Carotid body denervation improves hyperglycemia in obese mice

Abstract

The carotid bodies (CBs) have been implicated in glucose abnormalities in obesity via elevation of activity of the sympathetic nervous system. Obesity-induced hypertension is mediated by insulin receptor (INSR) signaling and by leptin, which binds to the leptin receptor (LEPR^b) in CB and activates transient receptor potential channel subfamily M member 7 (TRPM7). We hypothesize that in mice with diet-induced obesity, hyperglycemia, glucose intolerance, and insulin resistance will be attenuated by the CB denervation (carotid sinus nerve dissection, CSND) and by knockdown of Lepr^b, Trpm7, and Insr gene expression in CB. In series of experiments in 75 male diet-induced obese (DIO) mice, we performed either CSND (vs. sham) surgeries or shRNA-induced suppression of Lepr^b, Trpm7, or Insr gene expression in CB, followed by blood pressure telemetry, intraperitoneal glucose tolerance and insulin tolerance tests, and measurements of fasting plasma insulin, leptin, corticosterone, glucagon and free fatty acids (FFAs) levels, hepatic expression of gluconeogenesis enzymes phosphoenolpyruvate carboxykinase (PEPCK) and glucose 6-phosphatase (G-6-Pase) mRNA and liver glycogen levels. CSND decreased blood pressure, fasting blood glucose levels and improved glucose tolerance without any effect on insulin resistance. CSND did not affect any hormone levels and gluconeogenesis enzymes, but increased liver glycogen level. Genetic knockdown of CB Lepr^b, Trpm7, and Insr had no effect on glucose metabolism. We conclude that CB contributes to hyperglycemia of obesity, probably by modulation of the glycogen-glucose equilibrium. Diabetogenic effects of obesity on CB in mice do not occur via activation of CB Lepr^b, Trpm7, and Insr.

NEW & NOTEWORTHY This paper provides first evidence that carotid body denervation abolishes hypertension and improves fasting blood glucose levels and glucose tolerance in mice with diet-induced obesity. Furthermore, we have shown that this phenomenon is associated with increased liver glycogen content, whereas insulin sensitivity and enzymes of gluconeogenesis were not affected.

INTRODUCTION

Obesity defined as the body mass index (BMI) ≥ 30 kg/m² is highly prevalent in the United States affecting 40% of the adult population (1, 2). Obesity is a major risk factor for cardiovascular morbidity and mortality (3). Obesity also causes insulin resistance and type 2 diabetes, which contributes to the cardiovascular burden of obesity (4). Mechanisms by which obesity induces insulin resistance include systemic inflammation, increased lipolysis with free fatty acid (FFA) release and altered adipokine production, especially decreased adiponectin (5). β Cell dysfunction and failure, which mediates the transition from insulin resistance to prediabetes and diabetes, is a result of inability to cope with the progression of insulin resistance, i.e., β cell burnout, which is precipitated by other phenomena such as an increase in FFA and lipotoxicity (5). Obesity activates the sympathetic nervous system (6, 7). SNS activation increases blood pressure, induces adipose tissue lipolysis, suppresses insulin secretion, and up-regulates insulin counter-regulatory hormones, glucagon, and corticosteroids accelerating progression of insulin resistance and type 2 diabetes (8–14).

It has been reported that obesity-induced SNS activation and insulin resistance are linked to the carotid bodies (CBs) (15–17), a polymodal metabolic sensor located in the bifurcation of the carotid arteries. CBs have been implicated in glucose sensing (18) and insulin resistance (19, 20). CB increased splanchnic SNA (21), which suppressed insulin secretion and increased hepatic glucose output (22–28). CB denervation suppressed hypertension and abolished insulin resistance and hyperglycemia in obese rats and in mice exposed to intermittent hypoxia (19, 22).

We have previously identified a molecular pathway in CB that mediates the development of hypertension in obesity (29, 30). Adipocyte-produced hormone leptin binds to its receptor LEPR^b abundantly expressed on type I (glomus) CB cells. The leptin-LEPR^b complex transcriptionally and post-transcriptionally regulates transient receptor potential channel subfamily M member 7 (TRPM7) to increase CB activity(30–32), which increases SNS activity and causes hypertension (33). Alternatively, Ribeiro et al. (19) showed in obese rats that CB can be activated via phosphorylation of insulin receptor (INSR).

Here we propose that the carotid body contributes to the development of glucose intolerance and insulin resistance in diet-induced obese (DIO) mice, similar to obese rats (19) and that CB denervation will effectively treat hyperglycemia and insulin resistance in this model. Furthermore, we propose that the leptin-LEPR^b-TRPM7 CB axis is involved in the pathogenesis of type 2 diabetes and insulin resistance. We performed baseline measurements of fasting blood glucose, intraperitoneal glucose tolerance test (GTT), and insulin tolerance test (ITT) in DIO mice followed by 1) carotid sinus nerve dissection (CSND) or sham surgery as we have done before (22, 30, 34); 2) knockdown of Lepr^b, Trpm7 or Insr genes in the CB using short hairpin RNA (shRNA) followed by GTT and ITT.

METHODS

Experimental Animals

In total, 75 diet-induced obese (DIO) and 15 lean male C57BL/6J mice from the Jackson Laboratory, Bar Harbor, ME were used in the study. DIO was induced by feeding mice with a high-fat diet (TD 03584, Teklad WI, 5.4 kcal/g, 35.2% fat, 58.4% of kcal from fat) starting at 6 wk of age for 14 wk. Age-matched C57BL/6J mice on a regular chow diet (lean mice) were used as a control for DIO mice. All mice had free access to food and water and were housed in the standard laboratory environment with the 12-h light/dark cycle (9:00 AM–9:00 PM lights on/9:00 PM–9:00 AM lights off). The study was approved by the Johns Hopkins University Animal Use and Care Committee and complied with the American Physiological Society Guidelines for Animal Studies. The study design is shown in Fig. 1.

Figure 1.

The study design in different experimental groups (A and B) before and after carotid sinus nerve dissection (CSND) or (C) before and after shRNA treatment. BP, Blood pressure; HR, heart rate; IPGTT, Intraperitoneal glucose tolerance test; ITT, insulin tolerance test.

Carotid Sinus Nerve Dissection

DIO and age-matched lean C57BL/6J mice underwent carotid sinus nerve dissection (CSND. DIO, n = 15; lean, n = 7) or sham surgery (DIO, n = 13; lean, n = 8). CSND was performed under 1–2% isoflurane anesthesia with body temperature maintained at 37°C, as we have previously described (22, 30, 34, 35). Briefly, the CSNs were bilaterally dissected at the points of branching from the glossopharyngeal nerve to the cranial pole of the carotid body. During and after surgery, buprenorphine at 0.05 mg/kg/day sc was administered to prevent discomfort. Sham surgery was performed in a similar manner except that the CSNs were not severed.

Trpm7 shRNA, Lepr^b shRNA, Insr shRNA Infection in C57BL/6J Mice

To investigate TRPM7, LEPRb, and INSR signaling, in CB we used Ad-mCherry-U6-m-TRPM7-shRNA (Trpm7 shRNA, 9 × 10¹¹ VP/mL), Ad-GFP-U6-m-Lepr^b-shRNA (Lepr^b shRNA, (8.2 × 10¹¹ VP/mL), Ad-GFP-U6-mIr-shRNA (Insr shRNA, 7.6 × 10¹¹ VP/mL) and control scrambled shRNA, Ad-mCherry-U6-scrmb-shRNA (control shRNA, 3.3 × 10¹² VP/mL or 4.4 × 10¹² VP/mL) from Vector Biolabs, Malvern, PA. shRNA (4 µL) was mixed with 1 µL of Matrigel matrix and applied to the carotid body area bilaterally, as we have previously described (33, 35). Matrigel matrix treated mice were recovered for 9–10 days before glucose testing.

Intraperitoneal Glucose Tolerance Test and Insulin Tolerance Test

Intraperitoneal glucose tolerance test (IPGTT) and insulin tolerance test (ITT) were performed 9–14 days after the intervention (CSND/sham surgery or shRNA transfection) in unanesthetized animals. Mice were acclimated to the testing conditions and the tail was cut in the morning before the fasting period to avoid stress. IPGTT was performed after a 5-h fast (8:00 AM–1:00 PM) by injecting 1 g/kg glucose intraperitoneally (ip) (36). Glucose levels were measured by tail-snip technique using a handheld glucometer (Accu-Check Aviva, Roche, Indianapolis, IN) at baseline and at 10, 20, 30, 60, 90, and 120 min after glucose injection. ITT was performed after a 2-h fast (37) by injecting 1 U/kg insulin (Humulin R, Eli Lilly, Indianapolis, IN) ip. Glucose levels were measured at baseline, and at 10, 20, 30, 40, 50, 60, 90, and 120 min post-injection. The area under the curve was measured with GraphPad Prism software (Boston, MA).

Blood Pressure Telemetry Implantation and Recording

BP telemetry implantation and recording were performed, as previously described (30, 32, 33). Briefly, mice were anesthetized with 1–2% isoflurane and implanted with transmitters from Data Sciences International (St. Paul, MN). PA-C10 telemeters (pressure transducer) were placed in the left femoral arteries of mice. Mice were allowed to recover 10–14 days before the baseline recording. BP was recorded for 24 h (9:00 AM to the next day 9:00 AM), and CSND or sham surgery was performed. Mice were allowed to recover 10–14 days and BP was measured again for 24 h. Signals were captured using PowerLabs 16/35 interfaced with LabChart Pro software from ADInstruments (Colorado Springs, CO).

Systolic Arterial Pressure and Pulse Interval Variability

We used the software Cardioseries v2.7 (www.danielpenteado.com) to evaluate the pulse interval (PI) and systolic arterial pressure (SAP) variability in the time and frequency domains as previously described by us and others (38–40). We sampled the time interval of 30 min during the light phase (3:00–4:00 PM) and 30 min during the dark phase (0:00–5:00 AM). Briefly, in the time domain, we examined the PI average values. SAP and PI variability were also studied in the frequency domain by spectral analysis. The power spectra were estimated by the modified periodogram and Welch protocol. All series were interpolated at 12 Hz (cubic spline) and divided into segments of 512 points. Segments containing artifacts or transients were excluded. Afterward, each certain segment was multiplied by a Hanning window, and the periodogram was projected. The PI spectra were integrated into low- (LF, 0.1–1 Hz) and high-frequency (HF, 1–5 Hz) bands, whereas the SAP spectra were integrated at LF band only. The power at the LF band was assessed in normalized units (nu), represented by LF/(LF + HF), whereas the power at the HF band was evaluated in absolute units. Efforts were made to ensure that the analyses were done when SAP and PI were stable.

Biochemical Assays

Terminal blood draws by cardiac puncture and tissue harvesting were performed under 1–2% isoflurane anesthesia 3–5 days after the last IPGTT/ITT after the 5-h fast (8:00 AM–1:00 PM). Plasma leptin and adiponectin levels were measured with ELISA kits from Millipore (Billerica, MA). Plasma glucagon was measured with ELISA kit from R&D Systems, Inc. (Minneapolis, MN) and plasma insulin was measured with ELISA kit from Crystal Chemical, Inc. (Wakefield, MA). Plasma corticosterone was determined with ELISA kits from Abcam (Boston, MA). Liver glycogen was measured with ELISA kit from Biovision (Milpitas, CA). Liver tissue (10 mg) was homogenized with 200 ul dH₂O on ice, and then boiled for 5 min to inactivate enzymes. The boiled samples were centrifuged at 13,000 rpm for 5 min to remove insoluble material and the supernatant was used for the assay.

Quantification of mRNA Levels in Liver Tissues

Total RNA was extracted from liver using Trizol (Life Technologies, Rockville, MD) and cDNA was synthesized using Advantage RT for PCR kit from Clontech (Palo Alto, CA). Real-time reverse-transcriptase PCR (RT-PCR) was performed with primers from Invitrogen (Carlsbad, CA) and Taqman probes from Applied Biosystems (Foster City, CA). Mouse 18S rRNA and glucose 6-phosphatase (G-6-Pase) mRNA have been measured with the Applied Biosystems premade primers and probes (41). GoTaq Probe qPCR Mix (Promega, WI) and fluorogenic probes in TaqMan Gene Expression Assays (ThermoFisher Scientific, MA) were used to quantify mRNA levels of mouse phosphoenolpyruvate carboxykinase (PEPCK, assay#Mm01247058_m1). The mRNA expression levels were referenced to 18S rRNA and the values were derived according to the 2^−ΔΔCt method (42).

In Vitro Transcription Amplification and Quantification of mRNA Levels in CB

CBs were collected from untransfected DIO mice, and from DIO mice at the end of the experiments with Insr, Leprb, Trpm7, or control shRNAs and snap-frozen before RNA isolation. Total RNA was extracted using Trizol plus PicoPure RNA isolation kit (ThermoFisher Scientific, MA) followed by reverse transcription using High Capacity cDNA kit (ThermoFisher Scientific, MA) on individual mouse CB samples. Because of the limited amount of isolated RNA from CB of individual mouse, in vitro transcription (IVT) amplification was performed using the Arcturus RiboAmp HS PLUS RNA Amplification Kit (Applied Biosystems, CA). Amplified RNA (aRNA) was subject to reverse transcription using High Capacity cDNA kit (ThermoFisher Scientific, MA). GoTaq Probe qPCR Mix (Promega, WI) and fluorogenic probes in TaqMan Gene Expression Assays (ThermoFisher Scientific, MA) were used to quantify mRNA levels of Insr (assay#Mm01211875_m1), Lepr^b (assay#Mm01265583_m1), Trpm7 (assay#Mm00457998_m1), and tyrosine hydroxylase (TH) (assay#Mm00447557_m1). Technical triplicates were included in qPCR assay. The PCR efficiencies of Lepr^b, Trpm7, Insr, and TH are 102%, 100%, 98%, and 102%, respectively. cDNA generated from the Universal Reference Mouse RNA (Qiagen) and no RT-control cDNA served as the positive and negative control of qPCR. The mRNA levels of Insr, Insr, Lepr^b, and Trpm7 gene expression levels were normalized to the expression level of TH, and the fold change relative to universal mouse reference RNA was calculated using 2^−ΔΔCt method.

Statistical Analysis

The sample size was calculated based on our previous data on the effect of CSND during intermittent hypoxia (22). CSND decreased fasting blood glucose from 176.4 ± 12.5 mg/dL (n = 10) to 143.6 ± 10.1 mg/dL (n = 24) yielding Cohen d of 2.9 and a sample size of 8 per group at 80% power.

All values are reported as means ± SE. Statistical significance for all comparisons was determined by two-way analysis of variance test with Bonferroni post-hoc correction for multiple comparisons. For IPGTT and ITT, we performed a repeated measures analysis of variance test, and significance was determined using Tukey’s post hoc test by GraphPad Prism software. For SAP and PI variability, Wilcoxon matched-pairs signed was used to compare differences within groups. A P value of <0.05 was considered significant.

RESULTS

All the mice were 20–24 wk of age at the time of the experiment (Fig. 1). Animals fully recovered after the surgery regaining weight (Table 1). CSND decreased mean arterial pressure (MAP) from 111.4 ± 1.2 mmHg to 104.2 ± 2. 0 mmHg (P < 0.01) (Fig. 2). Blood pressure showed strong circadian characteristics with the values significantly higher during the dark active phase, but CSND was effective throughout the light/dark cycle. During the light phase, when animals were predominantly asleep, MAP decreased from 105.9 ± 1.2 mmHg to 96.9 ± 2.3 mmHg (P < 0.01). During the dark phase, MAP decreased from 117 ± 1. 3 mmHg to 111.5 ± 2.2 mmHg (P = 0.05). Sham surgery had no significant impact on blood pressure. The heart rate overall was not affected by CSND.

Table 1.

Basic characteristics of diet-induced obese mice before and after CSND or Sham surgery

	SHAM				CSND
	DIO		Lean		DIO		Lean
N	13		8		15		7
Time	Before surgery	After surgery	Before surgery	After surgery	Before surgery	After surgery	Before surgery	After surgery
Age, wk	21.5 ± 0.6	24.0 ± 0.4	22	24	21.0 ± 0.5	24.0 ± 0.3	22	24
Body weight, g	45.8 ± 1.1	46.3 ± 1.1	30.4 ± 0.7	30.5 ± 0.7	45.2 ± 1.3	44.2 ± 1.5	29.6 ± 0.6	29.7 ± 0.4

There were no significant effects of CSND or Sham surgery. CSND, carotid sinus nerve dissection; DIO, diet‐induced obese mice.

Figure 2.

Effect of carotid sinus nerve dissection (CSND) on blood pressure and heart rate (HR) in diet-induced obese mice. A: the 24-h profile of mean arterial pressure (MAP) before and after CSND (N = 7) or Sham surgery (N = 5). B: 24-h average MAP and HR. BPM, beats per min. Shaded areas denote the dark phase. *P < 0.01, **P < 0.005.

Spectral analysis in the frequency domain of the SAP showed that LF power was significantly decreased by CSND in the light phase from 4.7 ± 1.0 to 2.9 ± 0.5 mmHg², P < 0.05, but not in the dark phase. Analyses of PI in the frequency domain showed no changes in LF, HF, and LF/HF in dark and light phase (P > 0.05, Fig. 3A). PI showed a significant increase by the CSND in the light phase from 107 ± 2 ms to 120 ± 4 ms by CSND, P < 0.05, but not in the dark phase (P > 0.05, Fig. 3B). Apparent contradiction of this finding to the lack of the CSND effect on the heart rate overall can be attributed to the selection of a 30 min interval in the PI analysis. The PI was measured in the middle of the restful light phase (3:00–4:00 PM), when mice were predominantly asleep.

Figure 3.

Effect of carotid sinus nerve dissection (CSND) on (A) low frequency (LF) of systolic arterial pressure (SAP) and LF/high frequency (HF) ratio of pulse interval (PI). B: PI in the time domain. *P < 0.05 between before and after CSND.

DIO mice exhibited fasting hyperglycemia with blood glucose levels of 160 ± 6 mg/dL, which was attenuated by CSND with fasting blood glucose levels declining to 131 ± 6 mg/dL (Fig. 4A, P < 0.01). As expected, obese mice at baseline showed fasting hyperglycemia and significantly impaired glucose tolerance and insulin resistance compared with lean mice (Fig. 4A). CSND did not significantly affect fasting plasma levels of insulin, leptin, adiponectin, and insulin counter-regulatory hormones glucagon and corticosterone (Fig. 4, B–F). There was no effect on such modifiers of insulin resistance as free fatty acids and triglycerides (5) (Fig. 4, G and H). Sham surgery had no effect on blood glucose. GTT demonstrated a significant improvement of glucose tolerance after CSND with the area under the curve (AUC) decreasing from 48,898 ± 3,515 mg × min/dL to 37,369 ± 2,409, P < 0.05 (Fig. 5, A and B). The beneficial effect of CSND on glucose tolerance persisted after adjusting for fasting glucose values (Fig. 5, C and D, P < 0.05). In contrast, insulin tolerance test showed no benefit of CSND on insulin sensitivity (Fig. 6).

Figure 4.

Effect of carotid sinus nerve dissection (CSND) on fasting blood glucose level (A), and fasting plasma levels of insulin (B), glucagon (C), corticosterone (D), leptin (E), adiponectin (F), free fatty acids (FFAs) (G) and triglycerides (TGs) (H). Each dot represents one mouse (n = 8 per group). A shows data before and CSND and sham surgery in diet-induced obese (DIO) mice and age-matched lean mice. B–H show data after CSND in DIO mice. **P < 0.01 between before and after CSND in diet-induced obese (DIO) mice. ‡P < 0.0001 fasting blood glucose level prior to surgery between DIO mice and age-matched lean mice.

Figure 5.

Effect of carotid sinus nerve dissection (CSND) on glucose tolerance in the intraperitoneal glucose tolerance test (IPGTT) in diet-induced obese mice (DIO) and age-matched lean mice. A: blood glucose levels in IPGTT without adjustments (left) and after subtraction of fasting levels (right) in DIO mice; *P < 0.05 between baseline and post-CSND values. B: IPGTT area under the curve (AUC) without adjustment (left) and after subtraction of fasting levels (right) in DIO mice; *P < 0.05, between baseline and post-CSND values. Each dot represents one mouse (n = 8 per group). C: blood glucose levels in IPGTT without adjustments (left) and after subtraction of fasting levels (right) in age-matched lean mice. ‡P < 0.0001 fasting blood glucose level between DIO mice and age-matched lean mice. D: IPGTT area under the curve (AUC) without adjustment (left) and after subtraction of fasting levels in age-matched lean mice. ‡P < 0.0001 between DIO mice and age-matched lean mice. Each dot represents one mouse (Sham, n = 8; CSND, n = 7).

Figure 6.

Effect of carotid sinus nerve dissection (CSND) on insulin tolerance test (ITT) in diet-induced obese mice (A) and age-matched lean mice (B). ‡P < 0.0001 glucose levels between DIO mice and age-matched lean mice prior to surgery. Blood glucose levels in absolute values (left) and as % of fasting blood glucose (right) (DIO, n = 8 per group; lean, sham n = 8, CSND n = 7).

Given that fasting blood glucose levels are predominantly determined by hepatic glucose output, which incorporates two metabolic processes, gluconeogenesis, and glycogenolysis, we examined key indexes of both. Two key enzymes of gluconeogenesis, glucose-6-phosphatase (G6P) and phosphoenolpyruvate carboxykinase (PEPCK) are transcriptionally regulated (43). CSND had no effect on G6P mRNA levels (Fig. 7A), whereas PEPCK expression was increased (not decreased as would be expected if gluconeogenesis was upregulated, Fig. 7B). Glycogen level in the liver was increased from 3.1 ± 1.3% of liver weight to 4.6 ± 1.3% (Fig. 7B) suggesting that glycogen metabolism was switched to glycogen synthesis, whereas glycogen degradation was suppressed.

Figure 7.

Effect of carotid sinus nerve dissection (CSND) on the liver gene expressions and glycogen levels: glucose-6-Phosphatase (G6P) (A), phosphoenolpyruvate carboxykinase (PEPCK) (B). ***P < 0.001 unpaired t test between sham and CSND groups. C: liver glycogen level per tissue weight (*P < 0.05, unpaired t test between sham and CSND groups). Each dot represents one mouse (n = 8 per group).

In separate sets of experiments, we suppressed Lepr, Trpm7, and Insr gene expression in the CB using shRNA transfection (Fig. 8). shRNA against Lepr^b, Trpm7, and Insr decreased mRNA level (vs. CON shRNA) by 71%, 81%, and 74% respectively, indicating the success of shRNA knockdown assays. In contrast, CON shRNA treatment did not affect gene expression compared with untreated mice (Fig. 8). GTT and ITT were tested after the transfection. No significant effect of Lepr, Trpm7, and Insr knockdown in the CB on fasting glucose level, glucose tolerance, or insulin resistance was detected (Fig. 8).

Figure 8.

The effect of knockdown of specific genes in the carotid bodies (CBs) of obese mice on blood glucose levels. The left panels show CB mRNA ratios of leptin receptor (Lepr^b, A), transient receptor potential channel subfamily M member 7 (Trpm7, B) or insulin receptor (Insr, C) to tyrosine hydroxylase (TH, A–C) after control (CON) shRNA (n = 8–9), target shRNA treatments (n = 9, n = 8, and n = 7 for A–C, respectively) and untreated group (n = 8, A–C). Each dot represents one mouse. The middle panels show results of the intraperitoneal glucose tolerance test and the right panels show results of the insulin tolerance test. *P < 0.05, **P < 0.005, ***P < 0.001, Tukey’s multiple comparisons test between CONshRNA and the target shRNA and untreated groups.

DISCUSSION

We have examined the role of the CBs in metabolic dysfunction induced by obesity in our DIO mouse model. The main finding of the study is that carotid body denervation attenuates fasting hyperglycemia of obesity and improves glucose tolerance. There was no effect on plasma insulin levels and insulin resistance. The attenuation of hyperglycemia after CB denervation was associated with an increase in liver glycogen level, whereas indexes of gluconeogenesis were unchanged. The LEPR^b-TRPM7 carotid body pathway, which plays a critical role in cardiovascular and respiratory effects of the carotid bodies in obesity, appeared not to contribute to glucose regulation.

As expected, DIO mice develop fasting hyperglycemia and impaired glucose tolerance (44, 45). Here, we have shown that obesity-induced hyperglycemia is mediated, at least in part, by the carotid bodies sensory input and attenuated by CB denervation. CSND decreased fasting blood glucose levels by nearly 30 mg/dL, but it did not reach previously reported levels in lean mice of a similar age (44), which indicates that the carotid body afferent input is an important but not the sole contributor to hyperglycemia in dietary obesity. In lean mice with unimpaired glucose values, CSND did not impact metabolic parameters suggesting “a floor effect.” Ribeiro et al. reported similar findings in obese rats and showed that these effects are caused by an increase in sympathetic nervous system activity and insulin resistance (19, 46). In contrast, our murine model showed no changes in fasting plasma insulin or the insulin tolerance test after CSND suggesting that CB does not contribute to insulin resistance in obese mice. CSND did not have an impact on mechanisms of insulin resistance: there was no change in body weight, insulin counter-regulatory hormones glucagon and corticosterone, no change in leptin or indexes of lipotoxicity, free fatty acids, and triglycerides (5, 47, 48). We have previously shown that CSND does not affect the size of epididymal fat and liver triglyceride content (22). A major factor of insulin sensitivity, adiponectin, was also unaffected (49). Thus, CSND attenuated hyperglycemia by insulin-independent mechanisms.

Improvement in fasting hyperglycemia after CSND likely reflects a decrease in hepatic glucose output. Hepatic glucose production is a key mechanism in preventing hypoglycemia during fasting mediated by gluconeogenesis, which is de novo glucose synthesis from lactate, glycerol, and amino acids (50, 51), as well as glycogen biosynthesis and degradation balance (35, 52–54). Gluconeogenesis is regulated by two transcriptionally regulated rate-limiting enzymes (55, 56), glucose-6-phosphatase (G-6-P), gene expression of which was unchanged, and phosphoenolpyruvate carboxykinase (PEPCK), which was increased rather than decreased by CSND. CB denervation did not affect plasma level of glucagon, a key mediator of gluconeogenesis (51, 57). This data implies that a beneficial effect of CB denervation on hyperglycemia is not mediated by a decrease in gluconeogenesis. In contrast, CSND caused a significant accumulation of liver glycogen, a nearly 50% increase from baseline (from 3.1% to 4.6% of the liver weight) suggesting a shift in the glycogen balance from degradation to biosynthesis.

The glycogen metabolism is regulated by the SNS, activation of which increases glycogen degradation and blood glucose levels (58–60)^. Direct measurement of SNS activity is a complex task, since plasma catecholamine levels indicate mostly spillover from sympathetic synapses rather than true activity (61). Ribeiro et al. (19) have demonstrated in a rat model that obesity increases CB afferent output via the carotid sinus nerve resulting in increased SNS activity, which leads to hypertension and CSND abolishes this effect. We have previously reported that CSND decreases SNS activity in the liver in mice exposed to intermittent hypoxia (22). Here, we have demonstrated that CSND attenuates hypertension induced by obesity in mice reproducing the previous report in obese rats (19). In addition, CSND increased the pulse interval (decreased the heart rate) during the light phase, when animals are likely asleep (Fig. 3). Taken together our data suggest that CBs contribute to hyperglycemia in mouse DIO via SNS activation and shifting glycogen metabolism toward degradation and that this effect is abolished by CSND.

What mechanism of CB activation in obesity may lead to hyperglycemia? We have previously demonstrated that high levels of circulating leptin in DIO activate LEPR^b in the glomus cells leading to transcriptional induction of TRPM7, which accounts, at least in part, for DIO-induced hypertension (30, 32, 33) and obesity hypoventilation (62). Studies in obese rats showed that hyperinsulinemia increased phosphorylation of insulin receptor in CB (19) increasing SNS activity and systemic insulin resistance. Deficiency of the glucagon-like peptide 1 pathway in CB (63) can also lead to systemic SNS activation and insulin resistance. However, our experiments with transcriptional downregulation of Lepr, Trpm7 and Insr in CB showed that these molecular pathways do not contribute to hyperglycemia. Our findings are consistent with Dr. Paton’s hypothesis that there are several populations of type I glomus cells in CB, which differentially regulate responses to hypoxia and different metabolic stimuli (64).

Our study had several limitations. First, although we confirmed a previous report on the beneficial effect of carotid body denervation on glucose metabolism in rats, molecular pathways by which obesity acts in CB to induce hyperglycemia have not been identified. Second, the efficacy of CSND has not been confirmed by the hypoxic challenge in this study. However, we have previously shown that in our hands this technique is remarkably consistent and the hypoxic responses are significantly attenuated (22); in this study, we demonstrated the effect of CSND on blood pressure. Third, although G6P and PEPCK gene expression are suggestive of changes in gluconeogenesis, only functional studies would lead to a definitive conclusion. Fourth, SNS activity or glycogen degradation and biosynthesis were not directly measured. Fifth, the role of spectral analysis in assessment of the autonomic nervous system activity is controversial (65, 66). Our spectral analysis data have to be interpreted with caution and in combination with the data on CSND effect on blood pressure in the current report and liver sympathetic activity in the prior report (22). Sixth, female obese mice were not studied due to profound weight loss after surgical interventions.

In conclusion, we have shown that CB denervation markedly attenuates hypertension, fasting hyperglycemia, and glucose intolerance in DIO mice without any effect on insulin resistance.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

This work was supported by the NIH R01 HL133100 to Drs. Polotsky and Sham. Dr. Shin was supported by the American Heart Association Career Development Award 19CDA34700025.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.-K.S., J.S.-K.S., and V.Y.P. conceived and designed research; M.-K.S. and W.-Y.T. performed experiments; M.-K.S., W.-Y.T., M.R.A., and V.Y.P. analyzed data; M.-K.S., M.R.A., J.S.-K.S., and V.Y.P. interpreted results of experiments; M.-K.S., M.R.A., and V.Y.P. prepared figures; M.-K.S., J.S.-K.S., and V.Y.P. drafted manuscript; M.-K.S., W.-Y.T., M.R.A., J.S.-K.S., and V.Y.P. edited and revised manuscript; M.-K.S., W.-Y.T., M.R.A., J.S.-K.S., and V.Y.P. approved final version of manuscript.