Introduction

The endocannabinoid system plays an important role in the control of food intake and metabolism and participates in the pathophysiology of obesity and type 2 diabetes [1, 2]. Two different endocannabinoid receptors, CB1R and CB2R [3, 4], have been identified. Peripheral CB1R is located in adipocytes, hepatocytes, skeletal muscle and pancreas [1, 5]. Overactivation of the endocannabinoid system has been reported in rodent models of obesity [1, 6, 7] and in obese and type 2 diabetic patients [8]. Moreover, overactivity of the endocannabinoid system in obese patients has been associated with enhanced visceral rather than subcutaneous fat mass [8–10]. These findings suggest that increased activity of the endocannabinoid system in visceral fat may be a key component in the development of insulin resistance. Furthermore, overactivation of CB1R in skeletal muscle promotes insulin resistance [11]; in this regard, pharmacological activation of CB1R activity diminishes mitogen-activated (MAP) kinase- and protein kinase B (PKB)-directed signalling [12]. In the liver, overactivation of CB1R enhances hepatic insulin resistance and lipogenesis [7], while the selective disruption of CB1R in hepatocytes protects mice fed a high-fat diet from liver steatosis and insulin resistance [13]. Moreover, CB1R hyperactivity in adipocytes facilitates fat accumulation and insulin resistance [14]. In cultured pancreatic islets exposure to endocannabinoids increases insulin release [15], suggesting that overactivity of the endocannabinoid system may contribute to the hyperinsulinaemia associated with insulin resistance.

CB2R is mainly located in the cells of the immune system and participates in the modulation of immune responses [16, 17]. In both mice and humans, CB2R is also present in the main peripheral organs responsible for the control of metabolism, including the liver [18], adipose tissue [5, 19–21], skeletal muscle [22] and pancreatic islets [23]. Obesity leads to increased expression of the cannabinoid receptor 2 gene (Cb2 [also known as Cnr2]) both in adipose tissue and liver in high fat fed and ob/ob mice [24]. However, this expression is predominantly located in the stromal vascular fraction of adipose tissue, while in liver CB2R is present only in the non-parenchymal cell fraction. CB2R deficiency decreases body weight gain during the feeding of a high-fat diet and prevents obesity-associated inflammation, insulin resistance and fatty liver [24]. Nevertheless, the role of long-term CB2R deficiency in the control of whole body metabolism has not been clarified. Here we used Cb2 knockout mice (Cb2 −/−) [25] to further evaluate the contribution of CB2R to the regulation of glucose metabolism in young and old mice.

Methods

Animals

Male knockout mice deficient in CB2R (Cb2 −/−) and wild-type littermates (Cb2 +/+) [25] on a C57Bl/6J background of different ages (2, 6 and 12 months) and maintained under a 12 h light–dark cycle (lights on at 08:00 hours) were used. Mice were fed ad libitum with a standard diet (Basal Purified Diet with 12% energy from fat; Test Diet, Richmond, IN, USA) or, when stated, a high-fat diet (Basal Purified Diet with 60% energy from fat; Test Diet). An additional group of C57Bl/6J wild-type mice were treated with the CB2R inverse agonist SR 144528 (Sanofi Recherche, Montpellier, France; for more information, see Electronic supplementary methods [ESM]). These mice received once daily i.p. injections, on consecutive days, for 28 days of either vehicle (1.8% DMSO, 0.0004% Tween 80 in distilled water) or vehicle plus antagonist (3 mg/kg body weight). Where stated, mice were fasted for 16 h. Body temperature was assessed anally with a thermometer in the morning either at room temperature or at 4°C (cold exposure). Animal procedures were conducted in accordance with the guidelines of the European Communities Directive 86/609/EEC regulating animal research and were approved by the local ethics committee.

Hormone and metabolite determination

Serum insulin and leptin were evaluated with ELISA kits (Crystal Chem, Chicago, IL, USA). Serum adiponectin levels were evaluated by radioimmunoassay (Linco Research, St Charles, MO, USA). Non-sterified fatty acids (NEFA) were measured by enzymatic colorimetric assay (NEFA C; Wako Chemicals, Neuss, Germany).

Food intake and leptin-induced anorexigenic effects

Mice were housed two per cage. Baseline food intake was measured for 4 weeks. Mice were habituated to the injection procedure (saline, i.p.) for 4 days. Leptin (Prepotech, London, UK) was administered twice a day at a dose of 0.5 μg/g for each injection (i.p.). The decrease in food intake was expressed as a percentage with respect to the baseline and as grams per mouse per day.

Lipopolysaccharide treatment

To study the inflammatory response after the acute administration of an external antigen, mice received an acute i.p. injection of lipopolysaccharide (LPS; 0.5 mg/kg in saline). Blood samples were obtained 3 and 12 h after treatment and TNF-α (3 h after injection) and IL-6 (12 h after injection) concentrations were measured by ELISA (Assay Designs, Ann Arbor, MI, USA).

Insulin and glucose tolerance tests

To test insulin tolerance, insulin was administered i.p. (0.75 U/kg; Humilin Regular; Eli Lilly, Indianapolis, IN, USA) to fed mice and tail blood was sampled 0, 15, 30, 45, 60 and 75 min later. For the glucose tolerance test, mice were food-deprived overnight (16 h) and then glucose (1 g/kg i.p.) was administered. Blood glucose was measured from the tail vein 0, 15, 30, 60, 90 and 120 min after glucose injection. Blood glucose was measured with a Glucometer Elite (Bayer Diagnostics, Tarrytown, NY, USA).

Insulin secretion from isolated islets

Pancreatic islets were isolated from Cb2 +/+ and Cb2 −/− mice by intraductal injection of collagenase P (1 mg/ml; Roche Molecular Chemicals, Mannheim, Germany) in Hank’ balanced salt solution (HBSS) for 10 min at 37°C. Digestion was stopped by washing three times with cold HBSS. Islets were picked by hand. Insulin secretion was measured during a 1 h incubation of islets at 37°C in 1 ml HBSS/0.5% BSA in the presence of 2.8 or 15 mmol/l glucose. After this time, insulin levels in media and insulin content in islets (after homogenisation in ethanol/HCl [10%] solution) were determined with an RIA kit (Linco Research, St Charles, MO, USA).

Gene expression analysis

Total RNA was obtained from skeletal muscle, white and brown adipose tissue and liver samples. To synthesise cDNA, 1 μg RNA was used (Omniscript kit; Qiagen, Hilden, Germany). Random primers (Invitrogen, Carlsbad, CA, USA) were used for the reaction in the presence of Protector RNase Inhibitor (Roche Molecular Biochemicals, Mannheim, Germany). RT-PCR was performed in a SmartCycler II (Cepheid, Maurens-Scopont, France) using the QuantiTect SYBR Green PCR Kit (Qiagen). Data were normalised to ribosomal protein S26 (RBS) gene values. Primer sequences are shown in ESM Table 1.

Immunohistochemical and morphometric analysis

White and brown adipose tissue, liver and pancreas were fixed for 24 h in formalin, embedded in paraffin and sectioned. To determine white and brown adipose tissue and liver morphology, sections were stained with haematoxylin/eosin. For detection of macrophages, white adipose tissue sections were incubated either with a rabbit anti-F4/80 antibody (Acris, Hidenhausen, Germany) or with a rabbit anti-macrophage galactose-specific lectin-2 (MAC-2) antibody (Cedarlane Laboratories, Burlington, ONT, Canada). For detection of insulin, pancreatic sections were incubated with a guinea pig anti-insulin antibody (Sigma, St Louis, MO, USA). Secondary antibodies were: peroxidase-conjugated goat anti-guinea pig (Dako, Glostrup, Denmark); biotinylated goat anti-rabbit (Pierce Biotechnology, Rockford, IL, USA); horseradish peroxidase-conjugated streptavidin (Molecular Probes, Leiden, the Netherlands). Morphometric analysis is described in the ESM.

PET imaging

2-[18F]Fluoro-2-deoxy-d-glucose (FDG) was used as a radiotracer and was synthesised by Barnatron (Barcelona, Spain) in accordance with the standard procedure [26]. Imaging was performed using a small animal PET scanner (rPET; Suinsa Medical Systems, Madrid, Spain). The methods used for the PET experiment are described in the ESM.

CB2R antagonist treatment in 3T3-L1 adipocytes and Oil Red O staining

Maintenance and differentiation of 3T3-L1 cells are described in the ESM. For treatment with the CB2R antagonist (SR 144528), stock solutions of drugs were prepared in DMSO at 10 mmol/l and stored at −20°C. The concentration of solvent in an assay never exceeded 0.1% (vol./vol.). This final concentration was without effect on assays. Cells were cultured in the presence of a final concentration of 100 nmol/l of SR 144528 or the same volume of DMSO. For Oil Red O staining, after being washed twice with PBS buffer, cells were fixed with 10% formaldehyde for 10 min. After two washes in PBS, cells were stained for 20 min with Oil Red O solution (0.18% in 60% isopropanol). Oil Red O was then removed and the cells were washed twice with PBS.

Hepatic and 3T3L1 cell triacylglycerol content

The triacylglycerol content of liver and 3T3L1 cells was determined by extracting total lipids from liver samples or cultured cells with chloroform–methanol (2:1 vol./vol.) as described in [27]. Triacylglycerols were quantified spectrophotometrically in the supernatant fractions using an enzymatic assay kit (GPO-PAP; Roche Diagnostics, Basel, Switzerland).

Western blot analysis

Tibialis muscles were homogenised in protein lysis buffer. Proteins (50 μg) were separated by 10% SDS-PAGE, transferred to nitrocellulose membranes and probed with primary antibodies against insulin receptor (IR), insulin receptor substrate 1 (IRS1), phosphorylated-Tyr, phosphoSer473-protein kinase B (phosphoSer473-AKT), AKT, phosphoSer9-GSK3B and GSK3 (Cell Signaling Technology, Danvers, MA, USA) and GLUT4 (Chemicon, Billerica, MA, USA) overnight at 4°C. Detection was performed using horseradish peroxidase-labelled anti-goat IgG or horseradish peroxidase-labelled anti-rabbit IgG (Dako) and ECL Plus Western Blotting Detection Reagent (Amersham, Arlington Heights, IL, USA).

Statistical analysis

All values are expressed as mean ± SEM. Differences between groups were compared by Student’s t test. A p value <0.05 was considered statistically significant.

Results

Cb2 −/− mice showed increased food intake, adipose tissue hypertrophy and obesity with age

To further determine the role of CB2R in regulating glucose metabolism and energy balance, Cb2 −/− mice were examined at different ages. Two-month-old Cb2 −/− male mice were normoglycaemic and showed body weight similar to that of age-matched wild-type (Cb2 +/+) mice (Fig. 1a and ESM Fig. 1a). However, at 6 months of age Cb2 −/− mice presented a modest but statistically significant greater body weight (about 6%) compared with wild-type mice, but with normal serum glucose, insulin and leptin levels (Fig. 1a and ESM Fig. 1b, c). These mice also showed a normal beta cell mass (ESM Fig. 1d, e). Moreover, a similar glucose-stimulated insulin release was revealed in islets isolated from Cb2 −/− and wild-type mice (ESM Fig. 1f). Twelve-month-old Cb2 −/− mice showed 25% greater body weight than wild-type mice (Fig. 1a). This increase was parallel to a two-fold increase in epididymal white adipose tissue (WAT) and interscapular brown adipose tissue (BAT) mass (Fig. 1b). Histological analysis of fat tissue from Cb2 −/− mice showed larger white adipocytes than those from wild-type mice (Fig. 1c). Moreover, in Cb2 −/− mice lipid droplets in brown adipocytes were larger than those in wild-type mice and, in some adipocytes, lipid deposition appeared unilocular (Fig. 1c). However, the expression of uncoupling protein 1 (Ucp1) and peroxisome proliferator-activated receptor gamma coactivator 1-alpha (Pgc1a) was not decreased (Fig. 1d) and that of adiponectin (Adipoq) was not increased (data not shown) in BAT from Cb2 −/− mice. These findings are not consistent with a switch of phenotype from a BAT- to a WAT-like phenotype in Cb2 −/− mice. Furthermore, Pgc1a, uncoupling protein 2 and 3 (Ucp2 and Ucp3) mRNA levels were normal in skeletal muscle (ESM Fig. 2a) and body temperature was not decreased either at room temperature or after 4 h of cold exposure (ESM Fig. 2b, c). These results suggest that adipose tissue fat accumulation in Cb2 −/− mice did not result from decreased thermogenesis and/or energy expenditure. The increase in WAT mass of 12-month-old Cb2 −/− mice was consistent with 65% higher circulating leptin levels than in age-matched wild-type mice (Fig. 1e). These Cb2 −/− mice showed a 40% increase in food intake (Fig. 1f). Moreover, 1 day after i.p. injection of leptin, Cb2 −/− mice displayed a smaller reduction in food intake than wild-type mice (Fig. 1f, g). These results indicate that the mice gradually developed obesity and hypertrophy of adipose tissue with age in association with increased food intake. CB1R activity in the hypothalamus has been related to the enhancement of food intake [6]. However, the expression of cannabinoid receptor 1 (Cnr1) was not modified in the cerebellum and hypothalamus of Cb2 −/− mice (Fig. 1h). Therefore, the enhanced food intake of Cb2 −/− mice was not due to compensatory changes in CB1R receptors in the brain areas regulating feeding behaviour. Parallel to increased fat mass, 12-month-old Cb2 −/− mice also showed an increase in the WAT expression of Cnr1 (Fig. 1h). Similarly, young mice also showed a moderate increase in Cnr1 expression (ESM Fig. 2d). Nevertheless, no difference in the expression of transient receptor potential vanilloid type 1 (Trpv1), which can also be activated by the endocannabinoid anandamide and can regulate lipogenesis [28], was revealed in adipose tissue of Cb2 −/− mice at 6 months of age (ESM Fig. 2e). Overactivity of CB1R has also been observed in the adipose tissue of obese rodents [14]. Our findings suggest that fat accumulation in WAT, at least in part due to enhanced food intake, probably contributed to increased Cnr1 expression in WAT. Consistent with the enhanced fat storage, old Cb2 −/− mice showed higher mRNA levels of the transcription factor sterol regulatory element transcription factor 1c (Srebf1c) in epididymal WAT (Fig. 1i), a key positive regulator of lipogenesis [29], which can also be induced by CB1R [7]. Obesity is often associated with increased circulating Non-sterified fatty acid (NEFA) levels [30]. However, although 12-month-old Cb2 −/− mice displayed increased body weight, they presented similar NEFA levels to wild-type mice (0.5 ± 0.06 vs 0.53 ± 0.05 mmol/l), suggesting that either lipolysis was not induced or that NEFA re-esterification was activated. Consistent with the latter, a 1.5-fold increase in expression in WAT of phosphoenolpyruvate carboxykinase (Pck1), a key regulatory enzyme of glyceroneogenesis, was observed in 12-month-old Cb2 −/− mice (Fig. 1i).

Fig. 1
figure 1

Enhanced body weight, food intake and hypertrophy of adipose tissue. a Body weight of 2-, 6- and 12-month-old Cb2 +/+ and Cb2 −/− mice. b Epididymal white adipose tissue (eWAT) and interscapular brown adipose tissue (iBAT) mass in Cb2 +/+ and Cb2 −/− mice. c Histological analysis of adipose tissue of Cb2 −/− mice, showing cell hypertrophy in WAT and increased lipid droplet size in BAT (×100). d Ucp1 and Pgc1a expression in BAT from 12-month-old Cb2 −/− and wild-type mice by quantitative PCR (AU, arbitrary units with respect to wild-type). e Serum leptin levels in Cb2 +/+ and Cb2 −/− mice evaluated by ELISA. f Food intake was measured in non-treated Cb2 +/+ and Cb2 −/− mice and in mice 1 day after leptin administration. Leptin was injected i.p. twice on a single day at a total dose of 1 μg/g. g Food intake. Results are percentages of baseline level and are mean ± SEM of 12 mice for each group. In f and g, +/+, Cb2 +/+; −/−, Cb2 −/−; *p < 0.05 Cb2 −/− vs wild-type in each treatment; † p < 0.05 after vs before leptin injection (one-way ANOVA). h RT-PCR quantification of Cnr1 mRNA in cerebellum (Cb), hypothalamus (HT) and eWAT of Cb2 +/+ and Cb2 −/− mice. i RT-PCR quantification of Srebf1c and Pck1 mRNA of Cb2 +/+ and Cb2 −/− mice in eWAT. j Immunohistochemical analysis of white adipose tissue in 12-month-old Cb2 +/+ and Cb2 −/− mice. Macrophages in adipose tissue of 12-month-old Cb2 +/+ and Cb2 −/− mice were detected after immunohistochemical analysis using F4/80-specific antibody (×200). Arrow indicates infiltrated macrophage. k Chronic inflammatory responses in Cb2 +/+ and Cb2 −/− mice. RT-PCR analysis revealed no differences in mRNA expression of Tnfa, Ccl2 and Il6 in eWAT tissue from Cb2 +/+ and Cb2 −/− mice. Data are mean ± SEM of 12 mice for each group. +/+, Cb2 +/+; −/−, Cb2 −/−. *p < 0.05 vs Cb2 +/+ (one-way ANOVA)

Full size image

Twelve-month-old obese Cb2 −/− mice did not develop adipose tissue inflammation

Obesity and type 2 diabetes are associated with chronic inflammation in WAT, characterised by increased production of cytokines, such as tumour necrosis factor-α (TNF-α), interleukin-6 (IL-6) and chemokine (C-C motif) ligand 2 (CCL2) [31]. In addition, CB2R, which is expressed in immune cells [32, 33], plays a major role in the modulation of obesity-associated inflammatory responses [24]. Despite adipose hypertrophy, no macrophage infiltration in WAT from obese 12-month-old Cb2 −/− mice was observed after immunohistochemical analysis with F4/80 (Fig. 1j). In agreement, quantitative PCR analysis of WAT revealed similar expression levels of Tnfa, Il6 and Ccl2 in Cb2 −/− obese mice compared with Cb2 +/+ mice (Fig. 1k). We also evaluated the inflammatory response of Cb2 −/− mice after the acute administration of an external antigen. Cb2 +/+ and Cb2 −/− mice were treated with an acute i.p. injection of LPS (0.5 mg/kg). Blood samples were obtained 3 and 12 h after treatment and TNF-α (3 h after injection) and IL-6 (12 h after injection) concentrations were measured by ELISA. However, no differences were observed between Cb2 +/+ and Cb2 −/− mice (TNF-α: wild type, 646 ± 101.85 vs Cb2 −/−, 1033 ± 483.3 pg/ml; IL-6: wild type, 6048 ± 1686 vs Cb2 −/−, 8822 ± 1269 pg/ml), indicating that Cb2 −/− mice were not protected from acute antigen-induced systemic inflammation.

Twelve-month-old Cb2 −/− obese mice revealed enhanced insulin sensitivity and increased glucose uptake

Twelve-month-old obese Cb2 −/− mice were normoglycaemic (Fig. 2a) and reached a greater hypoglycaemic response during an insulin tolerance test (0.75 IU/kg i.p.; Fig. 2b). In addition, increased skeletal muscle glucose uptake was observed when FDG-PET analysis was performed (Fig. 2c–e). PET imaging revealed slightly lower FDG absorption in basal conditions in the anterior paws of Cb2 −/− mice (Fig. 2c, d). However, Cb2 −/− mice displayed higher forepaw FDG absorption than wild-type mice after an insulin administration, consistent with increased glucose uptake in response to the hormone (Fig. 2c, e). Nevertheless, this was not associated with further increased phosphorylation levels of either protein kinase B (AKT; Fig. 2f, g) or its substrate glycogen synthase-kinase 3B (GSK3B; Fig. 2h, i) in skeletal muscle. Moreover, when phosphorylation of insulin receptor (IR) and insulin receptor substrate 1 (IRS1) was determined, no differences were observed between wild-type and Cb2 −/− mice (data not shown), indicating that the IR-AKT pathway was not further upregulated by insulin in Cb2 −/− mice. In addition, wild-type and Cb2 −/− mice showed similar GLUT4 protein levels (ESM Fig. 2f–h), indicating that enhanced skeletal muscle glucose uptake did not result from increased expression of this glucose transporter.

Fig. 2
figure 2

Enhanced insulin sensitivity and glucose uptake in skeletal muscle of Cb2 −/− mice. a Blood glucose levels in fed condition in 12-month-old Cb2 +/+ and Cb2 −/− mice. b Intraperitoneal insulin tolerance test. Insulin (0.75 IU/kg body weight) was injected intraperitoneally into fed Cb2 +/+ (white squares) and Cb2 −/− (black squares) mice and blood samples were taken from the tail vein at the times indicated. Results are percentages of blood glucose value at time 0 and are mean ± SEM of 12 animals for each group. c PET images showing forepaw FDG absorption in basal conditions and after insulin (0.5 IU/kg) administration in Cb2 +/+ and Cb2 −/− mice. d Forepaw FDG absorption (standardised uptake value, SUV) in basal condition after saline administration in Cb2 +/+ and Cb2 −/− mice. *p < 0.05 vs Cb2 +/+ (one-way ANOVA). e Insulin (0.5 IU/kg)-induced FDG absorption (SUV) in forepaws in Cb2 +/+ and Cb2 −/− mice. Results are percentage increases in FDG absorption when compared with saline (n = 4). +/+, Cb2 +/+; −/−, Cb2 −/−. *p < 0.05 vs Cb2 +/+ (one-way ANOVA). f, g Western blot analysis of phosphorylated (P)Ser473-AKT and total AKT using tibialis muscle lysates from wild-type (WT; Cb2 +/+) and Cb2 −/− mice: representative immunoblots (f) and densitometric analysis (g). Results are mean ± SEM and are percentage increases in PSer473-AKT after insulin injection when compared with the non-stimulated condition (n = 3). White columns, wild type; black columns, Cb2 −/−. h, i Western blot analysis of phosphorylated (P)Ser9-GSK3B and total GSK3B using tibialis muscle lysates from wild-type (WT) and Cb2 −/− mice: representative immunoblots (h) and densitometric analysis (i). Results are mean ± SEM and are expressed as percentage increases in PSer9-GSK3B after insulin injection when compared with the non-stimulated condition (n = 3). White columns, wild type; black columns, Cb2 −/−. j Quantitative PCR analysis of Cnr1 mRNA expression in skeletal muscle of Cb2 +/+ and Cb2 −/− animals (n = 4). Results are mean ± SEM. *p < 0.05 vs Cb2 +/+

Full size image

Therefore, in spite of weight gain with age, Cb2 −/− mice showed improved insulin sensitivity, which agreed with the lack of adipose tissue inflammation and the normal circulating NEFA levels. The lack of CB2R in Cb2 −/− mice also led to a 60% decrease in the expression of Cnr1 in skeletal muscle (Fig. 2j), which could contribute to enhanced insulin sensitivity [34]. This decrease in Cnr1 expression was also observed in 6-month-old non-obese Cb2 −/− mice (ESM Fig. 2i).

CB2R antagonism resulted in improved insulin sensitivity

To further investigate CB2R-mediated responses in vivo, 2-month-old mice were treated daily with i.p. injections of vehicle (control) or a selective CB2 antagonist (SR 144528) for 4 weeks. After this period, antagonist-treated mice showed a slight, non-statistically significant increase in body weight gain compared with controls (Fig. 3a), but normal food intake (Fig. 3b) and hypothalamic expression of Cnr1 (Fig. 3c). Moreover, these mice displayed no difference in epididymal WAT mass (430 ± 25.3 vs 460 ± 19.8 mg) and in Cnr1 expression in WAT (Fig. 3c). Similarly, antagonist treatment of 3T3L1 adipocytes did not increase triacylglycerol content (ESM Fig. 3a, b), consistent with a lack of increased adipogenesis. In addition, antagonist-treated mice showed normal circulating levels of NEFA (0.43 ± 0.04 vs 0.46 ± 0.05 mmol/l) and triacylglycerol (71.3 ± 5.6 vs 58.3 ± 5.9 mg/dl) and normal glucose tolerance (Fig. 3d). After 4 weeks of treatment, antagonist-injected mice showed improved insulin sensitivity compared with vehicle-injected mice (Fig. 3f). No changes in Cnr1 expression in skeletal muscle were observed (Fig. 3c), suggesting that decreased CB2R activation in skeletal muscle could be responsible for the increased insulin sensitivity in antagonist-treated wild-type mice.

Fig. 3
figure 3

Metabolic effects of CB2R antagonism in mice. a After 4 weeks of treatment with daily i.p. injections of vehicle (control; white columns) or the CB2 antagonist SR 144528 (grey columns), no difference was observed in body weight gain. b Food intake in antagonist-treated mice presented as g of food/day. c Antagonist treatment did not alter Cnr1 expression in epididymal WAT (eWAT), hypothalamus (HT) or skeletal muscle (Sk muscle), as shown by RT-PCR analysis. d Intraperitoneal glucose tolerance test. Awake overnight-fasted control (white circles) and antagonist-treated (grey circles) mice were given an i.p. injection of glucose (1 g/kg). Blood samples were taken at the times indicated from the tail vein and glucose levels were measured. e Intraperitoneal insulin tolerance test. Insulin (0.75 IU/kg body weight) was injected intraperitoneally into fed antagonist-treated mice. Blood samples were taken at the times indicated from the tail vein of the same animal. Results are percentages of blood glucose values at time 0 for each mouse and are mean ± SEM for eight animals in each group. White circles, control mice; grey circles, antagonist-treated mice. *p < 0.05 vs wild-type

Full size image

Cb2 −/− mice fed a high-fat diet showed reduced body weight gain and increased insulin sensitivity

To examine whether the lack of CB2R could prevent diet-induced insulin resistance, 2-month-old Cb2 −/− and Cb2 +/+ mice were fed a high-fat diet for 8 weeks. Both groups developed obesity, but Cb2 −/− mice on the high-fat diet showed reduced body weight gain (about 30% increase) compared with Cb2 +/+ mice on the high-fat diet (about 50% increase; Fig. 4a). Consistent with this, high fat fed Cb2 −/− mice displayed decreased fat mass (Fig. 4b) and adipocyte mean area compared with high fat fed wild-type mice as measured by morphometric analysis of epididymal WAT sections (Fig. 4c). Although Cb2 −/− mice fed a standard diet showed greater food intake than wild-type mice, no differences were observed between these groups of mice when fed a high-fat diet (Fig. 4d). Consistent with the similar food intake, Cnr1 expression in the hypothalamus was not different between Cb2 +/+ and Cb2 −/− mice on the high-fat diet (ESM Fig. 4a). Whole-body insulin sensitivity was also measured in an insulin tolerance test. High fat fed Cb2 −/− mice displayed similar insulin sensitivity and similar insulinaemia compared with Cb2 +/+ mice fed a standard diet (Fig. 4e, f). Moreover, high fat fed Cb2 −/− mice showed normal expression of Cnr1 in skeletal muscle (ESM Fig. 4a). WAT from high fat fed Cb2 −/− mice showed reduced macrophage infiltration and expression of the cytokine genes Ccl2 and Tnfa compared with wild-type mice (Fig. 5a–d). In addition, serum levels of glucose, triacylglycerol, leptin and adiponectin remained similar in the two groups of mice (ESM Fig. 4b–e). Hepatic triacylglycerol storage was also markedly reduced in high fat fed Cb2 −/− mice compared with high fat fed wild-type mice (Fig. 5e, f). Thus, the reduction in body weight gain in high fat fed Cb2 −/− compared with Cb2 +/+ animals was parallel to improved adipose tissue inflammatory state, normal insulin sensitivity and hepatic lipid content. These results indicate that the lack of a CB2R-mediated response reduced body weight gain during the feeding of a hypercaloric diet and protected mice from diet-induced insulin resistance.

Fig. 4
figure 4

Cb2 −/− mice fed a high-fat diet showed decreased body weight gain and normal insulin sensitivity. Two-month-old male wild-type and Cb2 −/− mice were fed a high-fat diet for 8 weeks. a, Body weight gain. White squares, wild type, standard diet; black squares, Cb2 −/−, standard diet; white triangles, wild type, high-fat diet; black triangles, Cb2 −/−, high-fat diet. (b) Epididymal fat pad weight (eWAT). In b–d and f, white columns, wild type; black columns, Cb2 −/− (also applies to c, d and f) Chow, standard diet; HFD, high-fat diet. c Frequency distribution of adipocyte cell area from eWAT of wild-type and Cb2 −/− mice fed a high-fat diet. Results are mean ± SEM of >3,500 adipocytes per group of mice. d Food intake in wild-type and Cb2 −/− mice fed chow or a high-fat diet. e Insulin tolerance test. Insulin (0.75 IU/kg body weight) was injected i.p. into wild-type and Cb2 −/− mice. Blood samples were taken from the tail vein at the times indicated and glucose levels were determined. White triangles, wild type, standard diet; white squares, wild type, high-fat diet; black squares, Cb2 −/−, high-fat diet. f Serum insulin was measured in fat-fed wild-type and Cb2 −/− mice. Results are mean ± SEM of ten mice for each group. *p < 0.05 Cb2 −/− vs wild-type, † < 0.05 chow fed vs high fat fed wild-type, ‡ p < 0.05 high-fat fed wild-type vs high fat fed Cb2 −/−

Full size image
Fig. 5
figure 5

Two-month-old Cb2 −/− mice presented decreased adipose tissue inflammation compared with wild-type (WT) mice when both groups were fed a high-fat diet for 8 weeks. a Representative MAC-2 immunostaining of epididymal WAT sections from wild-type and Cb2 −/− mice. MAC-2-stained cells (brown) surrounding adipocytes form crown-like structures (CLS). b The number of MAC-2-positive CLS per field was counted manually and the number of CLS per 1,000 adipocytes was used as a measure of adipose tissue macrophage content. Three mice per group were analysed. White columns, wild type; black columns, Cb2 −/− (also applies to c, d and f). *p < 0.05 vs wild-type mice on the high-fat diet; † p < 0.1 HFD Cb2 −/− vs HFD wild type. c, d Gene expression in adipose tissue from Cb2 −/− mice was analysed by measuring relative mRNA levels using quantitative PCR for the proinflammatory cytokines Ccl2 (c) and Tnfa (d). *p < 0.05. e, f Hepatic steatosis was prevented in high fat fed Cb2 −/− mice. e Representative sections stained with haematoxylin/eosin of liver from wild-type and Cb2 −/− mice fed either a standard (chow) or a high-fat diet for 8 weeks. f Triacylglycerol content in the liver of wild-type and Cb2 −/− mice. Data are mean ± SEM of eight mice for each group. *p < 0.05

Full size image

Discussion

Cb2 −/− mice fed a standard diet displayed increased body weight gain and adipose tissue hypertrophy with age (older than 6 months), which was associated with increased food intake and hyperleptinaemia. In contrast, 2-month-old high fat fed Cb2 −/− mice displayed reduced body weight gain and decreased fat pad mass compared with high fat fed wild-type mice. However, during high-fat feeding, no differences in food intake were observed, as described previously [24]. These results may suggest that while differences in food intake could explain long-term enhanced body weight during standard diet feeding, the lack of CB2R protected Cb2 −/− mice from body weight gain during high fat feeding in young animals. Both CB1R and CB2R are present in white adipose tissue [35] and upregulation of CB1R activity in adipose tissue has been associated with increased lipogenesis and decreased lipolysis in obese humans and rodents [9, 10, 36]. Two- to 4-month-old Cb2 −/− mice presented normal Cnr1 expression in adipose tissue in both the standard and the high-fat feeding condition, suggesting that improved body weight in high fat fed animals was probably not due to decreased CB1R activity. However, adipose tissue of old Cb2 −/− mice (older than 6 months) showed a gradual overexpression of Cnr1 that increased in parallel to epididymal fat pad weight, suggesting that enhanced CB1R levels may be a consequence of the progressive adipose tissue hypertrophy. Treatment with a CB2R antagonist neither altered Cnr1 expression in white adipose tissue nor increased fat pad mass, and these mice showed normal food intake. Moreover, the CB2R antagonist did not alter lipogenesis in isolated adipocytes, whereas the CB1R antagonist did [8]. This suggests that adipose tissue hypertrophy in Cb2 −/− mice resulted from a long-term adaptive response to the constitutive lack of CB2R, which involved increased food intake. In addition, although standard-fed Cb2 −/− mice showed greater food intake, Cnr1 expression was not increased in the hypothalamus, suggesting that lack of CB2R may directly induce signals that increase appetite. This phenotype was opposite to that of cb1 −/− mice, which are leaner than cb1 +/+ mice and show reduced food intake [37].

In spite of the development of obesity, consistent with the results in high fat fed mice, insulin sensitivity was enhanced in 12-month-old Cb2 −/− mice, as revealed by an improved insulin tolerance test. Moreover, PET imaging revealed an enhancement of skeletal muscle insulin-mediated glucose uptake in Cb2 −/− mice. Similarly, CB1R antagonists facilitate skeletal muscle insulin-induced glucose uptake [38] and NEFA oxidation [34] and CB1R agonists can decrease insulin-induced responses in muscle cells [11]. Furthermore, a CB1R receptor inverse agonist increases glucose uptake both in muscle isolated from ob/ob mice and in cultured L6 muscle cells [38, 39]. The CB1R inhibition in muscle cells increases the signalling through the phosphatidyl inositol 3 (PI3) kinase pathway [38]. However, in skeletal muscle from insulin-treated Cb2 −/− mice no further increase in either AKT or GSK3B phosphorylation levels were observed, suggesting that downstream signals may be involved in increasing insulin sensitivity. Thus, CB1R/CB2R activation may regulate the Ras/Raf/ERK cascade downstream of AKT while having little or no effect on components of the PI3 kinase/AKT pathway. Since skeletal muscle contains both CB1R and CB2R [22], improvement in insulin sensitivity could be related to the absence of CB2R and/or the downregulation of CB1R in skeletal muscle of Cb2 −/− mice. Moreover, mice treated with CB2R antagonist showed increased insulin sensitivity with no changes in body weight or Cnr1 expression. Thus, increased insulin sensitivity resulted from lack of CB2R in Cb2 −/− mice rather than decreased Cnr1 expression. Similarly, when fed a high-fat diet, Cb2 −/− mice did not develop insulin resistance. This may have resulted from the fact that, although high fat fed Cb2 −/− mice became obese, they showed lower body weight gain compared with high fat fed Cb2 +/+ mice. Indeed, high fat fed Cb2 −/− mice gained about 30% in body weight, whereas Cb2 +/+ mice gained around 50%. High fat fed Cb2 −/− mice displayed similar insulin sensitivity compared with standard-fed lean mice. Similarly, Deveaux et al. [24] have shown a reduction in body weight gain and improved insulin sensitivity in high-fat-fed Cb2 −/− mice. In addition, high-fat-fed Cb2 −/− mice showed reduced hepatic steatosis, which would also have contributed to improved insulin sensitivity. The absence of inflammatory responses in WAT could also be involved in the improvement of peripheral insulin sensitivity in Cb2 −/− mice [24]. This agrees with the enhanced insulin sensitivity that was previously reported in obese mice with reduced WAT inflammation, such as Tnf-α (also known as Tnf)−/− [40], CCL2 or Mcp-1 −/− [41] and osteopontin−/− mice [42]. The reduced adipose tissue macrophage infiltration and cytokine expression in high fat fed Cb2 −/− mice was not associated with any alteration of Cnr1 expression, suggesting that the lack of CB2R is responsible for the reduced inflammation. The increased glucose uptake described in Cb2 −/− mice, together with the similar food intake compared with wild-type mice, may result in decreased body weight gain during high fat feeding.

In summary, this study reveals that genetic ablation of CB2R led to improved insulin sensitivity during both age-related and diet-induced insulin resistance. Thus, we demonstrated that CB2R plays an important role in glucose metabolism by modulating skeletal muscle insulin sensitivity and the inflammatory response in the adipose tissue. We provide new data confirming the key role of CB2R in the pathophysiology of obesity and type 2 diabetes, which may open new therapeutic approaches to counteract insulin resistance by using CB2R ligands.