Defective Natriuretic Peptide Receptor Signaling in Skeletal Muscle Links Obesity to Type 2 Diabetes

Obesity is a major risk factor of type 2 diabetes (T2D) and cardiovascular diseases (1,2). Although multiple hypotheses have been proposed, the link between obesity and the risk of T2D is still poorly understood. Over the last decade, several large cohort studies reported an inverse association between plasma natriuretic peptide (NP) levels and BMI (3,4), and the risk of T2D (5,6). Therefore, dysregulation of the NP system, referred to as the “NP handicap,” might be an important factor in the initiation and progression of metabolic dysfunction, making NPs potential candidates linking obesity and T2D (7–10).

NPs, including atrial NP (ANP) and brain NP (BNP), are mainly known as heart hormones secreted in response to cardiac overload and mechanical stretch in order to regulate blood volume and pressure (11,12). ANP and BNP classically bind to a biologically active receptor A (NP receptor A [NPRA]) that promotes cGMP signaling (13). They are also quickly cleared from the circulation and degraded through NP clearance receptor (NPRC). The NPRA-to-NPRC ratio therefore controls the biological activity of NP at the target tissue level (14).

Besides their well-documented role in the cardiovascular system, several studies revealed a metabolic role of NP (15,16). Pioneering studies demonstrated a potent lipolytic role of these peptides in human adipose tissue (17,18), and more recent studies indicated they may play a role in the “browning” of human white fat cells (19) as well as in favoring fat oxidative capacity in human skeletal muscle cells (20). The underlying mechanism involves activation of cGMP signaling, induction of PGC1α (peroxisome proliferator–activated receptor γ coactivator-1α), and enhancement of mitochondrial respiration. Together these studies argue for an important role of NP in the regulation of whole-body energy metabolism. The lipolytic effect of NP is absent in mice naturally expressing high levels of NPRC in adipose tissue (19,21). However mice overexpressing BNP are protected from diet-induced obesity and insulin resistance, which suggests that the protective effect of NP is achieved by targeting other metabolic tissues such as skeletal muscles (22). We therefore hypothesized that a downregulation of NPRA and/or an upregulation of NPRC in skeletal muscle could contribute to the “NP handicap” and provide a novel pathophysiological and mechanistic link between obesity and T2D.

In the current study, through a comprehensive set of experiments in humans, mouse models of obesity and T2D, and human primary skeletal muscle cells, we demonstrated a pathophysiological link between obesity-induced insulin resistance and T2D and defective skeletal muscle NPR signaling. In addition, increasing circulating BNP levels in diabetic and high-fat diet (HFD)–fed mice improved blood glucose control and insulin sensitivity. These effects were accompanied by improved muscle insulin signaling resulting from reduced lipotoxic lipid pressure and elevated lipid oxidative capacity.

Muscle biopsy samples from lean, obese with normal glucose tolerance, obese with impaired glucose tolerance (IGT), and obese subjects with T2D were obtained from three independent clinical studies. Study 1 included nine young lean and nine young obese subjects (Fig. 1A–D) (23). Study 2 included four middle-aged obese subjects with T2D and six with IGT at baseline and in response to 12 weeks of calorie restriction to induce weight loss and improve metabolic health (Fig. 1E and F) (24). Study 3 included 21 subjects with normal glucose tolerance but a wide range of body fat (Supplementary Fig. 1) (25). The clinical characteristics of the subjects are summarized in Supplementary Table 1. All volunteers gave written informed consent and the protocol was approved by an institutional review board. Studies were performed according to the latest version of the Declaration of Helsinki and the Current International Conference on Harmonization guidelines. Samples of vastus lateralis weighing 60–100 mg were obtained by muscle biopsy using the Bergström technique, blotted, cleaned, and snap frozen in liquid nitrogen (26). Insulin sensitivity was measured by hyperinsulinemic-euglycemic clamp after an overnight fast (27). An intravenous catheter was placed in an antecubital vein for infusion of glucose and insulin during the clamp. A second catheter was placed retrograde in a dorsal vein of the contralateral hand for blood withdrawal. The hand was placed in a plastic heated box at ∼60°C for arterialization of venous blood. Three blood samples were drawn before the initiation of insulin and glucose for the clamp and during the last 30 min of the clamp. A primed infusion of regular insulin (80 mU ⋅ min⁻¹ ⋅ m⁻²) was initiated and continued for 2 h. Plasma glucose was clamped at 90 mg/dL in all participants. Arterialized plasma glucose was measured at 5-min intervals and a variable infusion of exogenous glucose (20% solution) was given to maintain plasma glucose concentration. Plasma glucose was analyzed with a YSI 2300 STAT glucose analyzer (YSI Inc., Yellow Springs, OH), and plasma insulin was measured using an ultrasensitive ELISA kit (ALPCO Diagnostics, Salem, NH). Glucose disposal rate was adjusted by kilograms of fat-free mass. Body composition (considering a three-compartment model) was determined using a total-body dual-energy X-ray absorptiometer (DPX, Software 3.6; Lunar Radiation Corp., Madison, WI).

Five-week-old male diabetes-prone, obese db/db mice of the C57BL/KsJ-lept^db-lept^db strain with their nondiabetic lean littermate control db/+ were used. For HFD studies, we used regular C57BL/6J male mice (JANVIER LABS). The mice were housed in a pathogen-free barrier facility (12-h light/dark cycle) with ad libitum access to water and food. After weaning, db/db and db/+ mice were fed a normal chow diet (A04; SAFE Diets) for 4 weeks. C57BL/6J mice were fed for 16 weeks either a normal chow diet (10% energy as fat, D12450J; Research Diets, Inc., New Brunswick, NJ) or HFD containing 60% kcal from fat (D12492; Research Diets, Inc.). All experimental procedures were approved by a local institutional animal care and use committee and performed according to INSERM guidelines for the care and use of laboratory animals.

Mice were randomly assigned to receive a saline vehicle (NaCl 0.9%) and/or chronic rat/mouse BNP1-32 (B9901; Sigma-Aldrich) at a rate of 5 or 10 ng/kg/min. Treatments were chronically administered intraperitoneally with mini-osmotic pumps (model 1004; Alzet, Cupertino, CA) (28). Mini-pumps were placed after 12 weeks of HFD and treatment was administered for 4 weeks in C57BL/6J mice and at 6 weeks of age in db/db mice. Body weight was measured weekly and body composition was assessed by quantitative nuclear magnetic resonance imaging (EchoMRI 3-in-1 system; Echo Medical Systems).

Six hour–fasted mice were injected intraperitoneally with a bolus of d-glucose at 2 g/kg (Sigma-Aldrich, Saint-Quentin Fallavier, France) and insulin 0.5 units/kg (Insuman Rapid; Sanofi, Paris, France) for glucose and insulin tolerance tests (GTT and ITT), respectively (29). Blood glucose levels were monitored from the tip of the tail with a glucometer (Accucheck; Roche, Meylan, France) at 0, 15, 30, 45, 60, and 90 min after injection. Radiolabeled GTTs were performed as previously described (29).

After an overnight fast, mice were decapitated and blood was collected into tubes containing EDTA and protease inhibitors. Organs and tissues were rapidly excised and snap frozen in liquid nitrogen before being stored at −80°C. Blood glucose was assayed using the glucose oxidase technique (Biomérieux, Paris, France), and plasma insulin was measured using an ultrasensitive ELISA kit (ALPCO Diagnostics). Plasma BNP was measured using the RayBio BNP Enzyme Immunoassay Kit (RayBiotech, Inc., Norcross, GA). HbA_1c and fructosamines were determined using a PENTRA 400 multianalyzer.

Satellite cells from rectus abdominis biopsies of healthy subjects with normal glucose tolerance (age 34.3 ± 2.5 years, BMI 26.0 ± 1.4 kg/m², fasting glucose 5.0 ± 0.2 mmol/L) were grown in DMEM supplemented with 10% FBS and growth factors (human epidermal growth factor, BSA, dexamethasone, gentamycin, fungizone, and fetuin) as previously described (23,30). Myotubes were differentiated up to 5 days and were treated with 100 nmol/L human ANP (A1663; Sigma-Aldrich) or BNP (B5900; Sigma-Aldrich) every day for the last 3 days.

Pulse-chase experiments to determine lipolytic flux and oleate incorporation into total lipids, triacylglycerols (TAGs), and diacylglycerols (DAGs) by thin-layer chromatography were performed as previously described (31). Incorporation rates were normalized to total protein content in each well. Palmitate oxidation rates were measured as described previously (25).

TAG and DAG content were measured by gas chromatography and ceramide and sphingomyelin species by high-performance liquid chromatography–mass spectrometry after total lipid extraction as previously described for mouse and human muscle tissues (25,29).

Soleus and gastrocnemius skeletal muscles, white and brown adipose tissues, and myotubes were homogenized in a buffer containing 50 mmol/L HEPES, pH 7.4, 2 mmol/L EDTA, 150 mmol/L NaCl, 30 mmol/L NaPO₄, 10 mmol/L NaF, 1% Triton X-100, 10 µL/mL protease inhibitor (Sigma-Aldrich), 10 µL/mL phosphatase I inhibitor (Sigma-Aldrich), 10 µL/mL phosphatase II inhibitor (Sigma-Aldrich), and 1.5 mg/mL benzamidine HCl. Tissue homogenates were centrifuged for 25 min at 15,000g, and supernatants were stored at −80°C. Solubilized proteins (30–40 µg) were run on a 4–20% SDS-PAGE (Bio-Rad, Hercules, CA), transferred onto nitrocellulose membrane (Hybond ECL; Amersham Biosciences), and blotted with the following primary antibodies: NPRA (Abcam), NPRC (Sigma-Aldrich), Akt, phospho-Akt Ser473, p-Thr180/Tyr182-p38 MAPK, p38 MAPK, HSL, phospho-HSL Ser660, and phospho-HSL Ser565 (all Cell Signaling Technology, Beverly, MA). Subsequently, immunoreactive proteins were blotted with secondary horseradish peroxidase–coupled antibodies and revealed by enhanced chemiluminescence reagent (SuperSignal West Dura or SuperSignal West Femto; Thermo Scientific), visualized using the ChemiDoc MP Imaging System, and data analyzed using the Image Laboratory 4.1 version software (Bio-Rad). GAPDH (Cell Signaling Technology) and α-tubulin (Sigma-Aldrich) were used as internal controls for skeletal muscle and myotubes, and β-actin (Cell Signaling Technology) was used as internal control for adipose tissues.

Total from tissues and primary myotubes were processed for RNA extraction using the RNeasy RNA mini kit (Qiagen GmbH, Hilden, Germany). After reverse transcription of total RNA (1 µg), samples were analyzed on a StepOnePLus real-time PCR system (Applied Biosystems). All primers were obtained from Applied Biosystems and presented in Supplementary Table 5. All expression data were normalized by the 2^(ΔCt) method using 18S as internal control.

Statistical analyses were performed using GraphPad Prism 5.0 for Windows (GraphPad Software Inc.). Normal distribution and homogeneity of variance of the data were tested using Shapiro-Wilk and F tests, respectively. One-way ANOVA followed by Tukey post hoc tests and Student t tests were performed to determine differences between groups, interventions, and treatments. Two-way ANOVA followed by Bonferonni post hoc tests were applied when appropriate. Linear regression was performed after log transformation of nonparametric data. The false discovery rate for multiple testing was controlled by the Benjamini-Hochberg procedure with P_adj. values ≤0.05 as threshold. All values in figures and tables are presented as mean ± SEM. Statistical significance was set at P < 0.05.

Muscle NPRA protein expression was investigated in human vastus lateralis biopsies of healthy volunteers with varying degrees of body fat and insulin sensitivity. We observed that muscle NPRA protein was inversely related to body fat (Fig. 1A and Supplementary Fig. 1A), BMI, fasting insulin, and indices of insulin resistance (Supplementary Table 2). In addition, muscle NPRA correlated positively with whole-body insulin sensitivity measured by euglycemic-hyperinsulinemic clamp (Fig. 1B) and the insulin sensitivity index (Supplementary Fig. 1B) and negatively with total muscle saturated ceramide content (Fig. 1C). Importantly, muscle NPRA protein content was significantly reduced (−65%) in obese subjects when compared with age-matched lean subjects (Fig. 1D). Conversely, muscle NPRA protein was upregulated (1.8-fold) together with insulin sensitivity (+37%, 5.4 ± 0.6 vs. 7.4 ± 1.1 mg/min/kg for pre- and postcalorie restriction, respectively, P = 0.03) in obese subjects with IGT in response to diet-induced weight loss (Fig. 1E). Finally, muscle NPRC protein content was unchanged in obese versus lean individuals with normal glucose tolerance (0.41 ± 0.08 vs. 0.29 ± 0.07 arbitrary units [A.U.], not significant) but increased significantly in obese individuals with IGT and T2D (Fig. 1F). The ratio of NPRA to NPRC protein was significantly reduced in obese versus lean subjects (2.1 ± 0.3 vs. 3.6 ± 0.2 A.U., respectively, P = 0.0005) and increased in obese subjects in response to calorie restriction (0.38 ± 0.16 vs. 0.14 ± 0.04 A.U., respectively, P = 0.08). Together this suggests that skeletal muscle NPR expression relates to insulin sensitivity in humans and is altered in obesity and T2D.

Since both skeletal muscle and adipose tissue are known as key target tissues of NP, both in humans and mice, we further examined NPR expression in metabolic tissues of chow-fed versus HFD-fed mice. In line with human data, we found a significant downregulation of NPRA protein in skeletal muscle (Fig. 2A and D), as well as in white (Fig. 2B and D) and brown fat (Fig. 2C and D) of HFD-fed mice. No significant change in NPRC protein content was found in skeletal muscle and brown fat, whereas NPRC protein decreased significantly in white fat (0.48 ± 0.08 vs. 0.14 ± 0.04 A.U. for chow and HFD, respectively, P < 0.05). Plasma BNP levels were unchanged in HFD-fed mice compared with chow-fed mice (Fig. 2E). Collectively, as in humans, our data indicate a reduced NPRA expression in skeletal muscle of obese mice.

Since muscle NPRA is associated with insulin sensitivity in humans, we assessed the effect of acute and chronic BNP infusions on glucose tolerance and insulin sensitivity in chow-fed and HFD-fed mice. Acute intraperitoneal BNP injection did not affect fasting blood glucose levels over a time course of 30 min (Supplementary Fig. 2A) and had no effect on glucose excursion during an intraperitoneal GTT (Supplementary Fig. 2B). No effect of acute BNP injection was also seen on glucose disposal in skeletal muscle (Supplementary Fig. 2C). We further assessed the influence of acute NP treatment on basal and insulin-stimulated glucose uptake in human primary skeletal muscle cells. No effect of increasing doses of ANP and BNP on glucose uptake was observed (Supplementary Fig. 3). From these data we concluded that NPRA signaling does not acutely modulate glucose uptake in skeletal muscle.

Based on a previous study (32), we next infused BNP at a dose of 5 ng/kg/min, which raised plasma BNP levels by ∼40% (data not shown). BNP-treated mice had a similar body weight (Fig. 3A) and body composition (Fig. 3B) after saline and BNP treatment. Chronic BNP treatment significantly reduced fasting blood glucose levels in mice fed an HFD for 8 or 12 weeks (Fig. 3C). Lower blood glucose in the fasting state was also accompanied by improved glucose tolerance (Fig. 3D) despite no change in fasting and peak insulin at 15 min during the intraperitoneal GTT (Fig. 3E). In conclusion, whereas acute BNP treatment has no effect on insulin sensitivity, chronic BNP treatment improves glucose tolerance in HFD-fed mice.

We next examined NPR expression in metabolic tissues from leptin receptor–deficient mice (db/db) that become spontaneously obese and T2D by the age of 8 weeks. In line with data in human skeletal muscle and HFD-fed mice, NPRA protein was downregulated in white (Fig. 4B–D) and brown fat (Fig. 4C and D) of db/db mice compared with control db/+ mice. In agreement with data in individuals with IGT/T2D (Fig. 1E), we noted a remarkable upregulation of NPRC in skeletal muscle (Fig. 4A–E), as well as in white (Fig. 4B–E) and brown (Fig. 4C–E) fat of db/db mice. Overall the NPRA-to-NPRC protein ratio was markedly downregulated in muscle and fat of db/db mice (Fig. 4F) and was associated with dramatically lower levels of plasma BNP in db/db mice (−80%, P < 0.05) (Fig. 4G). This was also associated with a remarkable downregulation of p38 MAPK phosphorylation in skeletal muscle of db/db mice (−55%, P < 0.001) (Supplementary Fig. 4). Importantly, muscle NPRC was negatively correlated with plasma BNP levels (Supplementary Table 3). These changes in NPR signaling and plasma NP characterized an “NP handicap” of db/db mice. No association was found between white and brown fat NPRC protein and plasma BNP levels (data not shown). However, muscle NPRC was positively related to fasting blood glucose, insulin, and HbA_1c (Supplementary Table 4), again suggesting a link between defective skeletal muscle NPR signaling and impaired glucose control. Collectively, these data suggest that obesity and T2D are accompanied by profound changes in NPR expression and signaling in skeletal muscle, which may contribute to reduced plasma BNP levels.

We next studied the influence of chronic (4 weeks) BNP infusion on blood glucose control in db/db mice. BNP was infused at a dose of 10 ng/kg/min to induce a nearly twofold increase in plasma BNP levels with the goal of rescuing the “NP handicap.” Despite no change in body weight (Fig. 5A) and composition (Fig. 5B), BNP-treated db/db mice displayed significantly improved blood glucose control, with reduced fasting plasma glucose (−21%) (Fig. 5C) and HbA_1c (−17%) (Fig. 5D). This improved blood glucose control occurred in the absence of noticeable changes in fasting insulin (Fig. 5E). In addition, insulin tolerance (Fig. 5F) and insulin responsiveness (area above the curve during the ITT, +36%, P = 0.08) were improved in BNP-treated mice. In summary, chronic BNP treatment improves blood glucose control and peripheral insulin sensitivity in obese diabetic mice independently of changes in body weight, thus suggesting a direct effect of NP on metabolic organs.

We next studied the mechanism by which chronic BNP treatment improved blood glucose control and muscle insulin sensitivity in both HFD-fed and db/db mice. Insulin sensitivity is inhibited by the accumulation of toxic lipids such as DAGs and ceramides in skeletal muscle and liver (33,34). No significant change in total DAG and ceramides was found in the liver of BNP-treated db/db (Supplementary Fig. 5A and B) and HFD-fed mice (Supplementary Fig. 6A and B). No change as well in mRNA levels of genes involved in fat oxidation and glucose metabolism was observed after BNP infusion in the liver of db/db (Supplementary Fig. 5C) and HFD-fed mice (Supplementary Fig. 6C). Similarly, no change in the expression level of classical thermogenic genes in brown and white fat depots was observed in BNP-treated db/db (Supplementary Fig. 7) and HFD-fed mice (Supplementary Fig. 8). No change in Ucp1 mRNA levels was noted as well in inguinal white adipose tissue (data not shown). However, we observed a muscle-autonomous improvement of insulin-mediated Akt (46%, P = 0.02) and p38MAPK phosphorylation (278%, P = 0.06) (Fig. 6A), which was paralleled by a reduced content of total and main species (data not shown) of ceramides (−17%) (Fig. 6B) as well as total and main species (data not shown) of sphingomyelin (−19%) (Fig. 6C) in skeletal muscle of BNP-treated HFD-fed mice, as well as a reduced content of total ceramides in db/db mice (52.4 ± 4.4 vs. 40.0 ± 5.1 ng/µg protein for db/+ and db/db mice, respectively, P < 0.05). The content of total and subspecies of DAGs was also reduced in BNP-treated HFD-fed mice (ANOVA P < 0.05) (Fig. 6D). This lower lipotoxic pressure was paralleled by an upregulation of muscle palmitate oxidation rate (+46%) (Fig. 6E) and of PGC1α mRNA levels in HFD-fed mice (Fig. 6F) and in db/db mice (+32%, P = 0.08). Collectively, the data indicate that chronic BNP treatment improves insulin sensitivity in skeletal muscle by reducing lipotoxicity and upregulating fat oxidative capacity in a PGC1α-dependent manner in obese and diabetic mice.

We previously demonstrated a functional NPR signaling in human primary myotubes (20). Because NP are known to activate lipolysis in human adipocytes (35,36), we studied here the acute effect of NP treatment on lipid metabolism. Acute treatment of myotubes with BNP did not influence lipid storage, endogenous TAG-derived fatty acid (FA) release (i.e., lipolysis) (Supplementary Fig. 9A), and endogenous TAG-derived FA oxidation (Supplementary Fig. 9B). We further tested whether NP could activate one of the rate-limiting enzymes of lipolysis. Acute BNP treatment of human myotubes did not influence hormone-sensitive lipase phosphorylation either on the activating Ser660 residue (Supplementary Fig. 9C) or on the inhibitory Ser565 residue (Supplementary Fig. 9D). In contrast, chronic treatment with NP for 3 days robustly reduced total lipid accumulation and total TAG and DAG content (one-way ANOVA P < 0.001) (Fig. 7A–C). In line with ex vivo muscle data in mice (Fig. 6), reduced lipid accumulation was concomitant with an upregulation of palmitate oxidation rate (+27 and +19%, respectively, for ANP and BNP treatment) (Fig. 7D), and a significant induction of PGC1α gene expression (Fig. 7E), which was independent of PPARδ activation (Fig. 7F). No change in myogenic gene expression and differentiation of myoblasts into myotubes was observed in response to chronic NP treatment (data not shown). Based on the findings that muscle NPRA protein relates inversely to saturated ceramide content in human skeletal muscle (Fig. 1C) and that chronic BNP treatment reduces ceramide content in skeletal muscle of HFD-fed mice (Fig. 6B), we assessed the influence of chronic NP treatment on ceramide content in human primary myotubes. No significant effect of NP treatment on the content of total ceramides and various ceramide species (Supplementary Fig. 10) was noticed in the basal condition with FA-free BSA treatment. When myotubes were challenged overnight with 500 μmol/L of palmitate/BSA to induce ceramide production (2.7-fold, P = 0.001), we observed a significant decrease of ∼30% in total and various ceramide species analyzed in response to chronic ANP and BNP treatment (Fig. 7G). In summary, chronic NP treatment protects against lipotoxicity by upregulating lipid oxidative capacity in human primary myotubes.

Although longitudinal prospective studies indicated that high baseline levels of plasma NP confer a reduced risk of developing T2D (5,6), no study so far had demonstrated a mechanistic link between NP biological activity and T2D. We believe our data provide the first evidence that NPRA signaling in skeletal muscle is necessary for the maintenance of long-term insulin sensitivity by regulating lipid oxidative capacity and metabolism (Fig. 8). Our data show for the first time that muscle NPRA signaling is impaired in the context of obesity and glucose intolerance in humans and mice. We also provide evidence that upregulation of NPRC in muscle tissue can contribute to the “NP handicap” observed in T2D. Last but not least, increasing NP levels in obese and diabetic mice, with the goal of rescuing the “NP handicap” and so a normal NPRA signaling tissue response, markedly improves blood glucose control and insulin sensitivity in skeletal muscle.

We first observed a significant positive association between muscle NPRA protein and insulin sensitivity measured by clamp in humans, at a dose that mainly reflects skeletal muscle insulin sensitivity. This observation is consistent with the negative association that we found between muscle NPRA and body fat, and between muscle NPRA and muscle total saturated ceramide content, two factors negatively influencing whole-body and muscle insulin sensitivity (33,34). To our knowledge, this is the first study reporting an association between skeletal muscle NPRA signaling and insulin sensitivity. This indicates that besides plasma NP levels, NPR signaling in skeletal muscle may influence insulin sensitivity. Additionally, muscle NPRA protein was dramatically downregulated in obese individuals while increased in response to diet-induced weight loss and related improvement in insulin sensitivity. Although the biological factors modulating muscle NPRA protein content were not investigated in the current study, the data suggest that muscle NPRA behaves as a determinant of insulin sensitivity. Moreover, upregulation of muscle NPRC as glucose tolerance deteriorates in obese subjects with IGT and T2D can further repress biological activation of muscle NPRA and contribute to the “NP handicap” in the long-term. Considering that muscle mass represents up to 40% of total body weight, even a moderate increase in muscle NPRC expression could largely reduce plasma NP levels by an increased rate of clearance. Muscle NPRC might be induced by high blood insulin levels in obese subjects as glucose tolerance worsens independently of blood glucose concentrations as previously shown in adipose tissue (37). Although obese control and IGT/T2D were not age matched, increased expression of NPRC in skeletal muscle appeared independent of age since no correlation between age and muscle NPRC protein was found. Importantly, these findings in human muscle were largely replicated in obese diabetic mice. NPRC protein content was increased in skeletal muscle, white fat, and brown fat of obese diabetic mice, but only muscle NPRC protein negatively correlated with plasma BNP levels, reflecting that an increased plasma BNP clearance by the muscle can contribute to the “NP handicap” observed in these mice. Our data are in line with other studies demonstrating that elevated NPRC mRNA levels in white fat relate to metabolic dysfunction in mice and humans (22,38,39). Our data also provide a mechanistic understanding of the tight link observed between the NP handicap and insulin resistance independently of obesity in humans (40). The “NP handicap” concept is supported by the fact that the half-life of NP in the blood circulation is substantially increased in NPRC knockout mice and the biological activity of NP significantly increased in target tissues (41). Importantly, the altered NPRA-to-NPRC protein ratio in skeletal muscle was accompanied by a marked alteration of p38 MAPK phosphorylation in db/db vs. db/+ mice, thus indicating a potential signaling defect, considering that p38 MAPK is recognized as a canonical downstream molecular effector of the NPR signaling pathway (19).

Despite the observed link between muscle NPRA and insulin sensitivity, acute injection of BNP had no impact on fasting blood glucose, glucose tolerance, and muscle insulin sensitivity in mice. Furthermore, no acute effect of NP on glucose uptake was observed in human primary myotubes. These findings are in agreement with at least one human study reporting no acute effect of BNP on insulin sensitivity and insulin secretion (42). Altogether these data indicate that NP signaling does not acutely modulate skeletal muscle glucose uptake in vivo. We therefore performed chronic BNP infusion studies in HFD-fed and obese diabetic db/db mice to assess the long-term influence of BNP treatment on blood glucose control and insulin sensitivity. BNP was preferred for infusion studies as it has a higher half-life than ANP (14). Strikingly, chronic BNP infusion, at doses mimicking a physiological increase of the peptide and targeted to rescue the “BNP handicap” and/or a normal tissue NPRA signaling response, very significantly improved blood glucose control in both mouse models of obesity-induced glucose intolerance and T2D. We observed >20% reduction in fasting blood glucose levels as well as >15% decrease in HbA_1c, which is clinically meaningful and strongly reduces the risk of T2D complications (43). Reduced blood glucose levels during fasting and upon oral glucose challenge occurred in the absence of changes in blood insulin levels, indicating an improved metabolic clearance of glucose and insulin sensitivity. These findings are in agreement with other studies showing that increasing plasma BNP levels either pharmacologically (32) and/or genetically (22) improves glucose tolerance in obese mice. Preliminary evidence from our laboratory indicates that ANP-knockout mice are insulin resistant under normal chow diet compared with their wild-type littermates (data not shown), again arguing for a direct physiological link between NP signaling and insulin sensitivity.

Improved blood glucose control and insulin sensitivity were independent of significant changes in total DAGs and ceramides in liver, neither with noticeable changes in expression level of key metabolic genes in liver and thermogenic genes in white and brown fat of BNP-treated db/db and HFD-fed mice. However BNP-treated obese mice had an increased insulin-mediated Akt activation in skeletal muscle. Because Akt activation and phosphorylation are inhibited by lipotoxic lipids such as ceramides and DAGs (33,34), we measured ceramides and DAGs in skeletal muscle. In agreement with the negative correlation found in humans between muscle NPRA and ceramide content, we found a reduced level of total and main species of ceramides as well as main species of sphingomyelin in muscle of both HFD-fed and db/db mice chronically treated with BNP. Ceramides inhibit Akt activation and are produced de novo from saturated FAs and from sphingomyelin degradation (33). We also observed reduced muscle total DAG levels in BNP-treated mice. Interestingly, the reduced muscle lipotoxic lipid level was accompanied by a significant upregulation of muscle fat oxidative capacity and PGC1α gene expression. To demonstrate that elevated lipid oxidative capacity can reduce lipid accumulation, we chronically treated human primary myotubes with NP and showed increased palmitate oxidation rates and robustly reduced total lipid, TAG, and DAG accumulation. Chronic NP treatment also prevented palmitate-induced ceramide production in human primary myotubes. Although the precise mechanism was not investigated, it is likely that NP treatment reduces de novo ceramide production by increasing palmitate oxidation. We also show that NP-mediated elevated lipid oxidation involved the induction of PGC1α, which was independent of PPARδ activation. We and others previously described a cGMP-dependent induction of PGC1α gene expression by NP in white fat and skeletal muscle cells (19,20). PPARδ can be activated by lipid ligands derived from endogenous TAG lipolysis (31,44). In contrast to what has been shown in human fat cells (35,36), acute NP treatment of human primary myotubes did not influence the rate of lipolysis and TAG-derived FA oxidation or HSL phosphorylation at key regulatory sites.

Our findings provide the first evidence that NPRA signaling in skeletal muscle is pivotal for the maintenance of long-term insulin sensitivity by regulating lipid oxidative capacity through a PGC1α-dependent pathway. We also provide strong evidence that NPR signaling in skeletal muscle relates to insulin sensitivity and is disrupted in humans and mice with obesity and diabetes. Increasing plasma BNP levels in obese diabetic mice remarkably improves blood glucose control and could prove a novel therapeutic avenue to alleviate obesity-related insulin resistance.

Acknowledgments. The authors are very grateful to Max Lafontan (I2MC, Toulouse, France) for helpful discussions and critical reading of the manuscript. The authors thank Justine Bertrand-Michel and Aude Dupuy (Lipidomic Core Facility, INSERM, UMR1048 [part of Toulouse Metatoul Platform]) for lipidomic analysis, advice, and technical assistance, and the Anexplo Mouse Phenotyping and Animal Care Facility cores.

Funding. This study was supported by grants from the National Research Agency (ANR-12-JSV1-0010-01) and Société Francophone du Diabète (C.M.) and Fondecyt 11090007-Chile (J.E.G.). D.L. is a member of Institut Universitaire de France.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.C. and C.M. researched data and edited and wrote the manuscript. P.-M.B., I.K.V., C.L., K.L., M.-A.M., V.B., E.M., G.T., A.C.R., J.E.G., D.R.J., S.R.S., and D.L. researched data and edited the manuscript. C.M. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.