2023 Volume 70 Issue 7 Pages 663-675
Visceral fat-based metabolic syndrome has a strong impact on atherosclerotic cardiovascular disease (CVD), clustering diabetes, dyslipidemia, hypertension, hyperuricemia, and non-alcoholic fatty liver disease (NAFLD). Adiponectin, a protein specifically secreted by adipocytes, circulates abundantly in the human bloodstream, but its concentration decreases under pathological conditions such as visceral fat accumulation. Extensive clinical evidence has demonstrated that hypoadiponectinemia is associated with the development of CVD and chronic organ diseases. Although several binding partners of adiponectin, such as AdipoR1/2, have been identified, how adiponectin exerts its multiple beneficial effects on various organs remains to be fully elucidated. Recent progress in adiponectin research has revealed that adiponectin accumulates on cardiovascular tissues by binding to a unique glycosylphosphatidylinositol-anchored T-cadherin. The adiponectin/T-cadherin complex enhances exosome biogenesis and secretion, which may contribute to the maintenance of cellular homeostasis and tissue regeneration, particularly in the vasculature. Xanthine oxidoreductase (XOR) is a rate-limiting enzyme that catabolizes hypoxanthine and xanthine to uric acid. XOR produces reactive oxygen species in the reaction process, suggesting that XOR is involved in the pathological mechanism underlying CVD progression. Recent findings from clinical and laboratory studies have shown strong positive correlations between plasma XOR activity and liver enzymes. Furthermore, especially in NAFLD conditions, excessive hepatic XOR leaked into the bloodstream accelerates purine catabolism in the circulation, using hypoxanthine secreted from vascular endothelial cells and adipocytes, which can promote vascular remodeling. In this review, we focused on the cardiovascular significance of adipose-derived adiponectin and liver-derived XOR in the development of CVD associated with metabolic syndrome.
The number of subjects with overweight and obesity is increasing worldwide over the past 40 years, also in East Asian countries including Japan [1, 2]. Obesity is a common basis for chronic metabolic disorders, such as type 2 diabetes, dyslipidemia, hypertension, hyperuricemia, and non-alcoholic fatty liver disease (NAFLD), all of which are major independent risk factors for atherosclerotic cardiovascular disease (CVD). In this regard, fat distribution has a significant impact on the development of CVD and systemic organ diseases. Our group was the first in the world to develop a computed tomography (CT)-based fat analysis method, and demonstrated that intra-abdominal visceral fat accumulation was closely associated with metabolic diseases [3-5]. These lines of clinical data provided the groundwork for the concept of “visceral fat syndrome,” namely “metabolic syndrome,” as a clustering of multiple cardiovascular risk factors [6]. Subsequent progress in research on the biological and molecular mechanisms has identified that metabolic and endocrine dysfunction of adipocytes plays a critical role in the development of this syndrome and its associated cardiovascular complications [7].
Effective strategies for the prevention of CVD related to metabolic syndrome have emerged as an important issue in extending disability-free and healthy life expectancy. In recent years, enormous efforts have been made worldwide to investigate the underlying molecular mechanisms of obesity-related cardiometabolic disorders [8]. To find effective therapeutic targets for metabolic syndrome, we have been exploring the biology of the pathological significance connecting adipocyte dysfunction to the cardiovascular system, with a particular interest in bioactive molecules circulating in the bloodstream. This review mainly focuses on adipose-derived adiponectin and liver-derived xanthine oxidoreductase (XOR), which have been implicated in the development of CVD in metabolic syndrome.
Studies elucidating gene expression profiles in human adipose tissue have revolutionized the concept of “adipocytokines/adipokines,” which are biologically active mediators produced mainly in adipocytes [9, 10]. Among them, we identified a novel gene, “adipose most abundant gene transcript-1 (apM-1),” from a human adipose gene expression profile analysis [11] and named the product of apM1 “adiponectin” because it was specifically and abundantly secreted from adipocytes and had unique collagen-like domains and nectin-like adhesion properties [11, 12]. Adiponectin, a circulating protein derived from adipocytes, plays a critical role in the maintenance of systemic homeostasis and has two unique properties: (1) It circulates in very high concentrations of about 5–30 μg/mL, which is three to six orders of magnitude higher than those of conventional and classical hormones or cytokines such as leptin, insulin, and interleukins; and (2) although it is specifically produced by adipocytes, circulating adiponectin levels are inversely correlated with body fat mass, especially visceral fat accumulation. The protein structure of adiponectin consists of an N-terminal signal sequence, a non-homologous or hypervariable region, a collagen domain containing 22 collagen repeats, and a C-terminal C1q-like globular domain. The N-terminal hypervariable region of adiponectin contains a single cysteine residue (Cys-39 in mice and Cys-36 in humans), which is essential for its multimeric formation by disulfide bonds in the trimer [13, 14]. Circulating adiponectin is mainly the medium molecular weight (MMW, hexamer, approximately 180 kDa) and high molecular weight (HMW, dodecamer or octadecamer, approximately 360 to 540 kDa) forms. The HMW form of adiponectin, in particular, has been implicated in protective associations against various human diseases.
Circulating adiponectin levels are influenced by several factors at the level of transcriptional regulation and protein secretion. Notably, adiponectin concentrations significantly decreased with body weight/body mass index (BMI), particularly with visceral adiposity, regardless of its restricted expression in adipocytes [12, 15]. Although the mechanisms underlying such obesity-related adiponectin reduction have yet to be elucidated in detail, several factors associated with adipose tissue inflammation in obese conditions, including tumor necrosis factor-α (TNF-α) [16], interleukin-6 (IL-6) [17], and reactive oxygen species (ROS) [18], have been reported to reduce adiponectin mRNA levels and suppress adiponectin production in adipocytes. In contrast, among transcriptional factors, peroxisome proliferator-activated receptor (PPAR)-γ has been identified as the strongest positive regulator of adiponectin. PPAR-γ agonists, used clinically as insulin sensitizers in diabetic patients, have been shown to significantly increase circulating adiponectin levels in both rodents and humans via a PPAR-response element (PPRE) on the adiponectin gene promoter [16, 19, 20], which may partially explain the preventive effect of this drug class on atherosclerosis [21, 22].
1.2 Receptors/binding partners of adiponectinPrevious clinical studies have shown that low levels of circulating adiponectin are one of the risk factors for type 2 diabetes and CVD [23-27], suggesting links between hypoadiponectinemia and the progression of cardiometabolic disorders. Along with these clinical observational studies, numerous experimental studies, including ours, have demonstrated its protective properties, such as anti-diabetes [28, 29], anti-atherosclerosis [30, 31], anti-fibrosis [32-34], and anti-inflammation [35, 36] in systemic organs. Adiponectin knockout mice show no obvious phenotype at steady state, but when mice are exposed to conditions that mimic lifestyle-related diseases, they exhibit severe symptoms [28, 32, 33]. However, it is not fully understood how adipocyte-derived adiponectin exerts multiple beneficial effects on various organs distant from adipose tissue, and whether these adiponectin actions can be explained by receptor-mediated signaling.
Yamauchi T et al. reported AdipoR1 and AdipoR2 as the specific receptors for adiponectin [37]. AdipoR1 and AdipoR2 were discovered as molecules that bind to recombinant globular adiponectin produced by bacteria, using cells transfected with human skeletal muscle cDNA library. AdipoR1 is abundantly expressed in skeletal muscle, whereas AdipoR2 is mostly restricted to the liver. Besides, mice lacking AdipoR1 and AdipoR2 resembled the phenotype of adiponectin-deficient mice, including insulin resistance [38]. AdipoR1/2 have seven transmembrane domains like G protein-coupled receptors (GPCRs), but their N- and C-termini are oriented intracellularly and extracellularly, respectively, opposite to those of normal GPCRs, suggesting that AdipoR1/2 does not bind to G proteins. In addition, the crystal structure of human AdipoR1/2 was recently demonstrated, revealing that its seven transmembrane helices are structurally distinct from normal GPCRs, surrounding a large cavity with three conserved histidine residues and a zinc ion coordination [39]. On the other hand, it remains unclear whether AdipoR1/2 also binds to HMW multimeric adiponectin or how it transduces signaling after binding.
T-cadherin was first identified as a binding molecule for adiponectin by the Lodish group using an in vitro binding assay, in which T-cadherin bound to the hexameric and HMW forms of adiponectin, but not to its trimeric and globular forms [40]. Unlike other cadherin family members, T-cadherin is a unique GPI-anchored cadherin that lacks an intracellular or transmembrane domain. Cadherins are initially synthesized as immature forms with a prodomain and then the prodomain is removed; thus, mature cadherins usually lack this domain [41]. However, T-cadherin exists on the cell surface in both the mature form (100 kDa), consisting of five extracellular cadherin repeat (EC) domains, and the premature form (130 kDa), consisting of a pro-domain with EC domains [42]. We have developed a protein G-based T-cadherin capture assay system to evaluate its binding to adiponectin, revealing that T-cadherin bound specifically and with high affinity to hexameric and HMW forms of adiponectin with a dissociation constant of approximately 1.0 nM. Furthermore, the EC1 and EC2 domains of T-cadherin were shown to be the essential sites for binding to adiponectin [43, 44]. This capture assay system also allowed the purification of physiologically native adiponectin from serum or plasma samples.
1.3 T-cadherin-dependent tissue accumulation of adiponectin and its protective effects on CVDIn recent years, we and others have demonstrated that T-cadherin is critical for adiponectin tissue accumulation [45-49]. T-cadherin protein is dominantly expressed in the mouse aorta, heart, and skeletal muscle. Similar to the tissue distribution of T-cadherin, adiponectin protein was abundantly detected in these T-cadherin-expressing organs, although these tissues did not express adiponectin mRNA [46]. Immunostaining also showed co-localization of adiponectin and T-cadherin, even in mice lacking AdipoR1 or AdipoR2 [45]. In contrast, such adiponectin protein accumulation disappeared in T-cadherin knockout mice or after administration of the T-cadherin cleaving enzyme, phosphatidylinositol-specific phospholipase C (PI-PLC), to wild-type mice [46]. Furthermore, circulating levels of adiponectin were 3- to 4-fold higher in T-cadherin knockout mice than in wild-type mice [46, 48], and this increase was exclusively observed in the hexameric and HMW forms of adiponectin [46]. These findings clearly indicate that T-cadherin is a key molecule in the tissue accumulation of adiponectin.
T-cadherin knockout mice exhibited impaired cardiovascular phenotypes, and the results obtained from these mice suggested a protective role for adiponectin/T-cadherin in the cardiovascular system [45, 47, 48]. T-cadherin deficiency resulted in severe cardiac phenotypes in mouse models of transaortic constriction-induced cardiac hypertrophy and myocardial ischemia-reperfusion [48]. Additionally, in a hindlimb ischemia model, T-cadherin knockout mice failed to restore blood flow, similar to adiponectin knockout mice [45]. Administration of exogenous adiponectin rescued these impaired phenotypes in adiponectin knockout mice but not in T-cadherin knockout mice. In a series of recent studies, we investigated the role of adiponectin/T-cadherin system in atherosclerosis and vascular injury. Immunoelectron microscopy analysis revealed the precise localization of adiponectin protein in the healthy aorta and atherosclerotic lesions [50]. In healthy aortas of wild-type mice, adiponectin signals were detected only in vascular endothelial cells (ECs). In contrast, in atherosclerotic lesions of apolipoprotein E (ApoE)-deficient mice fed a high-cholesterol diet, adiponectin existed not only in ECs but also on the surface of vascular smooth muscle cells (SMCs) with a synthetic/proliferative phenotype that migrated into the subendothelium. In T-cadherin and ApoE double-knockout (Tcad/ApoE-DKO) mice, such adiponectin proteins in ECs and synthetic/proliferative SMCs completely disappeared, and atherosclerotic plaque formation and neointimal formation induced by carotid artery ligation were significantly worsened compared to ApoE-KO mice [47]. Local environmental cues, such as inflammatory stimuli, are well known to induce phenotypic dedifferentiation of SMCs from the quiescent contractile type to the synthetic/proliferative type, leading to the development of atherosclerosis [51, 52]. Therefore, in pathological conditions such as atherosclerosis, SMCs can recruit adiponectin by expressing T-cadherin which might keep them from dedifferentiation.
In addition to large vessels such as the aorta and the carotid artery, we also observed T-cadherin-dependent localization of adiponectin in the microvasculature of the retina [53] and kidney [54]. In the retina, immunofluorescence results clearly showed adiponectin accumulation on the vascular endothelium of retinal arterioles, which was completely diminished in T-cadherin-knockout mice. Interestingly, such retinal adiponectin progressively decreased after the onset of streptozotocin (STZ)-induced diabetes, possibly due to the reduction of T-cadherin expression in the retinal vascular endothelium under diabetic and/or hyperglycemic conditions. In addition, adiponectin deficiency resulted in the severe phenotype of the early features of diabetic retinopathy, as represented by the breakdown of the blood-retinal barrier (BRB) and increased vascular permeability [53]. This finding indicated a protective role of adiponectin/T-cadherin against the development of diabetic microvascular complications, as well as atherosclerotic macrovascular diseases. In the kidney, neither adiponectin nor T-cadherin was observed in the glomerular capillaries, whereas T-cadherin-dependent adiponectin was detected in the pericytes, subsets of tissue-resident mesenchymal stem/progenitor cells (MSCs) positive for platelet-derived growth factor receptor beta (PDGFRβ), surrounding peritubular capillaries [54]. Renal pericytes have been assumed to have numerous functions, including regulation of microvascular blood flow, maintenance of vascular permeability by supporting ECs, and repair of the damaged region by transforming into myofibroblasts [55-58]. In an acute renal ischemia-reperfusion (I/R) model, both adiponectin-knockout mice and T-cadherin-knockout mice exhibited the more progressive phenotype of renal tubular damage and increased vascular permeability, reflecting severe pericyte loss [54]. This result suggested that adiponectin plays an important role in protecting against renal vascular injury by binding to T-cadherin.
In genome-wide association studies (GWAS), several independent groups have shown a strong correlation between the T-cadherin gene and circulating adiponectin levels, in addition to metabolic syndrome and CVD [59-63], indicating the significance of T-cadherin as a physiological binding partner of adiponectin also in humans. On the other hand, adiponectin has been shown to positively regulate the T-cadherin protein. Adiponectin knockout mice showed markedly decreased T-cadherin protein levels in systemic organs, while T-cadherin mRNA levels were comparable in adiponectin knockout and wild-type mice. This decrease in T-cadherin protein was reversed by adiponectin supplementation in adiponectin knockout mice [46]. We have recently developed a novel T-cadherin enzyme-linked immunosorbent assay (ELISA) that can measure 3 forms of soluble T-cadherin: a 130-kDa form with a prodomain, a 100-kDa mature form, and a 30-kDa prodomain in human serum, by which a 130-kDa form of circulating T-cadherin levels were found to be positively correlated with plasma adiponectin concentrations in Japanese patients with type 2 diabetes [64]. These findings indicate that adiponectin is required for the stabilization of the T-cadherin protein in tissues and cells. By contrast, chronic hypoadiponectinemia might reduce tissue T-cadherin levels, thus reducing adiponectin accumulation and impairing the adiponectin-mediated protective system.
1.4 The adiponectin/T-cadherin system and exosome biogenesisGrowing evidence indicates that adiponectin protects various organs and cells in a multifunctional manner via T-cadherin. However, the mechanism by which adiponectin exerts its effect after binding to cell surface T-cadherin has yet to be fully elucidated. T-cadherin is a GPI-anchored protein without an intracellular domain; therefore, the conventional ligand-receptor mechanism cannot explain how the adiponectin/cadherin complex transduces intracellular biological signals. Certain GPI-anchored proteins are known to be internalized into cells by caveolin-mediated endocytosis pathways, accumulated in the multivesicular body (MVB), transported to the plasma membrane, and released extracellularly as exosomes [65-67]. Exosomes are membranous nanovesicles of endocytic origin released by various cell types of different organisms and possess two major functions: (1) exosomes contain proteins, lipids, mRNAs, microRNAs (miRNAs), and mitochondrial DNA (mtDNA) that mediate cell-to-cell communication [68, 69]; and (2) exosomes extracellularly release several intracellular proteins and metabolites, such as ceramides, amyloid-β, and prion proteins, that are unnecessary or harmful to the cell, and maintain homeostasis [65, 70, 71].
We previously demonstrated that both adiponectin and T-cadherin were packaged in exosomes, and that the physiological levels of adiponectin enhanced exosome biogenesis and secretion via T-cadherin [72, 73]. In T-cadherin-expressing cells, such as cultured ECs and pericytes, adiponectin significantly increased the number of cell-secreted exosomal particles in a dose-dependent manner within the physiological range of circulating adiponectin concentrations, but this increase was blunted by T-cadherin knockdown [54, 72]. Immunoelectron microscopy analysis also showed that adiponectin/T-cadherin increases exosomes by partially promoting exosome de novo biosynthesis in MVBs, as demonstrated by double knockdown of Ras-related in brain 7 (Rab7) (a key factor in the fusion of MVBs and lysosomes) and Rab27a (a key factor in exosome secretion) [72]. Furthermore, adiponectin treatment enhanced cellular ceramide flux into exosomes while decreasing intracellular ceramide accumulation, which was T-cadherin dependent [72]. In vivo, adiponectin signals were detected on the aortic endothelial cell surface as well as in intracellular MVBs by immunoelectron microscopy [50, 72]. Notably, plasma exosome levels were significantly reduced in mice lacking adiponectin or T-cadherin, suggesting that the adiponectin/T-cadherin system is a novel and unique factor that determines the amount of circulating exosomes.
Intracellular accumulation of ceramide, a potential intracellular lipid mediator, is assumed to be involved in the pathogenesis of diabetic microvascular complications and atherosclerosis through the process of lipotoxicity, including inflammation, ROS, and cell death [74]. Importantly, we previously reported that adiponectin exerted ceramide efflux activity through enhanced exosomal release, followed by a reduction in intracellular ceramide levels in cultured endothelial cells, but that these changes were abrogated by T-cadherin knockdown [72]. Therefore, the T-cadherin-dependent effects of adiponectin on exosome release may contribute to cellular maintenance through an alternative disposal pathway to lysosomes. Besides, we recently reported that adiponectin enhanced the therapeutic efficacy of adipose-derived MSCs by stimulating exosome release from transplanted MSCs in a rodent model of cardiac hypertrophy [75]. Taken together, the T-cadherin-mediated effects of adiponectin on exosome release from ECs or pericytes may play an important role in maintaining vascular homeostasis and protecting against vascular diseases.
As described above, unlike ordinary hormones, adiponectin circulates in high concentrations in the bloodstream and steadily accumulates on cardiovascular cells by binding to T-cadherin. Moreover, adiponectin might exert its various beneficial effects by promoting exosome biosynthesis in a T-cadherin manner. Collectively, adiponectin is unique among other secretory proteins in that it predominantly regulates the amount of systemically circulating exosomes, leading to the maintenance of whole-body homeostasis and cardiovascular protection (Fig. 1). Further research is warranted for therapeutic applications of the adiponectin/T-cadherin system.
Working model of the adiponectin/T-cadherin system in cardiovascular protection
Adiponectin is specifically secreted from adipocytes and circulates abundantly in the bloodstream at the steady state and accumulates on the cell surface of vascular component cells, such as endothelial cells, pericytes, and proliferative smooth muscle cells, by binding to T-cadherin. This T-cadherin-dependent accumulation of adiponectin stabilizes the T-cadherin protein itself and protects against vascular injury and atherosclerosis. However, chronic hypoadiponectinemia, mainly caused by visceral fat accumulation or diabetic conditions, can reduce T-cadherin protein levels on the plasma membrane, leading to decreased beneficial effects of adiponectin on the cardiovascular system. After adiponectin binds to T-cadherin, some of their complexes on the plasma membrane are endocytosed into the cell and transported to multivesicular bodies (MVBs), where exosomes are synthesized. The exosomes, packaging the adiponectin/T-cadherin complexes, are then released from the cells. Adiponectin enhances exosome biogenesis and secretion from the cells via T-cadherin, in a dose-dependent manner of adiponectin within a range of physiologically circulating concentrations, which can influence the amount of circulating exosomes. Exosomes contain various intracellular components such as proteins, lipids, mRNAs, microRNAs (miRNAs), and mitochondrial DNA (mtDNA). Since pro-inflammatory lipid mediators such as ceramides are also contained in exosomes, the adiponectin/T-cadherin system contributes to maintaining cellular homeostasis by providing an alternative disposal pathway to lysosomes through enhanced exosomal release from cells. Besides, the upregulation of circulating exosome particles may lead to cardiovascular protection and tissue repair in distant organs.
Hyperuricemia, defined as a serum uric acid concentration of more than 7.0 mg/dL, is closely associated with visceral fat accumulation [76], and has been considered a potential risk factor for life-threatening complications, such as CVD and chronic kidney disease (CKD), not only a risk factor of gout [77-80]. Serum uric acid levels are regulated primarily by renal and extrarenal (intestinal) excretion and de novo synthesis in the liver; thus individuals with hyperuricemia comprise a heterogeneous population, including reduced renal/extrarenal excretion type, overproduction type, and mixed type [81]. The reduced renal/extrarenal excretion type of hyperuricemia is often related to the genetic predisposition for uric acid transporters such as ATP-binding cassette subfamily G member 2 (ABCG2), glucose transporter 9 (GLUT9), and urate transporter 1 (URAT1) [82]. On the other hand, the overproduction type is associated with visceral fat-based metabolic syndrome, including type 2 diabetes mellitus, hypertension, dyslipidemia, and NAFLD, in which XOR is thought to play a central role [83].
XOR is a rate-limiting enzyme that catalyzes the production of xanthine from hypoxanthine and uric acid from xanthine, and is the pharmacological target of anti-hyperuricemic agents such as allopurinol, febuxostat, and topiroxostat. This enzyme is a homodimeric protein with a relative molecular mass of 290 kDa, composed of three domains: one molybdopterin cofactor containing a molybdenum atom (Moco), non-identical iron–sulfur clusters (2Fe-2S), and one flavin adenine dinucleotide (FAD) as cofactors [84]. Moreover, previous studies have demonstrated that XOR is involved not only in the production of uric acid, but also in the pathogenesis of CVD [85-87]. XOR can exist in two forms, xanthine dehydrogenase (XDH) and xanthine oxidase (XO), and is initially synthesized as the XDH form. Its protein structure is then converted to the XO form under pathophysiological conditions, such as tissue hypoxia [88, 89], or after being released into the circulation [90]. Through the process of purine catabolism, XDH consumes NAD+ as a cofactor to form NADH, while XO uses oxygen to produce superoxide anions (O2–) and hydrogen peroxide. In vitro, XO has been shown to attach to negatively charged glycosaminoglycans on the apical surface of ECs [91, 92]. Therefore, the generation of XO-dependent vascular ROS has been suggested as one of the underlying mechanisms responsible for endothelial dysfunction associated with hyperuricemia [93, 94].
Certain experimental studies have provided evidence that XOR is involved in the progression of atherosclerosis. Inhibition of XOR in macrophages was reported to prevent foam cell formation [95], cholesterol crystal-induced ROS generation [96], and IL-1β secretion mediated by the Nod-like receptor family, pyrin domain containing 3 (NLRP3) inflammasome [97], whereby the XOR inhibitors, febuxostat and allopurinol, attenuated atherosclerotic plaque formation in ApoE-knockout mice [95, 96]. Febuxostat also ameliorated acute ischemia/reperfusion renal injury in mice by promoting the synthesis of adenine nucleotides, including ATP, through the hypoxanthine-guanine phosphoribosyl transferase (HGPRT)-mediated salvage pathway from HX to inosine monophosphate (IMP) [98]. On the other hand, it remains controversial whether uric acid lowering therapy with XOR inhibitors is effective in preventing the development of CVD in clinical settings. In two randomized trials, treatment with allopurinol slowed the progression of carotid intima-media thickness (IMT) in patients with asymptomatic hyperuricemia and type 2 diabetes [99] or recent ischemic stroke [100]. However, no significant difference in the effect of XOR inhibitors on carotid atherosclerosis or cardio-ankle vascular index (CAVI) has been reported in recent clinical studies [101, 102]. In the Febuxostat for cerebral and caRdiovascular Events prevention study (FREED), the primary composite event rate (all-cause mortality, cerebral, cardiovascular, and renal events) was significantly lower than in the febuxostat group versus the control group, but this result was mainly driven by differences in renal impairment [103]. In addition, a recent prospective, randomized, open-label, blinded study of 5,721 subjects with ischemic heart disease showed no difference in the rate of the primary outcomes of non-fatal myocardial infarction, non-fatal stroke, or cardiovascular death between patients receiving allopurinol and those receiving usual care [104]. Because clinical trials in hyperuricemic patients often comprise confounding factors such as obesity, hypertension, insulin resistance, and dyslipidemia, these inconsistent results with XOR inhibitors underscore the importance of clarifying the underlying mechanisms by which XOR can affect the cardiovascular system.
2.2 XOR expression and purine metabolism in adipose tissuesAccumulating clinical evidence has implicated a close relationship between hyperuricemia and visceral adiposity/metabolic syndrome. Our group has previously reported that there was a significant positive correlation between visceral fat area and serum uric acid concentrations in subjects who underwent annual health check-ups, and furthermore, the one-year change in visceral fat area was also associated with that in serum uric acid levels [105]. In fact, the prevalence of metabolic syndrome increases significantly with increasing serum uric acid [106]. These results suggest that hyperuricemia can be a surrogate marker of metabolic syndrome, potentially reflecting altered purine metabolism in adipose tissue. For this reason, we have conducted a series of studies that focus on XOR expression and the purine catabolic pathway, particularly in obese adipose tissue. In rodents, adipose tissue was one of the major organs with abundant expression of XOR, and adipose tissue of obese mice had higher XOR activities than that of lean mice. This led to increased uric acid production and secretion from obese adipose tissue via XOR [107], which raises the possibility that adipose tissue-derived uric acid contributes to the elevated circulating uric acid levels associated with obesity, at least in mice.
On the other hand, there are species differences in the tissue distribution profiles of the XOR gene between humans and mice, and in contrast to murine adipose tissue, XOR expression and activity in human adipose tissue were extremely low, with restricted expression in the liver and small intestine [108]. Consistent with this finding, both human subcutaneous and mesenteric adipose tissue secreted relatively high amounts of hypoxanthine, instead of xanthine and uric acid, suggesting that the end product of purine catabolism in human adipocytes was hypoxanthine [108]. Further metabolomic analysis using cultured human adipocytes revealed that intracellular production of hypoxanthine was significantly increased under hypoxic conditions mimicking the local environment of adipocyte hypertrophy during the development of obesity, accompanied by enhanced de novo purine synthesis through activation of the pentose phosphate pathway [108]. Besides, a recent clinical study in the general population showed that the plasma hypoxanthine concentration was positively associated with BMI, which indicated increased hypoxanthine release from obese adipose tissue in humans [109]. Taken together, these findings suggest that human adipose tissue secretes purine metabolites into the bloodstream, primarily as hypoxanthine, a substrate of XOR, rather than uric acid, which may be enhanced by obese conditions.
2.3 Increased circulating XOR activity in NAFLD conditions and its potential roles in CVDXOR protein circulates in the bloodstream and growing evidence indicates a potential link between increased plasma XOR activity and cardiometabolic disorders. Due to limited XOR expression in human organs, mainly liver and intestine, XOR activity in human plasma is typically much lower than that in rodents. However, recent methodological advances using the liquid chromatography/triple quadrupole mass spectrometry (LC/TQMS) method have allowed accurate quantitative measurements of plasma XOR activity in humans [110], and several cross-sectional studies in healthy individuals and the general population showed significant correlations between plasma XOR activity and BMI, insulin resistance (homeostasis model assessment of insulin resistance, HOMA-IR), hyperuricemia, and liver dysfunction [111, 112]. In a population-based cohort, an annual change in plasma XOR activity was independently associated with changes in body weight and liver enzymes [113]. We previously reported that, across 2 weeks of hospitalized treatment, plasma XOR activity in patients with type 2 diabetes was strongly and independently correlated with serum liver transaminases (aspartate aminotransferase, AST and aspartate aminotransferase, ALT) [114]. Similarly, in morbidly obese subjects who underwent metabolic/bariatric surgery, post-operative changes in plasma XOR activity were closely related to changes in AST and ALT, but not to changes in BMI or other clinical parameters, including serum uric acid. In addition, we also confirmed that purine catabolism, at least from hypoxanthine to xanthine, was significantly accelerated in the plasma per se of subjects with high XOR activity [115]. This result might be supported by the previous observation that plasma XOR activity was an independent predictor of circulating xanthine concentrations in the general population [109]. Collectively, combined with the much lower expression of human XOR in adipose tissue than in the liver, it is likely that the majority of XOR protein in the human bloodstream is derived from physiological hepatocyte turnover, and that elevated plasma XOR activity is primarily caused by excessive leakage of hepatic XOR associated with the chronic liver injury such as NAFLD rather than by obesity per se.
One of the most important questions arising from these findings is whether such an increase in plasma XOR has pathological significance beyond being a biomarker of liver damage. Several reports suggested an association of circulating XOR, rather than uric acid, with cardiovascular complications. Plasma XOR activities have been reported to be associated with hypertension [116], vascular endothelial dysfunction [117], and heart failure with preserved ejection fraction (HFpEF) [118], independent of uric acid levels. Accumulating clinical data indicate that NAFLD, a typical disease of abnormal ectopic fat accumulation associated with metabolic syndrome, increases the risk of CVD independent of established cardiovascular risk factors [119-121]. Although the underlying mechanism of this association has yet to be fully elucidated, several factors related to NAFLD, including insulin resistance, dietary intake, altered lipid metabolism, the gut microbiome, and the pro-inflammatory state, are thought to play a role in the pathogenesis of atherosclerosis [122-124]. In this context, we have recently provided new insights into the potential involvement of increased circulating XOR activity in such crosstalk between NAFLD and CVD progression. The pterin assays, which can measure XO activity and total XOR (XO + XDH) activity separately with and without the addition of methylene blue instead of NAD+ as the electron acceptor, showed that circulating XOR increased under NAFLD conditions was predominantly the ROS-generating XO form in both mice and humans, while XDH activity was largely unaffected. Diet-induced non-obese NAFLD model mice exhibited markedly elevated plasma XOR activity, accompanied by the development of vascular neointima formation, both of which were suppressed by treatment with topiroxostat, a selective XOR inhibitor [115]. In vitro, human ECs secreted hypoxanthine in a similar manner to human adipocytes, reflecting the absence of XOR. Furthermore, liver-derived XOR induced ROS production by catabolizing hypoxanthine released from ECs and promoted SMC proliferation with phenotypic modulation [115].
Arterial stiffness is generally recognized as a predictor of cardiovascular morbidity and mortality, and its increase reflects the remodeling of arterial structures caused by the proliferation of SMCs and connective tissue in the diffuse medial layer of the vessel wall [125, 126]. A previous cross-sectional study reported a significant positive correlation between plasma XOR activity and CAVI in patients with type 2 diabetes and liver dysfunction [127]. In addition, we have recently reported that arterial stiffness parameters (CAVI and brachial-ankle pulse wave velocity, baPWV) were significantly improved after 24 weeks of treatment with topiroxostat in hyperuricemic hypertensive patients with higher baseline ALT levels, which was accompanied by suppression of elevated plasma XOR activity [128]. These clinical and experimental results suggest that increased plasma XOR activity may be implicated in the pathogenesis of vascular neointima proliferation and atherosclerosis, particularly in NAFLD patients.
Cardiovascular damage in hyperuricemia associated with visceral adiposity/metabolic syndrome has been thought to result from complex pathogenesis, including uric acid itself and uric acid crystals. Now, circulating XOR, independent of its enzymatic product of uric acid, is emerging as a novel bioactive molecule that can be involved in CVD progression. We propose that high plasma XOR activity is directly induced by liver disease conditions, such as NAFLD, and accelerates purine catabolism in the circulation, which may be associated with the pathogenesis of endothelial dysfunction and vascular injury. Given this finding, XOR inhibitors may have the potential as a therapeutic option for the prevention of atherosclerotic disease in patients with liver dysfunction who are assumed to have high plasma XOR activity (Fig. 2). However, to date, no clinical trials have been conducted from this perspective. Therefore, in order to clarify whether XOR inhibitors can prevent or delay cardiovascular complications associated with increased plasma XOR activity, future large-scale prospective studies enrolling not only patients with hyperuricemia and gout but also those with NAFLD are warranted in the future.
Working hypothesis of promoting vascular remodeling by increased XOR activity in the circulation
In humans, the liver is the primary organ that produces the xanthine oxidoreductase (XOR) protein, and the majority of the XOR in the human bloodstream may be derived from physiological hepatocyte turnover in the steady state. However, in chronic liver disease conditions, such as non-alcoholic fatty liver disease (NAFLD), plasma XOR activity (mainly as the XO form) is dramatically increased due to excessive leakage of hepatic XOR from the damaged liver into the circulation. Hypoxanthine (HX), a substrate for XOR, is released from certain human cells, including adipocytes and vascular endothelial cells (ECs), and subsequently catabolized to at least xanthine (Xan) by markedly high plasma XOR activity. Such local purine catabolic reactions in the circulation can generate vascular reactive oxygen species (ROS) and might be involved in the initial cascade of vascular injury and atherosclerosis, represented by endothelial dysfunction and vascular smooth muscle cells (SMCs) proliferation. Therefore, suppression of elevated plasma XOR by XOR inhibitors may have the potential as a therapeutic option for the prevention of atherosclerotic disease beyond their uric acid (UA)-lowering effects, particularly in patients with NAFLD.
Numerous studies on metabolic syndrome have revealed its biological and clinical significance. However, there are still a number of issues with regard to the clinical application of the research findings to date. Further investigation of adiponectin and XOR will help to refine the understanding of the pathogenetic processes of metabolic syndrome and open a new field of therapeutic strategies for its cardiovascular complications.
All authors declare no conflict of interest in this manuscript.
We wish to thank all members of the III laboratory (Adiposcience Laboratory), Department of Metabolic Medicine, Osaka University, for their contributions to the research.
