THE JOURNAL • RESEARCH • www.fasebj.org Calorie restriction prevents diet-induced insulin resistance independently of PGC-1-driven mitochondrial biogenesis in white adipose tissue Rosario Pardo,*,1 Maria Vilà,*,1 Luis Cervela,* Marina de Marco,* Pau Gama-Pérez,† Alba González-Franquesa,†,2 Lucia Statuto,* Ramon Vilallonga,*,‡ Rafael Simó,§,{ Pablo M. Garcia-Roves,† and Josep A. Villena*,{,3 *Laboratory of Metabolism and Obesity and §Unit of Diabetes and Metabolism, Vall d’Hebron-Institut de Recerca, Universitat Autònoma de Barcelona, Barcelona, Spain; †Department of Physiological Sciences, Institut d’Investigació Biomèdica de Bellvitge, Universitat de Barcelona, L’Hospitalet de Llobregat, Barcelona, Spain; ‡Metabolic and Bariatric Surgery Unit, European Accreditation Council for Bariatric Surgery Center of Excellence, Vall d’Hebron University Hospital, Barcelona, Spain; and {Centro de Investigación Biomédica en Red de Diabetes y Enfermedades Metabólicas Asociadas, Instituto de Salud Carlos III, Barcelona, Spain ABSTRACT: Calorie restriction (CR) exerts remarkable, beneficial effects on glucose homeostasis by mechanisms that are not fully understood. Given the relevance of white adipose tissue (WAT) in glucose homeostasis, we aimed at identifying the main cellular processes regulated in WAT in response to CR in a pathologic context of obesity. For this, a gene-expression profiling study was first conducted in mice fed ad libitum or subjected to 40% CR. We found that the gene network related to mitochondria was the most highly upregulated in WAT by CR. To study the role that increased mitochondrial biogenesis plays on glucose homeostasis following CR, we generated a mouse model devoid of the coactivators peroxisome proliferator-activated receptor g coactivator 1 (PGC-1)a and PGC-1b spe- cifically in adipocytes. Our results show that mice lacking PGC-1s in adipocytes are unable to increase mitochondrial biogenesis in WAT upon CR. Despite a blunted induction of mitochondrial biogenesis in response to calorie deprivation, mice lacking adipose PGC-1s still respond to CR by improving their glucose homeostasis. Our study demonstrates that PGC-1 coactivators are major regulators of CR-induced mitochondrial biogenesis in WAT and that increased mitochondrial biogenesis and oxidative function in adipose tissue are not required for the im- provement of glucose homeostasis mediated by CR.—Pardo, R., Vilà, M., Cervela, L., de Marco, M., Gama-Pérez, P., González-Franquesa, A., Statuto, L., Vilallonga, R., Simó, R., Garcia-Roves, P. M., Villena, J. A. Calorie restriction prevents diet-induced insulin resistance independently of PGC-1-driven mitochondrial biogenesis in white adipose tissue. FASEB J. 33, 2343–2358 (2019). www.fasebj.org KEY WORDS: oxidative metabolism • adipocytes • glucose homeostasis • mitochondria Dysregulation of white adipose tissue (WAT) function, as These metabolic alterations lie beneath the development of it occurs in obesity, is associated with a variety of meta- severe obesity-associated comorbidities, such as cardio- bolic derangements, including insulin resistance, glucose vascular disease, certain types of cancer, and type 2 dia- intolerance, hyperlipidemia, and hepatic steatosis (1). betes (T2D), which result in increased morbidity and ABBREVIATIONS: ACO2, aconitase 2; AKT, protein kinase B; AL, ad libitum; Arg1, arginase 1; ATP5b, mitochondrial ATP synthetase b subunit; BAT, brown adipose tissue; CIDEA, cell death-inducing DNA fragmentation factor a-like effector A; COL, collagen; COXII, cytochrome C oxidase subunit II; COXIV, cytochrome C oxidase subunit IV; CR, calorie restriction; CYPA, cyclophilin A; ECM, extracellular matrix; ETS, electron-transfer system; FCCP, carbonylcyanide-4-(trifluoromethoxy)-phenyl-hydrazone; GO, gene ontology; GTT, glucose tolerance test; HF, high fat; ITT, insulin tolerance test; M1, classically activated macrophages; M2, alternatively activated macrophages; mtDNA, mitochondrial DNA; NDUFB9, NADH:ubiquinone oxidore- ductase subunit B9; oxphos, oxidative phosphorylation; PGC-1, peroxisome proliferator-activated receptor g coactivator 1; PGC-1a/b-FAT-DKO, peroxisome proliferator-activated receptor g coactivator 1a/b double-knockout mice; Ppargc, peroxisome proliferator-activated receptor g coactivator; qRT-PCR, quantitative RT-PCR; SDHB, succinate dehydrogenase complex subunit B; Sirt1, Sirtuin 1; T2D, type 2 diabetes; TZD, thiazolidinedione; UCP1, uncoupling protein 1; WAT, white adipose tissue; WT, wild type 1 These authors contributed equally to this work. 2 Current affiliation: The Novo Nordisk Center for Basic Metabolic Research, Københavns Universitet, Copenhagen, Denmark. 3 Correspondence: Laboratory of Metabolism and Obesity, Unit of Diabetes and Metabolism, Vall d’Hebron-Institut de Recerca, Passeig Vall d’Hebron, 119-129, Barcelona 08035, Spain. E-mail: josep.villena@vhir.org doi: 10.1096/fj.201800310R This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information. 0892-6638/19/0033-2343 © FASEB 2343 mortality among obese individuals. In this regard, the process in response to CR. With the consideration of the prevalence of T2D has grown steadily in close association tight correlation between mitochondrial mass and insulin with the rise of obesity, and currently, T2D has become a sensitivity reported in humans and animal models major health problem of epidemic proportions (2). [reviewed in Zamora and Villena (8)], we aimed at in- As a key endocrine and lipid storage organ, WAT plays vestigating first, the mechanisms that regulate mito- a central role in the control of glucose homeostasis. In an chondrial biogenesis in response to CR and second, obesogenic environment, reduced energy expenditure and whether the increase in mitochondrial mass and oxida- excessive calorie intake favor the accumulation of triglyc- tive function in WAT is necessary for the improvement in erides in WAT, ultimately leading to hypertrophic and glucose homeostasis exerted by CR. dysfunctional adipocytes. Several cellular processes in WAT contribute to adipocyte-specific and whole-body insulin resistance. For example, sustained nutrient over- MATERIALS AND METHODS load of adipocytes in obese subjects results in an altered endocrine function and the inability of fat cells to store Animal studies lipids further in an efficient manner, leading to ectopic lipid deposition and impaired insulin signaling (3). In- Adipose-specific peroxisome proliferator-activated receptor terestingly, mitochondrial dysfunction has also been sug- g coactivator 1 (PGC-1)a/PGC-1b double-knockout mice gested to contribute to insulin resistance by favoring the (PGC-1a/b-FAT-DKO) were obtained by crossing double- homozygous peroxisome proliferator-activated receptor g coac- accumulation of lipid intermediates that disrupt proper tivator (Ppargc)1aflox/flox, Ppargc1bflox/flox mice with transgenic insulin signaling (4). In addition, activation in hypertro- mice expressing Cre recombinase under the control of the phic adipocytes of stress signaling pathways, such as en- adipocyte-specific adiponectin promoter (AdipoQ-Cre) (9). doplasmic reticulum stress, enhanced reactive oxygen The generation of Ppargc1aflox/flox and Ppargc1bflox/flox mice has species production, or hypoxic responses, results in in- been previously described, and mice had been backcrossed creased cell death and triggers a chronic low-degree in- for at least 10 generations with C57BL6/J mice to ensure a flammatory response in WAT characterized by increased homogeneous genetic background (10, 11). Mice were housed in a temperature-controlled environment at 21°C, subjected recruitment of immune cells, enhanced macrophage po- to a 12/12 h light/dark cycle and fed a standard diet larization toward the classically activated macrophages (2018 Tecklad Global 18% Protein Rodent Diet; Harlan Labo- (M1) proinflammatory type, and augmented secretion of ratories, Indianapolis, IN, USA), unless otherwise stated. proinflammatory cytokines (5). The secretion of in- In all experiments, Cre-negative PGC-1a/b-FAT-DKO litter- flammatory cytokines (e.g., TNF-a, IL-1b, IL-6) by both mates were used as wild-type (WT) controls. For the gene- adipocytes and infiltrated macrophages leads to the acti- expression profiling study, WT C57BL6/J male mice were used. vation of the JNK and IKK-NF-kB pathways that directly For the CR experiments, all mice were individually caged and impair insulin signal transduction and cause insulin re- started receiving a high-fat (HF) diet (45% kcal fat, 35% kcal sistance (6). Although numerous studies have provided carbohydrates, 20% kcal protein; Research Diets, New Brunswick, compelling evidence that supports a role of adipose lip- NJ, USA) at the age of 7 wk. At the age of 8 wk, mice were otoxicity and inflammation in the pathogenesis of insulin randomly assigned to ad libitum (AL) or CR groups. During the resistance, the contribution of other processes, such as im- following week, CR animals received 80% of the food eaten by paired mitochondrial function, remains to be elucidated. their AL counterparts. Afterwards, CR mice started receiving 60% Nutritional interventions have proven very efficient of the food eaten by AL animals. Mice were subjected to 40% CR for a period of 12 wk. in reducing excessive fat accumulation and reverting All procedures involving animals were conducted according obesity-associated diseases, including insulin resistance. to the European Union Ethical Guidelines and approved by the Calorie restriction (CR), defined as a 20–50% reduction in Animal Experimentation and Ethics Committee of Vall d’Hebron calorie intake, is the only lifestyle intervention known to Research Institute (12/11 CEEA). extend lifespan in a wide variety of species, from yeast to primates (7). The mechanisms by which CR delays aging are not fully understood, but the remarkable effects that Microarray analysis CR has on health—delaying the onset and progression of numerous pathologies, including insulin resistance and Gene-expression profiling in WAT of AL and CR mice was per- T2D—may contribute to extend a lifespan. In this regard, formed as previously described (12). In brief, total RNA was given the difficulty of patients to adhere to a long-term CR isolated from inguinal WAT of C57BL6/J WT mice fed AL or subjected to CR (n = 5) with Trizol reagent, further purified us- regime, the identification of the mechanisms and processes ing the RNeasy Mini kit (Qiagen, Germantown, MD, USA) and that mediate the beneficial effects of CR is crucial for the then used to synthesize sense single-stranded DNA with the identification of new target molecules and the develop- Ambion WT Expression kit (Thermo Fisher Scientific, Waltham, ment of new therapies based on drugs that mimic CR. MA, USA), following the manufacturer’s instructions. Labeled However, the key processes modulated by CR in WAT single-stranded DNA was hybridized onto Mouse Gene 1.0 ST have not been fully characterized, and their relevance with arrays (Thermo Fisher Scientific). The arrays were scanned, and regard to the control of glucose homeostasis remains to be the resulting images were processed with the Microarray Anal- ysis Suite 5.0 software package (Thermo Fisher Scientific). The defined. data obtained were analyzed by the Unit of Statistics and Bio- In this study, we have performed a gene-expression informatics from the Vall d’Hebron Research Institute (Barcelona, profiling analysis in WAT of mice subjected to CR and Spain) using Bioconductor software (https://www.bioconductor. identified mitochondrial biogenesis as the most upregulated org/) to generate a list of genes differentially expressed in WAT of 2344 Vol. 33 February 2019 The FASEB Journal x www.fasebj.org PARDO ET AL. CR mice. A gene-enrichment analysis was performed using ADP and the presence of 2 mM malate and 10 mM pyruvate the Database for Annotation, Visualization and Integrated (PML) in figures]. Oxidative phosphorylation (oxphos) capacity, Discovery (DAVID) Bioinformatics Database functional an- OXPHOSP, was quantified by the addition of ADP (5 mM), notation tool (https://david.ncifcrf.gov/summary.jsp) (13) and the followed by additions of glutamate (10 mM) for additional Gene Set Enrichment Analysis (GSEA) method (http://software. oxidation of reduced substrates through complex I [oxphos broadinstitute.org/gsea/index.jsp) (14) to identify the main bi- capacity in the presence of 5 mM ADP and 10 mM glutamate ologic processes and gene networks regulated by CR in WAT. (PMGP)] and succinate (complex II linked; 10 mM) for con- vergent electron flow through both NADH-linked and succinate-linked pathways [oxphos capacity in the presence of Real-time quantitative RT-PCR 5 mM ADP, 10 mM glutamate, and 10 mM succinate (PMGSP)]. Subsequently, carbonylcyanide-4-(trifluoromethoxy)-phenyl- Total RNA was used to synthesize cDNA with SuperScript II RT hydrazone (FCCP) was titrated to achieve maximum flux (Thermo Fisher Scientific) and oligo-dT. Gene expression was through the electron-transfer system (ETSE; respiratory state determined by real-time quantitative RT-PCR (qRT-PCR) using determined in the presence of FCCP), fueled by the oxidation of SYBR green dye and gene-specific primers as described in Pardo substrates through complex CI and CII in the presence of et al. (11). 0.5 mM FCCP [(PMGSE); 0.5 mM]. Finally, we inhibited electron transport through complex I [maximum flux through the ETS fueled by the oxidation of substrates through CII in the pres- Western blot ence of 0.5 mM FCCP (SE)] and III by the sequential addi- tion of rotenone (0.5 mM) and antimycin A (2.5 mM), respectively. The remaining oxygen flux after inhibition with antimycin Protein extracts from inguinal WAT or interscapular brown ad- A (O2 flux independent of ETS) was subtracted from the ipose tissue (BAT) were prepared in homogenization buffer values of each of the previous steps. Oxygen flux values (50 mM Tris-HCl, 150 mM NaCl, 1% Nonidet P-40, 5 mM EGTA, were expressed relative to tissue wet weight per second 5 mM EDTA, 1 mM PMSF) containing protease inhibitors. WAT (JO2, pmol O2/mg/s). or BAT protein extracts (15–30 mg) were subjected to electro- phoresis in a 12% or 15% SDS-PAGE and transferred to a PVDF membrane. Immunodetection was performed with specific an- tibodies against aconitase 2 (ACO2; ab71440, 1:1000; Abcam, Serological analysis Cambridge, United Kingdom), cytochrome C oxidase subunit IV (COXIV; 4844, 1:1000; Cell Signaling Technology, Danvers, Serum was obtained from blood of PGC-1a/b-FAT-DKO mice MA, USA), succinate dehydrogenase complex subunit B (SDHB; (food withheld for 5h) and WT littermates either fed AL or sub- ab84622, 1:1000; Abcam), NADH:ubiquinone oxidoreductase jected to CR. Insulin was assayed with an ultra-sensitive mouse subunit B9 (NDUFB9; ab106699, 1:1000; Abcam), mitochondrial insulin ELISA kit (Crystal Chem, Zaandam, The Netherlands). ATP synthetase b subunit (ATP5b; ab85068, 1:1000; Abcam), Free fatty acids were determined with the NEFA-C kit (Wako a-tubulin (11H10, 1:1000; Cell Signaling Technology), and Chemicals, Neuss, Germany). Triglycerides and total cholesterol cyclophilin A (CYPA; BML-SA296, 1:5000; Enzo Life Sciences, were measured using colorimetric commercial kits (FAR Diag- Farmingdale, NY, USA). For detection of uncoupling protein nostics, Pescantina, Verona, Italy). 1 (UCP1; ab10983, 1:3000 Abcam), protein extracts (50 mg) from inguinal WAT were used. Horseradish peroxidase-conjugated goat anti-rabbit-purified IgG was used as a secondary antibody Glucose and insulin tolerance tests (172-1013, 1:10,000; Bio-Rad, Hercules, CA, USA). Glucose tolerance tests (GTTs) and insulin tolerance tests (ITTs) were performed as detailed in Enguix et al. (10). For GTTs, mice Mitochondrial DNA content first had food withheld for 12 h. After an intraperitoneal in- jection of 2 g/kg glucose, glycemia was determined at 0, 15, Relative quantification of mitochondrial DNA (mtDNA) in 30, 60, 90, and 120 min, using a Contour XT glucometer (Bayer, WAT was performed by qRT-PCR using 2 ng total DNA as a Leverkusen, Germany). ITTs were performed on mice that template and specific primers to amplify cytochrome C oxi- had food withheld for 5h and were injected with 0.9 U/kg dase subunit II (COXII; mtDNA) and receptor-interacting insulin, and blood glucose levels were determined at the same protein 140 (nuclear DNA). The detailed procedure has been time intervals specified for the GTT. described in Villena et al. (15). Analysis of insulin signaling High-resolution respirometry After withholding food overnight, mice were first anesthetized Mitochondrial respiration was assessed in freshly isolated WAT and then intravenously injected with 1 U/kg insulin. After 3 min, by high-resolution respirometry in an Oroboros Oxygraph-2k samples from WAT, liver, and skeletal muscle (gastrocnemius/ system (Oroboros Instruments, Innsbruck, Austria) (16). In brief, soleus) were obtained, snap frozen, and stored at 280°C for WAT was first mechanically permeabilized in respiratory me- subsequent analysis. Samples from the same tissues were dium (0.5 mM EGTA, 3 mM MgCl2, 60 mM K-lactobionate, obtained before the injection of insulin and were used as non- 10 mM KH2PO4, 20 mM HEPES, and 110 mM sucrose, pH 7.1) stimulated controls. Protein extracts from tissues were obtained using a PBI-Shredder SG3 (Pressure BioScience, South Easton, in RIPA buffer (50 mM Tris, pH = 8.0, 150 mM NaCl, 1% Nonidet MA, USA) in the presence of digitonin (8 mM). The equivalent of P-40, 0.5% sodium deoxycholate, 0.1% SDS) containing a pro- 80 mg wet tissue was added to the experimental chamber con- tease and phosphatase inhibitor cocktail. Protein extracts taining respiration media (described above), supplemented with (15–30 mg) were subjected to SDS-PAGE and immunoblotted 1 g/L fatty acid-free bovine serum albumin. LEAKL respiration with specific antibodies against total protein kinase B (AKT; was measured by the addition of NADH-linked substrates 9272, 1:1000; Cell Signaling Technology), AKT phosphorylated (complex I linked) malate (2 mM) and pyruvate (10 mM) in the at residue Thr308 (2965, 1:1000; Cell Signaling Technology), and absence of ADP [denoted as leak respiration in the absence of CYPA (BML-SA296, 1:5000; Enzo Life Sciences). CALORIE RESTRICTION AND ADIPOSE MITOCHONDRIA 2345 Statistical analysis diet rapidly gained weight and became obese, CR mice fed the same diet remained lean throughout the exper- Data are presented as means 6 SEM. Statistical significance was imental period (Fig. 1B). As expected, AL mice became assessed using an unpaired Student’s t test or an ANOVA, fol- glucose intolerant, whereas CR mice were protected lowed by Tukey post hoc test. from the deleterious effects of a HF diet on glucose metabolism (Fig. 1C). To gain insight into the potential mechanisms and bi- RESULTS ologic pathways affected by CR in WAT, we conducted a gene-expression profiling study in inguinal WAT of AL Effects of CR on gene expression in WAT and CR mice using DNA microarrays. We found 3740 genes differentially expressed in WAT of CR mice (P , To identify the main gene networks regulated by CR in 0.05), 1494 of which were upregulated and 2244 down- WAT, we subjected mice to a 40% CR regime over a regulated. Gene-enrichment analyses revealed that the 12- wk period (Fig. 1A). Whereas mice fed AL with a HF gene network related to mitochondria was the most highly over-represented among the upregulated genes (Fig. 2A and Table 1). On the other hand, gene ontology (GO) categories related to extracellular matrix (ECM) were highly over-represented among the genes downregulated in WAT in response to CR (Fig. 3A, left, and Table 1), al- though with lower statistical significance, we also found that GO categories related to the immune system (i.e., in- flammatory response, P = 2.8E29; chemotaxis, P = 9.4E27; wound healing, P = 1.8E25) were significantly over- represented among the genes downregulated by CR (Fig. 3B, left). qRT-PCR confirmed that mRNA expression of genes related to mitochondrial function, including genes of the oxphos system, the tricarboxylic acid cycle, and fatty acid oxidation, was highly increased in response to CR (Fig. 2B). The mRNA increase correlated with elevated levels of mitochondrial proteins (Fig. 2C). Such increase in the relative amount of mitochondrial proteins (mitochondrial protein content/protein unit of WAT) translates into a net increase in the absolute content of mitochondrial proteins in the inguinal WAT depot as a whole (mitochondrial protein content/ WAT depot), despite the reduction in WAT mass im- posed by CR (Supplemental Fig. S1A). Consequently, even if WAT mass is reduced in calorie-restricted mice, the increase in absolute mitochondrial protein content makes WAT a potential net contributor to oxidative metabolism in response to CR. Furthermore, we also found that CR increased mtDNA content by almost 3-fold (Fig. 2D), indicating that CR-induced mitochon- drial gene expression was accompanied by an increase of mitochondrial mass in WAT. Moreover, CR also caused a severe reduction in the expression of genes encoding for structural proteins of the ECM, such as collagens, or genes involved in the remodeling of the ECM, such as metallopeptidases (Fig. 3A, right). Fur- thermore, a generalized decreased in the expression of genes related to the immune system was observed Figure 1. CR prevents the development of diet-induced obesity in WAT of calorie-restricted mice (Fig. 3B, right). and glucose intolerance. At the age of 7 wk, individually caged These include both proinflammatory (Il-6, Ccl2, Ccl7) WT C57BL6/J mice started receiving a HF diet (45% kcal from and anti-inflammatory genes (Il-13, Il-33), as well fat), and 1 wk later, they were randomly assigned to AL or CR as markers of inflammatory M1 macrophages (Cd11c, groups. CR mice received 60% of the food eaten by AL animals Nos2) and anti-inflammatory M2 macrophages [arginase 1 for a period of 12 wk. A) Food intake of AL and CR mice. B) Body weight was measured weekly throughout the experimental (Arg1)]. period. C ) GTTs (2 g/kg) were performed on AL and CR mice Increased mitochondrial mass and function have that had food withheld for 12 h. Results are expressed as means 6 been proposed as main determinants of whole-body SEM (n = 10–12 animals/group). *P , 0.05, **P , 0.01. insulin sensitivity, but the role that mitochondria play 2346 Vol. 33 February 2019 The FASEB Journal x www.fasebj.org PARDO ET AL. Figure 2. CR induces mitochondrial biogenesis in WAT. A) Gene-set enrichment analysis identified genes related to mitochondria as being differentially upregulated in inguinal WAT of WT mice subjected to 40% CR. B) mRNA expression of mitochondrial genes in inguinal WAT was assessed by qRT-PCR. C ) Western blot of protein extracts from inguinal WAT of mice fed AL or subjected to 40% CR. The results obtained from 2 independent cohorts of mice are shown. Densitometric quantification of Western blots from the 2 cohorts of mice (n = 9 animals/group) is depicted (right). D) mtDNA content was assessed by qRT-PCR in total DNA isolated from inguinal WAT. E ) mRNA levels of transcriptional coactivators PGC-1a and PGC- 1b were assessed by qRT-PCR. Results are expressed as means 6 SEM (n = 5–6 animals/group). Acadl, acyl-CoA dehydrogenase long chain; Acadvl, acyl-CoA dehydrogenase very long chain; Cpt1b, carnitine palmitoyltransferase 1B; Uqcr10, ubiquinol- cytochrome C reductase, complex III subunit X. *P , 0.05, **P , 0.01. in the control of glucose homeostasis is still a matter the induction of mitochondrial biogenesis in response to of debate. For this reason, among other cellular CR. For this, we generated a new knockout mouse model processes regulated by CR, we decided to investigate simultaneously devoid of PGC-1a and PGC-1b in adipo- the contribution of mitochondria to the metabolic cytes (PGC-1a/b-FAT-DKO). The disruption of Ppargc1a effects of CR. and Ppargc1b genes resulted in an ;75–80% decrease in the expression of PGC-1a and PGC-1b mRNA in WAT and a .95% reduction in BAT (Fig. 4A). PGC-1a/b-FAT-DKO PGC-1 coactivators regulate CR-induced mice exhibited a 30–50% reduction in the basal mRNA mitochondrial biogenesis in WAT levels of mitochondrial genes in WAT (Fig. 4B) that cor- related with reduced levels of mitochondrial proteins The coordinated induction of mitochondrial genes in re- (Fig. 4C). We next examined by high-resolution respi- sponse to CR was associated with an increase in the ex- rometry how mitochondrial respiration in WAT was af- pression of PGC-1a and PGC-1b coactivators (Fig. 2E). fected by the lack of PGC-1s (Fig. 4D). We found that Because PGC-1s appears to regulate mitochondrial bio- adipose PGC-1a/b-FAT-DKO mitochondria exhibited a genesis redundantly (17), we first aimed at investigating to reduction in the electron flux through complex I (PMGP) in which extent both types of PGC-1 are indeed required for the coupled state, when pyruvate/glutamate were used as CALORIE RESTRICTION AND ADIPOSE MITOCHONDRIA 2347 TABLE 1. Gene-enrichment analysis of differentially expressed genes in WAT of mice subjected to 40% CR was performed using the DAVID Bioinformatics Database functional annotation tool Ontology GO ID GO term Genes (n) P (,1.0E210) Upregulated genes CC GO:0005739 Mitochondrion 250 5.5E2102 CC GO:0005743 Mitochondrial inner membrane 110 3.4E272 BP GO:0055114 Oxidation-reduction process 109 3.6E243 BP GO:0008152 Metabolic process 77 4.3E231 MF GO:0016491 Oxidoreductase activity 84 2.1E228 BP GO:0006629 Lipid metabolic process 73 3.3E228 CC GO:0070469 Respiratory chain 30 3.6E228 BP GO:0006631 Fatty acid metabolic process 42 2.9E225 CC GO:0005759 Mitochondrial matrix 40 7.4E221 CC GO:0005747 Mitochondrial respiratory chain 22 1.9E219 complex I BP GO:0006099 Tricarboxylic acid cycle 16 3.9E215 MF GO:0051287 NAD binding 17 2.2E211 Downregulated genes CC GO:007006 2 Extracellular exosome 343 4.4E245 CC GO:0031012 ECM 84 1.8E233 CC GO:0005578 Proteinaceous ECM 81 9.6E229 CC GO:0005576 Extracellular region 217 4.4E225 CC GO:0005615 Extracellular space 193 4.8E224 BP GO:0007155 Cell adhesion 88 1.6E219 BP GO:0001525 Angiogenesis 55 5.6E217 MF GO:0005515 Protein binding 366 4.3E215 MF GO:0008201 Heparin binding 39 7.1E214 BP GO:0030335 Positive regulation of cell migration 45 2.6E213 MF GO:0005178 Integrin binding 30 1.4E212 MF GO:0005509 Calcium ion binding 94 2.5E212 MF GO:0005518 Collagen binding 22 2.2E211 BP, biologic process; CC, cellular component; MF, molecular function. substrates in the presence of malate and ADP. Sequential In addition, we found that CR enhanced WAT oxida- addition of succinate allowed us to determine conver- tive function in WT mice, an effect that was impaired in gent electron flux through NADH- and succinate-linked PGC-1a/b-FAT-DKO mice (Fig. 5E). pathways (PMGSP), which was also impaired in WAT of PGC-1a/b-FAT-DKO mice. Likewise, maximum elec- tron transport capacity with substrates fueling the ETS Basal mitochondrial biogenesis and Ucp1 through ETS complexes CI and CII (PMGSE) or only expression are regulated by PGC-1 through CII after inhibiting CI with rotenone (SE) was re- coactivators in BAT but not by CR duced in WAT of PGC-1a/b-FAT-DKO mice. Altogether, these results support a crucial role for PGC-1 coactivators Consistent with the well-known role of PGC-1 coactivators in the maintenance of basal mitochondrial biogenesis in the thermogenic function of brown adipocytes (18), BAT and oxidative function in WAT. of PGC-1a/b-FAT-DKO mice showed an increase in the Given the fact that CR-induced mitochondrial bio- accumulation of triglycerides, a clear indication of poor genesis was paralleled by an increase in the expression of thermogenic activity (Fig. 6A). As expected, lack of PGC-1 PGC-1a and PGC-1b, we next investigated whether si- coactivators in brown adipocytes resulted in a severe de- multaneous lack of both coactivators could affect the crease in the expression of genes encoding for mitochon- process of mitochondrial biogenesis in response to CR. For drial proteins, including Ucp1 (Fig. 6B). The expression this, WT and PGC-1a/b-FAT-DKO mice were either fed levels of mRNA correlated with lower levels of mito- AL with a HF diet or subjected to a 40% CR for 12 wk. chondrial proteins (Fig. 6C). However, unlike in WAT, we PGC-1a and PGC-1b expression was confirmed to in- did not observed any induction by CR of the expression of crease in WAT of WT mice in response to CR (Fig. 5A) and genes encoding for PGC-1a, PGC-1b, or mitochondrial correlated with an increase in the expression of mito- genes, with the exception of CoxIV, in which its expression chondrial genes, both at the mRNA and protein levels was found to be mildly but significantly increased in re- (Fig. 5B, C and Supplemental Fig. S1B). However, en- sponse to food deprivation (Fig. 6D). Furthermore, Ucp1 hanced mitochondrial gene expression in response to mRNA levels were not affected by CR, although its basal CR was severely blunted in WAT of PGC-1a/b-FAT- expression was entirely dependent on PGC-1 coactivators. DKO mice (Fig. 5B, C and Supplemental Fig. S1B). Similar findings were observed at the protein levels, with Likewise, the increase in mtDNA content on CR was mitochondrial proteins, including UCP1, not being af- impaired in mice lacking PGC-1 coactivators (Fig. 5D). fected or even mildly reduced by CR (Fig. 6E and 2348 Vol. 33 February 2019 The FASEB Journal x www.fasebj.org PARDO ET AL. Figure 3. CR reduces the expression of genes related to ECM remodeling and inflammation. Genes related to ECM (A) and the inflammatory response (B) were found to be downregulated in WAT of mice subjected to CR. Gene expression was validated by qRT-PCR. Results are expressed as means 6 SEM (n = 5–6 animals/group). Adam12, ADAM metallopeptidase domain 12; Adamts5, ADAM metallopeptidase with thrombospondin type 1 motif 5; Col, collagen; Mmp19, matrix metalloproteinase 19. *P , 0.05, **P , 0.01. Supplemental Fig. S2). Altogether, these data suggest that detectable browning of WAT does not occur in re- that basal mitochondrial biogenesis in BAT is highly sponse to CR. dependent on PGC-1 coactivators but that brown adi- pocyte mitochondrial gene expression and thermogenic activity are not enhanced by CR, contrary to what we CR-induced mitochondrial biogenesis is not required for the improvement in glucose observed in WAT. homeostasis exerted by CR CR does not induce browning of WAT Increased mitochondrial biogenesis in WAT has been suggested to be instrumental in the amelioration of insulin It has been recently reported that CR induces the re- resistance by certain pharmacological or lifestyle inter- cruitment of brown adipocytes within WAT depots. It has ventions [reviewed in Zamora and Villena (8)]. The dra- also been suggested that such a browning process could matic induction of mitochondrial biogenesis in response contribute to the positive effects that CR has on glucose to CR suggests that enhanced mitochondrial oxidative homeostasis (19). In WT mice, despite careful examination function in WAT could be required by CR to exert its of histologic preparations of WAT, no cells with the char- beneficial effects on glucose homeostasis. acteristic morphology of brown adipocytes were found in PGC-1a/b-FAT-DKO mice did not exhibit differences mice subjected to CR (Fig. 7A). Nevertheless, at the mo- in body weight when either fed a regular chow diet (data lecular level, a very modest and variable increase in the not shown) or a HF diet (Fig. 8A), neither did they show Ucp1 and cell death-inducing DNA fragmentation factor differences in the weight of major organs, including dif- a-like effector A (Cidea) but not PR domain zinc finger ferent depots of WAT, BAT, liver, and gastrocnemius protein 16 (Prdm16) mRNA levels was observed in WAT of muscle (Fig. 8B). In WT and PGC-1a/b-FAT-DKO mice, CR mice (Fig. 7B). The mild induction of Ucp1 and Cidea CR intervention similarly reduced the weight of BAT, mRNA was completely blunted in PGC-1a/b-FAT-DKO liver, and all depots of WAT with the exception of epi- mice (Fig. 7C). Importantly, despite the moderate increase didymal WAT, which remained almost unaffected by CR in mRNA levels, we were unable to detect UCP1 protein (Fig. 8B). No differences were found in the gastrocnemius in WAT of WT and PGC-1a/b-FAT-DKO mice, irre- muscle, indicating that lean mass was not affected by the spective of the nutritional treatment (Fig. 7D), indicating lack of PGC-1s or by the nutritional intervention. CALORIE RESTRICTION AND ADIPOSE MITOCHONDRIA 2349 Figure 4. Adipose-specific lack of PGC-1a and PGC-1b coactivators leads to reduced mitochondrial gene expression and reduced mitochondrial respiration in WAT. A) mRNA expression of PGC-1a (left) and PGC-1b (right) in tissues of WT and PGC-1a/b-FAT-DKO mice was analyzed by qRT-PCR (n = 4–5 animals/group). B) Mitochondrial gene expression in inguinal WAT of WT and PGC-1a/b-FAT-DKO mice analyzed by qRT-PCR (n = 10–12 animals/group). C ) Western blots of mitochondrial proteins in inguinal WAT of WT and PGC-1a/b-FAT-DKO mice (left). Densitometric quantification of blots (right). D) Mitochondrial respiration in inguinal WAT was assessed by high-resolution respirometry (PML, PMGP, PMGSP, PMGSE, and SE). Results are expressed as means 6 SEM (n = 5 animals/group). JO2, oxygen flux value; Nd2, NADH dehydrogenase subunit 2. *P , 0.05, **P , 0.01. AL feeding with a HF diet induced severe glucose in- Consistent with a healthier metabolic status, both WT and tolerance and insulin resistance in both WT and PGC-1a/ PGC-1a/b-FAT-DKO mice exhibited a similar reduction b-FAT-DKO mice (Fig. 9A, D). Although it has been in fasting glucose, insulin, and lipid levels when subjected speculated that impaired WAT mitochondrial function to CR (Table 2). could lead to the development of insulin resistance, we did Although whole-body glucose homeostasis did not not observe significant differences in insulin sensitivity appear altered in PGC-1a/b-FAT-DKO mice, it is possible between WT and PGC-1a/b-FAT-DKO mice fed AL that impaired mitochondrial function in adipocytes could when an ITT was performed (Fig. 9D, F). Likewise, no lead to adipose tissue-specific insulin resistance. However, significant differences were found between WT and PGC- as shown in Fig. 9G, insulin administration led to a similar 1a/b-FAT-DKO mice fed AL in their capacity to clear increase in AKT phosphorylation in WAT of both WT glucose from circulation during an intraperitoneal GTT and PGC-1a/b-FAT-DKO mice, indicating that impaired (Fig. 9A, C). When subjected to CR, both WT and PGC-1a/ mitochondrial biogenesis and oxidative function do not b-FAT-DKO mice improved whole-body glucose toler- impair insulin signaling in WAT. Moreover, in agreement ance and insulin sensitivity to the same extent (Fig. 9A–F). with the general improvement in insulin sensitivity 2350 Vol. 33 February 2019 The FASEB Journal x www.fasebj.org PARDO ET AL. Figure 5. CR-induced mitochondrial biogenesis in WAT is dependent of PGC-1 coactivators. A, B) mRNA expression of PGC-1a and PGC-1b coactivators (A) and genes encoding for mitochondrial proteins (B) in WAT of WT and PGC-1a/b-FAT-DKO mice fed AL or subjected to CR were analyzed by qRT-PCR (n = 10–12 animas/group). C ) Western blots of mitochondrial proteins in inguinal WAT from 2 independent cohorts of mice. Densitometric analysis of blots from the 2 cohorts of mice is shown in the graph. D) mtDNA content in WAT. E ) Mitochondrial function in inguinal WAT of AL and CR mice was assessed by high-resolution respirometry (PML, PMGP, PMGSP, PMGSE, and SE). Results are expressed as means 6 SEM (n = 6–8 animals/ group). *P , 0.05, WT vs. PGC-1a/b-FAT-DKO mice; #P , 0.05, AL vs. CR; **P , 0.01 WT vs. PGC-1a/b-FAT-DKO mice; ##P , 0.01, AL vs. CR. observed, CR notably improved insulin-stimulated AKT and enhancing insulin-dependent glucose uptake in skel- phosphorylation in WAT of WT animals. Despite the in- etal muscle, liver, and WAT (20, 21). Still, the precise ability of PGC-1a/b-FAT-DKO mice to increase mito- mechanisms by which CR improves diet-induced insulin chondrial mass and oxidative function in response to CR, resistance have not been defined. WAT of mice lacking PGC-1 coactivators normally re- Several studies in rodents and humans have reported spond to insulin stimulation by increasing AKT phos- an increase in mitochondrial gene expression in both hu- phorylation to the same extent as WT. Similar results were man and rodent tissues (22–25). However, the effect of CR obtained in liver and skeletal muscle (Fig. 9G). on mitochondrial biogenesis has remained controversial, and other studies have reported the lack of any change in mitochondrial protein content induced by low-caloric re- DISCUSSION gimes (26). Our gene-expression profiling study reveals that CR dramatically enhances the expression of genes CR is well known for having positive effects on glucose related to oxidative metabolism and promotes mitochon- homeostasis, improving whole-body insulin sensitivity, drial biogenesis in mouse WAT, joining ranks with other CALORIE RESTRICTION AND ADIPOSE MITOCHONDRIA 2351 Figure 6. PGC-1 coactivators but not CR regulate mitochondrial biogenesis in BAT. A) Gross morphologic aspect of interscapular BAT of PGC-1a/b-FAT-DKO mice and WT littermates (left). Histologic sections of BAT were stained with hematoxylin/eosin (right). B) Expression of mitochondrial genes in interscapular BAT of WT and PGC-1a/b-FAT-DKO mice (n = 6 animals/group) was assessed by qRT-PCR. C ) Western blots of mitochondrial proteins in BAT of WT and PGC-1a/b-FAT-DKO mice (left). Densitometric quantification of blots (right). D) mRNA expression of PGC-1a and PGC-1b coactivators and genes encoding for mitochondrial proteins, including Ucp1, in BAT of WT and PGC-1a/b-FAT-DKO mice fed AL or subjected to CR was analyzed by qRT-PCR (n = 4 animas/group). E ) Western blots of mitochondrial proteins in BAT of WT and PGC-1ab-FAT-DKO mice fed AL or subjected to CR (left). Densitometric analysis of blots is shown (right). Results are expressed as means 6 SEM. *P , 0.05, WT vs. PGC-1a/b-FAT-DKO mice; #P , 0.05, AL vs. CR; **P , 0.01. Cycs, cytochrome C, somatic. 2352 Vol. 33 February 2019 The FASEB Journal x www.fasebj.org PARDO ET AL. Figure 7. Analysis of the PGC-1-de- pendent browning of WAT in re- sponse to CR. A) Hematoxylin/eosin- stained histologic sections of inguinal WAT from AL and CR mice. B) mRNA expression of brown adipo- cyte markers in inguinal WAT from AL and CR mice was assessed by qRT-PCR. C ) mRNA expression of brown adipocyte markers in inguinal WAT of WT and PGC-1a/b-FAT-DKO mice, either fed AL or subjected to CR assessed by qRT-PCR. D) Western blot of UCP1 in protein extracts of inguinal WAT from WT and PGC-1a/ b-FAT-DKO mice fed AL or subjected to CR. Five micrograms of protein extracts from BAT was used as posi- tive control for UCP1. Fifty micro- grams of protein extracts from WAT of WT mice treated with 10 mg/kg rosiglitazone for 10 d was used as a positive control of WAT browning. Results are expressed as means 6 SEM (n = 6–7 animals/group). *P , 0.05, WT vs. PGC-1a/b-FAT-DKO mice; # P , 0.05, AL vs. CR. MWM, MW marker. Prdm16, PR domain zinc finger protein 16. studies to support a positive effect of CR on mitochondrial biogenesis induced by CR based on the observation that metabolism. The coordinated increase in mitochondrial PGC-1a mRNA levels are increased in rodents and hu- gene expression appears as a common signature of CR in mans subjected to mid- or long-term CR (23, 25, 27). many tissues, but its physiologic implications with regard However, only a recent study has provided experimental to the health benefits exerted by CR had not been ade- evidence that supports the involvement of PGC-1a in the quately addressed before. In this study, we show that CR- regulation of mitochondrial genes during CR in skeletal induced mitochondrial biogenesis in WAT depends on muscle. In that study, with the use of muscle-specific PGC- the transcriptional activity of PGC-1 coactivators, un- 1a knockout mice, Finley et al. (28) demonstrated that the equivocally demonstrating that PGC-1 acts as a master mitochondriogenic response observed in skeletal muscle regulator of mitochondrial biogenesis (17), also in re- upon CR was blunted in the absence of PGC-1a. In our sponse to low-calorie intake. Several studies had pointed study, the use of a mouse model simultaneously devoid of toward PGC-1a as a major regulator of mitochondrial PGC-1a and PGC-1b in adipocytes does not allow CALORIE RESTRICTION AND ADIPOSE MITOCHONDRIA 2353 Figure 8. Lack of PGC-1 coactivators does not alter body weight or adiposity. Body (A) and tissue (B) weight of WT and PGC-1a/ b-FAT-DKO mice, either fed AL or subjected to CR for 12 wk. All animals were fed the same HF diet (45% kcal fat, 35% kcal carbohydrates, 20% kcal protein). Results are expressed as means 6 SEM (n = 5–6 animals/group). ##P , 0.01, AL vs. CR. discrimination of which one of the two coactivators acts as drugs, have been shown to induce mitochondrial bio- the main driver of mitochondrial gene expression in re- genesis, specifically in WAT of genetically and diet- sponse to CR in WAT. Still, in white adipocytes, a co- induced insulin resistance and T2D animal models (33, operative contribution between PGC-1a and PGC-1b to 34), as well as in humans (35, 36), concomitantly with an the control of mitochondrial gene expression in response improvement in insulin sensitivity. Likewise, treat- to CR is very likely. Indeed, first, both PGC-1a and PGC- ment of rodent models with resveratrol or selective 1b mRNA levels are induced in WAT in response to CR; synthetic activators of Sirt1 has been reported to im- second, we have previously shown that PGC-1b appears prove insulin resistance in close association with an to have a dominant role over PGC-1a in the regulation of increase in mitochondrial biogenesis and function in mitochondrial gene expression in white adipocytes (10, WAT, skeletal muscle, and liver (37, 38). Based on this 11); and third, contrary to what it was previously assumed, tight association, it has been suggested that the increase PGC-1b also plays an important role in the dynamic reg- of mitochondrial mass in WAT and/or other tissues ulation of mitochondrial gene expression in response to would be one of the major mechanisms by which cues that prompt for increased mitochondrial oxidative CR contributes to the reversal of insulin resistance. function [i.e., thiazolidinediones (TZDs)] beyond its well- However, our data, showing that mice lacking PGC-1 known role in the maintenance of basal mitochondrial coactivators fail to increase adipose mitochondrial mass (10). biogenesis but still retain their full capacity to improve Our study provides experimental evidence that places whole-body glucose tolerance and insulin sensitivity in PGC-1 coactivators as central elements that integrate the response to CR, strongly support the notion that en- input of different regulatory signals to modulate mito- hanced mitochondrial function in adipocytes is not chondrial biogenesis in response to CR. Such signals acti- absolutely required for the positive effects that CR ex- vated by nutrient/energy deprivation include Sirtuin 1 erts on glucose homeostasis. These findings are in line (Sirt1), a NAD+-dependent deacetylase; AMPK, an AMP/ with a previous study in which by using a mouse model ADP-activated kinase; and eNOS, all of which are acti- with partial deficiency in mitochondrial function as a vated by CR. Thus, increase in the deacetylation (29, 30) result of the adipose-specific deletion of PGC-1b, we and phosphorylation (31) by Sirt1 and AMPK, re- demonstrated that rosiglitazone-induced mitochondrial spectively, would lead to the transcriptional activation of oxidative capacity in WAT was not essential for the insulin- PGC-1s, whereas increased NO production by eNOS sensitizing effects of TZDs (10). would enhance the expression of, at least, PGC-1a (25). The adipose-specific nature of our PGC-1a/b-FAT- The integration of all of these regulatory signals would DKO mouse model raises the question of whether CR result in a net boost of PGC-1 coactivators transcriptional could improve whole-body glucose homeostasis by en- activity and the consequent increase in mitochondrial hancing mitochondrial function in insulin target tissues biogenesis in response to CR. other than WAT. Although this appears plausible, a study Numerous studies have shown the existence of a tight in muscle-specific PGC-1a knockout mice demonstrated correlation between insulin sensitivity and mitochondrial that despite failure to increase mitochondrial mass and mass/oxidative function in skeletal muscle and WAT oxidative metabolism in skeletal muscle, muscle-specific [reviewed in Zamora and Villena (8)]. Consequently, mi- PGC-1a knockout mice responded to CR by improv- tochondrial oxidative capacity has been considered a ing glucose tolerance and insulin sensitivity to the strong predictor of insulin sensitivity (32). In this regard, it same extent as their WT counterparts (28). This is similar to is worth noting that several therapeutic interventions that what we described here for PGC-1a/b-FAT-DKO mice in improve glucose homeostasis are associated with in- WAT. Therefore, besides being a distinctive signature of creased mitochondrial biogenesis and oxidative function CR in many tissues, altogether, these studies demonstrate in WAT. For example, TZDs, a family of anti-diabetic a clear dissociation between CR-induced mitochondrial 2354 Vol. 33 February 2019 The FASEB Journal x www.fasebj.org PARDO ET AL. Figure 9. Impaired mitochondrial biogenesis in mice devoid of PGC-1 in adipose tissues does not prevent the amelioration of glucose homeostasis in response to CR. A, B) GTTs (2 g/kg) in WT and PGC-1a/b-FAT-DKO mice, either fed AL (A) or subjected to CR (B). C ) Area under the curve (AUC) of GTTs. D, E) ITTs (0.9 U/kg) in WT and PGC-1a/b-FAT-DKO mice, either fed AL (D) or subjected to CR (E) for 12 wk. F) Area under the curve of ITTs. G) Tissue-specific insulin signaling. Total and phosphorylated (p-) Akt was detected by Western blot in protein lysates of inguinal WAT, liver, and gastrocnemius muscle from WT and PGC-1a/b-FAT-DKO mice, either fed AL or subjected to CR, which were treated with an insulin bolus after having food withheld for 12 h. CYPA was used as loading control. Results are expressed as means 6 SEM (n = 5–6 animals/group). #P # 0.05, ##P # 0.01, AL vs. CR. biogenesis in WAT/muscle and the beneficial effects agents, such as TZDs, which we showed were capable of that this nutritional intervention exerts on glucose inducing detectable levels of UCP1 in WAT (9, 10). homeostasis. Moreover, induction of UCP1 mRNA in WAT in response If the induction of mitochondrial biogenesis and oxi- to CR appears blunted in PGC-1a/b-FAT-DKO mice, in dative function in WAT and muscle are not required to agreement with the relevant role of PGC-1 coactivators in ameliorate diet-induced insulin resistance in response to brown adipocyte differentiation and function (18). Still, CR, then the question is by which other means CR exerts its PGC-1a/b-FAT-DKO mice improve glucose homeostasis beneficial effects on glucose homeostasis? Very recently, it in response to nutrient deprivation to the same extent as has been reported that CR induces browning of WAT, an WT mice. Therefore, regardless of the magnitude of the effect that appears to be mediated by an enhancement of browning induced by CR, our results strongly support the type 2 immune response (19) and that has been associated notion that brown adipocyte recruitment in WAT is not with metabolic improvement during CR. Consistent with required for the effects of CR on glucose homeostasis. this observation, we were able to observe a moderate in- It is also well known that PGC-1 coactivators are re- crease in the mRNA levels of brown adipocyte markers, quired for proper BAT function, controlling mitochondrial such as Ucp1 and Cidea, in response to CR. However, de- biogenesis during the course of adipocyte differentiation spite the changes observed in mRNA, the failure to find and the expression of the thermogenic program in re- brown-like adipocytes and detectable levels of UCP1 sponse to low environmental temperatures (17). In agree- protein within WAT of CR mice rather suggest that the ment with the crucial role of PGC-1 coactivators in BAT, browning of WAT in response to CR is negligible and our PGC-1a/b-FAT-DKO mice exhibit a severe reduction irrelevant from a thermogenic point of view (39). Our data in the levels of mitochondrial proteins and UCP1 in BAT are further supported by a recent study in obese humans that leads to a massive accumulation of triglycerides that also reported the absence of browning in response to within brown adipocytes as a result of their incapacity CR (40). The lack of significant browning induced by CR to oxidize fatty acids properly into mitochondria and contrasts with the tangible browning capacity of other perform thermogenesis. Although reduced BAT mass and CALORIE RESTRICTION AND ADIPOSE MITOCHONDRIA 2355 TABLE 2. Serological parameters in WT and PGC-1a/b-FAT-DKO mice fed AL with a HF diet or subjected to 40% CR for 12 wk AL CR Parameter WT PGC-1a/b-FAT-DKO WT PGC-1a/b-FAT-DKO Glucose, mg/dl 203.5 6 6.4 222.2 6 11.5 161.3 6 14.4* 173.8 6 11.8* Triglycerides, mg/dl 55.17 6 3.37 49.42 6 3.02 45.42 6 2.69† 44.05 6 2.33 Free fatty acids, mM 0.71 6 0.06 0.73 6 0.09 0.62 6 0.08 0.59 6 0.05 Cholesterol, mg/dl 267.7 6 19.1 224.2 6 19.6 174.1 6 10.4* 154.7 6 9.4* Insulin, ng/ml 4.88 6 0.45 3.45 6 0.28‡ 1.80 6 0.22* 1.25 6 0.12* † Data are expressed as means 6 SEM. P # 0.05, AL vs. CR; ‡P # 0.05, WT vs. PGC-1a/b-FAT-DKO mice; *P # 0.01. impaired thermogenic activity have been correlated with least in part, by downregulating the expression of genes higher body weight and impaired insulin sensitivity in encoding for ECM proteins and reducing fibrosis. rodents and humans (41–43), our results show that dys- The anti-inflammatory effect of CR in WAT could also functional BAT in mice lacking PGC-1 coactivators is not be an important contributor to the improvement of glucose sufficient to bring about an obese or insulin-resistant homeostasis. It is well known that obesity-linked insulin phenotype. Importantly, contrary to what we observed in resistance is associated with low-degree chronic inflammation WAT, CR does not induce mitochondrial biogenesis in of WAT, which is characterized by increased recruitment of BAT nor does it increase UCP1 expression, suggesting that M1 macrophages and other immune cells, elevated levels of CR does not activate brown adipocyte thermogenesis. The proinflammatory cytokines, and the activation of NF-kB and lack of response of UCP1 expression to CR is consistent JNK pathways that directly inhibit insulin signaling with other studies reporting that UCP1 content or activity (51). The reduction in the expression of inflammatory in BAT is unchanged (44) or even reduced (19, 45) in rodent cytokines (i.e., Il-1b and Il-6) or chemokines (i.e., Ccl2 models subjected to long-term, low-calorie intake, an effect and Ccl7), together with the decrease, albeit not signif- that appears to be intended to increase calorie efficiency icant, in the levels of M1 proinflammatory macrophage and maintain energy balance. The lack of induction of markers (Cd11c and Nos2) that we observed in WAT of mitochondrial biogenesis and UCP1 protein levels in BAT mice subjected to CR is in line with other studies dem- in response to CR, strongly suggests that BAT does not onstrating that CR reduces WAT inflammation in ro- play a relevant role in mediating the metabolic effects of dents (52, 53) and humans (54). Recent evidence suggests CR. that CR enhances polarization of macrophages toward the Our study identified genes related to ECM remodeling alternative M2 anti-inflammatory phenotype (19), a switch and the immune system, as 2 gene networks significantly that appears to contribute significantly to the improvement downregulated by CR in WAT. The microenvironment of insulin sensitivity (55). Contrary to what has been re- provided by ECM proteins is essential for proper devel- ported by Fabbiano et al. (19), the expression of anti- opment and function of adipose tissue, as ECM remodel- inflammatory cytokines (i.e., Il-13 and Il-33) and M2 ing is fundamental for WAT to adapt to changing nutritional markers (i.e., Arg1) was reduced by CR in our study. This conditions (46). Thus, whereas ECM proteins are required argues against an enhancement of macrophage polariza- for proper differentiation of precursors cells into fully func- tion by CR, but it certainly indicates that reduced calorie tional adipocytes, excessive deposition of ECM proteins intake decreases macrophage recruitment induced by HF (i.e., fibrosis) restrains migration of precursor cells and feeding in WAT. The discrepancy between our study and adipocyte hypertrophy, compromising adipose tissue the one from Fabbiano et al. (19) is unclear, but the different expandability. Regarding the potential role of ECM in dietary regime (HF diet vs. regular chow) could have metabolic diseases, it has been shown that increased contributed to the different outcomes regarding not only expression in WAT of ECM genes, including collagen macrophage polarization but also browning of WAT. (COL)6A3 and matrix metalloproteinase 9, associates Nevertheless, the evidences provided on the immune re- with increased body mass and tissue inflammation in sponse to CR suggest that attenuation of diet-induced in- humans (47, 48), 2 major risk factors for the develop- flammation in WAT could contribute to the improvement ment of insulin resistance and T2D. Consistent with an of glucose homeostasis in response to CR. important role of adipose ECM in determining insu- In summary, our study has identified mitochondrial lin sensitivity, mice that overexpress endotrophin, a biogenesis as the main process positively regulated by CR cleavage product of COL6A3, develop WAT fibrosis in WAT. With the use of a new mouse model that simul- and exhibit insulin resistance when fed a HF diet, an taneously lacks the expression of PGC-1a and PGC-1b effect that is reversed after endotrophin neutralization coactivators specifically in adipocytes, we have demon- (49). Likewise, obese COL6A3 knockout mice show an strated that PGC-1 coactivators control the process of improvement in insulin sensitivity (50). Based on the mitochondrial biogenesis in WAT in response to CR. positive effects that the reduction of ECM in WAT has Furthermore, we have shown that the induction of mito- on whole-body glucose homeostasis, it is plausible to chondrial biogenesis and enhancement of mitochondrial speculate that CR may improve insulin resistance, at oxidative function in WAT are not required for the positive 2356 Vol. 33 February 2019 The FASEB Journal x www.fasebj.org PARDO ET AL. effects that CR exerts on glucose homeostasis. These adipose tissue is independent of peroxisome proliferator-activated findings support the notion that mitochondrial function in receptor g coactivator-1a. PLoS One 6, e26989 12. Lasheras, J., Vilà, M., Zamora, M., Riu, E., Pardo, R., Poncelas, M., WAT is dissociated from insulin sensitivity, both under Cases, I., Ruiz-Meana, M., Hernández, C., Feliu, J. E., Simó, R., basal conditions and in response to CR. The mechanisms Garcı́a-Dorado, D., and Villena, J. A. (2016) Gene expression by which CR improves adipose and whole-body insulin profiling in hearts of diabetic mice uncovers a potential role of estrogen-related receptor g in diabetic cardiomyopathy. Mol. Cell. sensitivity still remain to be defined, but our findings that Endocrinol. 430, 77–88 pathways related to ECM and the immune response are 13. Huang, W., Sherman, B. T., and Lempicki, R. A. (2009) Systematic also modulated by CR would deserve further exploration and integrative analysis of large gene lists using DAVID bioinformatics to unveil their relevance in the metabolic effects exerted by resources. Nat. Protoc. 4, 44–57 14. Mootha, V. K., Lindgren, C. M., Eriksson, K. F., Subramanian, A., low-calorie intake. Sihag, S., Lehar, J., Puigserver, P., Carlsson, E., Ridderstråle, M., Laurila, E., Houstis, N., Daly, M. J., Patterson, N., Mesirov, J. P., Golub, T. R., Tamayo, P., Spiegelman, B., Lander, E. S., Hirschhorn, J. N., Altshuler, D., and Groop, L. C. (2003) PGC-1alpha-responsive genes ACKNOWLEDGMENTS involved in oxidative phosphorylation are coordinately down- regulated in human diabetes. Nat. Genet. 34, 267–273 The authors thank Anastasia Kralli (Johns Hopkins 15. Villena, J. A., Hock, M. B., Chang, W. Y., Barcas, J. E., Giguère, V., and University, Baltimore, MD, USA) for the Ppargc1aflox/flox and Kralli, A. (2007) Orphan nuclear receptor estrogen-related receptor Ppargc1bflox/flox mice and helpful discussions. This work was alpha is essential for adaptive thermogenesis. Proc. Natl. Acad. Sci. USA supported by Grants SAF2011-23886 and SAF2012-39484 from 104, 1418–1423 Ministerio de Economı́a y competitividad (MINECO/FEDER; 16. Cantó, C., and Garcia-Roves, P. M. (2015) High-resolution respi- rometry for mitochondrial characterization of ex vivo mouse tissues. Spain) to J.A.V. and Grant PI15/00701 from Instituto de Salud Curr. Protoc. Mouse Biol. 5, 135–153 Carlos III (ISCIII) to P.M.G.-R. The authors declare no 17. Villena, J. A. (2015) New insights into PGC-1 coactivators: redefining conflicts of interest. their role in the regulation of mitochondrial function and beyond. FEBS J. 282, 647–672 18. Uldry, M., Yang, W., St-Pierre, J., Lin, J., Seale, P., and Spiegelman, B. M. (2006) Complementary action of the PGC-1 coactivators in mitochon- AUTHOR CONTRIBUTIONS drial biogenesis and brown fat differentiation. Cell Metab. 3, 333–341 19. Fabbiano, S., Suárez-Zamorano, N., Rigo, D., Veyrat-Durebex, C., R. Pardo, M. Vilà, L. Cervela, M. de Marco, P. Gama-Pérez, Stevanovic Dokic, A., Colin, D. J., and Trajkovski, M. (2016) Caloric restriction leads to browning of white adipose tissue through type 2 A. González-Franquesa, and L. Statuto performed re- immune signaling. Cell Metab. 24, 434–446 search and analyzed data; R. Vilallonga and R. Simó 20. Reaven, E., Wright, D., Mondon, C. E., Solomon, R., Ho, H., and contributed with material/reagents and reviewed the Reaven, G. M. (1983) Effect of age and diet on insulin secretion and drafted manuscript; P. M. Garcia-Roves designed research, insulin action in the rat. Diabetes 32, 175–180 analyzed data, contributed to the discussion, and reviewed 21. Wetter, T. J., Gazdag, A. C., Dean, D. J., and Cartee, G. D. (1999) Effect of calorie restriction on in vivo glucose metabolism by individual the manuscript; and J. A. Villena conceived of and supervised tissues in rats. Am. J. Physiol. 276, E728–E738 the project, analyzed data, and wrote the paper. 22. Barger, J. L., Vann, J. M., Cray, N. L., Pugh, T. D., Mastaloudis, A., Hester, S. N., Wood, S. M., Newton, M. A., Weindruch, R., and Prolla, T. A. (2017) Identification of tissue-specific transcriptional markers of caloric restriction in the mouse and their use to evaluate caloric re- REFERENCES striction mimetics. Aging Cell 16, 750–760 23. Civitarese, A. E., Carling, S., Heilbronn, L. K., Hulver, M. H., 1. Guilherme, A., Virbasius, J. V., Puri, V., and Czech, M. P. (2008) Ukropcova, B., Deutsch, W. A., Smith, S. R., and Ravussin, E.; Adipocyte dysfunctions linking obesity to insulin resistance and type 2 CALERIE Pennington Team. (2007) Calorie restriction increases diabetes. Nat. Rev. Mol. Cell Biol. 9, 367–377 muscle mitochondrial biogenesis in healthy humans. PLoS Med. 4, e76 2. International Diabetes Federation (2015) IDF Diabetes Atlas, In- 24. Higami, Y., Pugh, T. D., Page, G. P., Allison, D. B., Prolla, T. A., and ternational Diabetes Federation, Brussels, Belgium Weindruch, R. (2004) Adipose tissue energy metabolism: altered 3. Pellegrinelli, V., Carobbio, S., and Vidal-Puig, A. (2016) Adipose tissue gene expression profile of mice subjected to long-term caloric re- plasticity: how fat depots respond differently to pathophysiological striction. FASEB J. 18, 415–417 cues. Diabetologia 59, 1075–1088 25. Nisoli, E., Tonello, C., Cardile, A., Cozzi, V., Bracale, R., Tedesco, L., 4. Lowell, B. B., and Shulman, G. I. (2005) Mitochondrial dysfunction Falcone, S., Valerio, A., Cantoni, O., Clementi, E., Moncada, S., and and type 2 diabetes. Science 307, 384–387 Carruba, M. O. (2005) Calorie restriction promotes mitochondrial 5. Chawla, A., Nguyen, K. D., and Goh, Y. P. (2011) Macrophage- biogenesis by inducing the expression of eNOS. Science 310, 314–317 mediated inflammation in metabolic disease. Nat. Rev. Immunol. 11, 26. Hancock, C. R., Han, D. H., Higashida, K., Kim, S. H., and Holloszy, 738–749 J. O. (2011) Does calorie restriction induce mitochondrial 6. Wellen, K. E., and Hotamisligil, G. S. (2005) Inflammation, stress, and biogenesis? A reevaluation. FASEB J. 25, 785–791 diabetes. J. Clin. Invest. 115, 1111–1119 27. López-Lluch, G., Hunt, N., Jones, B., Zhu, M., Jamieson, H., Hilmer, 7. Bordone, L., and Guarente, L. (2005) Calorie restriction, SIRT1 and S., Cascajo, M. V., Allard, J., Ingram, D. K., Navas, P., and de Cabo, R. metabolism: understanding longevity. Nat. Rev. Mol. Cell Biol. 6, (2006) Calorie restriction induces mitochondrial biogenesis and 298–305 bioenergetic efficiency. Proc. Natl. Acad. Sci. USA 103, 1768–1773 8. Zamora, M., and Villena, J. A. (2014) Targeting mitochondrial 28. Finley, L. W., Lee, J., Souza, A., Desquiret-Dumas, V., Bullock, K., biogenesis to treat insulin resistance. Curr. Pharm. Des. 20, 5527–5557 Rowe, G. C., Procaccio, V., Clish, C. B., Arany, Z., and Haigis, M. C. 9. Eguchi, J., Wang, X., Yu, S., Kershaw, E. E., Chiu, P. C., Dushay, J., (2012) Skeletal muscle transcriptional coactivator PGC-1a mediates Estall, J. L., Klein, U., Maratos-Flier, E., and Rosen, E. D. (2011) mitochondrial, but not metabolic, changes during calorie restriction. Transcriptional control of adipose lipid handling by IRF4. Cell Metab. Proc. Natl. Acad. Sci. USA 109, 2931–2936 13, 249–259 29. Gerhart-Hines, Z., Rodgers, J. T., Bare, O., Lerin, C., Kim, S. H., 10. Enguix, N., Pardo, R., González, A., López, V. M., Simó, R., Kralli, A., Mostoslavsky, R., Alt, F. W., Wu, Z., and Puigserver, P. (2007) and Villena, J. A. (2013) Mice lacking PGC-1b in adipose tissues reveal Metabolic control of muscle mitochondrial function and fatty acid a dissociation between mitochondrial dysfunction and insulin oxidation through SIRT1/PGC-1alpha. EMBO J. 26, 1913–1923 resistance. Mol. Metab. 2, 215–226 30. Kelly, T. J., Lerin, C., Haas, W., Gygi, S. P., and Puigserver, P. (2009) 11. Pardo, R., Enguix, N., Lasheras, J., Feliu, J. E., Kralli, A., and Villena, GCN5-mediated transcriptional control of the metabolic coactivator J. A. (2011) Rosiglitazone-induced mitochondrial biogenesis in white PGC-1beta through lysine acetylation. J. Biol. Chem. 284, 19945–19952 CALORIE RESTRICTION AND ADIPOSE MITOCHONDRIA 2357 31. Jäger, S., Handschin, C., St-Pierre, J., and Spiegelman, B. M. (2007) 44. Villarroya, F., Felipe, A., and Mampel, T. (1986) Brown adipose tissue AMP-activated protein kinase (AMPK) action in skeletal muscle via activity in hypocaloric-diet fed lactating rats. Biosci. Rep. 6, 669–675 direct phosphorylation of PGC-1alpha. Proc. Natl. Acad. Sci. USA 104, 45. Rothwell, N. J., and Stock, M. J. (1982) Effect of chronic food 12017–12022 restriction on energy balance, thermogenic capacity, and brown- 32. Bruce, C. R., Anderson, M. J., Carey, A. L., Newman, D. G., Bonen, A., adipose-tissue activity in the rat. Biosci. Rep. 2, 543–549 Kriketos, A. D., Cooney, G. J., and Hawley, J. A. (2003) Muscle 46. Pope, B. D., Warren, C. R., Parker, K. K., and Cowan, C. A. (2016) oxidative capacity is a better predictor of insulin sensitivity than lipid Microenvironmental control of adipocyte fate and function. Trends status. J. Clin. Endocrinol. Metab. 88, 5444–5451 Cell Biol. 26, 745–755 33. Rong, J. X., Qiu, Y., Hansen, M. K., Zhu, L., Zhang, V., Xie, M., 47. Henegar, C., Tordjman, J., Achard, V., Lacasa, D., Cremer, I., Okamoto, Y., Mattie, M. D., Higashiyama, H., Asano, S., Strum, J. C., Guerre-Millo, M., Poitou, C., Basdevant, A., Stich, V., Viguerie, N., and Ryan, T. E. (2007) Adipose mitochondrial biogenesis is Langin, D., Bedossa, P., Zucker, J. D., and Clement, K. (2008) Adipose suppressed in db/db and high-fat diet-fed mice and improved by tissue transcriptomic signature highlights the pathological relevance rosiglitazone. Diabetes 56, 1751–1760 of extracellular matrix in human obesity. Genome Biol. 9, R14 34. Wilson-Fritch, L., Nicoloro, S., Chouinard, M., Lazar, M. A., Chui, 48. Pasarica, M., Gowronska-Kozak, B., Burk, D., Remedios, I., Hymel, P. C., Leszyk, J., Straubhaar, J., Czech, M. P., and Corvera, S. (2004) D., Gimble, J., Ravussin, E., Bray, G. A., and Smith, S. R. (2009) Mitochondrial remodeling in adipose tissue associated with obesity Adipose tissue collagen VI in obesity. J. Clin. Endocrinol. Metab. 94, and treatment with rosiglitazone. J. Clin. Invest. 114, 1281–1289 5155–5162 35. Boden, G., Homko, C., Mozzoli, M., Showe, L. C., Nichols, C., and 49. Sun, K., Park, J., Gupta, O. T., Holland, W. L., Auerbach, P., Zhang, N., Cheung, P. (2005) Thiazolidinediones upregulate fatty acid uptake and Goncalves Marangoni, R., Nicoloro, S. M., Czech, M. P., Varga, J., oxidation in adipose tissue of diabetic patients. Diabetes 54, 880–885 Ploug, T., An, Z., and Scherer, P. E. (2014) Endotrophin triggers 36. Bogacka, I., Xie, H., Bray, G. A., and Smith, S. R. (2005) Pioglitazone adipose tissue fibrosis and metabolic dysfunction. Nat. Commun. 5, induces mitochondrial biogenesis in human subcutaneous adipose 3485 tissue in vivo. Diabetes 54, 1392–1399 50. Khan, T., Muise, E. S., Iyengar, P., Wang, Z. V., Chandalia, M., Abate, 37. Feige, J. N., Lagouge, M., Canto, C., Strehle, A., Houten, S. M., Milne, N., Zhang, B. B., Bonaldo, P., Chua, S., and Scherer, P. E. (2009) J. C., Lambert, P. D., Mataki, C., Elliott, P. J., and Auwerx, J. (2008) Metabolic dysregulation and adipose tissue fibrosis: role of collagen Specific SIRT1 activation mimics low energy levels and protects VI. Mol. Cell. Biol. 29, 1575–1591 against diet-induced metabolic disorders by enhancing fat oxidation. 51. Reilly, S. M., and Saltiel, A. R. (2017) Adapting to obesity with adipose Cell Metab. 8, 347–358 tissue inflammation. Nat. Rev. Endocrinol. 13, 633–643 38. Lagouge, M., Argmann, C., Gerhart-Hines, Z., Meziane, H., Lerin, C., 52. Park, C. Y., Park, S., Kim, M. S., Kim, H. K., and Han, S. N. (2017) Daussin, F., Messadeq, N., Milne, J., Lambert, P., Elliott, P., Geny, B., Effects of mild calorie restriction on lipid metabolism and Laakso, M., Puigserver, P., and Auwerx, J. (2006) Resveratrol improves inflammation in liver and adipose tissue. Biochem. Biophys. Res. mitochondrial function and protects against metabolic disease by Commun. 490, 636–642 activating SIRT1 and PGC-1alpha. Cell 127, 1109–1122 53. Wang, Q., Perrard, X. D., Perrard, J. L., Mansoori, A., Raya, J. L., 39. Nedergaard, J., and Cannon, B. (2013) UCP1 mRNA does not Hoogeveen, R., Smith, C. W., Ballantyne, C. M., and Wu, H. (2011) produce heat. Biochim. Biophys. Acta 1831, 943–949 Differential effect of weight loss with low-fat diet or high-fat diet re- 40. Barquissau, V., Léger, B., Beuzelin, D., Martins, F., Amri, E. Z., Pisani, striction on inflammation in the liver and adipose tissue of mice with D. F., Saris, W. H. M., Astrup, A., Maoret, J. J., Iacovoni, J., Déjean, S., diet-induced obesity. Atherosclerosis 219, 100–108 Moro, C., Viguerie, N., and Langin, D. (2018) Caloric restriction and 54. Meydani, S. N., Das, S. K., Pieper, C. F., Lewis, M. R., Klein, S., Dixit, diet-induced weight loss do not induce browning of human sub- V. D., Gupta, A. K., Villareal, D. T., Bhapkar, M., Huang, M., Fuss, P. J., cutaneous white adipose tissue in women and men with obesity. Cell Roberts, S. B., Holloszy, J. O., and Fontana, L. (2016) Long-term Reports 22, 1079–1089 moderate calorie restriction inhibits inflammation without impairing 41. Cypess, A. M., Lehman, S., Williams, G., Tal, I., Rodman, D., Goldfine, cell-mediated immunity: a randomized controlled trial in non-obese A. B., Kuo, F. C., Palmer, E. L., Tseng, Y. H., Doria, A., Kolodny, G. M., humans. Aging (Albany N.Y.) 8, 1416–1431 and Kahn, C. R. (2009) Identification and importance of brown 55. Kang, K., Reilly, S. M., Karabacak, V., Gangl, M. R., Fitzgerald, K., adipose tissue in adult humans. N. Engl. J. Med. 360, 1509–1517 Hatano, B., and Lee, C. H. (2008) Adipocyte-derived Th2 cytokines 42. Hamann, A., Flier, J. S., and Lowell, B. B. (1996) Decreased brown fat and myeloid PPARdelta regulate macrophage polarization and in- markedly enhances susceptibility to diet-induced obesity, diabetes, sulin sensitivity. Cell Metab. 7, 485–495 and hyperlipidemia. Endocrinology 137, 21–29 43. Himms-Hagen, J. (1985) Defective brown adipose tissue thermogenesis Received for publication February 12, 2018. in obese mice. Int. J. Obes. 9 (Suppl 2), 17–24 Accepted for publication September 4, 2018. 2358 Vol. 33 February 2019 The FASEB Journal x www.fasebj.org PARDO ET AL.