Pro‐contractile effects of perivascular fat in health and disease

Abstract

Perivascular adipose tissue (PVAT) is now recognized as an active player in vascular homeostasis. The expansion of PVAT in obesity and its possible role in vascular dysfunction have attracted much interest. In terms of the regulation of vascular tone and blood pressure, PVAT has been shown to release vasoactive mediators, for instance, angiotensin peptides, reactive oxygen species, chemokines and cytokines. The secretory profile of PVAT is altered by obesity, hypertension and other cardiovascular diseases, leading to an imbalance between its pro‐contractile and anti‐contractile effects. PVAT adipocytes represent an important source of the mediators, but infiltrating immune cells may become more important under conditions of hypoxia and inflammation. This review describes recent advances in the effects of PVAT on the regulation of vascular tone, highlighting the evidence for a pro‐contractile action in health and disease. The role of the endothelium, vascular smooth muscle, immune cells and probably perivascular nerves in PVAT function is also discussed.

Linked Articles

This article is part of a themed section on Molecular Mechanisms Regulating Perivascular Adipose Tissue – Potential Pharmacological Targets? To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v174.20/issuetoc

Abbreviations

ADCF

adipocyte‐derived contractile factor

ADRF

adipocyte‐derived relaxing factor

Ang

angiotensin

DOCA

deoxycorticosterone acetate

MCP‐1

monocyte chemoattractant protein‐1

PVAT

perivascular adipose tissue

RANTES

regulated on activation, normal T cell expressed and secreted

Tables of Links

TARGETS
GPCRs a	Enzymes d
Adrenoceptors	ACE
AT1 receptor	Akt
Chemerin receptor (ChemR23)	AMPK
ETA receptor	COX
Voltage‐gated ion channels b	eNOS
BKCa (KCa1.1) channel	ERK
Kv7 channels	mTOR
Nuclear hormone receptors c	PKC
PPARγ	Rho kinase

LIGANDS
Adiponectin	IL‐10
Ang 1‐7	Insulin
Ang II	Leptin
Adrenaline	MCP‐1
Chemerin	NO
cGMP	Noradrenaline
ET‐1	PGE2
Hydrogen peroxide	RANTES
Hydrogen sulphide	TNFα
IL‐6	TXA2
IL‐8

These Tables list key protein targets and ligands in this article that are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Southan et al., 2016), and are permanently archived in the Concise Guide to PHARMACOLOGY 2015/16 (^a,b,c,dAlexander et al., 2015a,b,c,d).

Introduction

Perivascular adipose tissue (PVAT) was originally thought to only provide structural support and was thus routinely removed in vessel contractility studies. However, the growing prevalence of obesity, characterized by excessive adipose tissues, and the realization that adipose tissue acts as a complex paracrine and endocrine organ (Ahima and Flier, 2000; Gustafson et al., 2007) have drawn attention to a functional role for PVAT, which might also provide a mechanistic link between obesity and vascular dysfunction. PVAT is now recognized as a specialized fat depot that surrounds most blood vessels, releasing diffusible factors that modulate local vascular reactivity and inflammatory status and, as a result, may contribute to the pathophysiological changes seen in cardiovascular diseases, diabetes and obesity (reviewed in Yudkin et al., 2005; Szasz et al., 2013; Gil‐Ortega et al., 2015; Fernandez Alfonso et al., 2017). Indeed, the Framingham Heart Study shows that a higher volume of PVAT around thoracic aorta is associated with metabolic risk factors and a higher prevalence of cardiovascular disease in volunteers (Lehman et al., 2010; Britton et al., 2012). Numerous mechanisms have been suggested to underlie the crosstalk between PVAT and vascular cells, but the regulation of PVAT function, particularly the balance between its beneficial and deleterious effects, remains poorly understood. In 1991, Soltis and Cassis reported that, in rat aorta, PVAT potentiated contractions to a sympathomimetic, but reduced those to noradrenaline due to its reuptake by adrenergic nerves in PVAT. Subsequently, PVAT was also found to reduce responses to other vasoconstrictors, leading to the suggestion of PVAT‐derived or adipocyte‐derived relaxing factors (ADRFs). Much of the research has since focused on identifying these relaxing factors and establishing their vascular actions (see Gollasch, 2012; Withers et al., 2014, for reviews). However, there is also evidence that PVAT produces contractile factors, which were initially termed perivascular adipocyte‐derived contractor factors (Gao, 2007) and later adipocyte‐derived contractile factors (Meyer et al., 2013). This adds to the complex scenario of anti‐contractile versus pro‐contractile properties of PVAT. In this review, we intend to provide an update on the balance between relaxant and contractile effects of PVAT, highlighting a potential shift from an anti‐contractile action of PVAT in health to a pro‐contractile effect in obesity and related cardiovascular diseases.

Composition of PVAT

Adipose tissue surrounding blood vessels is not physically separated from the vascular wall by a fascial layer, providing access for its paracrine effects. In general, brown adipocytes are larger in size, with smaller oil droplets and larger numbers of mitochondria than white adipocytes, which store triglycerides. Morphological and gene expression analysis indicate that whilst perivascular adipocytes often resemble white adipocytes, they are distinct from visceral and subcutaneous fat and display characteristics of both white and brown adipocytes: sometimes referred to as beige adipocytes. For instance, perivascular adipocytes of human coronary arteries are smaller and irregularly shaped, with fewer differentiation markers but higher expression of some brown adipocyte‐related genes, than subcutaneous adipocytes (Chatterjee et al., 2009). The precise phenotype appears to depend on the vascular region and species (Szasz et al., 2013; Gil‐Ortega et al., 2015). Adipocytes around thoracic aorta are more similar to brown adipocytes, at least in rodents (Gálvez‐Prieto et al., 2008a; Padilla et al., 2013 ). However, adipocytes from the abdominal aorta and mesenteric arteries are closer to white adipocytes in both rodents and humans (Henrichot et al., 2005; Police et al., 2009; Padilla et al., 2013).

Importantly, adipose tissues are dynamically regulated, showing cellular and metabolic plasticity. Sustained obesity is associated with increases in the size and/or number of PVAT white adipocytes (Marchesi et al., 2009; Ketonen et al., 2010; Ma et al., 2010; Greenstein et al., 2009). These are accompanied by functional changes, including an altered secretion pattern of PVAT (Chatterjee et al., 2009; Greenstein et al., 2009; Ketonen et al., 2010). Conversely, an increased proportion of brown to white adipocytes (browning of adipose tissues) promotes thermogenesis and might represent a protective mechanism against metabolic diseases (Pellegrinelli et al., 2016) and perhaps improve vascular function in obesity and atherosclerosis (Fitzgibbons et al., 2011; Chang et al., 2012).

In addition to adipocytes, PVAT contains other important cell types such as macrophages, T‐lymphocytes and fibroblasts, which may also contribute to its function. Indeed, infiltration of immune cells into PVAT is characteristic of disease states associated with vascular inflammation (Omar et al., 2014; Pellegrinelli et al., 2016). The expansion of PVAT is also likely to involve the generation of pre‐adipocytes from resident mesenchymal stem cells and maturation of pre‐adipocytes (Pellegrinelli et al., 2016). Moreover, PVAT is also innervated by sympathetic nerves (Bulloch and Daly, 2014; Darios et al., 2016), which could stimulate the browning of PVAT. However, the interactions among perivascular adipocytes, immune cells and nerves in vascular regulation remain poorly defined. Adipocytes, which are the main cellular component of PVAT, are known to release vasoactive substances (e.g. ADRF and ADCF), but immune cells and sympathetic nerves might serve as additional sources of these molecules (e.g. Gao et al., 2006; Lumeng et al., 2007; Dashwood and Loesch, 2011; Nguyen et al., 2011). Where possible, we will highlight the likely cellular source(s) of vasoactive substances within PVAT.

Evidence for contractile factors from PVAT

Similar to adipocytes in other anatomical locations, increasing evidence suggests that PVAT secretes bioactive molecules, including adipokines and other cytokines that regulate cardiovascular function. A number of these diffusible factors can induce direct vasocontraction and may be referred to as PVAT‐derived or adipocyte‐derived contractor factors (denoted ADCFs herein), which are highlighted in Table 1. Much of the evidence comes from contractility studies using isolated arteries with and without PVAT, combined with isolated PVAT and its conditioned media under physiological conditions.

Table 1.

PVAT‐derived contractile factors

Contractile factor	PVAT expression	Effect	Vascular bed	Reference
Ang II	Ang II protein; angiotensinogen and ACE mRNA in PVAT adipocytes	↑ Sympathetic contraction	Rat superior mesenteric artery	Lu et al., 2010
	Ang II protein, angiotensinogen, ACE and chymase mRNA		Rat thoracic aorta	Galvez‐Prieto et al., 2008a
			Rat (resistance) mesenteric artery
Superoxide	Superoxide; NADPH oxidase protein in PVAT adipocytes	↑ Sympathetic contraction (via tyrosine kinase and ERK, but independent of NO)	Rat superior mesenteric artery	Gao et al., 2006
Catecholamines	Noradrenaline and adrenaline in PVAT adipocytes	Contraction	Rat thoracic aorta	Ayala‐Lopez et al., 2014
			Rat superior mesenteric artery
	Noradrenaline from sympathetic nerve endings	Contraction	Rat thoracic aorta	Soltis and Cassis, 1991
Prostanoids	TXA2 in PVAT‐conditioned buffer; COX‐1 and COX‐2 mRNA	↑ Agonist‐induced contraction (independent of NOS or ETA).	Mouse thoracic aorta (only in monogenic obesity and diet‐induced obesity)	Meyer et al., 2013
	PGE2 in PVAT (or PVAT‐conditioned buffer)	Contraction	Rat mesenteric artery	Mendizabal et al., 2013;
TNF‐α	TNF‐α protein in PVAT adipocytes	↑ Contraction to eNOS inhibition	Human small arteries from visceral fat (enhanced in obese patients)	Virdis et al., 2015
		↑ ET‐1 and NADPH oxidase‐derived superoxide in vascular cells
IL‐6	IL‐6 in PVAT adipocytes		Human coronary artery (enhanced with high‐fat diet)	Chatterjee et al., 2009
Chemerin	Chemerin protein in PVAT adipocytes	Contraction   ↑ Agonist‐induced contraction (enhanced by endothelial removal or NOS inhibition)	Rat thoracic aorta   Rat superior mesenteric artery   (enhanced responses in DOCA‐salt hypertensive but not diet‐induced obese or SHRSP rats)	Watts et al., 2013
	Chemerin protein in PVAT adipocytes	Contraction   ↑ Agonist‐induced contraction	Human resistance mesenteric artery	Watts et al., 2013
	Chemerin protein in PVAT adipocytes	↑ Sympathetic contraction	Rat superior mesenteric artery	Darios et al., 2016
Leptin	PVAT‐conditioned buffer	↑ Agonist‐ and depolarisation‐induced contraction (via increased voltage‐gated Ca2+ entry)   ↑ Smooth muscle proliferation	Pig coronary artery (enhanced in diet‐induced obesity)	Owen et al., 2013; Noblet et al., 2016

↑, increase; ➔, lead to; eNOS, endothelial NOS; ET_A, endothelin ET_A receptor; SHRSP, stroke‐prone spontaneously hypertensive rat.

This table shows studies that demonstrate PVAT production of mediators, which either induce direct contraction or potentiate contractions to other vasoconstrictors.

Adipocytes are known to express a local renin–angiotensin–aldosterone system (RAAS), including angiotensinogen and angiotensin converting enzyme (ACE) for the synthesis of the potent vasoconstrictor, angiotensin II (Ang II; Karlsson et al., 1998; Cassis et al., 2008). The expression of RAAS components can vary depending on the composition and location of adipose tissues (Cassis et al., 1988; Engeli et al., 1999; Galvez‐Prieto et al., 2008a; Riedel et al., 2016). PVAT is thought to express all components of RAAS and that PVAT‐derived Ang II promotes contractions through the activation of AT₁ receptors in rat mesenteric arteries (Lu et al., 2010). Gao and co‐workers proposed that Ang II acts indirectly by stimulating superoxide radical production from NADPH oxidase in PVAT adipocytes or the vascular wall itself (Gao et al., 2006; Lu et al., 2008). Ang II has also been shown to play a role in the local inflammation associated with hypertension and obesity, stimulating the infiltration of immune cells, including T‐lymphocytes and macrophages, into PVAT and th eproduction of reactive oxygen species (Police et al., 2009; Guzik et al., 2007; Mikolajczyk et al., 2016). However, the importance of PVAT as a source of Ang II in the control of vascular tone and blood pressure, particular in hypertension and obesity, remains to be established. Moreover, it is likely that the production and function of PVAT‐derived Ang II show regional heterogeneity (Galvez‐Prieto et al., 2008a). Ang II can also further exacerbate PVAT dysfunction, since AT₁ receptor activation has been shown to reduce th ebrowning of adipose tissue and promote adipocyte hypertrophy, insulin resistance and weight gain in mice with high fat‐induced obesity (Graus‐Nunes et al., 2017).

In the initial study by Soltis and Cassis (1991), PVAT greatly enhanced contractions to electrical field stimulation or to the indirect sympathomimetic tyramine in rat aorta, suggesting sympathetic nerve activity is involved in this effect of PVAT. The presence of sympathetic nerves has been demonstrated in PVAT of human saphenous veins (Dashwood and Loesch, 2011). Also, the involvement of sympathetic innervation in the regulation of vascular tone and blood pressure is well established. An elevated sympathetic activity is also associated with hypertension, including obesity‐associated hypertension (Thalmann and Meier, 2007); however, the interaction between PVAT and local sympathetic activity in healthy and disease conditions has not been scrutinized. Sympathetic activity and the subsequent release of catecholamines are known to regulate lipolysis and the proliferation and differentiation of adipocytes activation through the stimulation of α‐ and β‐adrenoceptors. Recent evidence has also suggested that adipocytes and alternatively activated macrophages in adipose tissues may synthesize and release noradrenaline and adrenaline (Nguyen et al., 2011; Vargovic et al., 2011). Noradrenaline and the enzymes involved in its synthesis have been detected in PVAT adipoctyes in thoracic aorta and superior mesenteric arteries, where PVAT enhances contraction via α₁‐adrenoceptors, (Ayala‐Lopez et al., 2014). PVAT‐dependent contractions to the sympathomimetic tyramine have also been reported in these arteries (Soltis and Cassis, 1991; Ayala‐Lopez et al., 2014). Tyramine is traditionally used to release catecholamines from sympathetic nerve endings, but the possibility that it has an effect on adipocytes or immune cells in PVAT cannot be excluded.

Another factor that may contribute to the contractile effects of PVAT is the adipokine, chemerin (Table 1). Chemerin, in particular chemerin‐9, evokes direct vasocontraction and enhances agonist‐induced contractions via its GPCR, ChemR23, in rat and human arteries. Moreover, these effects are exaggerated in thoracic aorta and mesenteric arteries with reduced endothelium‐dependent relaxation, a phenomenon often found in hypertension and obesity (Watts et al., 2013). A follow‐up study by the same group (Darios et al., 2016) has also shown that PVAT‐derived chemerin potentiates sympathetic contraction through its receptor, which is co‐localized with tyrosine hydrolase in sympathetic nerves of rat superior mesenteric artery. Direct application of chemerin to isolated aorta or mesenteric artery also augments agonist‐induced contraction in a manner dependent on endothelin ET_A receptors and ERK activation (Lobato et al., 2012) and increases systolic blood pressure in mice (Kunimoto et al., 2015). Thus, chemerin might play a particularly important role in some forms of hypertension and obesity.

In addition to chemerin, cytokines derived from PVAT might also increase vascular tone. For instance, TNF‐α and IL‐6 are known to enhance contractions, probably via up‐regulation of endothelin signalling or reduced NO production and endothelium‐dependent relaxation, especially in obese patients (Greenberg et al., 1985; Orshal and Khalil, 2004; Virdis et al., 2015). High‐fat diet has also been shown to promote IL‐6 expression in human coronary PVAT (Chatterjee et al., 2009).

Aortic and small mesenteric PVAT also releases contractile COX products, including TXA₂ and PGE₂ (Meyer et al., 2013; Mendizabal et al., 2013). In the same vascular regions, contractile responses to prostanoids and the expression of their receptors are enhanced in obese mice (Traupe et al., 2002) and diabetic rats (Ishida et al., 2012). Interestingly, however, significant amounts of PVAT‐derived TXA₂ and PGE₂ are also detected in healthy controls, suggesting a possible physiological role for PVAT (Meyer et al., 2013). Prostanoids might also mediate PVAT‐induced endothelial dysfunction in both normotensive and hypertensive rats (Mendizabal et al., 2013), contributing to its pro‐contractile effects. It remains to be clarified how the secretory pattern of various prostanoids is altered in pathophysiological conditions.

Taken together, these findings indicate that PVAT is capable of releasing various contractile factors, which elicit direct vasocontraction or enhance nerve‐ or agonist‐mediated contractions by acting on the vascular smooth muscle. These factors appear, at least partly, to be active even in normal, healthy circumstances especially in larger arteries. Previous studies have reported elevated systemic levels of angiotensin II, superoxide, catecholamines, contractile prostanoids, TNF‐α, chemerin and leptin in hypertension, diabetes and obesity (Brunner et al., 2005; Gu et al., 2015), but PVAT is yet to be established as a major source of these mediators. There is, however, evidence pointing to an increased responsiveness to chemerin, TNF‐α and prostanoids in aorta and resistance arteries (Traupe et al., 2002; Ishida et al., 2012; Meyer et al., 2013; Watts et al., 2013; Virdis et al., 2015). In addition to acute vasocontraction, sustained elevation of some of the ADCF, such as superoxide, Ang II and TNF‐α, might stimulate vascular smooth muscle growth and arterial stiffness (Almabrouk et al., 2014; Fleenor et al., 2014; Kunimoto et al., 2015; Noblet et al., 2016; also reviewed by Miao and Li, 2012, Aroor et al., 2013 and Villacorta and Chang, 2015), commonly found in atherosclerosis, hypertension and ageing. In line with this, the expression of chemerin in PVAT is positively correlated with atherosclerosis in human aorta and coronary artery (Spiroglou et al., 2010). The vascular remodelling effect of PVAT is also associated with endothelial dysfunction, a hallmark of cardiovascular diseases. A reduction in the endothelium‐dependent relaxation of vascular smooth muscle would exaggerate the pro‐contractile effects of PVAT and this will be explored further in the following section.

Evidence for PVAT‐induced endothelial dysfunction

The vascular endothelium is critical for maintaining cardiovascular homeostasis, and its dysfunction is considered an early sign or predictor of cardiovascular diseases, including those associated with obesity and diabetes (Brunner et al., 2005). Endothelial dysfunction can manifest, for example, as reduced endothelium‐dependent relaxation, endothelium‐dependent contraction, leukocyte adhesion and reduced anti‐coagulation properties. In Table 2, we highlight some of the studies demonstrating the inhibitory effect of PVAT on responses to endothelium‐dependent relaxants, which could enhance vasocontraction and might be particularly relevant for hypertension linked to obesity and diabetes. Where possible, the specific PVAT‐derived mediators and disease conditions involved are also indicated in Table 2.

Table 2.

Inhibitory effect of PVAT on endothelium‐dependent relaxation

PVAT‐derived factor	Proposed mechanism of inhibition	Endothelium‐dependent relaxant affected	Vascular bed	Reference
Resistin	↓ IRS‐1 and PI3K activity in endothelium ➔ ↓ eNOS activity	Insulin	Mouse aorta Mouse mesenteric artery	Gentile et al., 2008
	↑ Endothelial superoxide ➔ ↓ eNOS expression	Bradykinin	Pig coronary artery	Kougias et al., 2005
Adiponectin	↓ Adiponectin in PVAT ➔ ↓ AMPK and Akt phosphorylation	Insulin	Mouse resistance artery from skeletal muscle (only in genetic model of obesity and type 2 diabetes)	Meijer et al., 2013
Unknown	↓ AMPK phosphorylation ➔ ↑ mTOR phosphorylation ➔ ↓ eNOS expression	Acetylcholine	Rat thoracic aorta and mesenteric artery (only in diet‐induced obesity)	Ma et al., 2010
Leptin	↑ Leptin in PVAT and ↑ leptin receptor expression ➔ ↑ PKCβ activity in vascular cells	Bradykinin	Pig coronary artery (only in obesity with metabolic syndrome)	Payne et al., 2010
Unknown	↑ PKCβ‐mediated eNOS phosphorylation ➔ ↓ endothelial NO	Bradykinin	Dog coronary artery	Payne et al., 2008; Payne et al., 2009
Unknown	↓ Endothelial Ca2+ signal	Acetylcholine Methacholine	Rat coronary septal artery	Aalbaek et al., 2015
Unknown (but independent of superoxide, prostanoids, ET‐1 and AT1)	↑ Endothelial caveolin‐1 ➔ ↓ NO production	Acetylcholine	Rat thoracic aorta	Lee et al., 2014
Superoxide   Hydrogen peroxide   Leptin   MCP‐1	↑ Superoxide, hydrogen peroxide, leptin and MCP‐1 in PVAT	Acetylcholine	Mouse abdominal aorta (only in diet‐induced obesity)	Ketonen et al., 2010
Visfatin	↑ NADPH oxidase activity in vascular cells ➔ ↓ endothelial NO	Bradykinin or acetylcholine	Rat and human resistance mesenteric arteries	Vallejo et al., 2011
TNF‐α   Adiponectin	↑ TNF‐α and ↓ adiponectin in PVAT ➔ NADPH oxidase activation ➔ ↓ eNOS expression ➔ ↓ basal endothelial NO   ↑ superoxide and ET‐1 in vascular cells		Human small arteries from visceral fat (enhanced effects in obesity)	Virdis et al., 2015
RANTES	↑ RANTES ➔ ↑ T‐lymphocytes in PVAT but not visceral fat   ↑ Ang II‐induced superoxide production in vascular cells	Acetylcholine	Mouse thoracic and abdominal aorta (enhanced in Ang II‐induced hypertension)	Mikolajczyk et al., 2016

↑, increase; ↓, decrease; →, lead to; AT₁, Ang II receptor type 1; eNOS, endothelial NOS; ET‐1, endothelin‐1; IRS‐1, insulin receptor substrate‐1; mTOR, mechanistic target of rapamycin; PI3K, phosphoinositide 3‐kinase.

An inhibitory effect of PVAT is often demonstrated by studying the effects of PVAT on responses to endothelium‐dependent relaxants that are applied to isolated arteries or isolated vascular cells. The up‐ or down‐regulation of PVAT‐derived factors are thought to exaggerate the reduction in endothelium‐dependent relaxation in disease states. However, in some studies, the diffusible factors responsible for the inhibitory effects of PVAT on endothelial function are not identified

A primary mechanism of action for PVAT is to reduce NO production or bioavailability, although NO‐independent signalling pathways may also be compromised. Given the physical distance between PVAT and the endothelium particularly in conduit arteries, it is thought that mediators released by PVAT are involved. They include NADPH oxidase‐derived reactive oxygen species (superoxide and hydrogen peroxide) and pro‐inflammatory cytokines (leptin, TNF‐α, IL‐6, resistin and visfatin) (Greenstein et al., 2009; Marchesi et al., 2009; Ketonen et al., 2010; Payne et al., 2010; Vallejo et al., 2011; Aghamohammadzadeh et al., 2016). Importantly, targeting dysregulation of these PVAT factors, which accompanies adipocyte hypertrophy in obesity and metabolic syndrome can improve endothelial function (Marchesi et al., 2009; Aghamohammadzadeh et al., 2016). Circulating visfatin levels may also predict the extent of endothelium‐dependent, flow‐mediated dilation in patients with atherosclerosis and diabetes (Romacho et al., 2013). These findings support the clinical importance of PVAT dysfunction in vascular health. Indeed, oxidative stress and increased production of pro‐inflammatory cytokines, as well as endothelial dysfunction, have been closely linked to the pathophysiology of obesity, hypertension, atherosclerosis and insulin resistance.

Moreover, PVAT can also reduce endothelium‐independent relaxation. In many studies, the presence of PVAT has no significant effect on relaxation to NO donors (Payne et al., 2008; Ma et al., 2010; Vallejo et al., 2011; Lee et al., 2014). However, Tune and co‐workers (Owen et al., 2013; Noblet et al., 2015) have shown that PVAT inhibits distinct subtypes of K⁺ channels in coronary smooth muscle of lean versus diet‐induced obese pigs. Another adipokine, nesfatin‐1, has also been shown to reduce smooth muscle cGMP production in mesenteric arteries and increase arterial blood pressure in rats (Yamawaki et al., 2012).

Evidence for relaxant factors from PVAT

In contrast to the aforementioned (pro)‐contractile actions, numerous PVAT‐derived mediators are vasorelaxants and therefore exert anti‐contractile effects; these have been the focus of a number of excellent reviews (e.g. Gollasch, 2012 and Withers et al., 2014). PVAT relaxants include adiponectin, omentin, leptin, Ang 1–7, NO, hydrogen peroxide and hydrogen sulphide (Dubrovska et al., 2004; Gao et al., 2007; Lee et al., 2009; Gil‐Ortega et al., 2010; Payne et al., 2010; Schleifenbaum et al., 2010). Again, isolated tension recording and bioassay experiments have been instrumental in establishing an anti‐contractile action of PVAT in arteries from rodents and humans. Of note, the presence of PVAT reduces contraction to some, but not all, vasoconstrictors (Soltis and Cassis, 1991; Lohn et al., 2002; Verlohren et al., 2004; Gao et al., 2005b; Malinowski et al., 2008; Greenstein et al., 2009). Under physiological conditions, adipocytes are thought to be the main cellular source of these factors, which are sometimes referred to as ADRFs (Soltis and Cassis, 1991; Lohn et al., 2002; Verlohren et al., 2004).

Diverse signalling mechanisms have been proposed, including endothelial NO release, cGMP generation, reactive oxygen species and opening of various K⁺ channel subtypes, but independent of COX products or sympathetic nerves (Gollasch, 2012 and Withers et al., 2014). Accumulating evidence suggests that the anti‐contractile effect of PVAT relies on smooth muscle K⁺ channels, specifically the activation of voltage‐gated K⁺ channels (K_V7) and Ca²⁺‐activated K⁺ channels (BK_Ca) through endothelium‐independent and ‐dependent pathways respectively. Interestingly, in healthy rat coronary septal arteries, increases in PVAT also reduce Rho kinase‐dependent Ca²⁺ sensitivity in vascular smooth muscle (Aalbaek et al., 2015). This contrasts with the observation that PVAT from pig coronary artery enhances vasocontraction via Rho kinase (Owen et al., 2013).

As for a pathological role, a loss or reduced relaxant effect of PVAT is often reported in disease states. In spontaneously hypertensive rats, there is a loss of anti‐contractile effect in mesenteric arteries possibly due to a down‐regulation of K_V7 channels in vascular smooth muscle or reduced PVAT‐induced production of Ang 1–7, one of the ADRF candidates (Galvez et al., 2006; Galvez‐Prieto et al., 2008b; Li et al., 2013). In experimental models of obesity and metabolic syndrome, increases in PVAT‐derived leptin (Payne et al., 2010; Ketonen et al., 2010), superoxide, hydrogen peroxide (Gao et al., 2005a; Ketonen et al., 2010; Rebolledo et al., 2010; Aghamohammadzadeh et al., 2016) or free fatty acid (Sun et al., 2013) also play a role in aorta, mesenteric or subcutaneous arteries. Although PVAT produces the vasorelaxants adiponectin and NO, obesity is associated with a reduction in PVAT‐derived adiponectin and reduced endothelial NO release and bioavailability, partly due to oxidative stress (cf. Figure 2). For example, in rodent mesenteric arteries, this can be prevented by superoxide dismutase and catalase, which remove superoxide and hydrogen peroxide respectively, or antioxidants (Marchesi et al., 2009; Aghamohammadzadeh et al., 2016). It has been shown in models of hypertension and obesity that, despite an up‐regulation of leptin, there is an impairment of leptin‐induced NO release from the endothelium (Beltowski et al., 2003; Rahmouni et al., 2005; Galvez‐Prieto et al., 2012). This implicates PVAT in vascular leptin resistance, which exacerbates the cardiovascular complications associated with obesity.

Other data suggest that, in obesity and diabetes, a down‐regulation of PVAT‐derived adiponectin might lead to an up‐regulation of superoxide and TNF‐α and reduced endothelial NO production and relaxation (cf. Table 2; Virdis et al., 2015; Hou et al., 2016; Nacci et al., 2016). It is often unclear why PVAT‐derived relaxants are down‐regulated in disease states, but hypoxia in PVAT might be a contributing factor (Withers et al., 2011). A local reduction of PVAT adiponectin (Virdis et al., 2015; Aghamohammadzadeh et al., 2016) without concomitant changes in its circulatory levels is also evident in human obesity (Dreier et al., 2016). Of particular relevance to insulin resistance and diabetes, adiponectin is known to activate AMP‐activated protein kinase (AMPK), a key intracellular energy sensor that improves insulin sensitivity, and modulate adipocyte metabolism and inflammation (see Almabrouk et al., 2014 for reivew). AMPK in endothelium and vascular smooth muscle also regulates vascular tone and remodelling (Ma et al., 2010; Meijer et al., 2013; Almabrouk et al., 2014) and probably contributes to the cardiovascular benefits of the anti‐diabetic drugs, glitazones, which are PPARγ agonists and AMPK activators. Interestingly, a recent study suggests that AMPK in PVAT is required for the secretion of adiponectin in mouse aorta (Almabrouk et al., 2016), providing a molecular mechanism for crosstalks with hypoxia and other PVAT‐derived vasoactive substances that activate or inhibit AMPK (Almabrouk et al., 2014; Virdis et al., 2015).

Anti‐contractile versus pro‐contractile effects of PVAT

The coexistence of pro‐contractile and anti‐contractile actions of PVAT may seem contradictory, but such dual effects have also been demonstrated in the same arteries and within the same studies (Soltis and Cassis, 1991; Lohn et al., 2002; Ketonen et al., 2010; Li et al., 2013; Aalbaek et al., 2015). Indeed, some of the PVAT‐derived factors such as leptin, TNF‐α, IL‐6 and hydrogen peroxide are known to have both contractile and relaxant properties (Brian and Faraci, 1998; Orshal and Khalil, 2004; Thakali et al., 2006; Virdis et al., 2015). The up‐regulation or down‐regulation of these factors can also compromise endothelial function (cf. Table 2). It is therefore not surprising that the balance between pro‐contractile and anti‐contractile function and how it is altered in disease conditions is under increasingly intense investigations.

Both relaxant and contractile actions of PVAT effects are detectable in healthy conditions, at least in thoracic aorta and mesenteric and coronary artery (Soltis and Cassis, 1991; Dubrovska et al., 2004; Verlohren et al. 2004; Payne et al., 2010). Many studies have proposed a predominantly anti‐contractile action in health, although it is possible that the net effect on vascular tone depends on the anatomical location and experimental conditions used. As discussed in the previous section, systemic arteries (including mesenteric artery and thoracic and abdominal aorta) and coronary arteries often show a reduced production or responsiveness to PVAT‐derived relaxants or other vasorelaxants in hypertension, obesity and diabetes. This, together with an underlying contraction induced by PVAT, which can also be enhanced in some forms of hypertension and obesity (cf. Table 1), would promote a net contractile action of PVAT (Figure 1). This may result in sustained vasoconstriction. Indeed, PVAT dysfunction is correlated with raised arterial blood pressure in obese rats (Aghamohammadzadeh et al., 2016). Genetic deletion of PPARγ in mouse perivascular adipocytes during development results in the absence of PVAT and hypotension, indicating a key role for PVAT in the regulation of blood pressure (Chang et al., 2012). However, we are yet to fully understand how PVAT function transitions from health to disease and how best to reverse the adverse effects of PVAT. In the case of obesity, weight loss through bariatric surgery or caloric restriction might reduce PVAT‐mediated inflammation and improve NO bioavailability, resulting in normalized blood pressure (Aghamohammadzadeh et al., 2013; Bussey et al., 2016).

Figure 1.

Regulation of vascular tone by PVAT in health and disease. (A) PVAT releases a diverse group of bioactive and diffusible substances, including leptin, adiponectin, Ang II, Ang 1–7, catecholamines, reactive oxygen species, NO, hydrogen sulphide, cytokines such as TNF‐α and IL‐6, and chemokines such as MCP‐1 and RANTES. These mediators modulate vascular tone through a paracrine action on the endothelium, vascular smooth muscle and immune cells. The chemokines and cytokines regulate the migration of immune cells into PVAT, and activated macrophages and lymphocytes within PVAT can also release additional cytokines. (B) In healthy conditions, PVAT tends to exert a net anti‐contractile effect. Pathophysiological stimuli for example in obesity, hypertension and diabetes alter the secretion pattern of PVAT, leading to increased pro‐contractile and decreased anti‐contractile actions. This imbalance is characteristic of PVAT dysfunction in disease states. Other changes in PVAT composition and function include adipocyte hypertrophy, infiltration of macrophages and lymphocytes, and inflammation within PVAT and vascular cells. ADV, adventitia; EC, endothelium; SM, smooth muscle.

Interactions between adipocytes and immune cells in PVAT

Aside from vascular reactivity, many of the PVAT‐derived mediators are also critical players in vascular inflammation. Evidence suggests that a pro‐inflammatory phenotype of PVAT is a common feature of hypertension, obesity, insulin resistance and atherosclerosis (Chatterjee et al., 2009; Almabrouk et al., 2014; Omar et al., 2014; Mikolajczyk et al., 2016). Adipocytes are the main component in PVAT, but immune cells such as macrophages and T‐lymphocytes also play an important role in regulating PVAT function and provide an alternative source of vasoactive mediators. As part of the pathological remodelling of adipose tissues, obese rodents and humans have a higher PVAT mass and adipocyte hypertrophy (Greenstein et al., 2009; Marchesi et al., 2009; Ma et al., 2010). The hypertrophied PVAT probably exceeds the diffusion limit of oxygen and suffers from hypoperfusion, leading to local hypoxia (Hosogai et al., 2007; Greenstein et al., 2009). The hypoxic state is linked to an increased expression of the chemokine monocyte chemoattractant protein‐1 (MCP‐1 also known as CCL2 ) in PVAT, which in turn promotes the recruitment and infiltration of macrophages, which act as a major source of TNF‐α (see Gustafson et al., 2007, for a review; Ketonen et al., 2010). From models of obesity, hypertension and metabolic syndrome, it has been demonstrated that PVAT also stimulates the recruitment of monocytes and lymphocytes in arteries by up‐regulating of the expression of the chemokines IL‐8 and RANTES and superoxide in vascular cells (Table 2) (Henrichot et al., 2005; Marchesi et al., 2009; Mikolajczyk et al., 2016). Figure 2 illustrates how the dysregulation of PVAT‐derived factors might occur in these disease states.

Figure 2.

Proposed mechanisms of PVAT dysfunction in obesity, diabetes and hypertension. PVAT dysfunction is characterized by changes in its secretory pattern and increased occurrence of activated macrophages and lymphocytes in PVAT. In addition to adipocytes, activated immune cells within PVAT also release additional cytokines. In obesity and Type 2 diabetes, PVAT dysfunction is probably triggered by adipocyte hypertrophy and increases in PVAT mass. PVAT expansion is not a prerequisite for PVAT dysfunction, since the size of adipocytes and overall PVAT mass may be reduced in some forms of hypertension in the absence of obesity.

At the same time as there is an up‐regulation of pro‐inflammatory mediators from PVAT (e.g. TNF‐α and IL‐6), there is also a down‐regulation of anti‐inflammatory mediators (e.g. adiponectin and IL‐10), from adipocytes and macrophages (Greenstein et al., 2009; Chatterjee et al., 2009; Lumeng et al., 2007). The resultant pro‐inflammatory phenotype has been linked to the loss of PVAT‐induced relaxation, and this deficit may be partially reversed by TNF‐α antagonists or IL‐6 antibodies (Greenstein et al., 2009; Ozen et al., 2015; Aghamohammadzadeh et al., 2016). Endothelium‐dependent relaxation is also likely to be compromised by PVAT‐derived TNF‐α and reactive oxygen species (Virdis et al., 2015). In macrophage‐deficient mice, the ability of hypoxia to inhibit PVAT‐mediated relaxation is greatly reduced, supporting a key role for these macrophages (Withers et al., 2011). In addition, there is an accumulation of leukocytes in PVAT, so much so that a deficit in P‐selectin glycoprotein ligand‐1, a ligand essential for leukocyte attachment and rolling at the endothelium, prevents the endothelial dysfunction and inflammation mediated by PVAT in obese mice (Wang et al., 2012).

In contrast to the conditions in obesity, the size of PVAT adipocytes and PVAT mass are often reduced in experimental models of hypertension, including spontaneously hypertensive and DOCA–salt hypertensive rats (Galvez et al., 2006; Ruan et al., 2010). Despite this, a recent study has demonstrated that Ang II‐induced hypertension increases PVAT‐mediated expression of RANTES, resulting in increased T‐lymphocyte infiltration and impaired endothelium‐dependent relaxation (Mikolajczyk et al., 2016; Figure 2). It should be noted that whilst adipokines such as chemerin, adiponectin and leptin are released from PVAT adipocytes, many of the PVAT‐derived chemokines, cytokines and reactive oxygen species might be produced by multiple cell types within PVAT, including adipocytes and immune cells (see also Szasz et al., 2013 and Pellegrinelli et al., 2016, for reviews). Furthermore, vascular cells are responsible for some of the pro‐ and anti‐inflammatory mediators produced by PVAT and also express receptors for these mediators. Therefore, the interplay among adipocytes, infiltrated immune cells and vascular cells needs to be explored further.

PVAT dysfunction in humans

The past 10 years has seen a growing interest in PVAT dysfunction. While the current data on human PVAT remain limited, they broadly agree with those obtained in animal models. For example, the anti‐contractile effects of PVAT are compromised in small arteries of patients with metabolic syndrome or obesity (Greenstein et al., 2009; Aghamohammadzadeh et al., 2013), and that human PVAT shows a distinctive pattern of expression of pro‐inflammatory mediators, including IL‐6, MCP‐1 and leptin, compared with subcutaneous adipose tissue (Rittig et al., 2012; Mauro et al., 2013). However, the functional significance of various ADCFs and ADRFs may differ. For example, adiponectin appears to play a more important role in PVAT relaxation in humans than in rodents (Fesus et al., 2007; Greenstein et al., 2009; Meijer et al., 2013). In patients undergoing coronary bypass surgery, initial experiments suggest that a saphenous vein graft with intact adventitia and PVAT, as opposed to a conventional PVAT‐free graft, reduces vasospasm and potentially improves its patency (Dashwood et al., 2009). Thus, it might be concluded that PVAT exerts a predominantly anti‐contractile effect in humans. In contrast, individuals with more PVAT in their brachial artery have a diminished hyperaemic blood flow (Rittig et al., 2008), suggesting a basal contractile effect of PVAT. Further characterization of PVAT‐derived factors from different vascular regions is needed. Thus far, mechanistic studies have been performed on the more accessible vessels from volunteers, namely, the internal thoracic artery (Gao et al., 2005a; Malinowski et al., 2008), small arteries in gluteal fat (Greenstein et al., 2009; Aghamohammadzadeh et al., 2013) and saphenous veins (Dashwood et al., 2009).

The Framingham Heart Study reported a correlation between periaortic fat mass and hypertension, and diabetes, irrespective of the body mass index, but a causal relationship is yet to be established (Lehman et al., 2010; Britton et al., 2012). In addition to reduced body weight, bariatric surgery in severely obese patients has been shown to restore PVAT‐induced relaxation, improve the inflammatory cytokine profile and NO bioavailability and reduce macrophage infiltration and systolic blood pressure (Aghamohammadzadeh et al., 2013). More recently, in diet‐induced obese rats, calorie restriction and sustained weight loss has similarly been found to reverse PVAT‐mediated vascular damage (Bussey et al., 2016). These data support the concept that PVAT dysfunction contributes to the pathogenesis of obesity and metabolic syndrome. An increase in the circulatory levels of chemerin has also been correlated with impaired endothelial function and increased arterial stiffness in hypertensive patients (Gu et al., 2015). The specific role played by PVAT relative to other fat depots merits further investigations, particularly in view of the differential responses to high‐fat diet in white and brown adipocytes (Fitzgibbons et al., 2011).

When assessing data from animal and human experimental studies, it is also important to consider the effect of ageing, an independent risk factor for cardiovascular diseases. Ageing exacerbates aortic PVAT dysfunction, with increases in oxidative stress and macrophage infiltration and a pro‐inflammatory secretion pattern of cytokines and chemokines (Bailey‐Downs et al., 2013; Mauro et al., 2013; Fleenor et al., 2014). This effect is at least partly mimicked by medium conditioned with PVAT from aged aorta and is accompanied by endothelial dysfunction in aorta, especially in diet‐induced obese mice (Bailey‐Downs et al., 2013). Thus, PVAT may also contribute to the endothelial dysfunction and vascular remodelling seen in ageing. Moreover, PVAT‐mediated relaxation is inhibited in ageing mice (Agabiti‐Rosei et al., 2017), hinting at an overall pro‐contractile action of PVAT during ageing. Further investigations are needed to clarify the function of PVAT in healthy versus pathological ageing. It is also worth noting that the age of rodents used in models of obesity varies (Lutz and Woods, 2012) and is generally younger than that of subjects involved in clinical studies.

Conclusion

In this review, we have focused on the effects of PVAT on vascular tone regulation. It is apparent that PVAT exerts both contractile and relaxant actions through the release of autocrine/paracrine factors from adipocytes and infiltrating inflammatory cells (Figure 1). PVAT is therefore an integral part of vascular function, including crosstalk with the endothelium, smooth muscle, immune cells and perivascular nerves. The balance between pro‐contractile and anti‐contractile effects maybe tissue‐specific, but its modulation by obesity and hypertension induces a shift towards a pro‐contractile, pro‐inflammatory and pro‐oxidative phenotype (Figure 2). This PVAT dysfunction may also occur in other obesity‐related disorders, including metabolic syndrome, diabetes and atherosclerosis. However, despite a much better understanding of the structure and function of PVAT, there are still many unanswered questions. The molecular mechanisms that regulate PVAT quantity and composition, and secretion of vasoactive factors in health and disease, which probably exist in a continuum, remain elusive. Although there is experimental evidence for PVAT dysfunction in the pathogenesis of hypercontractility in disease states, it is also possible that it has a protective and adaptive role in vascular homeostasis. Moreover, the function of PVAT relative to the systemic influence of visceral and subcutaneous fat remains to be clarified. Given the heterogeneity of PVAT function in different anatomical regions and species, more studies in human tissues are required.

Conflict of interest

The authors declare no conflicts of interest.

Ramirez, J. G. , O'Malley, E. J. , and Ho, W. S. V. (2017) Pro‐contractile effects of perivascular fat in health and disease. British Journal of Pharmacology, 174: 3482–3495. doi: 10.1111/bph.13767.