Obesity and adipokines: effects on sympathetic overactivity

Abstract

Excess body weight is a major risk factor for cardiovascular disease, increasing the risk of hypertension, hyperglycaemia and dyslipidaemia, recognized as the metabolic syndrome. Adipose tissue acts as an endocrine organ by producing various signalling cytokines called adipokines (including leptin, free fatty acids, tumour necrosis factor-α, interleukin-6, C-reactive protein, angiotensinogen and adiponectin). A chronic dysregulation of certain adipokines can have deleterious effects on insulin signalling. Chronic sympathetic overactivity is also known to be present in central obesity, and recent findings demonstrate the consequence of an elevated sympathetic outflow to organs such as the heart, kidneys and blood vessels. Chronic sympathetic nervous system overactivity can also contribute to a further decline of insulin sensitivity, creating a vicious cycle that may contribute to the development of the metabolic syndrome and hypertension. The cause of this overactivity is not clear, but may be driven by certain adipokines. The purpose of this review is to summarize how obesity, notably central or visceral as observed in the metabolic syndrome, leads to adipokine expression contributing to changes in insulin sensitivity and overactivity of the sympathetic nervous system.

Christopher Minson is a Professor of Human Physiology at the University of Oregon. He has a long-standing interest in sympathetic and vascular regulation in humans. Primary areas of research include investigating the effects of sex steroids on sympathetic and vascular function, and understanding neural–vascular interactions in the skin. Current projects include investigating the role of the sympathetic nervous system in obesity and insulin resistance, particularly related to sex steroid regulation of endocrine function in women, and how alterations in normal function may contribute to clinical conditions such as the polycystic ovary syndrome.

Introduction

Obesity is a growing world epidemic resulting in massive increases in medical costs. It has been estimated that medical costs for obese patients is 42% higher than normal weight patients and the annual financial burden from obesity on the medical system in the United States alone is estimated to be as high as $147 billion annually (Finkelstein et al. 2009). In addition to cost, there are a cluster of well known co-morbidities associated with obesity, including but not limited to cardiovascular disease, type 2 diabetes, osteoarthritis and specific cancers (Dixon, 2009). Obesity also contributes to cognitive impairment and increased risk for dementia later in life, independent of co-morbidities such as cardiovascular disease and diabetes (Whitmer et al. 2008).

The purpose of this review is to present an overview of our current understanding of the relationships between obesity, insulin resistance and sympathetic outflow. A review of this topic is timely in that obesity has recently been recognized to contribute to cardiovascular mortality and morbidity through an increased sympathetic drive leading to end organ damage and hypertension (Lambert et al. 2010). Additionally, an elevated sympathetic outflow contributes to the decline of insulin sensitivity and increased mortality from diseases like advanced heart failure (Munhoz et al. 2009) and post-acute thromboembolic cerebrovascular accident (Sander et al. 2001). Examining the different factors that may contribute to increased sympathetic outflow could help in the development of targets for treatment in an effort to reduce the consequences of obesity. This paper will review the effects of increased expression of key cytokines released from adipocytes, termed adipokines, including how they work independently and together to create a state of insulin resistance and chronic sympathetic overactivity.

Obesity and sympathetic overdrive

Increased body fat deposition has been specifically correlated to sympathetic overdrive at rest, with resting levels of muscle sympathetic nerve activity (MSNA) in the obese reported as greater than 50% higher in some studies (Alvarez et al. 2002; Sivenius et al. 2003; Grassi et al. 2004). Elevated levels of MSNA are associated with obesity-induced subclinical organ damage to the heart, blood vessels, and kidneys in young subjects, even in the absence of hypertension (Lambert et al. 2010). To date, there have been four competing hypotheses put forth to explain the aetiology of the metabolic syndrome, and sympathetic overactivity seems to play a fundamental role in all of them (Straznicky et al. 2008). This strongly suggests the sympathetic nervous system (SNS) and factors that contribute to its overactivity in obesity should be considered a target for intervention early in the process of metabolic syndrome development.

Activation of the SNS is not equally targeted to a given body type or a specific organ system. Individuals with central obesity demonstrate augmented sympathetic outflow when compared to non-central adiposity body types (Alvarez et al. 2004). This has been observed via MSNA recordings of post-ganglionic sympathetic nerve fibres (Huggett et al. 2004), urinary noradrenaline (NA) excretion (Lee et al. 2001) and renal NA spillover (Rumantir et al. 1999) even when hypertension is not present, although the presence of hypertension in metabolic syndrome results in a further augmentation of sympathetic activity (Huggett et al. 2004; Grassi et al. 2005). Within this central adiposity phenotype, activation of the SNS is not homogeneous. Specifically, SNS outflow to the skeletal muscle vasculature and the renal system is elevated, whereas activation to the heart is reduced (Vaz et al. 1997). SNS activation to the skin (Grassi et al. 1998a) and splanchnic regions (Vaz et al. 1997) remain unchanged in both forms of obesity. Importantly, the increase in MSNA observed in obese subjects results from an increase in the recruitment of previously silent fibres, and not from an increase in the firing probability of already active sympathetic vasoconstrictor fibres (Lambert et al. 2007). Of note, adrenal adrenaline secretion does not appear to be increased in obese humans (Morgan et al. 1993; Brunner et al. 2002).

The underlying cause of the observed SNS overdrive in obesity is not fully understood, but is likely to be multifactorial and starts as part of a homeostatic mechanism aimed at expending the excess energy as heat rather than additional storage of fat. The accumulation of body fat resulting from a positive energy balance was first shown in animal models to result in SNS activation to a variety of peripheral tissues as a key component of the counter-regulatory response (Landsberg & Young, 1981; Landsberg, 1990; Landsberg & Young, 1993). SNS activation is also seen in humans following overfeeding (Arone et al. 1995; Dulloo & Jacquet, 2001), and is decreased with fasting (Young & Landsberg, 1977). The chronic increase in basal SNA is presumably aimed at stimulating β-adrenergic thermogenesis to prevent further fat storage (Dulloo, 2002), but can also stimulate lipolysis to increase non-esterified free fatty acids (NEFAs), contributing to insulin resistance. Adipose tissue itself can act as an endocrine organ and express various adipokines, which may directly or indirectly increase sympathetic activity. A chronically elevated SNS outflow could in turn impair β-adrenergic signalling, reduce stimulation of metabolism, and evolve into a cycle that contributes to obesity, insulin resistance, and morphological changes in target tissues.

Insulin, insulin resistance and sympathetic outflow

It has been well documented that insulin can augment sympathetic outflow in animals via intracerebroventricular administration (Muntzel et al. 1994; Rahmouni et al. 2004). Sympathetic outflow increases upon the injection of insulin into the third cerebral ventricle of rats (Muntzel et al. 1994), with recent data identifying the arcuate nucleus, via the paraventricular nucleus of the hypothalamus, as the specific site at which insulin acts to increase the sympathetic nervous system and increase sympathetic baroreflex gain (Cassaglia et al. 2011). While very little insulin is produced in the central nervous system, central insulin receptors are found on the hypothalamus (Hopkins & Williams, 1997), and can cause a co-activation of the SNS through transport-mediated uptake across the blood–brain barrier of peripherally secreted insulin. In addition, the arcuate nucleus is unusual in that it contains highly permeable capillaries (Ciofi, 2011), such that insulin may directly activate receptors in this area without a specific transport mechanism (Dampney, 2011). In humans, evidence demonstrates that insulin release in response to a mixed meal (Fagius & Berne, 1994; Cox et al. 1995; Young et al. 2010) or during a hyperinsulinaemic euglycaemic clamp (Anderson et al. 1991; Berne et al. 1992; Vollenweider et al. 1993; Hausberg et al. 1995; Van De Borne et al. 1999) increases MSNA and enhances the arterial baroreflex gain of sympathetic nerve activity (Young et al. 2010). These findings may be taken to suggest hyperinsulinaemia may contribute to the sympathetic overdrive in obesity.

There is a problem with the theory that excess insulin drives the sympathetic overactivity in obesity, however. Specifically, obese subjects do not appear to retain their sensitivity to the stimulatory effects of insulin on the SNS. In lean subjects, MSNA increased 94% in response to euglycaemic hyperinsulinaemia, much greater than the 9% increase observed in age-matched obese subjects (Vollenweider et al. 1994). Similarly, patients with insulin-resistant metabolic syndrome also have demonstrated blunted sympathetic responses to increased plasma insulin following a glucose load (Straznicky et al. 2009a, b), consistent with central insulin resistance in this population. So what factors are causing the SNS to remain elevated in obesity?

There is emerging evidence that certain adipokines expressed in central obesity may contribute to the state of sympathetic overdrive. In addition, certain adipokines when overexpressed may contribute to insulin resistance, ultimately resulting in a hyperinsulinaemic state in concert with greater sympathetic outflow. Insulin resistance is the eventual failure of glucose transporter-4 (GLUT-4) to translocate to the cell membrane to uptake blood glucose into peripheral tissues, predominately skeletal muscle (DeFronzo et al. 1981). It has been demonstrated that the failure is due to alterations in a cellular signalling mechanism, as insulin resistant individuals have normal GLUT-4 levels (Garvey et al. 1998; Shulman, 2000). When GLUT-4 fails to translocate or become activated and uptake blood glucose into peripheral tissues, blood glucose remains elevated. These abnormal blood glucose levels continue to trigger insulin release (Hedeskov, 1980), causing a hyperinsulinaemic state. Many adipokines expressed in both central and peripheral fat cells are implicated in creating this hyperinsulinaemic state through their deleterious effects on cellular insulin signalling, as displayed in Fig. 1.

Figure 1. The cellular effects of adipokines on insulin signalling.

Leptin has been shown to decrease phosphorylation of protein kinase C (PKC) which causes a failure of GLUT-4 translocation which then cannot bring glucose (glu) into the cell. C-reactive protein (CRP) can act in concert with interleukin-6 (IL-6) to decrease protein kinase B (PKB) phosphorylation, which will result in decreased GLUT-4 translocation. IL-6 can also inhibit the activation of insulin receptor substrate (IRS) by phosphatidylinositol 3-kinase (PI3K) and down-regulate the expression of the suppressor of cytokine signalling (SOCS) protein which normally serves a protective role in the effect of cytokines on cellular signalling. It is important to note that IL-6 only demonstrates deleterious effects on insulin signalling when chronically released. Free fatty acids (FFA) can impair one of the first steps in insulin signalling, tyrosine phosphorylation (TP) and impair activation of IRS by PI3K. Tumour necrosis factor-α (TNF-α) can impair insulin signalling by up-regulating serine/threonine phosphorylation, which has been associated with a decrease in GLUT-4 translocation, or by inhibiting phosphorylated tyrosine binding to IRS.

In addition to the effects a dysregulation of adipokines may have on insulin sensitivity resulting from impaired GLUT-4 translocation, chronically elevated SNS activity secondary to increased adipokine expression may reduce the efficiency by which GLUT-4 transports glucose into the cell. Sustained SNS activity on adipose α₁-adrenergic receptors can elevate intracellular calcium, which can then inhibit the action of GLUT-4 via activation of phospholipase-Cβ and impaired dephosphorylation of GLUT-4 (McCarty, 2004). When GLUT-4 is phosphorylated, it can still translocate to the cell membrane but its intrinsic activity is decreased, thus impairing the ability of the cell to uptake glucose (Reusch et al. 1991). Evidence for the positive relationship between intracellular calcium and insulin resistance has been demonstrated in rats (Jang et al. 2002) and in humans diagnosed with hyperparathyroidism (Kumar et al. 1994).

In obesity induced hyperinsulinaemia, the sympathetic nervous system may also be co-activated by the hypothalamic–pituitary–adrenal (HPA) axis. Co-activation has been observed by measuring MSNA (Anderson et al. 1991; Berne et al. 1992; Vollenweider et al. 1995), plasma NA (Rowe et al. 1981; O'Hare et al. 1989; Berne et al. 1992; Vollenweider et al. 1995), or NA spillover (Lembo et al. 1992), during direct intracerebroventricular injection of cortisol releasing hormone (Habib et al. 2000). When dexamethasone was used to acutely block cortisol releasing hormone stimulated by stress or hyperinsulinaemia, no increase in MSNA was observed (Stratakis & Chrousos, 1995), providing strong evidence the two brain centres can be co-activated in the presence of hyperinsulinaemia.

Another mechanism that could contribute to SNS hyperactivity is a polymorphism on the gly16 allele of the β₂-adrenergic receptor found on pancreatic β-cells. This particular polymorphism is associated with decreased insulin sensitivity, elevated blood pressure and NA levels, and a blunted response to β₂-adrenergic agonists (Masuo et al. 2005). Since these particular β₂-adrenergic receptors are implicated in augmenting the release of insulin, a decreased sensitivity would require additional sympathetic influence to achieve desired insulin levels. It is not clear whether diminished receptor response causes sympathetic overdrive, or if sympathetic overdrive, secondary to insulin resistance, causes decreased receptor sensitivity. Using either explanation, the end result may be that a decreased sensitivity of β₂-adrenergic receptors found on pancreatic β-cells is partially responsible for the sympathetic overdrive (Masuo et al. 2005).

In addition to the well documented neural SNS overactivity at rest, obese subjects demonstrate baroreflex impairments and blunted responses to sympathoexcitatory manoeuvres. Diminished cardiovagal baroreflex responses have been reported in normotensive obese humans in a number of studies (Grassi et al. 1995; Alvarez et al. 2002; Beske et al. 2002), with a further attenuation observed in obese hypertensive patients (Grassi et al. 2000). The sympathetic baroreflex has also been reported to be reduced with obesity (Grassi et al. 1995), although this finding is not ubiquitous (Alvarez et al. 2002). The reason for the discrepancy in findings may be related to the degree of obesity in the studies. In the study by Alvarez et al. (2002), the subjects’ BMI was less than 28 kg m⁻² in both the low abdominal visceral fat and high abdominal visceral fat groups versus 40 kg m⁻² in the study by Grassi et al. (1995).

The exact cause of changes in baroreflex function with obesity is not entirely clear. In lean humans, insulin is known to increase the sympathetic baroreflex in response to a hyperinsulinaemic–euglycaemic clamp, and in response to a mixed meal (Young et al. 2010). Thus, decreased insulin-mediated sympathetic baroreflex function in the presence of insulin resistance may be expected. However, weight gain and obesity are also associated with increased stiffness of the central arteries in which the baroreceptors lie (Orr et al. 2008), suggesting obesity-related changes in baroreceptor signalling may be a contributing factor to sympathetic overactivity and reduced baroreflex responsiveness. In further support of this notion, it has been demonstrated that obesity is associated with accelerated atherosclerosis in the carotid artery (Lakka et al. 2001), conceivably in part through the actions of adipokines such as leptin, which is independently associated with a thickening of the carotid intima-media (Ciccone et al. 2001). In children, obesity is also associated with a decreased arterial compliance, presumably as an effect of both endothelial dysfunction and structural changes to vessels, and these are possibly early markers of atherogenesis (Tounian et al. 2001). Even in older, overweight humans without insulin resistance or diagnosed metabolic syndrome, carotid intima media thickness was greater than their normal weight peers (Lind et al. 2011). Blood glucose values are also associated with carotid intima-media thickening, even when these blood glucose values lie on the high end of what is considered normal (Thomas et al. 2004). In the Linosa study, an investigation of the metabolic syndrome in an isolated Italian community, insulin resistance correlated negatively with cardiovagal baroreflex gain, and the degree of this baroreflex impairment was associated with increased carotid artery thickness (Lucini et al. 2006). Adult obese Zucker rats also have demonstrated baroreflex-mediated changes in SNS activity similar to humans (Grassi et al. 2000). However, a recent study demonstrated that these changes were due to a central mechanism and not to deficits in the responsiveness of aortic baroreceptors to changes in arterial pressure (Huber & Schreihofer, 2010). Lastly, oxidative stress is known to be increased in obesity (Keaney et al. 2003) and may contribute to cardiovagal dysfunction. Ascorbic acid infusion has been shown to improve cardiovagal baroreflex sensitivity in older adults (Monahan et al. 2004), indicating a link between excess reactive oxygen species and the cardiovagal baroreflex. As discussed in greater detail below, reactive oxygen species may also have a role in sympathetic overactivity and autonomic dysfunction in obesity.

Weight loss through caloric restriction can reduce arterial stiffness in overweight/obese humans (Dengo et al. 2010), and has been demonstrated to increase the sensitivity of the cardiovagal baroreflex (Alvarez et al. 2005) and the sympathetic baroreflex (Grassi et al. 1998b). In a recent study in obese subjects assigned to a hypocaloric diet resulting in weight loss of 9% (Straznicky et al. 2011), MSNA and whole body NA spillover rate were dramatically reduced during active weight loss. However, during weight maintenance following weight loss, MSNA rebounded while NA spillover was preserved, suggesting organ-specific differentiation in SNS adaptation under conditions of negative versus stable energy balance. These intriguing findings demonstrate our knowledge on SNS in obesity, including what factors drive the overactivity and the role of the baroreflexes, is far from complete.

Independent and interactive effects of adipokines on sympathetic regulation

The following are several key cytokines and enzymes released from adipocytes which may alter SNS activity and impair insulin signalling. Among the most prominent in contributing to these observed symptoms in obesity and metabolic syndrome include leptin, non-esterified free fatty acids (NEFAs), tumour necrosis factor-α (TNF-α), interleukin-6 (IL-6), C-reactive protein (CRP), angiotensinogen and adiponectin (Fig. 2). A complicated relationship exists between these adipokines that we are just beginning to understand, as many of them work in concert with others to impair insulin signalling, create a state of inflammation, and/or interact to alter sympathetic regulation (Fig. 3).

Figure 2. Specific pathways by which adipokines augment sympathetic outflow in obesity.

Leptin is preferentially expressed in peripheral fat cells and both free and protein-bound leptin can impair glucose transporter-4 (GLUT-4) which creates a hyperinsulinaemic state and results in a co-activation of the sympathetic nervous system (SNS). Leptin can also act through a central mechanism. Free fatty acids (FFA) are equally expressed in central and peripheral obesity and in addition to causing a decrease in GLUT-4 translocation, can increase blood sugar and ultimately insulin levels through up-regulating gluconeogenesis in the liver. There is also evidence FFA can stimulate the SNS by a central mechanism. TNF-α is preferentially expressed in central obesity and through interacting with several key steps of insulin transduction, can cause a decrease in GLUT-4 translocation and ultimately increase blood glucose levels. Interleukin-6 (IL-6) is preferentially expressed in central obesity and stimulates neural SNS hyperactivity through effects on insulin signalling, direct HPA activation and in positive feedback fashion with β₂- and β₃-adrenoceptors on adipocytes. C-reactive protein (CRP) acts by impairing insulin signalling causing the resultant hyperinsulinaemia and co-activation of the neural SNS. Angiotensinogen expression is up-regulated in obesity which, in concert with the observed increase in renal SNS and resultant renin release, may act through mass action to cause a vasoconstriction and activation of the SNS by angiotensin II (ANG II) formation.

Figure 3. Co-expression and mechanisms of adipokines in obesity.

In obesity, key adipokines which can result in sympathetic overdrive are dysregulated (angiotensinogen, tumour necrosis factor-α, adiponectin, leptin, free fatty acids, interleukin-6 and C reactive protein) and can act as endocrine and paracrine hormones to up-regulate expression of other adipokines. In the case of TNF-α they can sensitize angiotensin II receptors which, when activated, can cause sympathetic outflow through a central mechanism. These adipokines then work to increase sympathetic outflow either through their effects on insulin sensitivity, or through central cardiovascular control centres directly.

Leptin

Leptin is a 16 kDa protein produced principally in adipocytes in proportion to the amount of fat mass, although it has also been shown to be released from the brain (Esler et al. 1998). Production of leptin from adipocytes acts in a negative feedback loop to suppress appetite and prevent weight gain, as postulated in the lipostatic theory of body-weight set point (Friedman, 2000). Leptin infusion has been observed to increase sympathetic outflow in a dose-dependent manner in Sprague–Dawley rats (Haynes et al. 1997a,b), similar to the pattern of sympathetic activation in human obesity. Some notable exceptions in humans are that adrenaline secretion rates are normal and sympathetic activity to the heart is not increased (Vaz et al. 1997; Esler et al. 2006), whereas these changes were observed with leptin in rats.

Leptin concentration may act centrally to impact SNS overactivity. Two key sites associated with autonomic regulation have been identified that have leptin receptors. Leptin receptors in the hypothalamic arcuate nucleus (Grill, 2006) appear to be responsible for regulating glucose homeostasis and body weight (Coppari et al. 2005), and may also act to signal SNS activity. There are also leptin receptors on the nucleus tractus solitarii (NTS; Mark et al. 2009), a site where leptin can cause direct SNS activation. Evidence for this exists in the fact that leptin is found in two forms: protein-bound leptin and unbound leptin. Protein-bound leptin can cross the blood–brain barrier (Brabant et al. 2000). Tank et al. (2003) reported that unbound leptin is not associated with MSNA, as opposed to protein bound leptin. Chronic leptin stimulation may be expected to result in brain leptin resistance. However, it has been suggested that brain leptin resistance may be selective, such that the effect on appetite is inhibited, whereas that on sympathetic activity is preserved (Mark et al. 1999). More recently, leptin has been shown in an animal model to act through cardiovascular control centres in the brain to specifically target sympathetic outflow to the kidney (Mark et al. 2009). Further, microinjection of leptin into the NTS has been shown to impair baroreflex control of heart rate (Arnold et al. 2009).

It has been demonstrated that leptin levels are more closely linked in humans to peripheral than central fat stores (Russell et al. 1998; Van Harmelen et al. 1998). Thus, the association of leptin to SNS activity is more controversial than in animal models. Some studies show a positive but not strong relationship between leptin and measures of whole body and regional sympathetic activity (Snitker et al. 1997; Monroe et al. 2000), while others do not (Narkiewicz et al. 2001; Alvarez et al. 2002). Alvarez et al. (2002) studied non-obese men and men with subcutaneous obesity but did not find a correlation between leptin and MSNA, despite the group with subcutaneous obesity having a 2.6-fold higher plasma leptin concentration. This underscores the concept that leptin may only contribute to sympathetic outflow in visceral obesity. Additionally, there may be an interplay between leptin and other adipokines associated with central obesity that ultimately drive the SNS overactivity. Leptin itself exerts a paracrine effect on fat cells, and its expression and secretion by fat cells can be induced by IL-6 and inhibited by TNF-α (Lau et al. 2005).

Non-esterified fatty acids

In humans, circulating levels of NEFAs are increased in obesity and augmented levels are inversely correlated with insulin sensitivity (Bruce et al. 1994; Perseghin et al. 1997). NEFAs act on all aspects of glucose homeostasis, from uptake at peripheral tissue to hepatic production and disposal. Elevated NEFA levels inhibit glucose uptake into peripheral tissues by impairing phosphatidylinositol kinase (PI3-kinase) activation and IRS-1 phosphorylation following insulin stimulation, which ultimately may impair insulin-mediated GLUT-4 translocation (Kruszynska et al. 2002). NEFA disrupting insulin signalling may also be due to impaired PI3-kinase activation due to excess protein kinase C (Griffin et al. 1999). Further supporting this notion is the evidence that PI3-kinase activation is lost as a result of a high fat diet in mice (Zierath et al. 1997).

As NEFA can cross the blood–brain barrier (Lam et al. 2005), it has been found to cause a central activation of MSNA in lean subjects (Florian & Pawelczyk, 2009) and reduce cardiovagal baroreflex in lean and obese subjects (Gadegbeku, 2002). NEFA also stimulates plasminogen activator inhibitor-1, which may also contribute to the association between increased plasma NEFA in obesity and the augmentation of MSNA (Kohler & Grant, 2000). However, this concept has been the subject of conflicting results, with another study demonstrating no significant effect of NEFA on MSNA or sympathetic baroreflex sensitivity (Monahan et al. 2007). Furthermore, whole body and renal noradrenaline spillover did not change during NEFA infusion (Grekin, 2005).

Tumour necrosis factor-α

TNF-α is a protein that acts as a cytokine with a primary role in stimulating the acute phase of inflammation. It also has roles in inducing apoptosis, and inhibiting viral replication and tumorigenesis. TNF-α is capable of crossing the blood–brain barrier (Banks et al. 1995) and may exert its effects in this capacity as well as peripherally and integratively with other adipokines. In obesity, it has been shown in both the animal (Hofmann et al. 1994; Hotamisligil et al. 1994) and human model (Hotamisligil et al. 1995; Kern et al. 1995) that TNF-α is over-expressed when compared with lean subjects. TNF-α has a positive relationship with waist-to-hip ratio but not BMI, suggesting a preferential expression in central adiposity (Tsigos et al. 1999). TNF-α will also stimulate lipolysis (Feingold et al. 1992), which will increase NEFA levels. These augmented NEFA levels and resultant impairment of insulin signalling can cause a hyperinsulinaemic state and possibly a co-activation of the SNS as previously discussed. TNF-α is also responsible for signalling increased IL-6 production (Berne et al. 1992), further underscoring the interplay between adipokines in SNS hyperactivity associated with obesity.

Interleukin-6

IL-6 is a protein that acts in the acute-phase reaction of the inflammation model and is considered to have pro-inflammatory and anti-inflammatory effects. IL-6 is produced from smooth muscle and helper T cells and as an adipokine. It is the interleukin associated with direct action on the hypothalamus by crossing the blood–brain barrier (Banks et al. 1995) to induce fever, and may cause a co-activation of the SNS through its effects on the HPA axis. IL-6 also releases acute phase proteins into the blood, including CRP. Paradoxically, IL-6 can also be released from muscular contraction, as appearance of IL-6 has been observed following exercise (Febbraio & Pedersen, 2005) and in this context can augment insulin sensitivity (Petersen & Pedersen, 2005).

In obesity, up to 30% of circulating IL-6 is secreted from adipocytes (Mohamed-Ali et al. 2001) and is preferentially secreted from visceral adipocytes at a rate 2–3 times that of subcutaneous adipocytes (Lang & Dobrescu, 1989). Rather than an over-expression of IL-6 from adipocytes, it has been demonstrated that IL-6 production is linked to the increasing number of adipocytes in obesity (Bastard et al. 1999). In the human and rat model, stimulation of β₂- and β₃-adrenoceptors on adipocytes results in increased production of IL-6 (Mohamed-Ali et al. 1997), which could conceivably act in a positive feedback loop with the SNS, potentially contributing to the observed elevations in obesity.

An increased production of IL-6 may drive SNS activity by acting centrally. In IL-6 knockout mice, there is a blunted heart rate and energy expenditure response to stress and exposure to cold, respectively (Wernstedt et al. 2006). With central administration of IL-6, an enhancement of uncoupling protein activity and an increase in oxygen consumption is seen, suggesting a direct link to SNS activation (Wallenius et al. 2002). Evidence for IL-6 activating the HPA axis is provided by the appearance of adrenocorticotrophic releasing hormone (Mastorakos et al. 1994) and corticosterone (Lenczowski et al. 1999) following IL-6 administration. This co-activation is normally considered a part of the inflammation process (Chrousos, 1995), but may partially explain the effects of increased IL-6 production in obesity and the correlated increase in sympathetic outflow. In this case, the SNS activation may be an indication of the chronic inflammatory state observed in central obesity.

IL-6 can also affect insulin sensitivity. Plasma IL-6 levels are significantly and inversely proportional to insulin sensitivity (Kern et al. 2001). Mechanistically this may be attributed to the hypothesis that cytokines such as IL-6 can induce cellular protein expression, through a specific family of proteins called ‘suppressors of cytokine signalling’ (SOCSs). SOCSs normally act to suppress the effects of cytokines on insulin transduction steps such as IRS-1 phosphorylation, PI3 activation or protein kinase B activation (Senn et al. 2003). Inhibition of any one of these steps will cause a failure of GLUT-4 translocation and result in a hyperinsulinaemic state (see Fig. 1). These SOCS proteins are normally expressed in a pulsatile fashion. However, in obesity and the resultant chronic elevation of IL-6, SOCS proteins are inhibited.

Conflicting with the role of IL-6 in down-regulating insulin transduction is the paradoxical finding that increasing IL-6 levels can also be correlated to the postexercise period and observed heightened insulin sensitivity. Presumably, IL-6 is being released due to muscle contraction as a myokine. It has been demonstrated that in this capacity, IL-6 can enhance the action of insulin (Petersen et al. 2005), promote glucose uptake into cells (Carey et al. 2008) and drive fatty acid oxidation in skeletal muscle (Kelly et al. 2004). This action is proposed to occur by IL-6 activating AMP kinase to promote glucose uptake and fatty acid metabolism, but is not fully understood at this time. The difference in observations may be explained in that the sustained IL-6 release in obesity and its interaction with other adipokines results in SNS activation, whereas with exercise there is a transient release of IL-6 which is expressed in the absence of other adipokines. IL-6 acts with TNF-α to increase the synthesis of C-reactive protein (CRP) (Yudkin et al. 1999), and it may be that CRP, under the control of IL-6 expressed from adipocytes, contributes to insulin resistance and SNS overactivity indirectly.

C-reactive protein

CRP is synthesized mainly in the liver. It is rapidly synthesized in response to trauma, infection and inflammation, and reduced quickly when these symptoms are ameliorated. In obesity, CRP has been shown to be chronically elevated and the two are correlated in a dose-dependent fashion, with no preferential expression from visceral or subcutaneous adipocytes (Ridker et al. 2003). Although CRP has not been reported to directly stimulate sympathetic outflow, the use of β-blockers lowers CRP, and there is a strong correlation between CRP and IL-6 levels in developing insulin resistance (Festa et al. 2002; Freeman et al. 2002). Leptin (Chen et al. 2006), IL-6 and TNF-α (Yudkin et al. 1999) can induce CRP expression, again highlighting the interplay between adipokines in obesity. CRP is also responsible for increased expression of the angiotensin I receptor in vascular smooth muscle, indicating that CRP may indirectly contribute to SNS outflow through the up-regulation of the angiotensin II pathway (Wang et al. 2003), as discussed below.

Angiotensinogen

Both the SNS and the renin–angiotensin–aldosterone system (RAS) are activated in obesity, and both systems can upregulate the action of the other. Sympathetic nerves stimulate renin release and the formation of angiotensin II, and angiotensin II can alter SNS outflow directly (Matsukawa et al. 1991), or can augment SNS responsiveness by potentiating release of noradrenaline from sympathetic nerve terminals and amplification of α₁-receptor mediated vasoconstriction. Part of this bidirectional augmentation may be through the actions of angiotensinogen, the production of which is elevated in obesity (Umemura et al. 1997; Van Harmelen et al. 2000), more so in central adiposity than peripheral. An increase in angiotensinogen could allow for greater synthesis of angiotensin II by a mass action shift with increased availability of angiotensinogen (Umemura et al. 1997) and augmented renin release secondary to higher renal sympathetic nerve traffic.

Increased MSNA might also be explained by the over-expression of TNF-α in central obesity, which sensitizes the effects of angiotensin II (Zera et al. 2008). Obese women have been shown to express angiotensin II receptors in adipocytes at twice the rate of normal weight women, further implicating the RAS system in the transduction of sympathetic outflow (Gorzelniak et al. 2002). Angiotensin II has also been shown to induce CRP expression in human aortic endothelial cells (Han et al. 2010). Taken together, these findings indicate the RAS is not only implicated in the observed sympathetic overdrive in obesity, but may provide a mechanism through which sympathetic overactivity leads to chronic hypertension. Lending support to this concept, when angiotensin II was inhibited for three months in obese, hypertensive humans, MSNA activity was reduced by 21% (Grassi et al. 2003). However, this was attributed to the concomitant improvement in insulin sensitivity rather than a direct effect on SNS outflow. In lean insulin-sensitive hypertensives, angiotensin II receptor blocker did not alter SNS activity (Struck et al. 2002).

Adiponectin

Adiponectin is a protein highly expressed in adipocytes (Maeda et al. 1996). It has been shown that adiponectin improves insulin sensitivity and glucose tolerance (Yamauchi et al. 2002), and is reduced in obese humans (Arita et al. 1999). Levels of adiponectin are inversely proportionate to circulating leptin concentration and insulin sensitivity (Matsubara et al. 2002). Furthermore, adiponectin levels are sex specific, favouring production in women (Nishizawa et al. 2002). Adiponectin is also attenuated in lean humans with type 2 diabetes (Hotta et al. 2000), suggesting a closer association with metabolic dysfunction rather than as an adipokine.

Administering adiponectin to overweight mice consuming a high fat, high sugar diet was associated with a decrease in NEFAs and plasma glucose (Fruebis et al. 2001), suggestive of an improvement in insulin transduction. There does not seem to be any direct, central mechanism by which adiponectin alters SNA, and thus any relationship seems to be in response to peripheral changes. Specifically, adiponectin acts to augment hepatic but not muscle insulin signalling.

Reactive oxygen species

Reactive oxygen species (ROS) are known to have important roles in cell signalling but can participate in the dysregulation of autonomic balance and control of blood pressure in humans. ROS production is normally countered by enzymatic and non-enzymatic reducing agents to prevent the associated negative effects, but in a disease state excessive ROS may be generated. Germane to this review, ROS are capable of modifying central activation of sympathetic nervous activity (Campese et al. 2004; Han et al. 2005), and may serve as an important link between adipokine expression and SNS overactivity. As an example, autonomic dysfunction in obesity may be related to angiotensin II as a source of ROS, as links have now been established between angiotensin II and superoxide signalling in the CNS relating to the regulation of SNS activity (Zimmerman et al. 2002; Campese et al. 2005; Gao et al. 2005; Han et al. 2005). In support of this concept, direct injections of superoxide dismutase, an enzyme capable of reducing ROS, into the rostral ventrolateral medulla of a porcine model reduces sympathetic nerve activity (Zanzinger & Czachurski, 2000). Likewise, direct injections of a superoxide dismutase mimetic into the medulla of rats produced the same quieting of the sympathetic nervous system, and these effects were partially independent of nitric oxide production in these vascular beds (Campese et al. 2004).

ROS production is noted to be elevated in central adipose tissue (Cancello et al. 2006), possibly resulting from increased infiltration of macrophages. Macrophage-induced oxidative stress may be influenced by dysregulation of other adipokines as well, creating a positive feedback loop with SNS activity. For example, adiponectin has been noted to have an inhibitory effect on cell adhesion molecules and down-regulation could conceivably promote monocyte infiltration (Kawanami et al. 2004). The production of ROS resulting from dysregulation of adipokines is an interesting possible link between chronic inflammation and SNS overactivity in obesity. Future studies should be directed at investigating these potentially important relationships.

Summary

There is growing evidence that suggests links between an unfavourable expression of adipokines to insulin resistance and an increase in neural SNS activation. In addition to sympathetic overdrive in central obesity, there is an observed blunted response of the sympathetic nervous system to excitatory manoeuvres. A better understanding of the relationships between adipokines, insulin resistance, and sympathetic activity at rest and in response to stressors is necessary if we are to develop targeted strategies to improve health and decrease risks associated with obesity.

Perspectives

Just as clinical diagnoses of visceral obesity, hypertension, dyslipidaemia and glucose intolerance have evolved to be recognized as functionally related in the metabolic syndrome, recognition that sympathetic overdrive may be a common thread to these pathologies is evolving. Designing strategies to treat the neural sympathetic overdrive may result in improved outcomes in certain patient groups. As an example, Schlaich et al. (2011) recently performed bilateral sympathetic renal denervation by endovascular radiofrequency in obese women with polycystic ovary syndrome, a disease characterized by SNS activation, insulin resistance, and blood pressure elevation. In these patients, baseline SNA was elevated ∼3-fold from normal values. Renal nerve ablation resulted in a reduction in overall SNA (measured by whole body NA spillover and MSNA), a moderate reduction in blood pressure and significant improvement in insulin sensitivity. This suggests selectively targeting SNA overdrive may have profound improvements on insulin resistance and possibly comorbidities associated with the metabolic syndrome. However, there remains much to learn about the impact of chronic SNA overactivity in humans, including some basic information on sex differences, the influences of the sex steroids, the role of inflammation, and how the well-known age-related increase in SNA and body fat might contribute to morbidity. An integrative approach to our thinking is necessary to understand how sympathetic overdrive may be a key underlying component to the constellation of diseases resulting from the metabolic syndrome, and may even have a causal role in insulin resistance. Integrative physiological research should continue to test hypotheses and interventions which establish novel therapies in circumventing or ameliorating the cyclical relationship of neural sympathetic overdrive, insulin resistance and visceral obesity. Improving our understanding of the integrative effects of certain adipokines and how they may alter the regulation of sympathetic tone and contribute to insulin resistance is paramount in this regard.