Vascular endothelial adiponectin signaling across the life span

Abstract

Cardiovascular disease risk increases with age regardless of sex. Some of this risk is attributable to alterations in natural hormones throughout the life span. The quintessential example of this being the dramatic increase in cardiovascular disease following the transition to menopause. Plasma levels of adiponectin, a “cardioprotective” adipokine released primarily by adipose tissue and regulated by hormones, also fluctuate throughout one’s life. Plasma adiponectin levels increase with age in both men and women, with higher levels in both pre- and postmenopausal women compared with men. Younger cohorts seem to confer cardioprotective benefits from increased adiponectin levels yet elevated levels in the elderly and those with existing heart disease are associated with poor cardiovascular outcomes. Here, we review the most recent data regarding adiponectin signaling in the vasculature, highlight the differences observed between the sexes, and shed light on the apparent paradox regarding increased cardiovascular disease risk despite rising plasma adiponectin levels over time.

INTRODUCTION

Adiponectin, a substance released primarily from adipose tissue in healthy adult humans, has emerged as a critical signaling molecule within the systemic vasculature. Its deficiency in plasma has been strongly linked to obesity, diabetes, and cardiovascular disease (CVD) (1). Hypoadiponectinemia has also been associated with endothelial dysfunction (2–4), the precursor to many CVDs. Elevated levels of adiponectin in adults are linked to improvement in brachial artery flow-mediated dilation (FMD) (5); however, a U-shaped association is observed with CVD risk and the adipokine. In older adults without CVD, increases in adiponectin levels up to 20 mg/L appear to be cardioprotective. Beyond this concentration, further increases correlate with higher incidence of CVD (6). Aside from concentration, the age of the individual also contributes to the response to adiponectin. This is illustrated by studies showing that elevated levels appear to promote vascular health in younger adults (40+ yr of age) (7, 8), whereas studies that focused on adults age 60 yr and older had an increased risk for CVD (8–10).

Regardless of one’s sex, levels of natural hormones also fluctuate over time and can promote or prevent CVD depending on the hormone and age of the individual. Both testosterone and estradiol have opposing effects on adiponectin levels (11), which may account for differences in CVD risk observed between the sexes and over time (Fig. 1). Adiponectin regulates the vascular endothelium by promoting a quiescent environment, however, significant knowledge gaps remain regarding how this relationship varies with age. Although much is known about the effect of adipose and adiponectin on vascular function (12–18), very little has focused on age- and sex-specific differences. The goal of this mini-review is to briefly summarize up-to-date knowledge regarding adiponectin and its known protective role in the endothelium. Insight into sex differences and changes throughout the life span will be discussed. Finally, we will focus on potential targets within the adiponectin signaling pathway that may allow for prevention of vascular inflammation, endothelial dysfunction, and associated disease regardless of age, sex, or gender.

Figure 1.

Plasma adiponectin and cardiovascular (CVD) risk over time. A: both low and high levels of plasma adiponectin are associated with CVD risk and mortality. B: time course of plasma adiponectin levels in both women and men over the life span. Adiponectin levels increase in women at puberty and decrease during the transition to menopause before rising again postmenopause. Men experience a decrease in plasma adiponectin at puberty that recovers over time; however, levels are lower compared with women.

ADIPONECTIN AND RECEPTORS

Adiponectin

Adiponectin, a 30-kDa protein discovered in 1995 (19), is mainly released by adipocytes found in white, beige, and brown fat depots; however, other reports have shown that other cells such as cardiomyocytes (20), osteoblasts (21), skeletal muscle cells (22), and possibly endothelial cells (23) are capable of releasing it as well. Its 244 amino acid sequence contains G-X-Y repeats, which facilitates adiponectin’s ability to oligomerize (19). Three oligomeric forms have been identified in vivo: a 12–18mer known as the high-molecular weight (HWM) form, a hexamer or middle molecular weight (MMW), and a trimeric or low-molecular weight (LMW) structure (24–27). A fourth form of adiponectin, globular adiponectin, is a COOH-terminus cleavage product of full-length adiponectin. Of all the structural forms of adiponectin, evidence suggests that decreased HMW adiponectin is most closely associated with poor cardiovascular outcomes (28, 29).

Adiponectin Receptors, AdipoR1 and AdipoR2

The effects of adiponectin are primarily regulated by two surface membrane receptors, adiponectin receptor 1 (AdipoR1) and adiponectin receptor 2 (AdipoR2). Although structurally similar to G protein-coupled receptors (GPCRs), AdipoRs have an intracellular NH₂-terminus and an extracellular COOH-terminus (30). Both receptors have been identified in human aortic endothelial cells; however, data support preferential expression of AdipoR1 (31). It is currently unknown whether AdipoR1 or R2 exist in the human microvascular endothelium; however, data from our group would suggest that this to be true as chronic exogenous administration of adiponectin (∼16 h) improves endothelial function in microvessels collected from patients with known coronary artery disease (CAD) (32). In this setting, the addition of adiponectin had a positive effect on the vascular endothelium; however, it remains unknown whether the aging process affects receptor expression and/or function as well as whether it differs among men and women.

T-Cadherin

T-cadherin, a glycophosphatidylinositol (GPI)-linked plasma membrane protein, has been implicated as a third receptor for adiponectin, specifically HMW adiponectin (33, 34). Pascolutti and colleagues (35) discovered that T-cadherin preferentially binds to the globular domain of HWM adiponectin. The fact that T-cadherin is highly expressed in the cardiovascular system begs the question of whether it has a critical role in adiponectin’s cardioprotective effects. Wang and colleagues (36) recently demonstrated for the first time that loss of T-cadherin is associated with endothelial dysfunction. Acetylcholine (ACh)-induced vasodilation of aortic rings from T-cadherin knockout mice was significantly impaired compared with wild-type littermates. Furthermore, levels of NO_x and phosphorylated Akt in the T-cadherin^−/− mice were decreased compared with control animals (36). Other studies have shown that decreased expression of T-cadherin impairs adiponectin localization to the vascular endothelium (37, 38). Despite an absent intracellular structure, the presence of T-cadherin appears to be critical for adiponectin signaling and a functioning endothelium; however, details regarding its exact role are lacking.

Although little is known regarding how age or sex affects the expression level of adiponectin receptors, some evidence points to estrogen as being a key regulator of T-cadherin. Using a quantitative real-time RT-PCR assay, Bromhead and colleagues (39) demonstrated that β-estradiol (E₂) regulates both transcriptional and posttranscriptional expression of T-cadherin. This suggests the possibility that loss of T-cadherin expression due to declining estrogen levels in menopause may result in decreased uptake or physiological action of plasma adiponectin. Therefore, despite having elevated plasma levels of adiponectin during the postmenopause period (40, 41), this may not truly reflect the bioavailability or action of this adipokine in the endothelium.

SEX AND AGE DIFFERENCES

In healthy, nonobese adult humans, plasma adiponectin concentrations are within the range of 2–10 µg/mL (42). Adiponectin levels have been shown to increase with age in healthy men and women (43), with greater levels observed in women (44). Interestingly, approximately half of the adiponectin pool in females consists of HMW compared with males where adiponectin is evenly divided among the three major forms. The significance of this is underscored by data from Pischon et al. (45) showing that elevated levels of HMW adiponectin and/or an increased HMW/total adiponectin ratio are associated with a lower risk of ischemic heart disease in women. Increases in both HMW adiponectin and the ratio were inversely related to body mass index (BMI), triglycerides, C-reactive protein, hemoglobin A1C, and low-density lipoprotein-C. Importantly, in postmenopausal women, the HMW/total adiponectin ratio is significantly lower when compared with early-menopausal women (46), potentially contributing to the increased risk of CVD postmenopause.

Plasma adiponectin levels in women are dynamic, with levels declining to their lowest during the menopausal transition and increasing again after menopause. The decrease in adiponectin during this time is also associated with increased androgens (40, 41). Although adiponectin levels rise in postmenopausal women, low plasma levels during this time in life strongly correlate with risk for metabolic syndrome in those with increased adiposity (47). In men, testosterone levels are inversely associated with adiponectin levels (48), and in transmales, adiponectin levels decrease following testosterone therapy (49). A decrease in adiponectin is also seen in pubertal boys, as their androgen levels increase, and by the end of puberty, males have lower adiponectin levels compared with females (50). Combs et al. (51) further tried to elucidate hormonal effects on adiponectin showing that female mice have higher levels of adiponectin in fat depots compared with male following sexual maturation, as seen in humans. Castration of neonatal male mice, but not adult mice, led to comparable adiponectin levels to the female mice. Interestingly, neonatal ovariectomy did not impact adiponectin levels, but adult ovariectomy further increased levels, similar to alterations seen in postmenopausal women. The researchers propose that neonatal testosterone levels might establish adiponectin set-point levels while estrogen dominates as the regulator of adiponectin later in life. It’s clear that the hormone changes experienced over time influence concentration of adiponectin and may account for differences in CVD risk we observe as individuals age. These preclinical studies have yet to be translated to humans and may have tremendous implications for hormone-replacement therapy for both postmenopausal women and the transgender population.

ADIPONECTIN AND VASCULAR CELL SIGNALING

Endothelial Cells

Adiponectin is a key player in promoting vascular homeostasis and is mainly categorized as an anti-inflammatory compound by increasing levels of nitric oxide (NO). Adiponectin binds both AdipoR1 and AdipoR2 on the extracellular (C) terminus (30), triggers calcium influx (52), and activates multiple downstream targets including AMPK (53), PPARα (54), and PI3K/Akt (55), which has been recently reviewed extensively by da Silva Rosa and colleagues (56). The link between AdipoR1/R2 and downstream pathway activation relies on an adapter protein, APPL1. APPL1 activates AMPK resulting in endothelial eNOS phosphorylation (57). Adiponectin-induced APPL1 binding also activates PI3K to phosphorylate eNOS (58, 59). In addition to direct increases in NO, studies performed in bovine aortic endothelial cells have shown that globular adiponectin stimulates binding of Hsp90 to eNOS, a critical process for proper eNOS homodimerization and function (60). Ex vivo experiments on human internal mammary arteries have shown that adiponectin levels increase tetrahydrobiopterin (BH₄) bioavailability, thus promoting eNOS coupling and NO-producing efficiency (61). Experiments using siRNA to decrease expression of APPL1 in human umbilical vein endothelial cells (HUVECs) have demonstrated markedly reduced adiponectin-dependent NO production primarily as a result of decreased AMPK phosphorylation (62).

Studies in both animals and humans have shown a strong correlation between hypoadiponectinemia and impaired endothelium-dependent dilation (2–4), suggesting that lack of adiponectin promotes endothelial dysfunction. In aortic rings from adiponectin knockout mice, Cao et al. (63) demonstrated that NO bioavailability was significantly reduced because of decreased phosphorylated eNOS. This was reversed upon administration of exogenous globular adiponectin. Similarly, both globular and full-length adiponectin have been shown to induce NO-dependent vasodilation in nondiabetic male and female Zucker lean rats. However, in the diabetic fatty rats, administration of adiponectin failed to restore endothelial function, which was attributed to a decrease in APPL1, suggesting resistance to adiponectin signaling (64). In patients with heart failure (8 men and 5 women), adiponectin levels are fivefold higher compared with healthy subjects, but the expression of PPARa and AMPK genes and AdipoR1 mRNA are reduced in the skeletal muscle. This further supports a potential functional resistance to adiponectin in diseased states (65). In the vasculature specifically, a high-fat diet in male rats has been shown to decrease AdipoR1 and R2 expression, with an initial increase in adiponectin levels. The same study also showed a reduction in AMPK/eNOS signaling in response to recombinant adiponectin (66). Alterations in downstream signaling may explain the paradoxical scenario of elevated risk in older individuals, postmenopausal women, and those with heart failure despite increased plasma adiponectin levels.

Smooth Muscle Cells

Although this review focuses on the vascular endothelium, one would be remiss to not comment on how adiponectin affects smooth muscle cells (SMCs) within the medial layer. Both AdipoR1 and R2, as well as T-cadherin, have been identified in human coronary artery SMCs (67). Data thus far indicate that adiponectin is involved in phenotypic modulation of SMCs. Adiponectin promotes the contractile phenotype or elongated SMCs that primarily consist of contractile filaments. This is in opposition to the synthetic phenotype where SMCs contain organelles involved in protein synthesis allowing to create machinery necessary for cell migration and proliferation. Cerosimo and colleagues (67) demonstrated that the presence of adiponectin prevented human coronary SMC migration and proliferation induced by high glucose/palmitate. Although one would expect a similar situation with T-cadherin with its presence being protective, this does not appear to be the case. Studies done by Frismantiene et al. (68) have concluded that knock down of T-cadherin in human aortic SMCs promotes the contractile phenotype, whereas overexpression results in the dedifferentiated synthetic/secretory state. It is possible that the migration of T-cadherin from the endothelium to SMCs damages the vascular wall in two ways: 1) by decreasing the uptake of adiponectin in endothelial cells and 2) by triggering the SMC to enter the synthetic state to promote proliferation.

ADIPONECTIN AND THE SPHINGOLIPID RHEOSTAT

Sphingolipids, critical lipid messengers within the endothelium that regulate the balance between NO and reactive oxygen species (ROS) production, are strongly influenced by adiponectin. In healthy adults, ceramide levels are inversely correlated with adiponectin levels (69). Ceramide accumulation in endothelial cells leads to increased ROS production through both activation of NADPH oxidases (70) and mitochondria (71). Our group has shown that chronic exposure to exogenous ceramide induces human microvascular endothelial dysfunction (72) and more recently that blocking metabolism of ceramide thus causing accumulation, results in similar damage to human arterioles (32). Furthermore, ceramide is a potent stimulus for the formation of extracellular vesicles and we have shown that intraluminal exposure to these vesicles in a microvessel from a healthy patient also induced endothelial dysfunction (73).

The effects of ceramide can be mitigated if hydrolyzed to sphingosine by ceramidase enzymes. Once converted to sphingosine, it can further be phosphorylated to form sphingosine-1-phosphate (S1P), a known activator of eNOS (74). Both AdipoR1 and AdipoR2 possess intrinsic ceramidase activity and binding of adiponectin increases hydrolysis of ceramide by more than 20-fold (75). The inhibition of adiponectin production has been shown to decrease sphingosine and S1P levels and blocking ceramide production leads to an increase in HMW adiponectin (69), thereby highlighting a strong interplay between adiponectin signaling and the sphingolipid rheostat. Obata et al. (76) also found that through the binding of T-cadherin, adiponectin triggered the formation and release of ceramide containing exosomes, thus lowering endothelial cell ceramide content. Adiponectin-induced release of ceramide-filled vesicles did not occur in endothelial cells from T-cadherin-knockout mice. As previously mentioned, T-cadherin lacks intracellular signaling domains therefore studies such as this add important insight into its role and relationship with adiponectin. Taken together, adiponectin may be critical in limiting the potential damaging effects of elevated ceramides in the endothelium.

Elevated plasma ceramide levels are highly associated with many cardiovascular diseases thought to arise from a dysfunctional endothelium including both atherosclerosis (77) and heart failure with preserved ejection fraction (HFpEF) (78, 79). Interestingly, elevated levels of adiponectin are typically observed in individuals with HFpEF (80). Knowing that adiponectin is extremely efficient at decreasing cellular ceramides creates a bit of a quagmire. A similar scenario is observed in women of advanced age. After accounting for potential confounders, circulating ceramide levels have been shown to increase postmenopause (81) despite rising adiponectin levels (40, 41). The increased ceramide levels observed in aging women are strongly associated with the loss of estrogen (81) and yet the increased plasma levels of adiponectin appear to have little effect in mitigating ceramide accumulation. The elevated adiponectin levels associated with menopause may suggest the possibility that decreased and/or dysfunctional T-cadherin or adiponectin signaling may contribute to this confusing scenario (Fig. 2). In addition, sex differences also exist in the prevalence of HFpEF, a disease that tends to favor women over men by 2:1 (82) and is associated with elevated levels of both adiponectin and ceramide. This begs the question of whether dysfunctional T-cadherin/adiponectin signaling that occurs in aged females places them at higher risk for development of this particular form of heart failure.

Figure 2.

The estrogen-adiponectin paradox in aging females. Plasma estrogen levels decline while adiponectin levels increase during the transition to menopause. T-cadherin is critical for the binding of adiponectin to adiponectin receptor (AdipoR) 1/2. Once receptors are activated, the adapter protein AAPL1 is required to activate various downstream pathways, including AMPK and PI3K/Akt, which results in an increase in endothelial nitric oxide (NO) synthase (eNOS)-generated NO. Activation of AdipoR1/2 dramatically increases hydrolysis of ceramide (Cer), a sphingolipid that induces endothelial dysfunction. The NO-generating pathways combined with shifting the sphingolipid rheostat toward sphingosine (Sph)-1-phosphate (S1P) promotes a healthy endothelium. Despite the rise in adiponectin during menopause, women have an elevated risk of cardiovascular disease, which may be due to dysfunctional T-cadherin, AdipoR1/2, and/or APPL1 signaling (gray). Both the loss of NO-stimulating pathways and decrease in AdipoR1/2 ceramidase activity contributes to decreased NO bioavailability and endothelial dysfunction.

TRANSLATIONAL STUDIES

Translational studies in humans have also highlighted the positive influence of adiponectin within the vascular endothelium. Perivascular adipose tissue (PVAT), an important source of adiponectin, has been shown to elicit NO-mediated vasodilation. In an elegant bioassay experiment, Greenstein and colleagues (83) demonstrated that medium collected from a human microvessel surrounded by PVAT-induced vasodilation in an acceptor vessel minus the surrounding adipose. This dilation was inhibited in the presence of either the nitric oxide synthase inhibitor N^G-monomethyl-l-arginine or an adiponectin blocking peptide suggesting that adiponectin was responsible for the rise in NO and dilation. Unfortunately, women were underrepresented in this study and comprised only 20%–30% of the subjects studied. Using an in vitro approach, our group has recently demonstrated that chronic exposure (16 h) to exogenous adiponectin is able to restore NO-mediated flow-induced dilation in human arterioles collected from patients with CAD (32); however, over 80% microvessels were collected from male patients.

There is a paucity of in vivo work linking adiponectin to vascular function. Although Torigoe and colleagues (84) were able to demonstrate that NO-mediated brachial artery FMD correlates with HMW adiponectin levels, the study was done entirely in young men (∼30-yr old). Yoo et al. (5) recently showed that endothelial levels of adiponectin, but not circulating levels, were associated with greater brachial artery FMD in older men and women. A regression analysis performed by Okui et al. (85) concluded that plasma adiponectin levels positively correlate with ACh-induced increases in coronary blood flow. Interestingly, compared with body mass index (BMI), triglycerides, and the insulin resistance index, only adiponectin concentration could independently predict ACh-induced fluctuations in coronary flow (85). Although women were included in the Okui study, females represented only one-third of the total participants. There is no doubt that alterations in hormones complicate measured outcomes in clinical studies; however, it is critical that effort is given to include women in these in vivo studies to better elucidate adiponectin signaling in both sexes and within different age groups.

CLINICAL PERSPECTIVES

For clinical scenarios and possibly appropriate age groups in which low adiponectin levels are associated with cardiovascular risk, adiponectin replacement therapy may offer benefit to patients. Because of the complexity of the intact adiponectin protein, the majority of pharmacological advances have been made with small molecule agonists of the adiponectin receptor, primarily peptides that are structurally similar to the active site of globular adiponectin (86). One of the first to be discovered was ADP355, a peptidomimetic that although was shown to increase phosphorylation of both eNOS and AMPK and these effects were mainly observed in the liver (87). AdipoRon, a nonpeptide small molecule discovered in 2013, is the first orally active agonist for both AdipoR1 and R2. Our group has shown that as with globular adiponectin, chronic treatment with AdipoRon is capable of restoring NO-mediated FID in human arterioles from patients with known CAD (32). Very recently, Caldwell et al. (88) showed the importance of adiponectin in coronary microvascular function in response to exercise, and Iwabu and colleagues (89) demonstrated in AdipoR-humanized mice (mice that express human AdipoR1 in muscle of AdipoR1/R2 double knockout mice) that orally administered AdipoRon increased insulin sensitivity and exercise endurance. Aside from these mentioned studies, little is known regarding how small molecule agonists or peptide mimetics that target AdipoR1/R2 affect the cardiovascular system and improve endothelial function, much less in age- or sex-specific groups.

Recently, a member of the plant protein family, osmotin, was discovered as an AdipoR1/R2 agonist (90). AdipoR1/R2 do share homology with the receptor for osmotin, known as PHO36 (90). Osmotin has been shown to induce AMPK phosphorylation through adiponectin receptors in C2C12 myocytes (90). This stress-activated phytopeptide is found fairly ubiquitously in plants, but particularly in tomatoes, potatoes, and peppers (91, 92). Early evidence of osmotin as a potential therapeutic in the prevention of cardiovascular disease has been promising, as it was recently shown that osmotin infusion significantly decreased the amount of atherosclerotic plaque development in ApoE^−/− mice (91) fed a high-fat diet. Endothelial effects of osmotin, including its potential to promote NO signaling similar to that of adiponectin, have yet to be determined but may serve as a potential therapy for hypoadiponectinemia or better yet, adiponectin resistance during advanced age or disease.

SUMMARY AND CONCLUSIONS

We experience a multitude of physiological changes as we age. All organ systems undergo transformation with time including the systemic vasculature. Age alone is a risk factor for CVD; however, this risk varies depending on sex and stage of life (93) (e.g., pre- vs. postmenopause). Low levels of adiponectin, an anti-inflammatory compound typically released from adipose tissue, have historically been associated with increased CVD risk despite being elevated in older individuals. It is becoming increasingly clear that the effect of adiponectin within the vasculature is strongly influenced by both age and sex. Increasing adiponectin to enhance vascular function has proven benefit in younger adult populations yet appears to be futile in older adults. Adiponectin signaling within the endothelium seems to confer resistance in disease states such as diabetes, obesity, and heart failure, where higher adiponectin levels correlate with worse outcomes. A similar resistance may occur in women, as adiponectin levels are elevated in postmenopause yet risk of CVD dramatically increases during this time, which may be due to altered downstream signaling of AMPK/eNOS. Adiponectin is highly efficient at reducing vascular-damaging ceramides, yet plasma ceramide levels also increase following menopause. Male and female sex hormones modulate the production of adiponectin, which may partly explain the difference in CVD risk in men compared with women as well as over time. Studies focused on human vascular adiponectin signaling across the life span are needed for males, females, and the transgender population. Insight into these mechanistic transformations may provide more precise therapeutic targets to mitigate cardiovascular disease for all.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants K08HL141562 (to J. K. Freed) and R38HL143561 (to K. E. Cohen).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.E.C. and J.K.F. drafted manuscript; K.E.C., B.K., G.S., J.J.M., and J.K.F. edited and revised manuscript; K.E.C., B.K., G.S., J.J.M., and J.K.F. approved final version of manuscript.