The underlying mechanism by which amassing of white adipose tissue in obesity regulates sympathetic nerve system (SNS) drive to the tissues responsible for glucose disposal, and causes insulin resistance (IR), remains unknown. We tested the hypothesis that high-fat (HF) feeding increases afferent impulses from white adipose tissue that reflexively elevate efferent nerve activity to skeletal muscle (SM) and adipose tissue to impair their local glucose uptake. We also investigated how salt-intake can enhance IR. HF-fed rats received a normal salt (0.4%) or high salt (4%) diet for 3 weeks. High-salt intake in HF fed rats decreased insulin-stimulated 2-deoxyglucose uptake by over 30% in white adipose tissue and SM, exacerbated inflammation, and impaired their insulin signaling and glucose transporter 4 (Glut4) trafficking. Dietary salt in HF fed rats also increased the activity of the adipose-cerebral-muscle renin–angiotensin system (RAS) axes, SNS, and reactive oxygen species (ROS). Insulin sensitivity was reduced by 32% in HF rats during high-salt intake, but was improved by over 62% by interruption of central RAS and SNS drive, and by over 45% by denervation or deafferentation of epididymal fat (all P<0.05). Our study suggest that a HF diet engages a sympathetic reflex from the white adipose tissue that activates adipose-cerebral-muscle RAS/ROS axes and coordinates a reduction in peripheral glucose uptake. These are all enhanced by salt-loading. These findings provide new insight into the role of a reflex initiated in adipose tissue in the regulation of glucose homeostasis during HF feeding that could lead to new therapeutic approaches to IR.
Obesity is characterized by excessive accumulation of white adipose tissue due to an imbalance between food intake and energy expenditure . People with obesity often have insulin resistance (IR) . Obesity increases the activity of the sympathetic nerve system (SNS) [3,4], especially in those with truncal obesity [3,5,6]. However, it is unclear how amassing visceral fat activates the SNS and how this contributes to IR.
Obese subjects with IR often have salt-sensitive hypertension . Moreover, high salt intake itself can exacerbate IR in obese subjects , while restriction of dietary salt can improve both hypertension and IR [7,8]. However, how salt per se relates to IR in obesity remains elusive.
Our recent studies have disclosed a rich neural network in chronic kidney disease (CKD) exacerbated by dietary salt intake that is initiated in the kidneys, and activates a renal–cerebral renin–angiotensin system (RAS) axis. One consequence of this was the development of IR [9,10]. The present study uses a model of diet-induced obesity to test the hypothesis that high fat (HF) feeding increases afferent impulses from white adipose tissue that reflexively elevates efferent nerve activity to skeletal muscle (SM) and adipose tissue to generate local reactive oxygen species (ROS) and angiotensin (Ang) II that impair their local glucose uptake and thereby promotes IR, and furthermore, that this process is enhanced by a high dietary salt intake.
Materials and methods
Four-week-old male Sprague–Dawley rats (Nanfang Hospital Animal Experiment Center, China) were housed in plastic cages (3–5 per cage) at 24 ± 1°C, 12/12-h controlled light conditions with ad libitum access to water and food, unless indicated. All animal procedures were approved by the Institutional Animal Ethics Committee, and were conducted at the Nanfang Hospital Animal Experiment Center and in accordance with NIH guidelines.
Five-week-old rats were fed a normal-fat (17% fat-derived calories, 3.7 kcal/g, AIN93, control) or HF diet (60% fat-derived calories, 5.0 kcal/g, TP23300, HF) for 9 weeks. Then, control or HF rats were randomly assigned to two groups (n=6 per group) receiving their allocated diets but with normal (0.4% NaCl) or high salt contents (4% NaCl) for 3 weeks. All diets were purchased from Trophic Animal Feed High-Tech Co., Ltd (Nantong, China).
HF rats received a HF, high salt (4% NaCl) diet for 3 weeks, and were randomly assigned to 11 groups (n=6 in each group), for the following treatments: (1) Intragastric (IG) vehicle (phosphate buffered saline, pH 7.4) as an IG control, or IG losartan (Sigma, MO, U.S.A.) at 1 or 500 mg/kg/day to block the angiotensin (Ang) II type 1 (AT1) receptors; (2) Intracerebroventricular (ICV) administration vehicle (artificial cerebrospinal fluid, aCSF) as an ICV control or ICV losartan at 1 mg/kg/day to block central RAS using Alzet osmotic minipumps (Durect Corp, CA, U.S.A.); (3) ICV clonidine (Sigma) to block central sympathetic outflow at 5.76 μg/kg/day using osmotic minipumps; (4) Denervation of epidydimal fat pads to block both efferent and afferent SNS signals, as described ; (5) Selective deafferentation of epididymal fat pads by injection of resiniferatoxin (RTX) (Sigma) into epididymal fat pads as previously described ; (6) IG tempo (Sigma) at 30 mg/kg/day to block oxidative stress; (7) IG hydralazine (Sigma) at 15 mg/kg/day to normalize blood pressure (BP) as a control for the effects of hypertension.
The doses of ICV losartan or clonidine were chosen according to preliminary studies in which intravenous injection of the doses administered ICV had no detectable effect on angiotensin II (Ang II)-induced increase in BP or adipose norepinephrine (NE) concentration. The accuracy of the ICV injection was confirmed by brain distribution of Evans Blue dye. Denervation of epididymal fat pads (Epi DN) was performed as described previously . The effectiveness of epididymal fat pads denervation was confirmed by a reduction in NE concentration in adipose tissue to <30% of the control levels (Supplementary Figure S7A).
Briefly, selective deafferentation of epididymal fat pads were performed by exposure of the fat pads and injection with RTX (20 pmol/μl, 8 μl/site, 4 sites for each fat pad) as described . The effectiveness of deafferentation was confirmed by a decrease in directly recorded epididymal adipose tissue afferent SNS activity, and by a reduction in the expression of adipose calcitonin gene-related peptide (CGRP), to <30% of the control (Supplementary Figure S7B–F). CGRP is a marker for afferent nerves .
Measurement of biochemical parameters and systolic BP
Serum levels of sodium, glucose, triglyceride and cholesterol were measured with an automated chemistry analyzer (AU480, Beckman Coulter, CA, U.S.A.), insulin by ELISA (Millipore), and free fatty acid (FFA) by ELISA (Biovision, CA, U.S.A.).
FFA release from isolated epididymal adipose tissue (explants) was determined as previously described . Briefly, tissue pieces (∼100 mg total weight) of freshly isolated epididymal fat pads were incubated in Dulbecco’s modified Eagle’s medium (Thermo Scientific, MA, U.S.A.) containing 2% fatty acid–free BSA (Sigma) at 37°C for 120 min. Samples of medium were collected after incubation and investigated for FFA content by a commercial kit (Biovision, CA, U.S.A.). FFA release in each sample was expressed as nmol/hour/100 mg tissue protein.
Systolic BP (SBP) was measured in conscious rats by telemetry using TA11PA-C10 probes (Data Sciences International, MN, U.S.A.) as described .
Analysis of tissue sodium content
Samples were vacuum dried for 72 h (Labconco Corporation, MO, U.S.A.), dissolved in HNO3 and the sodium contents were measured by inductively coupled plasma atomic emission spectrometry (ICP-AES, Thermo Scientific) as described .
Evaluation of IR
A hyperinsulinemic–euglycemic clamp study was performed in conscious and unstressed rats to assess IR . Hepatic glucose production was calculated by subtraction of the glucose infusion rate (GIR) from whole-body glucose disposal . At 120 min after the clamp, rats received an intravenous bolus of 2 [14C] deoxyglucose (2 [14C]DG; PerkinElmer, MA, U.S.A.) . Thirty minutes after the bolus, samples of blood, white adipose tissue (epididymal adipose tissue) and SM (gastrocnemius) were taken to determine tissue-specific glucose uptake .
Expression and translocation of glucose transporter 4
Following hyperinsulinemic–euglycemic clamps, samples of white adipose tissue (epididymal adipose tissue) and SM (gastrocnemius) were excised. Total protein, plasma membrane (PM) and total membrane (TM) extracts were prepared using a commercial Plasma Membrane Protein Extraction Kit (Abcam, MA, U.S.A.). Fractions were characterized using expression of Na+/K+-ATPase α1 subunit (1:400, 05-369, Millipore, CA, U.S.A.) as a marker for PMs . Glucose transporter 4 (Glut4) expression in total protein, PM and TM were analyzed by Western blot with a mouse anti-Glut4 antibody (1:500, 2213, Cell Signaling Technology, MA, U.S.A.). mRNA levels of Glut4 were analyzed by real-time PCR in tissue homogenates. The primers used are listed in Supplementary Table S3.
Insulin-stimulated insulin receptor substrate 1-associated phosphoinositide 3-kinase activity and phosphorylation of AS160
Following the hyperinsulinemic–euglycemic clamp study, epididymal adipose tissue and gastrocnemius were excised. Insulin-stimulated insulin receptor substrate (IRS) 1-associated phosphoinositide 3-kinase (PI3K) activity assay were performed in these tissues . Briefly, IRS1 was immunoprecipitated using a rabbit antibody a rabbit antibody against IRS1 (1:100, sc-559, Santa Cruz Biotechnology, TX, U.S.A.), and PI3K activity in the immunoprecipitates was assayed as previously described . Insulin-stimulated phosphorylation of AS160 in these tissues were determined by immunoprecipitation and Western blot as previously described  using a rabbit antibody against phosphorylated AS160 (1:500, Thr642) (4288, Cell Signaling Technology) or total AS160 (1:1000, 07-741, Millipore).
Evaluation of tissue inflammation
Macrophage infiltration in epididymal adipose tissue and gastrocnemius were imaged by Hematoxylin–Eosin (HE) and immunohistochemical staining using a monoclonal antibody specifically recognizing macrophage marker ED-1 (1:50, MCA341, Serotec, Oxford, U.K.) and quantified by a blinded observer counting the ED-1-positive cells in 20 randomly chosen areas (0.3 × 0.3 mm2).
Measurement of inflammatory mediators
The expression of tumor necrosis factor α (TNFα) in tissue homogenates was determined by Western blots using a rabbit anti-rat TNFα antibody (1:1000, GTX110520, GeneTex, CA, U.S.A.). mRNA levels of ED-1 and TNFα in tissue homogenates were analyzed by real-time PCR, using primers listed in Supplementary Table S3.
Evaluation of RAS expression
Western blot and real-time PCR
The expression of angiotensinogen (AGT) in tissue homogenates was determined by Western blots using a rabbit anti-rat AGT antibody (1:500, 28101, IBL, Gunma, Japan) and by real-time PCR, using primers listed in Supplementary Table S3.
The cerebral expressions of AGT and Ang II were semi-quantitated with immunohistochemical staining using rabbit anti-mouse AGT (1:200, A6279, Wuhan, China), and rabbit anti-Ang II antibody (1:800, T-4007, Peninsula Laboratories, CA, U.S.A.).
For localization of RAS in adipocytes or macrophages, epididymal adipose tissue sections were double-stained with the first primary antibody against AGT (1:200, 28101, IBL) or Ang II (1:400, T-4007, Peninsula Laboratories), and the second primary antibody against adiponectin (1:400, ab22554, Abcam) or ED-1 (1:50, MCA341, Serotec). For localization of RAS in SM, gastrocnemius sections were double-stained using the anti-AGT (1:200, 28101, IBL) or anti-Ang II (1:400, T-4007, Peninsula Laboratories) antibody as the first primary antibody, and anti-dystrophin (1:200, ab129996, Abcam) antibody or fluorescein-labeled Griffonia simplicifolia I lectin (1:50, L9381, Sigma, MO, U.S.A.) as the second primary antibody.
Liquid chromatography/mass spectrometry
Ang II concentrations in plasma and tissue homogenates were determined by liquid chromatography/mass spectrometry (LC/MS) . Samples were extracted and separated with C18 Sep-Pak cartridges (Thermo Scientific, MA, U.S.A.), followed by detection with Q-ToF mass analyzer in the ESI+-MS ion mode.
Evaluation of oxidative stress
Serum concentrations of 8-iso-prostane were quantified by Elisa (Enzo Life Science, NY, U.S.A.; Millipore, MA, U.S.A.). The expression of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase subunits Nox2 and p22phox in tissue homogenates were determined by Western blot using anti-Nox2 (1:5000, ab129068, Abcam) or p22phox (1:100, sc-271968, Santa Cruz Biotechnology) antibody.
Measurement of tissue perfusion
After the IR testing, the blood perfusion of epididymal fat pad and hindlimb adductor muscle group (adductor magnus and semimembranosus) were determined by contrast-enhanced ultrasound . Initially, an epididymal fat pad was exposed and kept moist with saline, and a hindlimb adductor muscle group was visualized in situ and identified by B-mode imaging . Contrast imaging in the fat pad and the muscle group was recorded during continuous infusion of microbubbles via the tail vein at 8 × 107/min . Infusion time versus video intensity (VI), which is the ‘brightness’ of pixels in the ultrasound image in decibels (dB), was fitted to an exponential function: y = A (1 − e−βt) . Where, t is the infusion time, A is the plateau, VI is an index of blood volume, and β is a measure of mean microbubble velocity . Perfusion was calculated by multiplying A and β .
Evaluation of SNS activity
Recording SNS activity
Afferent and efferent SNS activity to the hindlimb and the epididymal fat pad was measured as described previously . Initially, the single nerve bundle supplying the left hindlimb and epididymal fat pad was isolated and severed. The nerve proximal to the cut was used to record efferent SNS activity, and the afferent distal end of the nerve bundle was used to record afferent SNS activity. The nerve signal was amplified 5000 times with an ERS 100C amplifier and filtered with a band-pass between 1 and 3000 Hz . The amplified and filtered nerve signal was stored and analyzed on a personal computer and recording software (Acqknowledge, Biopac System, CA, U.S.A.). Afferent/efferent SNS activity was expressed as mean number of bursts per second.
The concentrations of NE in serum and tissue homogenates were assessed using an ELISA kit (Demeditec Diagnostics, Kiel, Germany).
Tissue expression of CGRP was assessed by immunohistochemistry staining using a rabbit anti-rat CGRP antibody (1:200, GTX16211, Gene Tex, CA, U.S.A.). Concentrations of CGRP in tissue homogenates were determined using an ELISA kit (Cayman Chemical, MI, U.S.A.). The expression of cerebral tyrosine hydroxylase (TH) in c-fos positive neurons was determined by double staining using a rabbit anti-rat TH (1:100, MAB318, Millipore) and a rabbit anti-rat c-fos (1:50, ab209794, Abcam) antibody .
Evaluating the role of adrenergic receptors in IR
Expression of adrenergic receptors
The expressions of adrenergic receptors (ARs) in homogenates of epididymal adipose tissue and gastrocnemius were determined by Western blot using antibodies against α1- (1:400, ab3462, Abcam), α2- (1:1000, CA1003, Cell Application, CA, U.S.A.), β1- (1:1000, ab3442, Abcam), β2- (1:1000, ab182136, Abcam) or β3-ARs (1:100, sc-1473, Santa Cruz Biotechnology).
Effect of ARs blockade
Prazosin (40 mg/kg/day, IG, Sigma) or atipamezole (2 mg/kg/day, subcutaneous administration using Alzet osmotic minipumps, Tocris Bioscience, Bristol, U.K.) was administrated to block α1- and α2-ARs, and carvedilol (90 mg/kg/day, IG, Sigma) to block β- and α1-ARs in HF rats under high-salt conditions for 3 weeks.
Continuous variables were expressed as mean ± SD. Differences among groups were tested by one-way ANOVA or unpaired t test with Bonferroni correction for multiple testing. Two-way ANOVA was used to test the differences in effects of salt-loading between control and HF rats. Statistical analysis was conducted with SPSS 17.0 for Windows. A value of P<0.05 was considered statistically significant. Power analysis for the independent t test was conducted to assess the number of animals necessary to obtain statistically significant data (α 0.05) for preselected powers 0.8 (nQuery version 8.2.1, Statistical Solutions Ltd., Ireland).
HF feeding impairs glucose uptake in adipose tissue and SM and results in IR that are enhanced by high salt intake
Rats fed high or normal fat diets were divided randomly into two groups receiving a normal- (0.4% NaCl) or high salt (4% NaCl) intake for 3 weeks. Insulin sensitivity, measured by a hyperinsulinemic–euglycemic clamp test, was impaired in HF fed rats. This was decreased further by a high salt intake (Figure 1A,B).
HF feeding impairs glucose uptake in adipose tissue and SM and results in IR that is enhanced by high salt intake
Compared with control rats, insulin-stimulated hepatic glucose production was increased significantly in HF fed rats, but was not changed further by salt-loading (Figure 1C). Insulin-stimulated glucose uptake in adipose tissue and SM was impaired markedly in HF fed rats. These were reduced further by high salt intake (Figure 1D,E). HF fed rats also had increased circulating NE, TNFα, 8-iso-prostaglandin F2α and FFA, enhanced the release of FFA from adipose tissue, and elevated SBP, which were increased further by high salt intake (Figure 1F–H; Supplementary Table S1). Serum adiponectin levels were decreased and serum leptin levels were increased in HF fed rat, but they were not changed further by high salt intake (Supplementary Table S1).
The sodium content of adipose tissue or SM was unaltered by salt-loading, but that of skin was elevated by approximately 50% by salt-loading in both HF and control rats (Supplementary Table S1).
HF diet activates adipose RAS and SNS, leading to inflammation and impaired Glut4 trafficking that are enhanced by high salt diet
HF diet significantly increased macrophage infiltration in adipose tissue that was increased further by high salt (Figure 2A,B).
HF diet activates adipose RAS and SNS, leading to inflammation and impaired Glut4 trafficking that are enhanced by high salt diet
AGT and Ang II were expressed mainly in adipocytes and infiltrating macrophages (Figure 2C–E). Both were unchanged by high salt in control rats, but were increased in HF fed rats and increased further by high salt intake (Figure 2C,D; Supplementary Figure S1A). This contrasts with circulating levels of Ang II which were increased in HF rats but were reduced noticeably by high salt intake (Supplementary Table S1).
HF feeding also increased the adipose tissue afferent and efferent sympathetic nerve activity (Figure 2F,G), up-regulated the expression of the NADPH oxidase subunits Nox2 and p22phox, and elevated protein levels of TNFα in adipose tissue, all of which were enhanced further by a high salt intake (Figure 2H; Supplementary Figure S1B,C).
The adipose expression and trafficking of glucose transporter (Glut) 4 was decreased in HF fed rats, and was reduced further by high salt intake. IRS1-associated PI3K activity and phosphorylation of AS160 in adipose tissue were decreased in HF fed rats, and were decreased further by a high salt intake (Figure 2I; Supplementary Figure S1D–G). Insulin-stimulated adipose tissue perfusion was impaired in HF fed rats, but was not reduced further by high salt intake (Supplementary Figure S1H).
HF diet activates RAS and SNS in SM, reduces muscle perfusion, and impairs Glut4 trafficking that are enhanced by high salt intake
Unlike adipose tissue, macrophage infiltration was not observed in SM of HF fed rats (Supplementary Figure S2A,B). However, the expression of AGT and Ang II, located in both myocytes and microvessels surrounding muscle fibers, were increased in HF fed rats (Figure 3A–C; Supplementary Figure S2C), accompanied by increased hindlimb efferent and afferent sympathetic nerve activity (Figure 3D,E), and up-regulated expression of NADPH oxidase subunits in SM, all of which were enhanced further by a high salt intake (Figure 3F).
HF diet activates RAS and SNS in SM, reduces muscle perfusion, and impairs Glut4 trafficking that are enhanced by high salt intake
The translocation of Glut4 was reduced in SM from HF rats and were decreased further by high salt intake (Figure 3G; Supplementary Figure S2D,E). HF diet impaired the insulin signaling (Supplementary Figure S2F,G) and insulin-stimulated hindlimb blood perfusion (Figure 3H) that were exacerbated by high salt.
HF diet up-regulates the expression of cerebral RAS and TH via signals from the innervated adipose tissue that are enhanced by high salt intake
HF diet increased the expression of the brain RAS in the paraventricular nucleus (PVN) and subfornical organ (SFO) that was increased further by high salt (Supplementary Figure S3A–E). The expression of the RAS in the cerebral cortex (as a control) was modest and independent of salt in both control and HF rats (data not shown). HF diet increased central sympathetic drive as indicated by increased generation of TH in c-fos-positive neurons in the rostral ventrolateral medulla (RVLM) and induced central oxidative stress indexed by up-regulation of Nox2. Both were accelerated by a high salt intake (Supplementary Figure S3F,G).
Remarkably, blockade of adipose SNS traffic by denervation of the epididymal fat pads or selective blockade of afferent SNS signals from the epididymal adipose by RTX alleviated the up-regulation of the brain RAS, TH, and Nox2 in HF rats receiving a high salt intake, while normalizing BP by hydralazine was ineffective (Supplementary Figure S3).
Activation of adipose RAS and impaired Glut4 trafficking in HF fed rats receiving high salt intake are prevented by blockade of an adipose-cerebral sympathetic reflex
Among HF fed rats receiving a high salt intake, blockade of cerebral AT1 receptors by ICV losartan, at a dose of 1/500 of the effective IG dose, decreased the adipose overexpression of AGT, Ang II, and Nox2, reduced the efferent sympathetic nerve activity to adipose tissue, alleviated the inflammation (Figure 4A–E; Supplementary Figure S4A–C), and improved Glut4 trafficking, insulin signaling, and insulin-stimulated glucose uptake in adipose tissue (Figure 4F,G; Supplementary Figure S4D–F).
Activation of adipose RAS and impaired Glut4 trafficking in HF rats receiving high salt intake are prevented by blockade of an adipose-cerebral sympathetic reflex
Blockade of central sympathetic outflow by ICV clonidine or blockade of SNS traffic by denervation or deafferentation of epididymal fat pads all suppressed the effects of high salt intake to enhance adipose inflammation, and improved glucose uptake in adipose tissue of these rats (Figure 4).
Moreover, blockade of systemic oxidative stress by IG tempol suppressed adipose SNS activity and inflammation, and improved glucose uptake in HF fed rats receiving a high salt intake, while normalizing BP by hydralazine was ineffective (Figure 4). Adipose inflammation was increased by 63% in HF fed rats receiving a high salt intake, while denervation or deafferentation of epididymal fat pads reduced adipose inflammation by 64 or 57%, and ICV losartan or clonidine reduced this by >63%. Insulin-stimulated glucose uptake in adipose tissue was reduced by 62% in HF fed rats receiving a high salt intake, but this was restored by ICV losartan or clonidine, or by denervation or deafferentation of epididymal fat pads (Figure 4).
Activation of the RAS/ROS and impaired Glut4 trafficking in the SM of HF fed rats during a high salt intake are prevented by blockade of an adipose-cerebral-skeletal muscle sympathetic reflex
ICV losartan given to block cerebral AT1 receptors in HF fed rats during a high salt intake down-regulated the overexpression of AGT, Ang II, and Nox2 of SM and reduced the sympathetic nerve activity to the hindlimb (Figure 5A–D), enhanced Glut4 trafficking, increased blood perfusion, and improved glucose uptake in SM (Figure 5E–G; Supplementary Figure S5).
Activation of the RAS/ROS and impaired Glut4 trafficking in the SM in HF fed rats during a high salt intake are prevented by blockade of an adipose-cerebral-skeletal muscle sympathetic reflex
ICV clonidine given to block sympathetic outflow or denervation or deafferentation of the epididymal fat pads, or blockade of oxidative stress by IG tempol in HF fed rats during high salt intake all inhibited the activation of the RAS, or oxidative stress, preserved muscle perfusion, and improved Glut4 trafficking and glucose uptake in SM (Figure 5A–G). In general, insulin-stimulated glucose uptake in SM of HF rats was decreased 55% by high salt intake but this was increased by >62% by ICV losartan or clonidine and 48% by Epi DN. Interestingly, blockade of adipose afferent sympathetic signals alone reduced efferent sympathetic nerve activity to the hindlimb and improved the glucose uptake in SM (Figure 5A–G).
Blockade of an adipose-cerebral sympathetic reflex in HF fed rats during high salt intake improves whole-body IR
Insulin sensitivity, measured by the euglycemic-clamp test, was reduced by 28% in HF fed rats during high salt intake but was improved by 65% by ICV losartan, 62% by ICV clonidine and by >45% by denervation or deafferentation of epididymal fat pads (Figure 5H). These treatments also decreased the circulating levels of NE, 8-iso-prostaglandin F2α, TNFα and FFA, and the release of FFA from adipose tissue (Supplementary Table S2).
ARs mediate IR in HF rats during high salt intake
β1- and β3-ARs were the major ARs localized in white adipose tissue, while α1- and β2-ARs were the major ARs localized in SM. Their expression was unaltered by dietary salt in either control or HF fed rats (Supplementary Figure S6A,B).
HF fed rats during high salt intake were administered prazosin over 3 weeks to block α1-ARs, or atipamezole to block α2-ARs, or carvedilol to block both β- and α1-ARs. Prazosin and carvedilol both improved insulin sensitivity in HF fed rats receiving high salt intake (Figure 6A). Blockade of β- and α1-ARs with carvedilol, but not α1-ARs alone with prazosin, alleviated cellular inflammation, and abnormal Ang II expression, Glut4 trafficking and glucose uptake in adipose tissue from HF fed rats during high salt intake (Figure 6B–E; Supplementary Figure S6D–F). Both carvedilol and prazosin improved SM perfusion, and reduced expression of Ang II, and improved Glut4 trafficking and glucose uptake (Figure 6F–I; Supplementary Figure S6G,H).
ARs mediate IR in HF rats under high salt conditions
The present study confirms that HF diet causes IR in rats by impairing glucose uptake in adipose tissue and SM. The new findings, as illustrated in Figure 7, are that an HF diet increases adipose tissue afferent sympathetic nerve activity that reflexively increases efferent sympathetic nerve activity to adipose tissue and SM that are promoted by a high salt intake. The increased efferent nerve activity to SM and adipose tissue impair their local glucose uptake.
Schematic diagram summarizes the steps that link an adipose reflex to IR during HF feeding
Our results disclose a rich neural network that provides a metabolic link among white adipose tissue and the tissues responsible for glucose disposal during HF feeding, that is itself exacerbated by dietary salt intake. We demonstrate that HF feeding increases the expression of the systems that generate ROS in adipose tissue. These may contribute to the activation of the afferent sympathetic nerves since ROS can reflexively enhance sympathetic reflex activity [24,25]. This local ROS response itself may be driven by impaired metabolic activity in fat . Moreover, ROS enhances metabolic disturbance in fat , thereby setting in place a positive feedback loop. Indeed, systemic administration of tempol corrected both the metabolic dysfunction in adipose tissue and the enhanced sympathetic tone in this model.
We reveal that activation of afferent nerves from the white adipose tissue are required to increase efferent nerve activity to SM that mediate the IR in that tissue. Similarly, we have reported recently that CKD activates renal afferent nerves that activate the central SNS  to increase peripheral SNS traffic that impairs glucose uptake in SM and results in IR in a rat model of CKD . This study now broadens our findings by demonstrating that a model of obesity from HF feeding reflexively activates the SNS by afferents from white adipose tissue that mediates much of the impaired peripheral glucose uptake and whole-body IR in obese rats. Thus, glucose uptake into adipose tissue and SM is increased more than 60% following interruption of this reflex by central administration of clonidine. Moreover, selective blockade of afferent signals from white adipose tissue by local deafferentation down-regulated central SNS outflow, corrected the increased SM SNS activity, and restored the impaired glucose uptake and perfusion in SM. This identifies adipose tissue as the prime initiator of the metabolic response to HF feeding and implies a critical role of the adipose reflex in the impairment of insulin sensitivity in this model.
AT1 receptor signaling in the brain plays a prominent role in the response to HF feeding. HF-fed rats during a high salt intake had increased numbers of TH-positive neurons in the RVLM, indicating increased noradrenergic activity in the center that is a gateway for activation of the SNS . An increased central sympathetic drive was confirmed by direct recording of increased efferent sympathetic nerve activity in the adipose tissue and SM. The activation of these interactive peripheral RAS/ROS systems, and the associated IR, were themselves induced by central activation of the sympathetic outflow in HF fed rats during a high salt intake, since central blockade of the SNS with ICV clonidine or adipose tissue denervation suppressed the activation of the local RAS/ROS/SNS systems and improved insulin sensitivity. This central SNS response is driven by activation of a central RAS, since ICV losartan decreased efferent SNS activity to adipose tissue and SM and attenuated all reflex responses to an HF diet, thereby improving IR. However, activation of the central RAS in HF fed rats contributed to the accompanying hypertension [29,30], since blockade of central AT1 receptors with losartan reduced cerebral sympathoexcitation and prevented hypertension . However, it was the SNS, rather than hypertension, that provided the link between the central and peripheral RAS, because normalization of BP with hydralazine failed to normalize peripheral glucose uptake. Thus, activation of afferent nerves from adipose tissue by HF feeding induces reflex activation of efferent sympathetic nerves that underlie much of the associated hypertension and IR.
Another important finding was that although the SNS triggers IR in SM and adipose tissue during HF feeding, the IR occurs by different mechanisms. It has been proposed that reduced glucose delivery can limit cellular glucose uptake into SMs . Indeed, we confirmed that α-AR blockade with prazosin restored SM blood flow and restored SM glucose uptake in parallel. This is in agreement with the reports of improved insulin sensitivity in patients treated with α1-AR blockers [33,34]. However, the impaired glucose uptake in adipose tissue required a combined α and β-AR blockade with carvedilol but not by α1-AR blockade with prazosin for its improvement. Presumably, efferent sympathetic nerves to adipose tissue activate β-receptors on adipocytes to inhibit cellular glucose uptake.
Local oxidative stress and RAS can impair insulin signaling and Glut4 trafficking in myocytes and adipocytes [35,36] and thereby can induce IR. We demonstrate that denervation of the adipose tissue in HF fed rats reduces their efferent SNS activity and reduces ROS and RAS in SM and adipose tissue, implicating a critical upstream role for the SNS in the pathophysiology of IR in obesity. This is confirmed by the finding that the administration of carvedilol prevented the local activation of ROS and RAS in adipose tissue, and in SM in this model. It is consistent with reports that combined α- and β-AR blockade reduce oxidative stress and inflammation [37,38].
In conclusion, we show that an adipose-brain-peripheral SNS reflex underlies impaired systemic glucose homeostasis during HF feeding by controlling their local RAS/ROS activity. We show further that this adipose reflex/RAS/ROS regulation is critical for the local insulin signaling and glucose uptake in adipose tissue and SM. Collectively, our data provide insight into the role of a reflex response initiated in adipose tissue that regulates glucose homeostasis. These findings disclose a pathway involved in adipose-brain-peripheral RAS/ROS activation during HF feeding (Figure 7) that may lead to the development of new therapeutic approaches to treat insulin-resistant individuals.
The underlying mechanism by which amassing of white adipose tissue in obesity regulates sympathetic drive to the tissues responsible for glucose disposal, and causes IR, remains unknown.
The present study demonstrates that an HF feeding engages a sympathetic reflex from the white adipose tissue that activates adipose-cerebral-muscle RAS/ROS axes and coordinates a reduction in peripheral glucose uptake. This process is promoted by salt-loading.
These findings disclose a pathway involved in adipose-brain-peripheral RAS/ROS activation during HF feeding that may lead to the development of new therapeutic approaches to treat insulin resistant individuals.
W.C. performed the experiments, analyzed the data, and drafted part of the manuscript. M.S. performed the experiments and analyzed the data. L.W., J.L. and Z.Y. performed biochemical experiments. Y.L. gave suggestions to the study design and paper writing. C.S.W. revised the manuscript. F.F.H. designed and financed the study, wrote and edited the manuscript.
This work was supported by the Major International (Regional) Joint Research Project of National Natural Science Foundation of China [grant number 81620108003 (to F.F.H.)]; the Guangzhou Regenerative Medicine and Health-Guangdong Laboratory Research Grant [grant number 2018GZR0201003 (to F.F.H.)]; the Program of Introducing Talents of Discipline to Universities, 111 Plan [grant number D18005 (to F.F.H.)]; the State Key Program of National Natural Science Foundation of China [grant number 81430016 (to F.F.H.)]; the Foundation for Innovation Research Groups of the National Natural Science Foundation of China [grant number 81521003 (to Y.L.)]; the National Natural Science Foundation of China [grant numbers 81570619, 81870473,81922014 (to W.C.)]; the NIH [grant numbers HL-68686, HL-134511, DK-049870, DK-036079 (to C.S.W.)]; the George E. Schreiner Chair of Nephrology and the Walters Family Chair of Cardiovascular Research (to C.S.W.); and the Smith-Kogard and Gildenhorn-Spiesman Family Foundation and the Georgetown University Hypertension Research Center (to C.S.W.).
The authors declare that there are no competing interests associated with the manuscript.
- Ang II
Akt substrate of 160 kDa
angiotensin II type 1 receptor
calcitonin gene-related peptide
chronic kidney disease
- Epi DN
denervation of epididymal fat pad
- Epi RTX
deafferentation of epididymal fat pad with resiniferatoxin
free fatty acid
glucose infusion rate
glucose transporter 4
insulin-stimulated insulin receptor substrate
nicotinamide adenine dinucleotide phosphate
reactive oxygen species
rostral ventrolateral medulla
sympathetic nerve system
tumor necrosis factor α
These authors contributed equally to this work.