AMPK pathway, which represents an opportunity for the use of synthetic peptides in developing agonists against AdipoRs. On the basis of the success of the first generation of AdipoR agonist ADP355, Otvos et al. (175) developed a second generation of the peptide, the dimeric ADP399. The improved ADP399 had a 20-fold increase in activity in comparison to the monomeric peptide ADP355 (175). Moreover, to counteract ADP355 and ADP399, a novel octapeptide (ADP400) was developed after several screenings (10, 175). ADP400 antagonizes endogenous adiponectin and acts as an inverse agonist. Therefore, at a similar dose as the agonist ADP399, the ADP400 peptide showed greater antagonist activity, representing an important target as a validation tool to further study adiponectin functions (175).

Although the adiponectin- and AdipoR1-bound structure is not available, the extracellular COOH-terminal domain has been identified to be critical for adiponectin binding to AdipoR1 (24).

Recently, Kim’s group (168) identified the ligand-binding regions in AdipoR1 and designed peptides that stimulate interaction in those regions. After several rounds of designing and screening, the BHD1028 was identified as the most potent peptide in activating AMPK (160). Even though further clinical studies are needed to validate its efficacy, BHD1028 already presents a potential drug candidate for treating diabetes and metabolic diseases. Finally, another peptide Pep70 has also been ultimately identified as an inhibitor of fibrotic responses and as a potential AdipoR1 agonist (176). However, it was concluded that more studies with Pep70 are necessary to achieve a higher level of bioactivity.

Adiponectin Gene Therapy

As indicated above, in Type 2 diabetes patients’ adiponectin levels are reduced. A study was designed to investigate whether an injection of plasmid DNA-encoding adiponectin promotes any therapeutic effects on increasing adiponectin levels (177). In a nonobese Type 2 diabetes mouse model, it was observed that they significantly had higher blood adiponectin levels and decreased glucose concentrations following adiponectin injections (177). Adiponectin adenoviral injections also decreased blood pressure levels and increased circulating adiponectin in obese hypertense mice and salt-fed Ad- KO mice (57). Furthermore, Dyck’s group concluded that injectable adiponectin gene therapy could reduce fat and improve insulin sensitivity in HFD-treated mice (178).

One of the most attractive forms of gene delivery is through synthetic polymer vectors because of its relative safety and practicality. The downside is its low efficiency; for this reason, successful applications have focused on therapy with genes that at low concentrations promote biological effects (179). As adiponectin is a highly abundant circulating protein, it represents a great candidate for this nonviral gene delivery. Banerjee et al. (180) recently developed a gene-based targeted nanoparticle formulation capable of improving insulin sensitivity in Type 2 diabetes by stimulating adiponectin production in adipose tissue and increasing its circulation levels in the body for an extended period. In their work, they used free amino groups on chitosan-oleic acid polymer conjugated to adipose homing peptide to enhance its internalization into adipose tissue. This delivery approach has been previously described and validated by various studies (181–183). Adiponectin gene therapy could be a clinically useful tool for the treatment of Type 2 obesity and endothelial dysfunction. However, more studies are needed as efficiency, stability, specificity, safety, and convenience are key factors that require careful consideration before the clinical application of any gene delivery system.

Inducing Adiponectin Expression/Secretion

Overcoming the reduced availability of circulating adiponectin levels in various disease states requires interventions that enhance adiponectin receptor expression. The expression of adiponectin receptors can be increased by lifestyle modification, such as diet, exercise, and pharmacological therapy (9). In obese patients, weight loss induced by hypocaloric diet and physical exercise promoted an increase in adiponectin levels in circulation, enhancing glucose utilization, lipid oxidation, and energy expenditure (82, 184–186). Both human and animal studies demonstrate that intense exercise training potentiates the increase in AdipoR1 in skeletal muscle (82, 187). Furthermore, in rodent models of Duchenne muscular dystrophy, moderate exercise plays a significant role in improving muscle performance and increasing serum adiponectin. Moreover, serum adiponectin could be used as a biomarker for the benefits of exercise in Duchenne muscular dystrophy (188). Collectively, these studies support the idea that consistent weight loss and lifestyle intervention, such as exercise, contribute to elevating adiponectin levels.

Thiazolidinediones (TZD), an antidiabetic drug, is also a great regulator of adiponectin expression. The major mechanism of TZD- induced adiponectin expression is through activation of proliferator-activated receptor gamma (PPARγ). In human and animal studies of subjects treated with TZD presented increased adiponectin levels (189–191), concurrent with improvement in insulin sensitivity (192). However, the major challenge in relying on TZD as an agent to increase adiponectin levels is its adverse effects, including but not limited to, heart failure and cardiac ischemia (193–195). Moreover, incretin hormones, such as GLP-1 analogs, which are used to increase incretin availability, also upregulate adiponectin expression both in vitro (196) and in clinical trials (197). Evidence from human studies demonstrates that by targeting the renin-angiotensin-aldosterone system through angiotensin-converting enzyme inhibitor (ACEi) and angiotensin blockers (ARBs) significantly increased adiponectin levels and enhanced adipogenesis via PPARγ activation (198). Altogether, there are multiple pathways that regulate adiponectin expression in circulation via stimulation of gene expression of adiponectin or adipogenesis and insulin sensitivity-enhancing mechanisms.

In addition to its effects on adiponectin expression and adipogenesis, PPARγ pathway also plays a critical role in regulating adiponectin multimerization and secretion. TZD treatment upregulated expression levels of ER chaperones Ero1-Lα and DsbA-L and, subsequently, promoted HMW adiponectin formation and release in adipocytes (125, 138). Consistent with this, pioglitazone intervention led to increased levels of all adiponectin multimers, with the largest fold increases in nonamer (9mer) and dodecamer (12mer) levels despite a disproportionate increase in the amount of octadecamer (18mer), in diabetic patients (199). On the other hand, the endoplasmic reticulum (ER) stress appears to play a causative role in obesity-induced adiponectin downregulation and -impaired adiponectin multimerization (135, 136). In support of this, weight reduction by dietary intervention or inhibiting ER stress using pharmaceutical approach resulted in a relative increase of HMW adiponectin in the circulation of humans and rodents, respectively (136, 200). In addition, the inter-trimer disulfide bond formation is essential for adiponectin multimerization and favors an oxidizing redox environment with sufficient amounts of redox couples for disulfide rearrangement (201). Therefore, redox environment in adipocytes is a potential target for future treatments against Type 2 diabetes (201). This is supported by the evidence that mitochondria dysfunction-induced ROS overproduction is associated with decreased levels of the HMW adiponectin in adipose tissue of caveolin-1 knockout mice (102). However, whether the redox environment is able to promote adiponectin multimerization and secretion in vivo necessitates future clarification.

Conclusions

Here, we propose that the transendothelial movement of adiponectin may be highly regulated by paracellular transport. However, very little is known about studies examining the paracellular movement of adiponectin in diabetes target tissues, such as muscle and liver. Adiponectin plays an important role in vasculature. It is well established that it acts as a vasodilator inducing eNOS activation and endothelial NO production and protects the endothelium from oxidative stress. Furthermore, it mediates the activation of anti-inflammatory and proinflammatory markers. A few studies investigating the role of adiponectin in the paracellular movement concluded that hyperglycemia significantly regulated tight junctions, which contributed to increase adiponectin flux in STZD diabetes-induced mice. Alternatively, glucocorticoid-mediated tightening of endothelial barrier reduced the flux of adiponectin across the endothelial monolayers. Further investigations from the therapeutic perspective are needed, as this may represent a possible mechanism of adiponectin physiological actions.

In summary, even though much is known about general vascular functions of adiponectin, most evidence is generated from genetically modified rodent models, and not from humans. Notably, human vascular physiology and structure are quite different from animals. Therefore, further studies are needed, including larger animals, in order to provide a deeper understanding and to confirm vascular functional properties of adiponectin.