Original Article Effects of Rimonabant (SR141716) on Fasting-Induced Hypothalamic-Pituitary-Adrenal Axis and Neuronal Activation in Lean and Obese Zucker Rats Christian Doyon, Raphae¨l G. Denis, Elena-Dana Baraboi, Pierre Samson, Jose´e Lalonde, Yves Deshaies, and Denis Richard

The effects of the cannabinoid-1 receptor (CB1) antagonist rimonabant on energy metabolism and fasting-induced hy- pothalamic-pituitary-adrenal (HPA) axis and neuronal ac- besity results from a prolonged energy imbal- tivation were investigated. Lean and obese Zucker rats ance during which intake exceeds expenditure. were treated orally with a daily dose of 10 mg/kg rimon- The difficulty to lose excess weight is tightly abant for 14 days. A comprehensive energy balance profile Olinked to the ability of the systems regulating based on whole-carcass analyses further demonstrated the energy balance to defend body weight. The complexity potential of CB1 antagonists for decreasing energy gain and redundancy within these systems, which involve an through reducing food intake and potentially increasing intricate network of peripheral signals and neuronal cir- brown adipose tissue thermogenesis. Rimonabant also re- cuits, constitute obstacles to finding potential targets for duced plasma glucose, insulin, and homeostasis model as- antiobesity treatments. Currently, one of the most prom- sessment of insulin resistance, which further confirms the ising targets for the pharmacological treatment of obesity ability of CB1 antagonists to improve insulin sensitivity. To test the hypothesis that rimonabant attenuates the effect is the cannabinoid-1 receptor (CB1). Rimonabant of fasting on HPA axis activation in the obese Zucker (SR141716), the first selective CB1 antagonist (1), acts as model, rats were either ad libitum–fed or food-deprived for a potent antiobesity agent when administered to diet- 8 h. Contrary to expectation, rimonabant increased basal induced obese mice (2). Rimonabant is presently in phase circulating corticosterone levels and enhanced the HPA III clinical trials for the treatment of obesity. The recently axis response to food deprivation in obese rats. Rimon- published results from clinical trials, known as Rimon- abant also exacerbated the neuronal activation seen in the abant in Obesity–Europe (3), Rimonabant in Obesity– arcuate nucleus (ARC) after short-term deprivation. In Lipids (4), and Rimonabant in Obesity–North America (5), conclusion, the present study demonstrates that CB1 indicate that rimonabant not only reduces body weight but blockade does not prevent the hypersensitivity to food also improves cardiovascular risk factors associated with deprivation occurring at the level of HPA axis and ARC obesity. activation in the obese Zucker rats. This, however, does not The precise mechanism responsible for the antiobesity prevent CB1 antagonism from exerting beneficial effects on effect of rimonabant remains unknown. It has been sug- energy and glucose metabolism. Diabetes 55:3403–3410, gested that the hypophagic effect of CB1 antagonists 2006 results from an attenuation of feeding-related reward processes (6,7) that could be under the modulation of hypothalamic centers regulating energy balance. Injection of the endocannabinoid anandamide in the ventromedial hypothalamic nucleus, an area rich in CB1 mRNA (8), increases food intake, and this effect is blocked by rimon- abant (9). Also, CB1 mRNA is co-expressed with hypotha- From the Merck Frosst/CIHR Research Chair in Obesity and Centre de lamic neuropeptides involved in the modulation of food recherche de l’Hoˆ pital Laval, Hoˆ pital Laval, Que´bec, Canada. intake, including the anorectic peptide corticotropin- C.D. is currently affiliated with the Department of Cellular and Molecular Medicine, Faculty of Medicine, University of Ottawa, Ottawa, Ontario, Canada. releasing factor (CRF) (10). The presence of CB1 in Address correspondence and reprint requests to Denis Richard, Direction CRF-positive neurons of the paraventricular hypothalamic de la recherche, Hoˆ pital Laval, 2725 chemin Sainte-Foy, Que´bec, Que´bec, nucleus (PVN) also suggests a possible connection be- Canada, G1V 4G5. E-mail: [email protected]. Received for publication 13 April 2006 and accepted in revised form 28 tween the cannabinoid system and the hypothalamic- August 2006. pituitary-adrenal (HPA) axis, the activity of which has a AgRP, agouti-related peptide; ARC, arcuate nucleus; BAT, brown adipose major impact on energy balance regulation (11). However, tissue; CB1, cannabinoid-1 receptor; CRF, corticotropin-releasing factor; HOMA-IR, homeostasis model assessment of insulin resistance; HPA, hypo- the relationship between the cannabinoid system and the thalamic-pituitary-adrenal; MCH, melanin-concentrating hormone; MCR4, HPA axis remains unclear because both cannabinoid ago- melanocortin receptor-4; NEFA, nonesterified fatty acid; NPY, neuropeptide Y; nists and antagonists have been reported to activate the POMC, proopiomelanocortin; PVN, paraventricular hypothalamic nucleus; HPA axis (12–16). SON, supraoptic nucleus; UCP1, uncoupling protein-1. DOI: 10.2337/db06-0504 The potential interaction between the cannabinoid sys- © 2006 by the American Diabetes Association. tem and the HPA axis was examined by subjecting Zucker The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance rats to a mild physiological stress (8-h daytime food with 18 U.S.C. Section 1734 solely to indicate this fact. deprivation) after 2 weeks of treatment with rimonabant.

DIABETES, VOL. 55, DECEMBER 2006 3403 RIMONABANT AND HPA/NEURONAL ACTIVATION

The obese Zucker rat is not only hypercorticosteronemic They were then transferred to a paraformaldehyde-borax solution containing (17,18) and hypersensitive to stress (17,19), but it also 10% sucrose at least 12 h before cutting 30-␮m-thick coronal sections using a exhibits hyperactivity of the cannabinoid system. Defec- sliding microtome (HM 440E; Microm, Walldorf, Germany). Brain sections taken from the olfactory bulb to the brainstem were allocated to six sequential tive leptin signaling in obese Zucker rats is associated with sets in 24-well tissue culture plates containing a cold sterile cryoprotecting elevated hypothalamic levels of the endocannabinoid solution (50 mmol/l sodium phosphate buffer, 30% ethylene glycol, and 20% 2-arachidonoylglycerol (20). Based on these observations, glycerol) and stored at Ϫ30°C. we hypothesized that rimonabant attenuates HPA axis In situ hybridization histochemistry. In situ hybridization histochemistry hyperactivity, which is associated with the development of was used to determine c-fos, CRF, CRF1 receptor, agouti-related peptide obesity in Zucker fa/fa rats (21). The hypothesis was (AgRP), neuropeptide Y (NPY), proopiomelanocortin (POMC), and melanin- concentrating hormone (MCH) mRNA levels on tissue sections taken from the addressed in a study further aimed at investigating the hypothalamus. Our method was largely adapted from that of Simmons et al. effects of rimonabant on hypothalamic neuronal activation (26). Briefly, brain sections (one of every six sections) were rinsed in sterile induced by short-term food deprivation and on hypotha- 0.05 mol/l potassium PBS treated with diethyl pyrocarbonate, mounted onto lamic mRNA levels of neuropeptides known for their poly-L-lysine coated slides, and dehydrated in 100% ethanol. The sections were involvement in the regulation of energy balance and HPA successively fixed for 20 min in paraformaldehyde (4%), digested for 30 min at axis activity. 37°C with proteinase K (10 ␮g/ml in 100 mmol/l Tris-HCl containing 50 mmol/l EDTA, pH 8.0), acetylated with acetic anhydride (0.25% in 0.1 mol/l trietho- lamine, pH 8.0), and dehydrated through graded concentrations (50, 70, 95, and 100%) of ethanol. After drying for at least 2 h, 100 ␮l hybridization RESEARCH DESIGN AND METHODS mixture, which contained an antisense 35S-labeled cRNA probe (107 cpm/ml), Lean (Fa/?) and obese (fa/fa) male Zucker rats, aged 7–8 weeks, were was spotted on each slide. Slides were sealed under a coverslip and incubated purchased from Charles River Laboratories (St. Constant, Que´bec, Canada). overnight at 60°C. The next day, coverslips were removed, and slides were All rats were cared for and handled according to the Canadian Guide for the rinsed four times with 4ϫ sodium chloride–sodium citrate (0.6 mol/l NaCl and Care and Use of Laboratory Animals, and the protocol was approved by the 60 mmol/l trisodium citrate buffer, pH 7.0), digested for 30 min at 37°C with Universite´ Laval Animal Care Committee. The animals were housed individ- RNAse-A (20 ␮g/ml in 10 mmol/l Tris-500 mmol/l NaCl containing 1 mmol/l ually in wire-bottom cages, allowed unrestricted access to water, and, unless EDTA), washed in descending concentrations of sodium chloride–sodium specified, fed ad libitum with a ground stock diet (Charles River Rodent Diet citrate (2ϫ, 10 min; 1ϫ, 5 min; 0.5ϫ, 5 min; and 0.1ϫ, 30 min at 60°C), and 5075; Ralston Products, Woodstock, Ontario, Canada). They were subjected to dehydrated through graded concentrations of ethanol. After2hofdrying, a 12-h-dark/12-h-light cycle (lights on between 0700 and 1900) and kept under slides were exposed on an X-ray film (Eastman Kodak, Rochester, NY) for ambient temperature (23 Ϯ 1°C). Rats were separated in groups of equal initial 20 h. Slides were defatted in toluene, dipped in NTB2 nuclear emulsion average weights within each genotype the day preceding the treatment period. (Eastman Kodak), and exposed 3 (NPY and POMC), 4 (CRF), 5 (c-fos, AgRP

Treated rats received a daily oral administration of 10 mg/kg rimonabant and MCH), or 22 (CRF1 receptor) days before being developed in D19 (SR141716; Sanofi-Aventis, Paris, France) for 14 days at 0830, except on the developer (Eastman Kodak) for 3.5 min at 14°C and fixed in rapid fixer last day where rimonabant was administered 6 h before death. This dose was (Eastman Kodak) for 5 min. Finally, tissues were rinsed in running water for previously shown to reduce body weight gain in the obese Zucker rat (22,23). 1–2 h, counterstained with thionin (0.25%), dehydrated through graded Each day, a solution containing rimonabant (2 mg/ml) and Tween-80 (2 ␮l/ml) concentrations of ethanol, cleared in toluene, and coverslipped with DPX was administered at a dose of 5 ml/kg body wt. Rats were weighed and food mounting medium (BDH; VWR, Mississauga, Ontario, Canada). Processed intake was measured daily throughout the experiment. Rats were killed slides were examined by darkfield microscopy using an Olympus BX51 between 1400 and 1700 in either an ad libitum–fed state or after an 8-h food microscope (Olympus America, Melville, NY). Images were acquired with an deprivation. Lean and obese rats were respectively anesthetized with 2 and 4 Evolution QEi camera and analyzed with ImagePro plus v5.0.1.11 (Media- ml of a mixture containing ketamine (20 mg/ml) and xylazine (2.5 mg/ml). Cybernetics, Silver Spring, MD). The system was calibrated for each set of Blood was collected by intracardial puncture into syringes coated with 0.5 analyses to prevent saturation of the integrated signal. Mean pixel densities mol/l EDTA (Sigma-Aldrich, St. Louis, MO), and rats were perfused intracar- were obtained by taking measurements from both hemispheres of one to four dially for 2 min with ice-cold isotonic saline. Brain and interscapular brown brain sections and subtracting background readings taken from areas imme- adipose tissue (BAT) were sampled immediately after the perfusion. BAT was diately surrounding the region analyzed. flash frozen in liquid nitrogen and stored at Ϫ86°C. Antisense 35S-labeled riboprobes. Complementary RNA probes were gen- Body gains in energy, fat, and protein. Carcasses were autoclaved at 125 erated from rat cDNA fragments for c-fos (Dr. I. Verma, The Salk Institute, La kPa for 15 min, homogenized in two volumes of water (w/v), and freeze-dried. Jolla, CA; GenBank accession no. V00727), CRF (Dr. K. Mayo, Northwestern

Carcass energy content was determined by adiabatic bomb calorimetry, University, Evanston, IL), CRF1 receptor (Dr. M.H. Perrin and Dr. W.W. Vale, whereas carcass protein was determined using a FP-2000 Nitrogen Analyzer The Clayton Foundation, La Jolla, CA; GenBank accession no. L24096), POMC (Leco, St. Joseph, MI) with 250–300 mg dehydrated carcasses. Nonprotein (Dr. B.T. Bloomquist, Bayer, West Haven, CT), and MCH (Dr. R. Thompson matter energy was obtained by subtracting protein energy from total carcass and Dr. S.J. Watson, University of Michigan, Ann Arbor, MI), the XhoI-XbaI energy. Values of 23.5 and 39.2 kJ/g were used for the calculation of the energy genomic fragment containing exon 2 of rat NPY (Dr. D.S. Larhammar, Uppsala content of protein and fat, respectively (24). Initial energy, fat, and protein University, Uppsala, Sweden), and a murine AgRP cDNA (Dr. M. Graham, contents of the carcasses were estimated from the live body weight of lean Amgen, Thousand Oaks, CA). Radiolabeled antisense riboprobes were syn- and obese rats with reference to a baseline group of rats (six per phenotype) thesized by incubating 250 ng linearized plasmid at 37°C for 60 min in the

killed at the beginning of the experimental period. Such estimates allow gains presence of 10 mmol/l NaCl, 10 mmol/l dithiothreitol, 6 mmol/l MgCl2,40 in energy, fat, and protein to be determined for the treatment period. Rats in mmol/l Tris (pH 7.9), 0.2 mmol/l ATP/GTP/CTP, [␣-35S]UTP, 40 units RNase

the initial group were identical in every respect (e.g., age and sex) to those of inhibitor (Roche Diagnostics), and 20 units T7 (c-fos, CRF1 receptor, NPY, and the experimental groups. Food efficiency was expressed as the ratio of energy AgRP), SP6 (CRF and MCH), or T3 (POMC) RNA polymerase (Promega, gain to digestible energy intake multiplied by 100. Madison, WI). The DNA templates were treated with 100 ␮l DNAse solution (1 ␮ ␮ ␮ Plasma determinations. Blood was harvested by cardiac puncture and l DNAse, 5 l 5 mg/ml tRNA, and 94 l 10 mmol/l Tris/10 mmol/l MgCl2). centrifuged (1,500g, 15 min at 4°C), and plasma was stored at Ϫ20°C until later Riboprobes were purified on RNeasy Mini Spin Columns (Qiagen, Mississauga, biochemical measurements. Plasma glucose concentrations were determined Ontario, Canada). using an automated glucose analyzer YSI 2,300 Stat Plus (YSI, Yellow Springs, Real-time quantitative RT-PCR. Total RNA was isolated from 60–90 mg OH). Commercially available radioimmunoassay kits were used to determine BAT using the RNeasy Lipid Tissue mini kit (Qiagen). On-column DNA plasma levels of corticosterone (MP Biomedicals, Toronto, ON), insulin, digestion was performed using the RNase-free DNase Set (Qiagen). First- leptin, and adiponectin (Linco Research, St. Charles, MO), whereas enzymatic strand cDNA was synthesized from 1 ␮g total RNA with Expand Reverse kits were used for triglycerides (Roche Diagnostics, Laval, Que´bec, Canada) Transcriptase and oligo(dT) (Roche Diagnostics) and diluted 1:25 with diethyl and nonesterified fatty acids (NEFAs; Wako Diagnostics, Richmond, VA). The pyrocarbonate–treated water. Rat uncoupling protein-1 (UCP1) amplicons homeostasis model assessment of insulin resistance (HOMA-IR) was calcu- were generated using the sense primer 5Ј-TGGTGAGTTCGACAACTTCC-3Ј lated using plasma glucose and insulin levels of food-deprived rats as and the antisense primer 5Ј-GTGGGCTGCCCAATGAATAC-3Ј (GenBank ac- previously described (25). cession no. NM_012682). Rat L27 amplicons were generated using the sense Brain preparation. Brains were essentially prepared as previously described primer 5Ј-CTGCTCGCTGTCGAAATG-3Ј and the antisense primer 5Ј-CCTTGC (18). After their removal, brains were fixed into a 4% paraformaldehyde-3.8% GTTTCAGTGCTG-3Ј (GenBank accession no. NM_022514). Amplification was borax solution for at least 7 days with frequent replacement of the solution. carried out using Platinum Taq polymerase (Invitrogen), CYBR Green I

3404 DIABETES, VOL. 55, DECEMBER 2006 C. DOYON AND ASSOCIATES

FIG. 1. Daily food intake (A and B) and cumulative weight gain (C and D) in lean (Fa/?) (A and C) and genetically obese (fa/fa)(B and D) Zucker rats treated with a daily oral administration of 10 mg/kg rimonabant. *Significant effect of rimonabant treatment as assessed by repeated- .group/14–13 ؍ measures ANOVA. P < 0.05, n (Cedarlane Laboratories, Hornby, Ontario, Canada), and a Rotor Gene 3000 290.2 Ϯ 4.6 g; rimonabant 290.0 Ϯ 5.1 g) or obese (control (Corbett Research, Sydney, Australia) with the following program: 2 min 463.0 Ϯ 9.7 g; rimonabant 468.0 Ϯ 5.3 g) rats. After an denaturing at 94°C, then 40 cycles of denaturation at 94°C for 20 s, annealing at 64°C for 20 s, extension at 72°C for 20 s, and emission measurement at 85°C initial weight loss, cumulative body weight gain of rimon- for 15 s. After the last cycle, the temperature was gradually increased from 72 abant-treated rats remained below that of controls to 99°C for the determination of a melting curve. The PCR reaction contained throughout the whole experiment (Fig. 1C and D). Treat- 5 ␮l diluted cDNA in a 20-␮l PCR reaction. Results were analyzed using ment with rimonabant reduced total energy gain, as as- Rotor-gene v6.0 software (Corbett Research). sessed by whole-carcass analyses, and this was mostly Statistics. Results are presented as means Ϯ 1 SE. Statistical differences in daily food intake and cumulative weight gain between control and rimon- accompanied by a reduction in fat gain in obese rats, abant-treated rats were determined within each genotype using a crossed- whereas lean rats showed a significant reduction in pro- nested design with repeated measurements. Cumulative weight gain data were tein gain (Table 1). Rimonabant reduced food efficiency, log-transformed, and multivariate normality was verified with Mardia’s test. and this reduction reached statistical significance in lean Statistical differences within each genotype were determined by Student’s t rats (Table 1). Rimonabant did not alter apparent energy test or two-way ANOVA. Data for corticosterone, insulin, and c-fos mRNA were log-transformed, whereas a square root transformation was used for expenditure (Table 1). BAT weight was increased in obese NEFA. Tukey’s multiple comparison tests followed two-way ANOVAs with rats but reduced in lean rats by rimonabant (Table 1). BAT significant interaction effect. Results were considered significant with P values UCP1 mRNA levels were significantly increased in rimon- Ͻ0.05. Statistical analyses were performed using SAS v9.1.3 software package abant-treated obese rats (Table 1). (SAS Institute, Cary, NC) or SigmaStat v2.0 software (SPSS, Chicago, IL). Metabolic plasma variables. Treatment with rimon- abant reduced circulating triglyceride levels in both lean RESULTS and obese rats (Table 2). NEFA levels were increased after Body weight, food intake, and energy balance. Treat- food deprivation but not significantly modified by rimon- ment with the CB1 antagonist rimonabant led to a transient abant. Circulating glucose levels of rimonabant-treated food intake reduction that remained significant for 10 days obese rats were reduced to values similar to those of lean in obese rats compared with 3 days in lean rats (Fig. 1A and untreated food-deprived obese rats. Food deprivation and B). Initial weight did not differ within lean (control reduced plasma insulin levels in both lean and obese rats,

DIABETES, VOL. 55, DECEMBER 2006 3405 RIMONABANT AND HPA/NEURONAL ACTIVATION

TABLE 1 Energy balance and body composition in lean (Fa/?) and genetically obese (fa/fa) Zucker rats after 14 days of daily oral administration of 10 mg/kg rimonabant Lean Obese Control Rimonabant Control Rimonabant DEI (kJ) 4,935 Ϯ 133 4,477 Ϯ 113* 8,356 Ϯ 318 6,901 Ϯ 314* Energy gain (kJ) 431 Ϯ 111 118 Ϯ 60* 2,223 Ϯ 371 1,034 Ϯ 225* Energy expenditure (kJ) 4,504 Ϯ 115 4,359 Ϯ 93 6,133 Ϯ 507 5,866 Ϯ 351 Food efficiency (%) 8.6 Ϯ 2.0 2.5 Ϯ 1.3* 26.8 Ϯ 4.5 15.1 Ϯ 3.4 Fat gain (g) 5.3 Ϯ 2.9 -1.3 Ϯ 1.6* 53.2 Ϯ 7.5 24.6 Ϯ 6.5* Protein gain (g) 9.54 Ϯ 0.42 7.15 Ϯ 0.74* 6.49 Ϯ 3.79 3.49 Ϯ 2.19 BAT weight (g) 0.57 Ϯ 0.02 0.48 Ϯ 0.02* 1.59 Ϯ 0.08 1.91 Ϯ 0.11* BAT UCP1 mRNA 13.0 Ϯ 1.6 15.2 Ϯ 1.4 5.8 Ϯ 1.0 11.3 Ϯ 2.5* Data are presented for ad libitum–fed rats (n ϭ 6–7/group) except for BAT weight and UCP1 mRNA (n ϭ 13–14/group). Digestible energy intake (DEI) represents 95.5% of total energy intake. BAT UCP1 mRNA levels are expressed as a ratio of L27 mRNA levels. *Student’s t test, P Ͻ 0.05. and these levels were further reduced by rimonabant. rats. Rimonabant also tended to increase POMC mRNA HOMA-IR was reduced in lean and obese rats treated with levels in the ARC of obese rats (P ϭ 0.05; Fig. 3D). c-fos rimonabant. Rimonabant significantly reduced circulating mRNA levels were elevated in the supraoptic nucleus leptin levels only in lean rats. Circulating adiponectin (SON) of ad libitum–fed rats, and these levels were levels were increased in rimonabant-treated obese rats. enhanced by rimonabant in obese rats (Fig. 3G). The HPA axis. Food deprivation increased plasma corti- costerone levels in obese rats (Fig. 2A). Treatment with rimonabant increased basal corticosterone levels and po- DISCUSSION tentiated the fasting-induced elevation in plasma cortico- The present results confirm the ability of CB1 antagonists sterone in obese rats. In addition, treatment with to reduce weight gain in both lean and obese rats rimonabant increased c-fos mRNA levels in the parvocel- (22,23,27). Here, we extend previous findings with a com- lular division of the PVN (Fig. 2B). Rimonabant did not plete energy balance profile based on carcass analyses. We alter CRF or CRF1 receptor mRNA levels in the PVN (Fig. show that the reduction in weight gain was largely ac- 2C and D). counted for by a reduction in fat gain, which is consonant Neuronal activation and neuropeptides in the arcu- with previous studies (2,28). The fact that rimonabant did ate nucleus, lateral hypothalamic area, and supraop- not significantly decrease energy expenditure, which tic nucleus. Food deprivation increased NPY (Fig. 3A and would be expected with a reduction in both total energy B) and c-fos (Fig. 3E and F) mRNA levels in the arcuate intake and energy gain, suggests that CB1 blockade ex- nucleus (ARC) of obese rats and MCH mRNA in the lateral erted a stimulating effect on thermogenesis. This is further hypothalamic area of lean rats (Fig. 3H). Treatment with supported by the observation that, in obese rats, rimon- rimonabant enhanced ARC NPY (Fig. 3A and B), AgRP abant increased BAT UCP1 mRNA levels, a widely used (Fig. 3C), and c-fos (Fig. 3E and F) mRNA levels in obese marker of thermogenic capacity. A recent study by Jbilo et

TABLE 2 Plasma glucose, lipids, and hormones in ad libitum–fed and 8-h–food-deprived lean (Fa/?) and genetically obese (fa/fa) Zucker rats after 14 days of daily oral administration of 10 mg/kg rimonabant Control Rimonabant AL FD AL FD Lean Glucose (mmol/l) 10.7 Ϯ 0.4 11.5 Ϯ 0.4 11.6 Ϯ 0.5 10.4 Ϯ 0.4 Insulin (nmol/l) 0.25 Ϯ 0.04 0.13 Ϯ 0.02* 0.15 Ϯ 0.02† 0.07 Ϯ 0.01*† HOMA-IR NA 11.6 Ϯ 2.2 NA 5.5 Ϯ 0.5† Triglycerides (mmol/l) 2.6 Ϯ 0.3 1.5 Ϯ 0.2* 1.2 Ϯ 0.2† 0.8 Ϯ 0.1*† NEFA (mmol/l) 0.08 Ϯ 0.01 0.13 Ϯ 0.01* 0.08 Ϯ 0.01 0.15 Ϯ 0.02* Leptin (ng/ml) 4.8 Ϯ 0.5 4.2 Ϯ 0.4 3.7 Ϯ 0.3† 3.4 Ϯ 0.1† Adiponectin (␮g/ml) 2.46 Ϯ 0.22 2.18 Ϯ 0.16 2.50 Ϯ 0.16 2.91 Ϯ 0.29 Obese Glucose (mmol/l) 24.0 Ϯ 1.3 11.8 Ϯ 0.9‡ 14.8 Ϯ 1.8§ 12.0 Ϯ 0.6 Insulin (nmol/l) 2.45 Ϯ 0.39 1.54 Ϯ 0.23* 1.48 Ϯ 0.28† 0.76 Ϯ 0.10*† HOMA-IR NA 132.7 Ϯ 19.7 NA 67.6 Ϯ 9.9† Triglycerides (mmol/l) 8.9 Ϯ 1.2 9.0 Ϯ 0.9 5.0 Ϯ 0.6§ 2.9 Ϯ 0.7‡§ NEFA (mmol/l) 0.21 Ϯ 0.04 0.57 Ϯ 0.06* 0.16 Ϯ 0.04 0.79 Ϯ 0.08* Leptin (ng/ml) 37.8 Ϯ 3.7 39.4 Ϯ 3.9 43.1 Ϯ 3.7 35.8 Ϯ 3.1 Adiponectin (␮g/ml) 3.76 Ϯ 0.15 4.71 Ϯ 0.27‡ 5.48 Ϯ 0.40§ 5.01 Ϯ 0.33 AL, ad libitum fed; FD, 8-h food deprived. †Significant main effect of rimonabant treatment and *significant main effect of food deprivation as assessed by two-way ANOVA or t test (HOMA-IR). When significant, only interaction results are shown: ‡significant effect of food deprivation within a specific treatment and §significant effect of rimonabant within a specific feeding status. P Ͻ 0.05, n ϭ 6–7/group.

3406 DIABETES, VOL. 55, DECEMBER 2006 C. DOYON AND ASSOCIATES

FIG. 2. Plasma corticosterone (A) and mean pixel density of the hybridization signal for c-fos (B), CRF (C), and CRF1 receptor (CRF-R1) (D) mRNA in the parvocellular division of the PVN of ad libitum–fed (AL) and 8-h–food-deprived (FD) lean (Fa/?) and genetically obese (fa/fa) Zucker rats after 14 days of daily oral administration of 10 mg/kg rimonabant. *Significant main effect of rimonabant treatment and †significant main .group/7–6 ؍ effect of food deprivation as assessed by two-way ANOVA. P < 0.05, n al. (29) showed that rimonabant activates several BAT hanced the HPA axis response to a daytime food depriva- genes involved in the regulation of mitochondrial activity. tion in obese Zucker rats. This enhanced activity of the Although increased BAT thermogenic activity likely con- HPA axis was associated with increased c-fos mRNA levels tributed to rimonabant-induced reduction in energy gain, in the parvocellular division of the PVN. Thus, the re- such reduction was obviously also attributable to the sponse of obese Zucker rats was similar to that of lean ICR transient reduction in energy intake. The present study mice, in which rimonabant was recently shown to poten- also further highlights the ability of CB1 antagonists to tiate the corticosterone response to restraint stress (16). improve glucose metabolism. In obese Zucker rats, which One possible mechanism for rimonabant-induced activa- are hyperinsulinemic (30) and glucose intolerant (31), tion of the HPA axis is through increased CRF release. rimonabant restored normal glycemia and reduced plasma Because CB1 is primarily a presynaptic receptor that levels of insulin and triglycerides. It is noteworthy that modulates neurotransmitter release (34), the presence of Ͼ these marked changes were observed at a time when CB1 mRNA in 50% of CRF neurons in the PVN suggests rimonabant no longer affected food intake. Our results are that cannabinoids may directly influence CRF release (10). consistent with other studies that showed improved Finally, the observation of increased HPA axis activity is plasma profiles in Zucker rats (22), diet-induced obese inconsistent with a reduced energy gain. Corticosteroids mice (2,32), and obese humans (3,4). have been shown to promote energy deposition (35–37) In the present investigation, we tested the possibility and to inhibit BAT thermogenesis (37–39). This raises the that rimonabant may attenuate the food deprivation-in- possibility that rimonabant may impede the downstream duced stress response in obese Zucker rats. This hypoth- effects of an activated HPA axis by obstructing the effects esis was based on the observation that the hyperactivity of of chronically elevated plasma corticosterone on energy the endocannabinoid system (20), which is thought to deposition and thermogenic activity. The enhancing effect increase the incentive value of food (7,33), may in part of rimonabant on central insulin sensitivity would repre- explain the hypersensitivity of obese Zucker rats to food sent one mechanism through which rimonabant may coun- deprivation (19). Our results demonstrated that rimon- teract the detrimental metabolic effects of an increased abant did not block the effect of food deprivation on HPA HPA axis activity. axis activation in obese Zucker rats. In fact, rimonabant Obese Zucker rats exhibited a predictable increase in increased basal circulating corticosterone levels and en- NPY mRNA levels in response to food deprivation, and this

DIABETES, VOL. 55, DECEMBER 2006 3407 RIMONABANT AND HPA/NEURONAL ACTIVATION

FIG. 3. Mean pixel density of the hybridization signal for NPY (A), AgRP (C), and POMC (D) and c-fos mRNA in the ARC (E), c-fos mRNA in the SON (G), and MCH mRNA in the lateral hypothalamic area (LH) (H) of ad libitum–fed (AL) and 8-h food deprived (FD) lean (Fa/?) and genetically obese (fa/fa) Zucker rats after 14 days of daily oral administration of 10 mg/kg rimonabant. Representative hybridization signals for NPY (B) and c-fos mRNA (F) in the ARC of obese rats. *Significant main effect of rimonabant treatment and †significant main effect of food deprivation as assessed by two-way .group/7–6 ؍ ANOVA. When significant, only interaction results are shown (bars), as assessed by Tukey’s multiple comparison tests. P < 0.05, n

3408 DIABETES, VOL. 55, DECEMBER 2006 C. DOYON AND ASSOCIATES was accompanied by an increase in c-fos mRNA levels in neuronal activation in the ARC is consistent with in- the ARC. The NPY response was not observed in lean rats, creased NPY and AgRP mRNA levels. Rimonabant also which are less sensitive to food deprivation than obese increased c-fos mRNA levels in the SON of ad libitum–fed rats (40). Nonetheless, lean rats showed a fasting-induced obese rats. The magnocellular neurosecretory cells of the elevation in MCH mRNA levels in the lateral hypothalamus SON secrete either arginine vasopressin or oxytocin (46), and reduction in c-fos mRNA levels in the SON. As and these cells were shown to respond to feeding (47) and previously observed (23,27,28), the hypophagic effect of hypertonicity (48). Verty et al. (49) demonstrated an rimonabant was transient despite continuous administra- interaction between CB1 and oxytocin receptors; rimon- tion of the CB1 antagonist. In obese rats, rimonabant abant attenuated the food and water intake induced by increased NPY and AgRP mRNA levels in both fed and tocinoic acid, an oxytocin receptor antagonist. Therefore, fasted rats. These increases, which were paralleled by rimonabant induced neuronal activation in brain struc- induction of the c-fos gene, likely occurred as a mecha- tures that are consistent with its effect on neuropeptides nism to compensate for the reducing effects of rimonabant involved in the regulation of energy balance. on energy stores. However, the inductions, which were In conclusion, our results further demonstrate the po- obtained at a time when food intake had normalized in tential of CB1 antagonists to improve energy metabolism both lean and obese rats, suggest that CB1 antagonists may in obese rats, in particular by reducing energy gain and block a possible NPY/AgRP-stimulated compensatory normalizing several plasma variables related to the meta- overfeeding. In agreement with this statement, previous bolic syndrome and diabetes. The study also demonstrates studies showed that interrupting the administration of a that CB1 blockade does not prevent the hypersensitivity to CB1 antagonist leads to compensatory overfeeding (27,28). food deprivation occurring at the level of HPA axis and Also, rimonabant was shown to attenuate NPY-induced ARC activation in the obese Zucker rats. This, however, sucrose drinking in rats (41) and overeating in satiated does not prevent CB1 antagonism from exerting beneficial mice (42). Together, these studies suggest that blockade of effects on energy and glucose metabolism. NPY action may have contributed to the rimonabant- induced hypophagia and/or inhibition of a compensatory overfeeding. ACKNOWLEDGMENTS A recent study (43) showed that AM251, an analog of C.D. has received a postdoctoral fellowship from Fonds rimonabant, reduces NPY release and attenuates cannabi- que´be´cois de la recherche sur la nature et les technologies. noid-induced NPY release from hypothalamic explants, sug- D.R. has received a grant from the Canadian Institutes of gesting that the effect of CB1 antagonists on food intake Health Research. may result from a reduction in NPY synaptic tone. How- We thank Sanofi-Aventis for their supply of rimonabant. ever, it is unlikely that cannabinoids act directly on We also thank Serge Simard for statistical advice, and Julie NPY/AgRP neurons. A colocalization study failed to detect Plamondon, Marie-Noe¨lle Cyr, Se´bastien Poulin, and the presence CB1 mRNA in NPY-positive neurons of the Hakima Zekki for technical assistance. ARC (10). Also, Di Marzo et al. (20) showed that the hypophagic effect of rimonabant remains present in NPY- deficient mice, suggesting that NPY and cannabinoids are REFERENCES independent orexigenic systems. Rimonabant-induced im- 1. Rinaldi-Carmona M, Barth F, He´aulme M, Shire D, Calandra B, Congy C, provement in insulin sensitivity represents a potential Martinez S, Maruani J, Ne´liat G, Caput D, Ferrara P, Soubrie´ P, Brelie`re JC, Le Fur G: SR141716A, a potent and selective antagonist of the brain mechanism whereby the drug could blunt the activity of cannabinoid receptor. FEBS Lett 350:240–244, 1994 the NPY/AgRP neurons. It is clear from our own HOMA-IR 2. Ravinet Trillou C, Arnone M, Delgorge C, Gonalons N, Keane P, Maffrand results and from those of other studies (3) that rimonabant has JP, Soubrie´ P: Anti-obesity effect of SR141716, a CB1 receptor antagonist, a strong positive action on insulin sensitivity. Whether this in diet-induced obese mice. Am J Physiol Regul Integr Comp Physiol action of rimonabant is also apparent at the brain level is a likely 284:R345–R353, 2003 possibility that certainly warrants further investigation. 3. Van Gaal LF, Rissanen AM, Scheen AJ, Ziegler O, Rossner S: Effects of the cannabinoid-1 receptor blocker rimonabant on weight reduction and Rimonabant tended to increase POMC mRNA, the pre- cardiovascular risk factors in overweight patients: 1-year experience from cursor of the anorexigenic neuropeptide ␣-melanocyte– the RIO-Europe study. Lancet 365:1389–1397, 2005 stimulating hormone but did not affect the MCH system in 4. Despre´s JP, Golay A, Sjo¨ stro¨ m L: Effects of rimonabant on metabolic risk obese rats. Involvement of these systems during the initial factors in overweight patients with dyslipidemia. N Engl J Med 353:2121– period of exposure, when food intake is reduced, cannot 2134, 2005 be excluded. A recent study demonstrated that rimon- 5. Pi-Sunyer FX, Aronne LJ, Heshmati HM, Devin J, Rosenstock J: Effect of rimonabant, a cannabinoid-1 receptor blocker, on weight and cardiometa- abant attenuated the feeding response induced by the bolic risk factors in overweight or obese patients: RIO-North America: a melanocortin receptor-4 (MCR4) antagonist JKC-363 (44). randomized controlled trial. JAMA 295:761–775, 2006 This result indicates that rimonabant may act downstream 6. Thornton-Jones ZD, Vickers SP, Clifton PG: The cannabinoid CB1 receptor from MCR4 and that it may block the orexigenic action of antagonist SR141716A reduces appetitive and consummatory responses MCR4 antagonists, including AgRP. Of relevance to the for food. Psychopharmacology 179:452–460, 2005 present study is the fact that a stimulated melanocortin 7. Kirkham TC, Williams CM: Endogenous cannabinoids and appetite. Nutr Res Rev 14:65–86, 2001 system, without concomitant activation of the MCH sys- 8. Cota D, Marsicano G, Lutz B, Vicennati V, Stalla GK, Pasquali R, Pagotto U: tem, constitutes a condition promoting UCP1-induced Endogenous cannabinoid system as a modulator of food intake. Int J Obes thermogenesis (45). Relat Metab Disord 27:289–301, 2003 Neuronal activation, as assessed by c-fos mRNA levels, 9. Jamshidi N, Taylor DA: Anandamide administration into the ventromedial in hypothalamic nuclei known for their involvement in the hypothalamus stimulates appetite in rats. Br J Pharmacol 134:1151–1154, regulation of energy balance and the response to stress, 2001 10. Cota D, Marsicano G, Tscho¨ p M, Gru¨ bler Y, Flachskamm C, Schubert M, argues in favor of an involvement of the hypothalamus in Auer D, Yassouridis A, Tho¨ ne-Reineke C, Ortmann S, Tomassoni F, the effects of rimonabant. Increased c-fos mRNA in the Cervino C, Nisoli E, Linthorst AC, Pasquali R, Lutz B, Stalla GK, Pagotto U: PVN is in line with increased HPA axis activity, whereas The endogenous cannabinoid system affects energy balance via central

DIABETES, VOL. 55, DECEMBER 2006 3409 RIMONABANT AND HPA/NEURONAL ACTIVATION

orexigenic drive and peripheral lipogenesis. J Clin Invest 112:423–431, 29. Jbilo O, Ravinet-Trillou C, Arnone M, Buisson I, Bribes E, Peleraux A, 2003 Penarier G, Soubrie´ P, Le Fur G, Galiegue S, Casellas P: The CB1 receptor 11. Dallman MF, Akana SF, Laugero KD, Gomez F, Manalo S, Bell ME, antagonist rimonabant reverses the diet-induced obesity phenotype Bhatnagar S: A spoonful of sugar: feedback signals of energy stores and through the regulation of lipolysis and energy balance. FASEB J 19:1567– corticosterone regulate responses to chronic stress. Physiol Behav 79:3– 1569, 2005 12, 2003 30. Zucker LM, Antoniades HN: Insulin and obesity in the Zucker genetically 12. Weidenfeld J, Feldman S, Mechoulam R: Effect of the brain constituent obese rat “fatty.” Endocrinology 90:1320–1330, 1972 anandamide, a cannabinoid receptor agonist, on the hypothalamo-pitu- 31. Triscari J, Stern JS, Johnson PR, Sullivan AC: Carbohydrate metabolism in itary-adrenal axis in the rat. Neuroendocrinology 59:110–112, 1994 lean and obese Zucker rats. Metabolism 28:183–189, 1979 13. Wenger T, Jamali KA, Juaneda C, Leonardelli J, Tramu G: Arachidonyl 32. Poirier B, Bidouard JP, Cadrouvele C, Marniquet X, Staels B, O’Connor SE, ethanolamide (anandamide) activates the parvocellular part of hypotha- Janiak P, Herbert JM: The anti-obesity effect of rimonabant is associated lamic paraventricular nucleus. Biochem Biophys Res Commun 237:724– 728, 1997 with an improved serum lipid profile. Diabetes Obes Metab 7:65–72, 2005 14. Manzanares J, Corchero J, Fuentes JA: Opioid and cannabinoid receptor- 33. Fride E, Bregman T, Kirkham TC: Endocannabinoids and food intake: mediated regulation of the increase in adrenocorticotropin hormone and newborn suckling and appetite regulation in adulthood. Exp Biol Med corticosterone plasma concentrations induced by central administration of 230:225–234, 2005 ⌬9-tetrahydrocannabinol in rats. Brain Res 839:173–179, 1999 34. Freund TF, Katona I, Piomelli D: Role of endogenous cannabinoids in 15. Wenger T, Ledent C, Tramu G: The endogenous cannabinoid, anandamide, synaptic signaling. Physiol Rev 83:1017–1066, 2003 35. Castonguay TW, Dallman MF, Stern JS: Some metabolic and behavioral activates the hypothalamo-pituitary-adrenal axis in CB1 cannabinoid re- ceptor knockout mice. Neuroendocrinology 78:294–300, 2003 effects of adrenalectomy on obese Zucker rats. Am J Physiol 251:R923– 16. Patel S, Roelke CT, Rademacher DJ, Cullinan WE, Hillard CJ: Endocan- R933, 1986 nabinoid signaling negatively modulates stress-induced activation of the 36. Cabanac M, Richard D: The nature of the ponderostat: Hervey’s hypothesis hypothalamic-pituitary-adrenal axis. Endocrinology 145:5431–5438, 2004 revived. Appetite 26:45–54, 1996 17. Guillaume-Gentil C, Rohner-Jeanrenaud F, Abramo F, Bestetti GE, Rossi 37. Strack AM, Bradbury MJ, Dallman MF: Corticosterone decreases nonshiv- GL, Jeanrenaud B: Abnormal regulation of the hypothalamo-pituitary- ering thermogenesis and increases lipid storage in brown adipose tissue. adrenal axis in the genetically obese fa/fa rat. Endocrinology 126:1873– Am J Physiol 268:R183–R191, 1995 1879, 1990 38. Arvaniti K, Ricquier D, Champigny O, Richard D: Leptin and corticosterone 18. Richard D, Rivest R, Naı¨mi N, Timofeeva E, Rivest S: Expression of have opposite effects on food intake and the expression of UCP1 mRNA in corticotropin-releasing factor and its receptors in the brain of lean and brown adipose tissue of lepob/lepob mice. Endocrinology 139:4000–4003, obese Zucker rats. Endocrinology 137:4786–4795, 1996 1998 19. Timofeeva E, Richard D: Functional activation of CRH neurons and 39. Galpin KS, Henderson RG, James WP, Trayhurn P: GDP binding to expression of the genes encoding CRH and its receptors in food-deprived brown-adipose-tissue mitochondria of mice treated chronically with cor- lean (Fa/?) and obese (fa/fa) Zucker rats. Neuroendocrinology 66:327–340, ticosterone. Biochem J 214:265–268, 1983 1997 40. Timofeeva E, Richard D: Activation of the central nervous system in obese 20. Di Marzo V, Goparaju SK, Wang L, Liu J, Ba´tkai S, Ja´rai Z, Fezza F, Miura Zucker rats during food deprivation. J Comp Neurol 441:71–89, 2001 GI, Palmiter RD, Sugiura T, Kunos G: Leptin-regulated endocannabinoids 41. Arnone M, Maruani J, Chaperon F, Thiebot MH, Poncelet M, Soubrie´P,Le are involved in maintaining food intake. Nature 410:822–825, 2001 Fur G: Selective inhibition of sucrose and ethanol intake by SR 141716, an 21. Bray GA: 1989 McCollum Award Lecture: Genetic and hypothalamic antagonist of central cannabinoid (CB1) receptors. Psychopharmacology mechanisms for obesity: finding the needle in the haystack. Am J Clin 132:104–106, 1997 Nutr 50:891–902, 1989 42. Poncelet M, Maruani J, Calassi R, Soubrie´ P: Overeating, alcohol and 22. Bensaid M, Gary-Bobo M, Esclangon A, Maffrand JP, Le Fur G, Oury-Donat sucrose consumption decrease in CB1 receptor deleted mice. Neurosci F, Soubrie´ P: The cannabinoid CB receptor antagonist SR141716 increases 1 Lett 343:216–218, 2003 Acrp30 mRNA expression in adipose tissue of obese fa/fa rats and in 43. Gamber KM, MacArthur H, Westfall TC: Cannabinoids augment the cultured adipocyte cells. Mol Pharmacol 63:908–914, 2003 release of neuropeptide Y in the rat hypothalamus. Neuropharmacology 23. Vickers SP, Webster LJ, Wyatt A, Dourish CT, Kennett GA: Preferential 49:646–652, 2005 effects of the cannabinoid CB receptor antagonist, SR 141716, on food 1 44. Verty AN, McFarlane JR, McGregor IS, Mallet PE: Evidence for an intake and body weight gain of obese (fa/fa) compared to lean Zucker rats. interaction between CB cannabinoid and melanocortin MCR-4 receptors Psychopharmacology 167:103–111, 2003 1 Endocrinology 24. Webster AJF: Energetics of maintenance and growth. In Mammalian in regulating food intake. 145:3224–3231, 2004 Thermogenesis. Girardier L, Stock M, Eds. London, Chapman and Hall, 45. Flier JS: Obesity wars: molecular progress confronts an expanding epi- 1983, p. 178–207 demic. Cell 116:337–350, 2004 25. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner 46. Armstrong WE: Hypothalamic supraoptic and paraventricular nuclei. In RC: Homeostasis model assessment: insulin resistance and beta-cell func- The Rat Nervous System. Paxinos G, Ed. San Diego, CA, Academic Press, tion from fasting plasma glucose and insulin concentrations in man. 1995, p. 377–390 Diabetologia 28:412–419, 1985 47. Timofeeva E, Picard F, Duclos M, Deshaies Y, Richard D: Neuronal 26. Simmons DM, Arriza JL, Swanson LW: A complete protocol for in situ activation and corticotropin-releasing hormone expression in the brain of hybridization of messenger RNAs in brain and other tissues with radiola- obese (fa/fa) and lean (fa/?) Zucker rats in response to refeeding. Eur beled single-stranded RNA probes. J Histotechnol 12:169–181, 1989 J Neurosci 15:1013–1029, 2002 27. Colombo G, Agabio R, Diaz G, Lobina C, Reali R, Gessa GL: Appetite 48. Naı¨mi N, Rivest S, Racotta I, Richard D: Neuronal activation of the suppression and weight loss after the cannabinoid antagonist SR 141716. hypothalamic magnocellular system in response to oropharyngeal stimuli Life Sci 63:PL113–PL117, 1998 in the rat. J Neuroendocrinol 9:329–340, 1997 28. Hildebrandt AL, Kelly-Sullivan DM, Black SC: Antiobesity effects of 49. Verty AN, McFarlane JR, McGregor IS, Mallet PE: Evidence for an

chronic cannabinoid CB1 receptor antagonist treatment in diet-induced interaction between CB1 cannabinoid and oxytocin receptors in food and obese mice. Eur J Pharmacol 462:125–132, 2003 water intake. Neuropharmacology 47:593–603, 2004

3410 DIABETES, VOL. 55, DECEMBER 2006