3700 Diabetes Volume 64, November 2015

Anne-Laure Poher,1 Christelle Veyrat-Durebex,2 Jordi Altirriba,1 Xavier Montet,3 Didier J. Colin,4 Aurélie Caillon,1 Jacqueline Lyautey,1 and Françoise Rohner-Jeanrenaud1

Ectopic UCP1 Overexpression in White Adipose Tissue Improves Insulin Sensitivity in Lou/C Rats, a Model of Obesity Resistance

Diabetes 2015;64:3700–3712 | DOI: 10.2337/db15-0210

Brown adipose tissue (BAT), characterized by the pres- epidemic is long-term dysregulation of energy balance. In view ence of uncoupling protein 1 (UCP1), has been described oftherelativelackofdrugssuppressing appetite, approaches as metabolically active in humans. Lou/C rats, originat- to increase energy expenditure are viewed as potential new ing from the Wistar strain, are resistant to obesity. We therapeutic options to treat obesity and metabolic complica- previously demonstrated that Lou/C animals express tions. Along this line, brown adipose tissue (BAT), now known UCP1 in beige adipocytes in inguinal white adipose tissue to be present in adult humans (2–4),isthefocusofmajor (iWAT), suggesting a role of this protein in processes interest, due to its role in inducing thermogenesis (5). The such as the control of body weight and the observed mitochondria of brown adipocytes are characterized by the b improved insulin sensitivity. A 3 adrenergic agonist was presence of uncoupling protein 1 (UCP1), which uncouples administered for 2 weeks in Wistar and Lou/C rats to oxidative phosphorylation from ATP synthesis, resulting in activate UCP1 and delineate its metabolic impact. The heat production (6). This process consumes substantial treatment brought about decreases in fat mass and

METABOLISM amounts of free fatty acids (FFAs) and glucose (7). improvements in insulin sensitivity in both groups. In BAT, UCP1 expression increased similarly in response to In rodents, brown adipocytes are found in discrete areas, the treatment in the two groups. However, the intervention such as interscapular, cervical, peri-renal, and intercostal “ ” induced the appearance of beige cells in iWAT, associated depots (8), which are referred to as classical BAT depots. In with a marked increase in UCP1 expression, in Lou/C white adipose tissue (WAT), brown-like cells, called beige or fi rats only. This increase was correlated with a markedly brite cells, express UCP1 (9). The existence of a speci cpre- enhanced glucose uptake measured during euglycemic- cursor, different from classical white or brown adipocytes, hyperinsulinemic clamps, suggesting a role of beige cells arising from smooth muscle cells that would differentiate in this process. Activation of UCP1 in ectopic tissues, such into beige adipocytes in WAT has been proposed (10). More- as beige cells in iWAT, may be an interesting therapeutic over, some studies suggest that, under specificconditions, approach to prevent body weight gain, decrease fat mass, most or all of white adipocytes transdifferentiate into beige and improve insulin sensitivity. adipocytes (11). Interestingly, whatever their developmental origin, white, beige, and brown adipocytes seem to greatly differ Obesity has reached epidemic proportions worldwide and in their function. As mentioned above, BAT is the effector has become a major global public health problem in recent organ of nonshivering thermogenesis that, by utilizing decades (1). It represents a considerable risk factor for the large quantities of glucose and lipids from the circulation, development of several comorbidities, among which is type can promote negative energy balance. Moreover, as recently 2 diabetes (1). Globally, the key component of the obesity reviewed by Peirce and Vidal-Puig (12), the role of BAT

1Laboratory of Metabolism, Division of Endocrinology, Diabetology, Hypertension Corresponding author: Anne-Laure Poher, [email protected]. and Nutrition, Department of Internal Medicine Specialties, Faculty of Medicine, Received 13 February 2015 and accepted 20 July 2015. University of Geneva, Geneva, Switzerland This article contains Supplementary Data online at http://diabetes 2Department of Cell Physiology and Metabolism, Faculty of Medicine, University .diabetesjournals.org/lookup/suppl/doi:10.2337/db15-0210/-/DC1. of Geneva, Geneva, Switzerland 3Division of Radiology, Faculty of Medicine, University of Geneva, Geneva, Switzerland © 2015 by the American Diabetes Association. Readers may use this article as 4MicroPET/SPECT/CT Imaging Laboratory, Centre for BioMedical Imaging, University long as the work is properly cited, the use is educational and not for profit, and Hospital of Geneva, Geneva, Switzerland the work is not altered. diabetes.diabetesjournals.org Poher and Associates 3701 activation might be broader than solely promoting negative administered. Blood samples were collected by tail nicking energy balance. Indeed, such activation was described to and were used for further analyses of plasma glucose and exert anti–type 2 diabetic effects (13,14), associated with insulin levels. improvements of dyslipidemia (rev. in 15). These effects are partly interrelated but can also be dissociated and EHC exerted by different UCP1-expressing adipocyte types. Rats were overnight fasted and anesthetized with pentobar- The Lou/C rat is a model of age- and diet-induced bital (50 mg/kg i.p.; Abbott Laboratories, Chicago, IL). The obesity resistance, which also exhibits a lower body fat glucose infusion rate (GIR) required to maintain euglycemia mass, increased leptin sensitivity, and improved insulin under insulin-stimulated conditions (18 mU/kg/min; Actrapid sensitivity compared with Wistar animals (16–18). We HM, Novo Nordisk, Bagsvaerd, Denmark) was determined also demonstrated that the most striking differences as previously described (19,20). At the end of the EHC, the between Lou/C and Wistar rats were the presence, in in vivo insulin-stimulated glucose utilization index inguinal WAT (iWAT) of the Lou/C group, of Ucp1 (16,17) of individual tissues was measured, using 2-deoxy-D- 3 m and of a marked Adrb3 overexpression (17). [1,2- H]glucose (30 Ci/rat, NET 328A; PerkinElmer, In the current study, we used the Lou/C rat as a model Schwerzenbach, Switzerland) (20). Rats were killed by – to investigate the impact of UCP1 activation on glucose decapitation and tissues stored at 80°C. The 2-deoxy-D- 3 – fi metabolism. To this end, various groups of Wistar and Lou/C [1,2- H]glucose speci c activity was measured in deproteinized rats were subcutaneously infused for 2 weeks with a b3 blood samples. Determination of tissue concentration of 3 adrenoreceptor agonist (CL-316243), and UCP1 in BAT and 2-deoxy-D-[1,2- H]glucose-6-phosphate allowed for the cal- WAT depots, as well as the insulin-stimulated glucose utili- culation of the in vivo glucose utilization index of individ- zation rate, of different tissues was determined. ual tissues and was expressed in nanograms of glucose per milligram of tissue per minute (21,22). RESEARCH DESIGN AND METHODS Plasma Measurements Animals and Diets Plasma glucose levels were measured by the glucose Two month-old male Lou/C and Wistar rats were pur- oxidase method (Glu; Roche Diagnostics GmbH, Rotkreuz, chased from Harlan UK Limited (Oxon, U.K.) and Charles Switzerland). FFA and triglyceride (TG) levels were deter- ’ River (L Arbresle, France), respectively. They were housed mined using a Wako Chemicals GmbH (Neuss, Germany) – in pairs under controlled conditions (22°C; light on 7:00 A.M. and a Biomérieux (Marcy l’Etoile, France) kit, respectively. 7:00 P.M.) and were allowed free access to water and Plasma leptin (Linco Research, Inc., St Charles, MO) levels diet (RMI; Hersteller, Essex, U.K.). Osmotic minipumps were measured using a double antibody radioimmuno- b (Alzet, Cupertino, CA) delivering a 3 agonist (CL-316243, assay kit. Plasma insulin levels were determined using 1 mg/kg/day; Tocris, Bristol, U.K.,) or NaCl 0.9% for 14 days an ELISA commercial kit (10-1250; Mercodia, Uppsala, were placed subcutaneously at the age of 12 weeks. One week Sweden). before and during the treatment, body weight and food in- take were measured daily (9:00 A.M.). A first colony was sub- Tissue Processing and RT-PCR mitted to a glucose tolerance test (GTT) at day 9 and a body RNA was reverse transcribed (M-MLV-RT; Invitrogen, fat composition analysis by MRI (Echo-MRI) at day 14. A Basel, Switzerland) and quantitative PCR (qPCR) was second colony was submitted to a computed tomography performed using the SYBR green PCR Master Mix (Applied (CT) scan on day 12 and a euglycemic-hyperinsulinemic clamp Biosystems, Warrington, U.K.) on a Stepone Plus machine (EHC) on day 14. A third colony underwent a positron emis- (Applied Biosystems). Primers (Supplementary Table 1) sion tomography (PET) scan coupled to a micro-CT on day were designed with the PrimerExpress software (Applied 14. A fourth colony was housed in pairs at 22°C or at Biosystems). Results were normalized to the expression 30°C in a climatic chamber (Meditest H 1300L; Froilabo, levels of the housekeeping gene, ribosomal protein S29 Meysieu, France). (Rps29). Except for EHC experiments, all rats were killed between 9:00 A.M. and 1:00 P.M.,usingisoflurane (Halocarbon Labora- Mitochondrial DNA Copy Number tories, River Edge, NJ) anesthesia and rapid decapitation. Quantification of mitochondrial DNA (mtDNA) copy num- Trunk blood was collected to measure the concentration of ber was achieved by qPCR. Briefly, DNA was extracted from various hormones and metabolites. Tissues were freeze iWAT using the DNeasy Blood and Tissue kit (Qiagen, clamped and stored at –80°C for determination of gene Düsseldorf, Germany). Nuclear and mtDNA copy numbers and protein expressions. The procedures were approved by were assessed by real-time PCR using primers targeted the ethics committee of our university and were in accor- toward the Cox1 gene (for mtDNA) and nuclear RNAseP dance with the Swiss guidelines for animal experimentation. (for nuclear DNA).

GTT Western Blotting Wistar and Lou/C rats were food deprived for 4 h (from Frozen tissues were homogenized in ice-cold RIPA buffer. Pro- 9:00 to 1:00 P.M.). A glucose load of 1.5 g/kg i.p. was tein levels were quantified using a BCA protein assay (Pierce, 3702 Ectopic UCP1 in Lou/C Rats Improves Metabolism Diabetes Volume 64, November 2015

Lausanne, Switzerland). Ten and 50 mg protein, respectively, lower than that of Wistar rats, but it was unaffected by was used for BAT and WAT SDS-PAGE. UCP1 antibody (Abcam, the treatment in both groups (Fig. 1A). Interestingly, food Cambridge, U.K.) was used at a concentration of 1/10,000 efficiency was similar in Wistar and Lou/C rats whether for BAT and 1/1,000 for WAT before the secondary antibody treated or not (Fig. 1A). Analysis of body composition by (anti-rabbit 1/5,000) was added. Housekeeping proteins MRI (EchoMRI) further showed that the b3 agonist treat- ERM (Santa Cruz Biotechnology, Santa Cruz, CA) and actin ment decreased the total fat mass and consequently in- (Millipore, Billerica, MA) were used at 1/2,000 and creased the lean body mass in both Wistar and Lou/C 1/100,000, with anti-goat (1/10,000) or anti-mouse (1/5,000) animals (Fig. 1B). In addition, CT scan analysis showed antibody, respectively. Detection was performed with an en- a significant treatment-induced decrease in the mass of all hanced chemiluminescence detection system (Amersham fat depots (interscapular BAT, subcutaneous WAT, and Biosciences, Amersham, U.K.). Signals were quantified using intra-abdominal WAT) in the two animal groups, except the PXi and the Genetools software from Syngene UK. for subcutaneous WAT of Lou/C rats (Fig. 1C). This was in keeping with a significant or with a trend toward a signif- Body Composition icant decrease in leptinemia in b3 agonist–treated Wistar An EchoMRI-700 quantitative nuclear magnetic resonance and Lou/C rats, respectively (Table 1). Qualitatively sim- analyzer (Echo Medical Systems, Houston, TX) was used to ilar results were obtained for plasma glucose, insulin, and measure body composition (total fat and lean body mass). TG levels, while plasma FFA concentrations were unaf- Rats were also scanned on a multidetector CT scanner fected by the treatment in both groups (Table 1). (Discovery 750 HD; GE Healthcare, Milwaukee, WI) to To evaluate the impact of the b3 agonist treatment on determine the volume of the various fat depots. glucose metabolism, a GTT was performed at day 9 of the study (Fig. 2A). The treatment had no impact on basal Light Microscopy glycemia in both groups (data not shown). As indicated by Tissues were fixed in formalin (10%) after dissection. the changes in glycemia from baseline values, as well as They were washed in PBS, dehydrated, cleared, and finally the areas under the curve (AUCs) during the whole tests, embedded in paraffin blocks, which were cut at 7 mm and glucose tolerance of Wistar and Lou/C rats was improved then stained with hematoxylin-eosin (Sigma-Aldrich, Buchs, by the treatment (Fig. 2A). With regard to insulin, the Switzerland). For immunohistochemistry, the same UCP1 treatment effect on the AUC during the GTT almost antibody as that used for Western blots was used at a con- reached statistical significance (P = 0.057), this effect be- centration of 1/100 before the secondary anti-rabbit Cy3 ing clearly more marked in the Lou/C than in the Wistar antibody (1/250) (Jackson ImmunoResearch). group (Fig. 2B). [18F]-fluorodeoxyglucose Glucose Uptake Peripheral insulin sensitivity was then evaluated by Measurement in Tissues performing EHC. Basal and steady-state (clamp) values of Rats were overnight fasted and anesthetized using glycemia and insulinemia are provided in Supplementary isoflurane. Glycemia was measured before the experi- Table 2. As previously reported (16,17), the GIR of Lou/C ment. Animals received a 3.5 mg/kg i.p. injection of rats was higher than that of Wistar animals, indicating furosemide (Lasix; Sanofi,Paris,France)toemptytheir improved insulin sensitivity (Fig. 2C). In both the Wistar bladder. [18F]-fluorodeoxyglucose ([18F]-FDG) was injected and the Lou/C group, the b3 agonist treatment increased via the pudendal vein (20 MBq/rat). Glucose uptake was mea- the GIR (Fig. 2C). As clearly depicted by the dynamic GIR sured with a microPET/CT (Triumph; TriFoil, Chatsworth, CA). changes, the values reached in Wistar-treated rats were A micro-CT analysis was simultaneously performed to localize similar to those of the untreated Lou/C group, while the the [18F]-FDG emission in different structures. Data were an- b3 agonist treatment further increased the GIR in Lou/C alyzed by the Osirix software (Pixmeo, Geneva, Switzerland). animals (Fig. 2C). Glucose uptake by individual insulin- sensitive tissues was then assessed, using the 2-deoxyglucose Data Analyses technique (20–22). The treatment had no impact on skeletal Results are expressed as mean 6 SEM. Comparison between muscle glucose uptake in both groups (Fig. 2D). It resulted in the four groups was performed by GraphPad Prism (San similar two- to threefold increases in glucose uptake in BAT Diego, CA), using the two-way ANOVA analysis followed of Wistar and Lou/C rats (Fig. 2D). Interestingly, glucose by a Tukey post hoc test. Correlation between the Ucp1 uptake by different WAT depots was not significantly mRNA expression and the glucose uptake index in WAT modified in Wistar but was markedly increased in Lou/C was calculated using the Pearson correlation coefficient. animals. This was observed in abdominal WAT (epididymal Statistical significance was established at P , 0.05. WAT [eWAT] and retroperitoneal WAT) and in the iWAT fat depot (Fig. 2D). RESULTS To try understanding the mechanisms underlying the b3 As previously reported (16,17), Lou/C had a lower body agonist treatment–induced increases in BAT and WAT weight than Wistar rats. The 2-week CL-316243 treat- glucose uptake, and in view of the well-known effect of ment had no impact on body weight gain in Wistar and CL-316243 on UCP1 (23), we measured the expression of this Lou/C animals (Fig. 1A). Total food intake of Lou/C was proteininthesetissues.ConsideringBAT,hematoxylin-eosin diabetes.diabetesjournals.org Poher and Associates 3703

Figure 1—CL-316243 decreases fat mass in both Wistar and Lou/C rats. A: Total body weight gain of Wistar and Lou/C rats measured after 14 days of CL-316243 treatment 1 mg/kg/day (CL) or saline solution (NaCl), total food intake measured after 9 days, and food efficiency calculated from Dbody weight gain divided by total food intake at 9 days. B: Determination of the percent lean and fat mass repartition using MRI. C: Determination of the volume of various fat depots by CT scan. Results are means 6 SEM of 8 experiments per group. Statistical significance assessed by two-way ANOVA: #strain effect; @treatment effect; $interaction between strain and treatment. *Significant difference by post hoc pairwise comparison between different conditions. Single, double, and triple symbols indicate P < 0.05, P < 0.01, and P < 0.001, respectively.

staining revealed less lipid inclusions and therefore acti- in Wistar and Lou/C rats regarding the expression of the vated tissue in the Wistar and the Lou/C treated groups Adrb3, which was markedly inhibited (Fig. 3B). (Fig. 3A). Accordingly, Ucp1 mRNA expression was in- Subsequently, BAT activity was evaluated by PET scan creased by fourfold by the treatment in the two groups analysis, using [18F]-FDG. Figure 3D shows micro-CT, PET (Fig. 3B). These results were confirmed by Western blot scan, and merged micro-CT and PET scan images of one analysis (Fig. 3C). The treatment also had similar effects representative animal per group. PET scan allowed for the

Table 1—CL-316243 improves metabolic parameters Wistar Lou/C NaCl CL-316243 NaCl CL-316243 Glycemia (mmol/L) 7.6 6 0.3 6.6 6 0.2†† 6.2 6 0.2 5.5 6 0.2‡ Insulin (ng/mL) 2.06 6 0.50 0.71 6 0.15†† 0.54 6 0.09 0.40 6 0.04 Leptin (ng/mL) 3.80 6 0.19 2.37 6 0.22††† 1.44 6 0.21 0.98 6 0.12 TG (mg/mL) 1.49 6 0.13 0.91 6 0.09††† 0.88 6 0.08 0.54 6 0.03‡ FFA (mmol/L) 0.25 6 0.02 0.23 6 0.01 0.42 6 0.05 0.32 6 0.03 Results are means 6 SEM of 8 experiments per group. Plasma was collected from rats receiving NaCl or CL-316243 treatment for 14 days starting at the age of 12 weeks. Statistical analyses were performed using two-way ANOVA (Tukey posttest) with ††,0.01, †††,0.001, Wistar NaCl vs. Wistar CL; ‡,0.05 Lou/C NaCl vs. Lou/C CL. 3704 Ectopic UCP1 in Lou/C Rats Improves Metabolism Diabetes Volume 64, November 2015

Figure 2—Improvement in insulin sensitivity in Lou/C rats is due to an increase in glucose uptake by WAT. A: DGlycemia during a GTT (1.5 g/kg i.p.) after 4 h of fasting; glucose AUC at 120 min. B: Plasma insulin levels during the GTT; insulin AUC after 60 min. C: GIR during (right panel) and at the end (left panel) of EHC in Wistar NaCl, Wistar treated with CL-316243 (CL), Lou/C NaCl, and Lou/C treated with CL-316243 rats. D: Tissue-specific insulin-stimulated glucose uptake. eWAT, epididymal WAT; rpWAT, retroperitoneal WAT; Qr, quadriceps red; Qw, quadriceps white. Results are means 6 SEM of 6–9 experiments. Statistical significance assessed by two-way ANOVA: #strain effect; @treatment effect; $interaction between strain and treatment. *Significant difference by post hoc pairwise comparison between different conditions. Single, double, and triple symbols indicate P < 0.05, P < 0.01, and P < 0.001, respectively. For the dynamic changes of Dglucose and insulinemia during the GTT (A and B, left panels), as well as of GIR as a function of time (C, right panel): *Wistar NaCl vs. Lou/C NaCl; †Wistar NaCl vs. Wistar CL-316243; ‡Lou/C NaCl vs. Lou/C CL-316243; §Wistar CL vs. Lou/C CL-316243. gluc, glucose. diabetes.diabetesjournals.org Poher and Associates 3705

Figure 3—UCP1 is equally overexpressed in BAT in both Wistar (W) and Lou/C (L) rats by CL-316243. A: Representative hematoxylin-eosin staining of BAT (7 mm). Scale bars = 50 mm. B: Ucp1 and b3 adrenoreceptor (Adrb3) mRNA expression in BAT. C: Protein expression by Western blot with relative quantification in BAT. D: Representative sagittal cuts of [18F]-FDG uptake measured by microPET. microCT was performed to localize glucose uptake in the tissues; quantification of glucose uptake in interscapular BAT (iBAT) (29), cervical BAT (cBAT), and paravertebral BAT (pvBAT) in control and CL-316243–treated Wistar and Lou/C rats. Results are means 6 SEM of 4–8 experiments. Statistical significance assessed by two-way ANOVA: @, treatment effect. *Significant difference by post hoc pairwise comparison between different conditions. Double and triple symbols indicate P < 0.01 and P < 0.001, respectively. CL, CL-316243.

detection of three main BAT depots: at the basis of the In line with our previous observation of measurable brain (cervical BAT), in the interscapular region, and along UCP1 expression in iWAT of Lou/C, but not of Wistar the vertebral column. As clearly revealed after quantifica- rats (16,17), we looked at the potential presence of tion of the signals in all the animals studied, the b3agonist UCP1-positive beige adipocytes in this tissue after the treatment increased glucose uptake to a similar extent in b3 agonist treatment. Examination of hematoxylin-eosin each of these three depots in Wistar and Lou/C rats, staining revealed the presence of multilocular adipocytes whereas the signals were absent under standard conditions in the b3agonist–treated Lou/C group only (Fig. 4A). (Fig. 3D). Immunofluorescence analysis further showed that these 3706 Ectopic UCP1 in Lou/C Rats Improves Metabolism Diabetes Volume 64, November 2015

Figure 4—UCP1 is overexpressed in the subcutaneous iWAT depot in Lou/C rats by CL-316243. A: Representative hematoxylin-eosin staining in iWAT (7 mm). Scale bars = 50 mm. iWAT sections were stained with an anti-UCP1 antibody. Hematoxylin staining was used to recognize structures. The sections were examined by fluorescence microscopy. Red indicates UCP1. B: Ucp1, Adrb3, and Pgc1a mRNA expression in iWAT. C: UCP1 protein expression measured by Western blot with relative quantification in iWAT. D: Representative axial cuts of [18F]-FDG uptake measured by microPET. microCT was performed to localize glucose uptake in the tissues; quantification of glucose uptake in iWAT of control and CL-316243 (CL)-treated Wistar and Lou/C rats. SUV, standard uptake value. E: Mitochondrial DNA copy number in iWAT. DNA copies of mitochondrial complex I (COX1, mtDNA) were normalized to the DNA levels of nuclear RNAseP (nDNA). diabetes.diabetesjournals.org Poher and Associates 3707 multilocular adipocytes expressed UCP1, demonstrating groups of Wistar and Lou/C rats at 22°C or at 30°C for the presence of beige cells in the iWAT of Lou/C treated 4 weeks. At 30°C, both food intake and body weight gain rats (Fig. 4A). In keeping with these results, Ucp1 mRNA were reduced in Wistar, as well as in Lou/C rats (Fig. 5A). expression was increased a hundred fold in iWAT of Lou/C This resulted in similar food efficiencies in response to the rats in response to the treatment. It was also induced in treatment under the two experimental conditions, with iWAT of Wistar treated animals, reaching values measured lower values in the Lou/C than in the Wistar groups (Fig. in the untreated Lou/C group (Fig. 4B). Western blot anal- 5B). Regarding BAT Ucp1 expression, it was lower in Lou/C ysis clearly showed that UCP1 protein was expressed in the than in Wistar animals, and it was reduced by the iWAT of the Lou/C treated group only (Fig. 4C). In line exposure at 30°C in the Wistar group (Fig. 5D). Such with the results obtained for Ucp1, the expression of Pgc1a conditions also brought about a decrease in BAT Adrb3 (Fig. 4B), of Tbx1 (Supplementary Fig. 1), and of various expression in Wistar and Lou/C rats (Fig. 5D). In iWAT, markers of mitochondrial biogenesis (Supplementary Fig. 1), Ucp1 expression was doubled in Lou/C compared with as well as mitochondrial DNA (Fig. 4E), were markedly Wistar at 22°C, although this did not reach statistical enhanced by the treatment in Lou/C rats (600% increase in significance, and it remained at the same values at 30°C the mtDNA copy number). Next, we measured the activity (Fig. 5E). In this tissue, expression of the Adrb3 was of iWAT by PET scan coupled with micro-CT and observed higher in Lou/C than in Wistar rats, with no effect of a higher glucose uptake in treated Lou/C rats compared the temperature (Fig. 5E). Similar observations were with the three other groups (Fig. 4D). To substantiate made for the GIR measured during EHC (Fig. 5C), indi- the existence of the previously proposed link between glu- cating that the improved insulin sensitivity of the Lou/C cose uptake/insulin sensitivity and Ucp1 expression in ad- group is maintained at 30°C and is therefore not linked to ipose tissue (24), we looked at the correlation between cold exposure, even of minor magnitude. these two parameters, considering all the animals studied. Finally, the expression of genes encoding for enzymes As shown on Fig. 4F, the data fitted a hyperbolic regression known to be involved in lipid metabolism was measured with a positive and significant correlation, suggesting that in BAT and iWAT. Acetyl CoA carboxylase represents the UCP1-positive beige cells in iWAT may be involved in the rate-limiting step in fatty acid (FA) synthesis. Carnitine improvement in insulin sensitivity in this tissue. With palmitoyl transferase-1a is the rate-limiting step of the regard to the expression of the Adrb3, it was markedly FA oxidation pathway, mediating FA transport from the overexpressed in iWAT of Lou/C compared with Wistar cytosol to the mitochondria. Medium- and large-chain rats, in which it was barely detectable. The b3agonist acyl-CoA dehydrogenases (MCAD and LCAD) are involved treatment downregulated the expression of the Adrb3 in in FA oxidation. PEPCK allows for the in situ production the Lou/C but not in the Wistar group (Fig. 4B). of glycerol-3-phosphate (glyceroneogenesis) and therefore Given that the insulin-induced glucose uptake was for the formation of TG. Hormone-sensitive lipase (HSL) increased by the b3 agonist, not only in iWAT, but also in is one of the main enzymes involved in lipolysis. In iWAT, the eWAT, we examined the potential presence of beige Acc1 expression was twofold higher in Lou/C than in adipocytes in this depot. Multilocularization of adipocytes Wistar rats in the absence of b3 agonist treatment. Fur- was observed in Lou/C-treated rats only (Supplementary thermore, it was more than doubled in the Lou/C group Fig. 2A), in keeping with the induction of a marked Ucp1 in response to the treatment, while it remained unchanged gene expression (by 160-fold) (Supplementary Fig. 2B). The in Wistar rats (Fig. 6A). As expected from its inhibitory mRNA expression of the Adrb3 was low in both Wistar and regulation by malonyl CoA, Cpt1a expression was lower in Lou/C rats, with no intergroup difference and no impact of Lou/C than in Wistar animals, with a trend toward a fur- the b3 agonist treatment (Supplementary Fig. 2B). ther decrease in the b3 agonist–treated group (Fig. 6A). It is well-known that thermoneutrality in rodents is On the contrary, the b3 agonist treatment resulted in an reached at ;28–30°C (24°C for rats and 30°C for mice) enhancement of Hsl, Lcad, and Mcad expression in Lou/C and that, although it is the usual temperature used in rats, without any change in the Wistar group (Fig. 6A). animal quarters, 22°C represents a cold stimulus increas- Similar results were obtained for the expression of Pepck ing UCP1 expression and activity (25). It was therefore (Fig. 6A). possible that the presence of UCP1-positive beige cells in In BAT, Acc1 expression was similar in Wistar and Lou/C WAT observed in the Lou/C group without any treatment untreated rats and it was reduced by the treatment in was a consequence of a better temperature sensing than the Lou/C group only (Fig. 6B). The expression of Cpt1a in Wistar rats. To address this issue, we maintained was under the detection limit, as a likely consequence of

F: Nonlinear regression with hyperbolic correlation between 2-DG glucose uptake and Ucp1 expression in iWAT. Results are means 6 SEM of 8 experiments per group. Statistical significance assessed by two-way ANOVA: #strain effect; @treatment effect; $interaction between strain and treatment. *Significant difference by post hoc pairwise comparison between different conditions. Single and triple symbols indicate P < 0.05 and P < 0.001, respectively. gluc, glucose. 3708 Ectopic UCP1 in Lou/C Rats Improves Metabolism Diabetes Volume 64, November 2015

Figure 5—Differences between Wistar and Lou/C rats are not due to a difference in thermoneutrality. A: DBody weight (BW) gain of Wistar and Lou/C rats measured between days 4 and 30 of 22°C (control) and 30°C exposure. Total food intake (FI) measured between days 4 and 30. B: Food efficiency calculated from Dbody weight gain divided by total food intake between days 4 and 30. C: GIR measured at the end of EHCs in Wistar and Lou/C rats housed at 22°C and at 30°C. D: Ucp1 and Adrb3 mRNA expression in BAT. E: Ucp1 and Adrb3 mRNA expression in iWAT. Results are means 6 SEM of 8–12 experiments. Statistical significance assessed by two-way ANOVA: #strain effect; @treatment effect; $interaction between strain and treatment. *Significant difference by post hoc pairwise comparison between different conditions. Double and triple symbols indicate P < 0.01 and P < 0.001, respectively. diabetes.diabetesjournals.org Poher and Associates 3709

Figure 6—The higher glucose uptake in iWAT is explained by an increase in FA synthesis in CL-treated Lou/C rats. A: Acc1, Cpt1a, Hsl, Pepck, Lcad, and Mcad mRNA expression in iWAT. B: Acc1 and Pepck mRNA expression in BAT. Results are means 6 SEM of 8 experiments per group. Statistical significance assessed by two-way ANOVA: #strain effect; @treatment effect. *Significant difference by post hoc pairwise comparison between different conditions. Single, double, and triple symbols indicate P < 0.05, P < 0.01, and P < 0.001, respectively. AU, arbitrary units; CL, CL, CL-316243; ctrl, control.

elevated Acc1 expression in this tissue. There was no resistance characterized by the presence of beige cells in iWAT difference in Hsl expression in Wistar and Lou/C rats (16,17). For this purpose, Lou/C and Wistar rats were treated whether treated or not (data not shown). In control for 2 weeks with a b3 agonist (CL-316243). The treatment animals, Pepck was less expressed in the Lou/C compared had no effect on food intake and body weight gain, while it with the Wistar group but it was increased by the treat- decreased the fat mass as well as the weight of interscapular ment in Lou/C rats to reach the expression level mea- BAT and of various WAT depots in the two groups. sured in Wistar rats (Fig. 6B). In terms of its effects on UCP1 expression, different types of results were obtained in BAT and WAT when DISCUSSION comparing Wistar and Lou/C rats. In BAT, we and others This study was undertaken to determine the impact of previously reported that Ucp1 mRNA and/or protein levels recruiting UCP1-expressing adipocytes in a model of obesity were lower in Lou/C than in Wistar rats at 22°C or at 25°C 3710 Ectopic UCP1 in Lou/C Rats Improves Metabolism Diabetes Volume 64, November 2015

(17,26), whereas, in the current study, this was observed for UCP1 protein at 22°C but not for the Ucp1 gene at 22°C or 30°C. Regarding the effects of the b3 agonist treatment, it increased Ucp1 and decreased Adrb3 expression in BAT of both groups to a similar extent. These results, confirmed by the measurement of UCP1 protein, closely correlated with the observation of similar increases in BAT glucose uptake in Wistar and Lou/C rats, as measured during EHC and by [18F]-FDG–PET scan analysis. UCP1-expressing beige cells are known to be present in various WAT depots (27,28). Chronic cold exposure was shown to recruit beige adipocytes in WAT, resulting in WAT “browning” (29,30). Interestingly, resistance to diet- Figure 7—Schematic representation of a beige adipocyte in iWAT induced obesity in rats and mice has been suggested to of b3 agonist–treated Lou/C rats, showing the occurrence of futile depend on the induction of beige adipocyte recruitment cycling between lipolysis and lipogenesis. Higher glucose uptake is converted to malonyl-CoA by acetyl CoA carboxylase 1 (ACC1), in WAT (16,31). Moreover, such increased recruitment of which is markedly increased. Together with hormone-sensitive beige cells has been shown to compensate for decreased lipase (HSL), this increases the amount of FA within the cell. These BAT thermogenesis, and the specific loss of beige adipo- FA are used by the mitochondria, as suggested by the increase in b cytes due to adipose tissue–specific deletion of a transcrip- the expression of enzymes involved in -oxidation (large-chain acyl-CoA dehydrogenase [LCAD] and medium chain acyl-CoA de- tional factor involved in Ucp1 expression (32), PRDM16, hydrogenase [MCAD]), despite a slight decrease in carnitine palmitoyl was shown to cause obesity (32,33). Beige adipocytes are transferase-1 (CPT1) (likely due to an increase in the production also strikingly involved in the regulation of glucose metab- of malonyl-CoA). Enhanced FA utilization induces the production fi of acetyl-CoA, as well as of a proton gradient between the two olism, as demonstrated by the bene cial effects on glucose mitochondrial membranes. Such a gradient is dissipated by UCP1, tolerance and insulin sensitivity observed in transgenic increasing heat production and decreasing ATP synthesis. Part of mice overexpressing PRDM16 in iWAT (32). mitochondrial acetyl-CoA can be transformed to oxaloacetate in the In the current study, we showed that the b3agonist cytosol (after its transformation to citrate), where it can be converted to glycerol-3-phosphate (G3P) via PEPCK (glyceroneogenesis), ulti- treatment affected the Lou/C group only, markedly increas- mately increasing lipogenesis. Yellow arrows represent lipogenic, ing UCP1 gene and protein expression, mitochondrial bio- whereas green ones represent lipolytic pathways. genesis, and the insulin-stimulated glucose uptake measured during EHC (subcutaneous and intra-abdominal fat depots) or using [18F]-FDG–PET scan analysis (subcutaneous fat de- pot) as previously reported (34). Furthermore, in iWAT from Lou/C- treated rats, a hyperbolic correlation between glucose With regard to the overall insulin-stimulated glucose uptake and Ucp1 mRNA expression could be observed. Al- utilization, as reflected by the GIR during EHC (under our though this does not allow conclusions about cause-effect experimental conditions of suppressed hepatic glucose relationships, these data suggest that the amount of beige production), we observed that it was similarly increased adipocytes in WAT depots can determine the overall tissue by the treatment in Wistar and Lou/C rats. Given that glucose utilization rate. Such conclusion is supported by skeletal muscle glucose utilization was unaltered by the the appearance of multilocular beige cells in iWAT of treatment in the two groups, only BAT and WAT glucose CL-316243–treated Lou/C rats, as determined by histological uptake could be held responsible for the treatment- means. Whether these beige cells result from activation of induced increase in GIR. By considering the total mass existing adipocytes or from transdifferentiation of white of BAT and WAT, as well as the mean glucose uptake in into brown cells is an important issue that will need to be these tissues, it was possible to approximate the percent addressed in future work. contribution of each of these two tissues to the increased As schematized by Fig. 7, the molecular mechanisms GIR. Thus, in Wistar rats, both BAT (;30%) and WAT underlying the b3 agonist–induced increase in glucose (;60%) contributed to the GIR enhancement in response metabolism in iWAT of Lou/C rats seem to involve induc- to the treatment. In contrast, in Lou/C rats, the treatment tions of increased FA synthesis (de novo lipogenesis doubled BAT glucose uptake but decreased the BAT mass [DNL]) and lipolysis (increased Acc1 and Hsl, respec- by twofold. Therefore, in these animals, the stimulatory tively), as well as of enhanced FA b-oxidation (increased effect of the treatment on GIR could only be attributed Mcad and Lcad). Additionally, the marked increase in the to increased glucose uptake in WAT depots containing expression of Pepck in iWAT of Lou/C treated rats sug- beige cells. It should, however, be mentioned that these gests the presence of increased glyceroneogenesis. Simi- calculations may underestimate BAT glucose uptake, as lar data of gene expression responsible for enzymes the reduction in tissue weight may reflect loss of lipids involved in lipid metabolism in adipose tissues were due to increased thermogenic activity induced by the b3 reported in normal mice treated with CL-316243 for 7 agonist treatment. These data are in keeping with previously days (35). reported results showing that chronic stimulation of the diabetes.diabetesjournals.org Poher and Associates 3711 sympathetic nervous system stimulates glucose uptake Francophone du Diabète, Montpellier, France, 26–29 March 2013; the 1st Annual inWATandBATwithoutanyeffectinskeletalmuscles Meeting of DIABAT, Berlin, Germany, 6 October 2012; the 2nd Annual Meeting of (36). The predominant role of beige cells in WAT glucose DIABAT, Zurich, Switzerland, 18 October 2013; the 3rd Annual Meeting of DIABAT, uptake is further supported by the previous observation Nice, France, 9 October 2014; and the 21st Meeting of the Association G2L2, Geneva, Switzerland, 12 April 2013. that the increased glucose utilization under sympathetic stimulation depends on UCP1 activation (37,38). References Altogether, the results of the current study extend 1. World Health Organization. Obesity and Overweight. Fact Sheet No. 311 other data showing that chronic cold exposure or b3 fi Geneva, World Health Org., 2014 agonist treatment exerts bene cial effects on glucose 2. Cypess AM, Lehman S, Williams G, et al. Identification and importance of – metabolism in mice with high-fat diet induced obesity brown adipose tissue in adult humans. N Engl J Med 2009;360:1509–1517 (rev. in 12) and normalizes hyperglycemia and hyper- 3. Virtanen KA, Lidell ME, Orava J, et al. Functional brown adipose tissue in triglyceridemia in mouse models of diabetes and dysli- healthy adults. N Engl J Med 2009;360:1518–1525 pidemia (39–41). They also extend previous data (42), 4. van Marken Lichtenbelt WD, Vanhommerig JW, Smulders NM, et al. Cold- supporting the notion of interdependency between glu- activated brown adipose tissue in healthy men. N Engl J Med 2009;360:1500–1508 cose uptake, lipid metabolism, and thermogenesis. As 5. Garcia CA, Van Nostrand D, Majd M, et al. Benzodiazepine-resistant “brown recently further discussed (43), a particularly important fat” pattern in positron emission tomography: two case reports of resolution with – role for the futile cycling between FA synthesis and temperature control. Mol Imaging Biol 2004;6:368 372 6. Nedergaard J, Cannon B, Lindberg O. Microcalorimetry of isolated oxidation linked to thermogenesis and glucose disposal mammalian cells. Nature 1977;267:518–520 is played by DNL. This may be particularly relevant in 7. Matthias A, Ohlson KB, Fredriksson JM, Jacobsson A, Nedergaard J, beige cells of WAT depots, in keeping with the obser- Cannon B. Thermogenic responses in brown fat cells are fully UCP1-dependent. vation that WAT is required for the full thermogenic UCP2 or UCP3 do not substitute for UCP1 in adrenergically or fatty scid-induced response to CL-316243 (44). To go a step further in the thermogenesis. J Biol Chem 2000;275:25073–25081 understanding of the mechanisms involved in the 8. Cinti S. The adipose organ: morphological perspectives of adipose tissues. effects of b3 agonists, several data indicate that DNL Proc Nutr Soc 2001;60:319–328 is required for the synthesis of several signaling lipids 9. Petrovic N, Walden TB, Shabalina IG, Timmons JA, Cannon B, Nedergaard J. with systemic effects (“lipokines”) supporting metabolic Chronic peroxisome proliferator-activated receptor gamma (PPARgamma) acti- homeostasis. The presence of these molecules in our vation of epididymally derived white adipocyte cultures reveals a population of different experimental groups could therefore be of in- thermogenically competent, UCP1-containing adipocytes molecularly distinct from classic brown adipocytes. J Biol Chem 2010;285:7153–7164 terest and will be investigated in the future. 10. Long JZ, Svensson KJ, Tsai L, et al. A smooth muscle-like origin for beige To conclude, induction of DNL coupled with a simulta- adipocytes. Cell Metab 2014;19:810–820 neous increase in FA oxidation as occurs during browning 11. Cinti S. Adipocyte differentiation and transdifferentiation: plasticity of the – – of WAT (CL-316243 treated normal mice [35]; CL-316243 adipose organ. J Endocrinol Invest 2002;25:823–835 treated Lou/C rats [present study]) appears to be important 12. Peirce V, Vidal-Puig A. Regulation of glucose homoeostasis by brown adi- for the regulation of glucose homeostasis and may therefore pose tissue. Lancet Diabetes Endocrinol 2013;1:353–360 be of clinical relevance for the treatment of obesity and type 13. Chondronikola M, Volpi E, Børsheim E, et al. Brown adipose tissue improves 2diabetesinhumans. whole-body glucose homeostasis and insulin sensitivity in humans. Diabetes 2014;63:4089–4099 14. Gunawardana SC, Piston DW. Reversal of type 1 diabetes in mice by brown Acknowledgments. The authors thank Professor Claes Wollheim and adipose tissue transplant. Diabetes 2012;61:674–682 Professor Jean-Paul Giacobino for critical comments on the manuscript. 15. Poher AL, Altirriba J, Veyrat-Durebex C, Rohner-Jeanrenaud F. Brown ad- Funding. This study was supported by grants from the Swiss National Foun- ipose tissue activity as a target for the treatment of obesity/insulin resistance. dation (Bern, Switzerland), Fonds De La Recherche Scientifique (31003A_134919 Front Physiol 2015;6:4 and 310030_160290), and the 278373 DIABAT Collaborative Project from the 16. Veyrat-Durebex C, Montet X, Vinciguerra M, et al. The Lou/C rat: a model of European Community (Diabat/FP7_HEALTH_2011_TWO_STAGE). A.-L.P. was re- spontaneous food restriction associated with improved insulin sensitivity and cipient of a grant from Fondation d’Endocrinologie (Geneva, Switzerland). decreased lipid storage in adipose tissue. Am J Physiol Endocrinol Metab 2009; Duality of Interest. No potential conflicts of interest relevant to this article 296:E1120–E1132 were reported. 17. Veyrat-Durebex C, Poher AL, Caillon A, Montet X, Rohner-Jeanrenaud F. Author Contributions. A.-L.P. carried out the animal, qPCR, Western blot, Alterations in lipid metabolism and thermogenesis with emergence of brown and immunofluorescence studies and drafted the manuscript. A.-L.P., C.V.-D., and adipocytes in white adipose tissue in diet-induced obesity-resistant Lou/C rats. F.R.-J. conceived the study, participated in its design and coordination, and drafted Am J Physiol Endocrinol Metab 2011;300:E1146–E1157 the manuscript. C.V.-D., J.A., A.C., and J.L. helped with the animal treatments. J.A. 18. Veyrat-Durebex C, Poher AL, Caillon A, et al. Improved leptin sensitivity carried out biostatistical analysis and drafted the manuscript. X.M. performed CT as a potential candidate responsible for the spontaneous food restriction of the scan and its analysis. D.J.C. performed [18F]-FDG. J.L. participated in qPCR experi- Lou/C rat. PLoS One 2013;8:e73452 ments. All authors read and approved the final manuscript. F.R.-J. is the guarantor 19. Terrettaz J, Assimacopoulos-Jeannet F, Jeanrenaud B. Severe hepatic and of this work and, as such, had full access to all the data in the study and takes peripheral insulin resistance as evidenced by euglycemic clamps in genetically responsibility for the integrity of the data and the accuracy of the data analysis. obese fa/fa rats. Endocrinology 1986;118:674–678 Prior Presentation. Parts of this study were presented in abstract form at 20. Vettor R, Zarjevski N, Cusin I, Rohner-Jeanrenaud F, Jeanrenaud B. In- the 49th Annual Meeting of the European Association for the Study of Diabetes, duction and reversibility of an obesity syndrome by intracerebroventricular Barcelona, Spain, 23–27 September 2013; the 2013 Congress of Société neuropeptide Y administration to normal rats. Diabetologia 1994;37:1202–1208 3712 Ectopic UCP1 in Lou/C Rats Improves Metabolism Diabetes Volume 64, November 2015

21. Ferré P, Leturque A, Burnol AF, Penicaud L, Girard J. A method to quantify 34. Park JW, Jung KH, Lee JH, et al. 18F-FDG PET/CT monitoring of b3 agonist- glucose utilization in vivo in skeletal muscle and white adipose tissue of the stimulated brown adipocyte recruitment in white adipose tissue. J Nucl Med anaesthetized rat. Biochem J 1985;228:103–110 2015;56:153–158 22. James DE, Burleigh KM, Kraegen EW. In vivo glucose metabolism in indi- 35. Mottillo EP, Balasubramanian P, Lee YH, Weng C, Kershaw EE, Granneman vidual tissues of the rat. Interaction between epinephrine and insulin. J Biol Chem JG. Coupling of lipolysis and de novo lipogenesis in brown, beige, and white 1986;261:6366–6374 adipose tissues during chronic b3-adrenergic receptor activation. J Lipid Res 23. Cinti S, Cancello R, Zingaretti MC, et al. CL316,243 and cold stress induce 2014;55:2276–2286 heterogeneous expression of UCP1 mRNA and protein in rodent brown adipo- 36. Liu X, Pérusse F, Bukowiecki LJ. Chronic norepinephrine infusion stimulates cytes. J Histochem Cytochem 2002;50:21–31 glucose uptake in white and brown adipose tissues. Am J Physiol 1994;266: 24. Orava J, Nuutila P, Lidell ME, et al. Different metabolic responses of human R914–R920 brown adipose tissue to activation by cold and insulin. Cell Metab 2011;14:272–279 37. Cannon B, Nedergaard J. Brown adipose tissue: function and physiological 25. Dulloo AG, Seydoux J, Girardier L. Dissociation of enhanced efficiency of fat significance. Physiol Rev 2004;84:277–359 deposition during weight recovery from sympathetic control of thermogenesis. 38. Inokuma K, Ogura-Okamatsu Y, Toda C, Kimura K, Yamashita H, Saito M. Am J Physiol 1995;269:R365–R369 Uncoupling protein 1 is necessary for norepinephrine-induced glucose utilization 26. Belouze M, Sibille B, Rey B, et al. Leanness of Lou/C rats does not require in brown adipose tissue. Diabetes 2005;54:1385–1391 higher thermogenic capacity of brown adipose tissue. Physiol Behav 2011;104: 39. Arbeeny CM, Meyers DS, Hillyer DE, Bergquist KE. Metabolic alterations 893–899 associated with the antidiabetic effect of beta 3-adrenergic receptor agonists in 27. Loncar D, Afzelius BA, Cannon B. Epididymal white adipose tissue after cold obese mice. Am J Physiol 1995;268:E678–E684 stress in rats. II. Mitochondrial changes. J Ultrastruct Mol Struct Res 1988;101: 40. Connor SC, Hughes MG, Moore G, Lister CA, Smith SA. Antidiabetic efficacy 199–209 of BRL 49653, a potent orally active insulin sensitizing agent, assessed in the 28. Young P, Arch JR, Ashwell M. Brown adipose tissue in the parametrial fat C57BL/KsJ db/db diabetic mouse by non-invasive 1H NMR studies of urine. J pad of the mouse. FEBS Lett 1984;167:10–14 Pharm Pharmacol 1997;49:336–344 29. Shibata H, Pérusse F, Vallerand A, Bukowiecki LJ. Cold exposure reverses 41. Ghorbani M, Claus TH, Himms-Hagen J. Hypertrophy of brown adipocytes in inhibitory effects of fasting on peripheral glucose uptake in rats. Am J Physiol brown and white adipose tissues and reversal of diet-induced obesity in rats 1989;257:R96–R101 treated with a beta3-adrenoceptor agonist. Biochem Pharmacol 1997;54:121– 30. Vallerand AL, Pérusse F, Bukowiecki LJ. Stimulatory effects of cold expo- 131 sure and cold acclimation on glucose uptake in rat peripheral tissues. Am J 42. Dulloo AG, Gubler M, Montani JP, Seydoux J, Solinas G: Substrate cycling Physiol 1990;259:R1043–R1049 between de novo lipogenesis and lipid oxidation: a thermogenic mechanism 31. Schulz TJ, Tseng YH. Brown adipose tissue: development, metabolism and against skeletal muscle lipotoxicity and glucolipotoxicity. Int J Obes Relat Metab beyond. Biochem J 2013;453:167–178 Disord 2004;28(Suppl. 4):S29–S37 32. Seale P, Conroe HM, Estall J, et al. Prdm16 determines the thermogenic 43. Solinas G, Borén J, Dulloo AG. De novo lipogenesis in metabolic homeo- program of subcutaneous white adipose tissue in mice. J Clin Invest 2011;121: stasis: More friend than foe? Mol Metab 2015;4:367–377 96–105 44. Grujic D, Susulic VS, Harper ME, et al. Beta3-adrenergic receptors on white 33. Kajimura S, Seale P, Tomaru T, et al. Regulation of the brown and white fat and brown adipocytes mediate beta3-selective agonist-induced effects on energy gene programs through a PRDM16/CtBP transcriptional complex. Genes Dev expenditure, insulin secretion, and food intake. A study using transgenic and 2008;22:1397–1409 gene knockout mice. J Biol Chem 1997;272:17686–17693