The high prevalence of obesity, estimated to exceed 25% in many developed countries including Australia (Australian Bureau of Statistics, 2015), is a serious public-health concern that confers excess mortality and morbidity due to its associations with chronic conditions such as type 2 diabetes (T2D) and cardiovascular disease (Prospective Studies Collaboration, 2009). It is therefore pertinent to elucidate potential mechanisms that link excess adiposity to cardiometabolic risk and to identify groups most likely to benefit from targeted lifestyle interventions.

The sympathetic nervous system (SNS) plays a pivotal role in both cardiovascular and metabolic regulation and perturbations in SNS activity are widely believed to contribute to the pathophysiology of the obese state. An early viewpoint, based on experimental hypothalamic models of obesity, clinical studies of venous and urinary norepinephrine concentration, and studies in Pima Indians (an ethnic population with low SNS activity) was that SNS underactivity was the basis of weight gain and the development obesity (Bray, 1990; Tataranni et al., 1997; Snitker et al., 2000). This was congruent with the known importance of the SNS in all aspects of daily energy expenditure. However, more recent studies which have applied more specific techniques such as direct microneurographic recordings of efferent sympathetic nerve traffic and isotope dilution, indicate that obesity is a state of chronic sympathetic overactivity and that there is organ specific differentiation in sympathetic outflow favoring skeletal muscle vasculature and the kidneys (Vaz et al., 1997; Grassi et al., 2004, 2005). Elevated SNS activity and impaired baroreflex sensitivity (BRS) is evidenced in obese, metabolic syndrome, and T2D cohorts when compared to lean controls (Lee et al., 2001; Huggett et al., 2003; Grassi et al., 2005) and the greater contribution of visceral, as opposed to subcutaneous fat accumulation, to sympathetic neural drive has been affirmed (Alvarez et al., 2002; Grassi et al., 2004).

Prospective epidemiological and offspring studies provide information on the chronology of pathogenic changes in the general and at risk populations. Various indices of SNS activity (baseline venous norepinephrine concentration, the norepinephrine response to cold pressor test, and resting heart rate) have been positively associated with future weight gain, insulin resistance and hyperglycemia in epidemiological studies with 5–18 years follow-up (Masuo et al., 2003; Flaa et al., 2008; Wulsin et al., 2015). Although seemingly counterintuitive, elevated SNS activity could predispose to weight gain via β-adrenoceptor desensitization and the promotion of central insulin resistance (Vollenweider et al., 1994; Julius et al., 2000). Notably, insulin-sensitive off-spring of T2D patients have elevated single-unit and multi-unit muscle sympathetic nerve activity (MSNA) compared with age- and body mass index (BMI)-matched controls with no family history of T2D (Huggett et al., 2006). This finding highlights that sympathetic overactivity is an early pathogenic phenomenon in subjects genetically predisposed to the development of metabolic disease. In established obesity several factors act in concert to maintain elevated sympathetic drive including hyperinsulinemia, obstructive sleep apnoea, baroreflex impairment, and activation of the renin-angiotensin system (Lambert et al., 2010). Sustained sympathetic activation is mechanistically involved in obesity-related target organ damage such as left ventricular diastolic dysfunction, left ventricular hypertrophy, and arterial remodeling via both blood pressure-mediated and blood pressure-independent effects (Schlaich et al., 2009). Vis-a-vis, plasma norepinephrine is an independent predictor of adverse cerebral and cardiovascular events (Yufu et al., 2014).

Weight loss through energy-restricted diet and/or physical activity represents first-line treatment for obesity and has proven sympathoinhibitory benefits (Straznicky et al., 2010). However, relatively little attention has focused on whether baseline metabolic status modifies the magnitude of change in SNS activity following weight reduction. Our group has previously demonstrated that hyperinsulinemic subjects derive significantly greater sympathoinhibition than normoinsulinemic subjects at matched weight loss (Straznicky et al., 2014). This concurs with the known central sympathoexcitatory effects of insulin (Anderson et al., 1991; Rahmouni et al., 2004). Other studies have shown greater improvement in glycemic and cardiac autonomic indices in T2D subjects compared to normal glucose tolerant (NGT) subjects after lifestyle induced and surgical weight loss (Solomon et al., 2010; Casellini et al., 2016). Hitherto, the relative benefits of weight loss in obese subjects with different stages of glycemic control on robust assessments of SNS activity are unknown. Moreover, there are no data regarding the effects of weight loss on MSNA or norepinephrine kinetics in T2D populations. The present study was conducted to compare the impact of 4-month hypocaloric diet (HCD) on direct recordings of efferent postganglionic MSNA and whole-body norepinephrine kinetics in obese subjects categorized as NGT, impaired glucose tolerant (IGT) and newly diagnosed T2D. We hypothesized that T2D subjects would derive the greatest sympathoinhibition at equivalent weight loss due to greater attenuation of glycemia and concomitant improvement in baroreflex function (Gerritsen et al., 2000).

In this study we sought to compare the benefits of equivalent weight loss, achieved through HCD, on SNS activity in un-medicated, matched obese subjects who were stratified according to their glucose tolerance, and who are representative of distinct metabolic sub-groups along the diabetes continuum. In accordance with previously published data, we found that T2D subjects experienced greater reductions in fasting and post-OGTT glucose levels and an increase in insulinogenic index compared with their non-diabetic counterparts (Solomon et al., 2010). However, the greater improvement in glycemic indices did not translate to any additive effect on spontaneous cardiac BRS nor on sympathoinhibition. Indeed, weight loss was associated with significant and similar attenuation in both neurophysiological and biochemical measures of sympathetic activity across the groups. To our knowledge, ours is the firsts study to demonstrate that moderate weight loss attenuates MSNA and whole-body norepinephrine spillover rate in treatment naïve T2D subjects. Pooled regression analysis indicated that reductions in hyperinsulinemia, gauged as the insulin response during OGTT, leptin and PRA, and enhancement of BRS were independently associated with sympathoinhibition within the cohort. These findings are elaborated below.

Only one other study has compared weight loss associated improvements in autonomic function in different obese subgroups. Casellini et al. (2016) measured time and frequency domain analysis of heart rate variability before and after surgical weight loss in non-diabetic, pre-diabetic, and T2D subjects. They reported that changes in cardiac autonomic function, representing both sympathetic and parasympathetic components, were significantly greater in T2D subjects. However, this study is not comparable to ours because of differences in participants' clinical status (their T2D subjects had long-standing drug-treated diabetes and non-diabetic subjects were normoinsulinemic); the mode and magnitude of weight loss (four times greater than that attained in our study); and the indirect quantification of cardiac autonomic tone. In contrast, our NGT group was hyperinsulinemic and we found no differentiation in sympathetic neural activity or spontaneous cardiac BRS at baseline, nor in the response to HCD, between our three groups. Previous cross-sectional studies have shown higher MSNA and lower BRS in T2D subjects vs. matched normoinsulinemic NGT subjects (Gerritsen et al., 2000; Huggett et al., 2003). This discrepancy is likely explained by the hyperinsulinemic status of our NGT group and our findings need to be interpreted in this context. Furthermore, categorization as NGT in our study was based only on 2-h glucose concentration, thus some NGT subjects may have had impaired fasting glucose, which could have influenced the results by minimizing between-group differences. Albeit, a recent study reported no differences in BRS between subjects categorized as having isolated impaired fasting glucose vs. those with NGT (Wu et al., 2014).

There is now ample clinical and experimental evidence that links hyperinsulinemia to augmented central sympathetic outflow (Anderson et al., 1991; Rahmouni et al., 2004). Distinct central signaling pathways have been identified that mediate insulin-induced regional sympathoactivation to skeletal muscle, adipose tissue and the kidneys (Rahmouni et al., 2003, 2004). In the present study, reduction in hyperinsulinemia was independently associated with attenuation in arterial norepinephrine concentration, which concurs with our previous findings in other cohorts (Straznicky et al., 2014). This observation also concurs with the demonstration that individuals with higher baseline plasma insulin derive the greatest improvement in macrovascular and microvascular endothelial function following weight loss (Bigornia et al., 2013). Furthermore, post-hoc analyses in the Swedish Obese Study indicated that higher baseline insulin concentration was associated with a more favorable outcome of bariatric surgery on cardiovascular events after a median follow-up of 15 years (Sjöström et al., 2012). Aside from its direct effects on central sympathetic outflow, hyperinsulinemia secondary to insulin resistance may also modulate BRS by both peripheral and central mechanisms. Peripherally, insulin resistance and attendant pathophysiologies (glucotoxicity, lipotoxicity, and inflammation) promote endothelial dysfunction, arterial stiffening, and impaired baroreceptor afferent responses to arterial wall stretch (Watkins et al., 2000; Grassi et al., 2005; Kim et al., 2006). In the brain, insulin increases the gain of baroreflex control of heart rate and MSNA (Pricher et al., 2008; Young et al., 2010). Therefore, central insulin resistance which is a characteristic of the metabolic syndrome, may reverse the normal effects of brain insulin to support optimal baroreflex control (Kaiyala et al., 2000). Consistent with these mechanisms, hyperinsulinemic subjects had lower baseline BRS compared with normoinsulinemic subjects in our study. It is of note that the observed dichotomy of changes in insulinogenic index between T2D and NGT subjects after HCD has been reported before. Solomon et al. (2010) demonstrated improved insulin secretion in T2D subjects following moderate dietary weight loss vs. reversal of compensatory hyperinsulinemia in nondiabetic individuals. Relief of extreme hyperglycemia (glucotoxicity) and enhanced secretion of the incretin hormone, glucose-dependent insulinotropic polypeptide, explained the improvement in pancreatic β-cell function in T2D subjects in their study.

Other factors that predicted sympathoinhibition in our study included reductions in plasma leptin, LDL-cholesterol, and activity of the renin-angiotensin system. In addition to its classic effects on appetite, the adipokine leptin is known to increase regional sympathetic nerve activity to the kidneys, adrenal gland, hindlimb, and adipose tissue in laboratory animals (Mark, 2013). The evidence for leptin-induced sympathoexcitation in humans is less strong, and is limited to correlative data, including our own showing an association with renal norepinephrine spillover rate in healthy men of differing adiposity (Eikelis et al., 2003) and the fact that leptin deficiency states are associated with sympathetic hypofunction (Ozata et al., 1999). Clinical leptin administration studies have yielded equivocal results (Mackintosh and Hirsch, 2001; Machleidt et al., 2013). A bidirectional interaction between the SNS and the renin-angiotensin-aldosterone system (RAAS) has been long appreciated (Grassi, 2001). RAAS activation in obesity has been attributed to increased synthesis of angiotensinogen by visceral fat, renal compression, and increased sympathetic outflows to the kidneys stimulating renin release. Elevated plasma LDL-cholesterol concentration is an established trigger for impairment of endothelial function and we have previously demonstrated both elevated MSNA and diminished endothelial function in young adults with hypercholesterolaemia compared to normal controls (Lambert et al., 2013). In the present study LDL-cholesterol decreased significantly in all three groups, likely in response to the reduced saturated fat content of the background diet which averaged −10 g per day, and changes in LDL independently related to reduction in MSNA.

Our study has a number of limitations which need to be noted. Firstly, it did not include a control arm, however we have previously demonstrated the reproducibility of MSNA and norepinephrine kinetics measurements in weight stable subjects over a 3-month period (Straznicky et al., 2010). Moreover, any effects of familiarization on sympathetic neural measurements would have been similar across the groups. Secondly, we did not assess body composition using a robust technique such as dual energy x-ray absorptiometry, so the impact of changes in fat depots on SNS parameters could not be determined. Thirdly, the study does not distinguish the effects of weight loss vs. negative energy balance from dietary restriction. Fourthly, subject numbers were relatively small and we may not have had adequate power to detect small differences between groups. Finally, we did not have any robust measures of fitness in order to ascertain whether changes in exercise habit contributed to sympathoinhibition. Exercise diaries indicated that on average our cohort walked 30 min per day at baseline, and that the NGT and IGT groups increased their average walking times by 12–14 min per day, whereas there were no increments in the T2D group during the intervention. Aerobic exercise has independent beneficial effects on SNS activity, endothelial function, baroreflex modulation, and insulin sensitivity, which may have contributed to our findings (Jennings et al., 1986). Results of the CALERIE study indicate that at identical energy deficit incorporation of exercise training with caloric restriction augmented improvements in cardiac sympathovagal balance compared to calorie restriction alone (deJonge et al., 2010).

In summary, our study demonstrates that moderate weight loss following HCD significantly attenuates SNS activity in obese subjects with diverse glycemic status. Concurrent reductions in hyperinsulinemia, leptin, LDL-cholesterol, and RAS were associated with the sympathoinhibition attained. These findings are of clinical importance in the context that elevated SNS activity contributes to the pathogenesis of hypertension, target organ damage and adverse cardiovascular and metabolic outcomes in both non-obese and obese populations (Masuo et al., 2003; Flaa et al., 2008; Schlaich et al., 2009). Additional studies are needed to address (1) the benefits of weight loss on autonomic function in hypoinsulinemic individuals with T2D; (2) the length of time that reduced sympathetic neural function is maintained after weight loss; and (3) the relationship between glycemic variability assessed by 24-h continuous monitoring and SNS activity.

NS, EL, JD, and GL conceived the study and contributed to the interpretation of the data. SP and NE performed laboratory analyses. NS, MG, CS, EL, DK, DH, and JM collected clinical data. NS performed statistical analysis and wrote the manuscript. All authors read and had final approval of the manuscript.

This study was funded by a Diabetes Australia Millennium Grant in type 2 diabetes to NS. JD and GL are supported by NHMRC Fellowships. We also wish to acknowledge the Victorian Government's Operational Infrastructure Support Program.