Targeted Disruption of the SUCNR1 Metabolic Receptor Leads to Dichotomous Effects on Obesity

G-protein–coupled receptors (GPCRs) constitute the major and most diverse group of membrane receptors in eukaryotes (1). Found in the plasma membrane, these receptors are tasked with the recognition and transmission of messages from the external environment. An ever-increasing number of GPCRs are now being identified as receptors for metabolites or energy substrates (2), expanding the repertoire of biological targets to diseases associated with the metabolic syndrome, including hypertension and type 2 diabetes (3). Metabolites such as lactate (4), ketone bodies (5), and α-ketoglutarate (6), known primarily to serve functions distinct to signaling, have been shown to act as ligands for GPCRs. Representative of this new class of receptor is Gpr91, a receptor for the Krebs cycle intermediate succinate, now termed Sucnr1 for succinate receptor (7). The Krebs, or citric acid, cycle occurs at the junction between glycolysis and oxidative phosphorylation (8) and is a key component of aerobic metabolism by providing reducing equivalents for the electron transport chain. Interestingly, succinate has been known for many years to increase in tissues during hypoxia (9–11) and may act as a surrogate marker for reduced oxygenation. Accordingly, Sucnr1 may be activated in times of hypoxic stress such as ischemic injury in liver (12) or retina (13) or during cardiomyocyte crisis (14). Moreover, dysregulated metabolism, akin to diabetes or the metabolic syndrome, also has the potential to increase peripheral blood concentrations of succinate to levels where activation of the Sucnr1 receptor will occur (15,16). Indeed, the low levels (∼2–80 μmol/L) of succinate found in blood plasma (15–18), in comparison with that required for activation (in the region of 180 μmol/L [2]), would suggest that Sucnr1 functions more as a sensor of metabolic (or oxidative) damage (19) rather than a physiological mediator of signaling.

Mounting evidence points to an important role for Sucnr1 during periods of metabolic dysfunction, such as diabetes-related hypertension, oxygen-induced retinopathy, and the metabolic syndrome (19). Interestingly, obesity is a key component of many metabolic-related diseases (20) and can lead to diabetes, hyperlipidemia, and cardiovascular disease (20,21). The effect of Sucnr1 deficiency on body composition and metabolism has not previously been studied. We reasoned that Sucnr1 could be a potential modulator of metabolism since a recent study indicated that this receptor participates in the breakdown (lipolysis) of stored triglycerol (TG) in fat (22). To test this hypothesis, we analyzed the expression of Sucnr1 in mouse adipose tissue and observed high expression in the white adipocyte compartment. To gain insight into the (patho)physiological role of Sucnr1, we created mice deficient for this gene using a knockout (KO)-first approach. We show that deficiency of the Sucnr1 gene has metabolic consequences for weight gain and glucose homeostasis. Sucnr1 KO mice have a modest lean phenotype under standard diet (SD) conditions, correlating with a reduction in the size of white adipose tissue (WAT) depots, increased energy expenditure, and an improved glucose clearance rate. Challenge with high-fat diet (HFD) also resulted in a significant decrease in body weight in Sucnr1 KO mice, but only during initial periods. At later periods, body weight of Sucnr1 KO mice was indistinguishable from wild-type (WT) littermates, but WAT content was increased. This appeared to coincide with a hyperglycemic profile and a reduction in insulin secretion from the pancreas. Collectively, our results point to a potentially important role for Sucnr1 in metabolic disease related with obesity.

To create Sucnr1-mutant mice, we obtained two Sucnr1-targeted embryonic stem (ES) cell clones from the European Mouse Mutagenesis Consortium (EUCOMM), designated as KO-first, conditional-ready ES cells (23). This flexible system allows both for gene-trap (GT) reporter constructs and also for conditional and null alleles to be generated by exposure to the site-specific recombinases Cre and flippase (www.knockoutmouse.org). ES cells were injected into C57BL/6 blastocysts, and chimeric males were bred to C57BL/6 females to produce germ line transmission. Heterozygous animals were intercrossed, and genotyping of F2 progeny detected homozygous mice at the expected Mendelian ratio. Genotyping was performed after weaning by PCR assay with genomic DNA extracted from tail biopsies using the REDExtract-N-Amp Tissue PCR kit (Sigma-Aldrich, Saint Louis, MO). Genotyping primers were primer 1, 5′-ACTATAAGCATCACTTCACCACTTCC-3′; primer 2, 5′-CAACGGGTTCTTCTGTTAGTCC-3′; and primer 3, 5′-AAGAACAAATGGTAGAACAGTCACGG-3′. A PCR product of ∼500 nucleotides (nt) was obtained from WT DNA, whereas mutant DNA produced an amplification product of ∼320 nt in size. These KO-first mice were designated GT mice. In order to remove the critical exon of the Sucnr1 gene, together with the neomycin cassette, GT mice were crossed with Sox2-Cre mice (supplied by Dr. Miguel Torres, Centro Nacional de Investigaciones Cardiovasculares [CNIC]) to produce whole-body KO mice. Genotyping was performed as before using the primer 4, 5′-ACTTCCAGTTCAACATCAGCCGCTACA-3′; primer 5, 5′-GTGATTCATCTGTATTATTAGATGAGC-3′; and primer 6, 5′-CAGCACAACCATCAGAGAAACAATGGACTC-3′. A PCR product of ∼650 nt indicated that recombination had occurred to produce a KO locus, and a product of ∼300 nt identified a WT locus. All experiments were carried out using these KO mice.

Mice were housed at the CNIC animal facility under pathogen-free conditions. Male mice were used for all studies. Mice were given ad libitum access to food and water unless otherwise noted in experimental methods. Environmental conditions were controlled to 12/12-h light/dark cycles. Animal studies were approved by the local ethics committee. All animal procedures conformed to European Union Directive 86/609/EEC and recommendation 2007/526/EC regarding the protection of animals used for experimental and other scientific purposes, enacted under Spanish law 1201/2005. Male Sucnr1 KO mice and WT mice (5–6 weeks old) obtained from heterozygous mating were fed an SD of rodent chow, containing 13.4% fat (16% energy; Diet 5K67, LabDiet, Saint Louis, MO) or an HFD containing 34.9% fat (60.9% energy; Diet 58Y1, TestDiet, Saint Louis, MO). Body weight was recorded weekly. Separate cohorts of animals on SD and HFD were used for metabolic cage analysis using a 16-chamber indirect calorimetry system (PhenoMaster; TSE Systems, Bad Homburg, Germany). Mice were individually housed in the chambers for 5 days; the first 2 days were used for acclimatization, and data were analyzed from the last 3 days. Oxygen consumption rate, carbon dioxide production rate, feeding, and total locomotive activity were measured concurrently for each mouse.

Imaging was performed in the Advanced Imaging Unit at CNIC. Mice on SD or HFD were preanesthetized by inhalation of isoflurane/oxygen (2–4%) and monitored by electrocardiogram and breathing rhythm (Model 1025; SA Instruments, Inc., NY). Animals were positioned supine in a customized bed with a built-in nose cone supplying inhalatory anesthesia (1–2%) and kept at 35–37°C by a warm airflow. Images were acquired with an Agilent/Varian 7 Tesla system equipped with a DD2 console and an active shielded 205/120 gradient insert coil with 130 mT/m maximum strength and a volume coil (Rapid Biomedical GmbH, Rimpar, Germany) for transmit/receive. For MRI, data were collected with a two- and three-dimensional (3D) spin-echo sequence. T1 weighted two-dimensional images (repetition time 380 ms, echo time 13 ms, number of averages 4) were acquired to visualize the fat qualitatively in high resolution. 3D T1 weighted spin-echo acquisition was also acquired to calculate the body fat–to–total body weight ratio using 100 ms repetition time, 8 ms echo time, 1 average. In both sequences, the field of view was 90 × 45 × 45 mm³ and the reconstruction matrix was 256 × 256 × 256. For MRS, the ¹³C and ¹H surface coil was placed over the epididymal fat. ¹³C spectra were acquired with a repetition time of 1,000 ms, 512 accumulations, and a spectral width of 18 kHz. The ¹H spectra were acquired with 1,000 ms of repetition time, 256 scans, and a spectral width of 10 kHz. Images were processed using OsiriX (Pixmeo, Geneva, Switzerland). Quantification of fat composition from 3D images was estimated applying a semiautomatic image thresholding segmentation plug-in included in the software and referenced to the total (body) volumes calculated by the same approach. ¹H and ¹³C spectra were processed with MestReNova (Mestrelab Research, Santiago de Compostela, Spain). An exponential line broadening (20 Hz for ¹³C spectrum and 3 Hz for ¹H spectrum) was applied before Fourier transformation. Individual contributions of the signals from carbonylic, olefinic, methylene, and methyl moieties were quantified by integration after automatic baseline correction. The same process was applied in ¹H spectrum. The areas under each peak were used to calculate the fractional contribution of saturated (130 ppm), monounsaturated (130 and 128 ppm), and doubly (poly) unsaturated (128 ppm) fatty acids, taking as a reference the area under the carbonylic peak (24). The areas under the peak in the ¹H spectrum, total lipids (1.3 ppm) and water peak (4.7 ppm), were used to calculate the percentage of total lipids with respect to water content (25).

Mouse epididymal adipose tissue was dissected, weighed, divided into equal pieces, and suspended in 400 μL Krebs-Ringer bicarbonate buffer (Sigma-Aldrich) containing 1% fatty acid-free BSA (Sigma-Aldrich). The tissue samples were treated with either PBS or 25 nmol/L isoproterenol or pretreated with 100 μmol/L succinate (for 10 min) or 100 ng/mL pertusis toxin for 4 h before isoproterenol treatment. Samples were incubated on a shaker at 37°C, and glycerol release was measured using a Glycerol Determination kit (Cayman Chemical Company, Ann Arbor, MI).

After an overnight fast, mice were injected intraperitoneally with succinate (1.2 mg/kg) or the adenosine A₁ receptor agonist (R)-N⁶-(2-phenylisopropyl)-adenosine (PIA; 0.15 nmol/g). Blood was collected before and 15 min after injection. Serum nonesterified fatty acid (NEFA) was measured using the NEFA-HR(2) kit from Wako Chemicals (Richmond, VA).

At the end of the study, mice were killed following an overnight fast. Blood samples were collected by cardiac puncture, and serum was separated by centrifugation. Total serum cholesterol, triglycerides, free fatty acids, hepatic aspartate aminotransferase (AST), and alanine aminotransferase (ALT) were determined using an automated analyzer (Dimension RxL Max Integrated Chemistry System; Siemens, Munich, Germany) and commercially available kits from Siemens (total cholesterol, triglycerides, ALT, and AST) and Wako Chemicals (free fatty acids). Diabetes-related biomarkers in serum were analyzed using the Bio-Plex Pro Mouse Diabetes standard 8-plex and adiponectin kits (Bio-Rad Laboratories Inc., Berkeley, CA). Selected tissues were collected and weighed. Tissues were fixed in neutral buffered formalin solution and processed for histological analysis or snap frozen in liquid nitrogen and stored at −80°C for RNA and protein extraction.

Tissues were fixed in 10% buffered formalin (VWR International, Radnor, PA), embedded in paraffin, and sectioned at 5 μm. For morphometric analysis, sections were stained with hematoxylin and eosin according to standard protocols. Adipocyte size (from WAT) and islet cell area and cell number (from pancreas) were determined with ImageJ software (National Institutes of Health). Sections from six animals per group were used, and at least 500 adipocytes per animal were measured from multiple fields. Islet cell area was quantified from at least four sections per pancreas and expressed as the percentage of total pancreas area. Islet cell density was assessed by counting the number of nuclei in a 1,720 μm² area from the same sections. Additional sections were subjected to heat-induced epitope retrieval in citrate-buffer, pH 6, followed by blocking of endogenous peroxidase activity with hydrogen peroxide prior to the immunostaining. Sections were incubated overnight with antibodies specific for Sucnr1 (Novus Biologicals, Littleton, CO), insulin (Cell Signaling Technology Inc., Danvers, MA), glucagon (Abcam plc, Cambridge, U.K.), Ki67 (Master Diagnostica, Granada, Spain), or caspase-3 (R&D Systems Inc., Minneapolis, MN). For Sucnr1 immunodetection, a tyramide signal amplification kit (Molecular Probes; Thermo Fisher Scientific, Waltham, MA) was used. For insulin/glucagon double staining, we used fluorescence-labeled secondary antibodies. Insulin, Ki67, and caspase-3 were detected using an automatic immunostainer (Dako Autostainer Plus; Dako, Glostrup, Denmark) with a horseradish-peroxidase-labeled polymer anti-rabbit as secondary antibody (EnVision-HRP; Dako) and the chromogenic substrate DAB (Dako), followed by counterstaining with Mayer’s hematoxylin.

For examination of phospho-Akt in WAT, skeletal muscle, and liver samples, fasted mice were injected intraperitoneally with saline or insulin (1.5 units/kg body mass). Extracts were prepared at 10 min postinjection as described (26). Briefly, tissues were homogenized in 500 μL ice-cold lysis buffer (1% Triton X-100, 150 mmol/L NaCl, 10 mmol/L Tris-HCl pH 7.5, 1 mmol/L EDTA, 1% Nonidet P-40, and complete mini protease inhibitor cocktail; Roche, Mannheim, Germany) with a mechanical homogenizer, and debris was cleared by centrifugation at 10,000g at 4°C for 10 min. The supernatant was collected and the quantity of protein was measured with the BCA kit (Thermo Fisher Scientific). Protein extracts (40 μg of protein) were examined by protein immunoblot analysis by probing with antibodies to Akt and phospho-Ser473 Akt (Cell Signaling Technology Inc.). Immunocomplexes were detected with the enhanced chemiluminescence reagent (Amersham, GE Healthcare Bio-Sciences AB, Uppsala, Sweden).

For the intraperitoneal glucose tolerance test (IPGTT), mice were fasted overnight (16–18 h). Conscious mice were then injected intraperitoneally with glucose (1.5 g/kg body weight). Blood glucose was monitored in samples obtained from tail punctures before and after glucose administration using a handheld glucometer (Ascencia Elite; Bayer AG, Leverkusen, Germany). To measure insulin levels in response to glucose, blood was collected at 0 and 10 min postinjection through the submaxilar vein, and the resulting serum was used to measure insulin by ELISA (Millipore Corporation, Billerica, MA). For the intraperitoneal insulin tolerance test (IPITT), mice were fasted for 2 h and injected with insulin (0.75 units/kg).

The stromal vascular fraction (SVF) was isolated from WAT by incubating the tissue for 1 h at 37°C in Krebs buffer (1 g tissue/3 mL solution) containing 2 mg/mL collagenase A (Roche Diagnostics, Basel, Switzerland) and 20 mg/mL fatty acid–free BSA (Sigma-Aldrich). Digested tissue was centrifuged at 250g, and the SVF pellet and adipocyte fractions were collected for RNA extraction. For preadipocyte differentiation, the same procedure was followed using subcutaneous WAT. Once the SVF fraction was collected, SVF cells were differentiated to adipocytes ex vivo using a modification of a published protocol (27). Cells were plated (50,000 cells/cm²) and maintained in culture medium (DMEM/F12, 8% FBS, 2 mmol/L L-glutamine, 50 units/mL penicillin, and 50 μg/mL streptomycin) for 5 days, and differentiation was induced by incubating with culture medium supplemented with 5 μg/mL insulin, 1 μg/mL dexamethasone, 25 μg/mL isobutylmethylxanthine, and 1 μmol/L rosiglitazone. After 2 days, cells were incubated for an additional 2 days in culture medium containing insulin and rosiglitazone only and were then maintained in culture medium supplemented with only insulin for the remaining duration of differentiation (a further 4 days). For flow cytometry analysis of adipocytes, the same enzyme dissociation procedure was followed for epididymal WAT digestion. Then adipocytes were recovered by centrifugation (250g, 10 min, 4°C), collected in 1 mL of ice-cold PBS containing 0.25 μL of Draq5 (BioStatus Limited, Leicestershire, U.K.) to stain the cells, and analyzed on a LSR Fortessa flow cytometer (Becton Dickinson, San Jose, CA). Data were analyzed with FlowJo 1.6 software.

Hepatic and WAT triglyceride content was measured using tissue from mice fasted overnight as previously described (28). Briefly, total lipids were extracted from samples (50 mg) by using an 8:1 mixture of chloroform and methanol (4 h). The extracts were mixed with 1 N sulfuric acid and centrifuged (10 min). Triglyceride content was measured with a kit purchased from bioMérieux Inc. (Marcy l’Etoile, France). Oil Red O staining and measurement of intracellular triglyceride levels were performed as previously described (27). Fixed cells or frozen liver sections (10 μm) were stained with Oil Red O (Sigma-Aldrich) for 15–20 min at room temperature. For quantification, cell monolayers were washed extensively with water to remove unbound dye, and 1 mL of isopropyl alcohol was added to the culture dish. After 5 min, the absorbance (510 nm) of the extract was measured by spectrophotometry.

Total pancreatic insulin content was measured as described (29). Pancreata were harvested from fed mice, Polytron homogenized (30 s) on ice in 5 mL cold 2 N acetic acid, boiled for 5 min, transferred to ice, and centrifuged at 15,000g and 4°C for 15 min. Supernatant extracts were then assayed for protein content and insulin content (ELISA, Millipore Corporation).

Total RNA extraction, cDNA synthesis, and quantification were performed by standard methods. The expression of mRNA was examined by quantitative reverse transcription–polymerase chain reaction (qRT-PCR) using a 7900 Fast Real Time thermocycler and Fast SYBR Green assays (Applied Biosystems, Thermo Fisher Scientific). Three technical replicates were performed per sample, and relative expression was normalized against mouse Srp14. Primer sequences are given in Supplementary Table 1.

Calculations were performed using Biogazelle qBase+ and GraphPad Prism 5.0 software (San Diego, CA) and reported as mean ± SEM. MRI and MRS data were reported as mean ± standard deviation. Statistical comparisons were made using repeated-measures ANOVA with Bonferroni posttests for comparisons between groups at individual time points and Student t test for independent samples, as appropriate. Significance was established as a value of P < 0.05.

Gene transcription analysis of C57BL/6 WT mouse tissues revealed high Sucnr1 expression in kidney and liver, with little or no expression in lung tissue (Fig. 1A). These results are consistent with previous work demonstrating high Sucnr1 expression in kidney medulla (7) and also hepatic stellate cells (12). Interestingly, significant expression was also detected in WAT from epididymal fat (Fig. 1A), ratifying the recent grouping of this gene to an adipose cluster in the adult mouse (22). Analysis of fractionated WAT demonstrated Sucnr1 expression in the mature adipocyte but not the stromal cell fraction, which contains preadipocytes (Fig. 1B). In contrast to WAT, Sucnr1 expression was minimal from brown adipose tissue (BAT) deposits (Fig. 1C).

To investigate the in vivo function of Sucnr1, we generated Sucnr1-mutant mice by using KO-first targeted ES cell clones obtained from EUCOMM. The targeting vector, directed to intron 1 of the Sucnr1 gene (Fig. 2A), contained a splice acceptor (SA) module from the engrailed-2 (En2) gene, followed by an internal ribosome entry site, which directs translation of a fusion reporter protein with β-galactosidase (LacZ), which, in turn, is dependent on transcription of the endogenous Sucnr1 promoter. A separate transcription unit contained a promoter-driven neomycin gene for antibiotic selection. These two modules were flanked by flippase recognition target sites, permitting flippase-mediated recombination. Additionally, exon 2 of the gene, which was included in the targeting vector, was flanked by LoxP sites, allowing conditional ablation of the exon following Cre-mediated recombination (30). A third LoxP site bordered the neomycin cassette (Fig. 2A), ensuring its removal with exon 2. Thus, depending on mating regimes, either whole-body KO mice or conditional KOs can be produced (23). In this work, we describe only results from the GT and whole-body KO mice.

Two targeted ES cells clones were injected into mouse blastocysts, and germ line transmission was confirmed after mating. All mice were on a C57BL/6 background. Mouse genotypes were evaluated by PCR using genomic DNA from tail biopsies, with primers designed to detect correctly targeted events (Fig. 2B, primers 1, 2, and 3). Homozygous offspring from heterozygous Sucnr1^GT/+ male and female intercrosses were viable and exhibited normal development. Whole-body KO mice, distinct from GT mice by the complete removal of exon 2, were produced following a crossing regimen with Sox2-Cre mice. Removal of exon 2 and the neomycin module was confirmed by PCR using primers 4, 5, and 6, giving the expected amplification products for WT and KO (Fig. 2C). Because intronic insertions of GT vectors may allow normal mRNA splicing to some degree, we analyzed endogenous Sucnr1 levels in GT mice to determine whether residual transcripts were expressed. Using primers e1 and e2, which bridge exons 1–2 (Fig. 2A), PCR amplification of cDNA from different tissues of WT mice yielded a single product with the expected size of ∼190 nt (Fig. 2D). Unexpectedly, amplification of cDNA from Sucnr1^GT/GT mice with the same primer pair yielded a larger single product of 300 nt, with no detectable WT Sucnr1 expression (Fig. 2D, left panel) suggesting that the SA is not skipped. Introducing an internal ribosome entry site–specific reverse PCR primer (p7), together with primer e1, resulted in the amplification of an intermediate-sized fragment (220 nt), consistent with the proper functioning of the SA module (Fig. 2D, middle panel). PCR across exons 1–2 in Sucnr1^KO/KO (Sucnr1 KO) mice failed to amplify a reaction product, consistent with the loss of exon 2 (Fig. 2D, right panel). To determine the origin of the unanticipated 300 nt product in Sucnr1^GT/GT mice, and exclude the possibility of PCR artifacts, reactions were performed on mRNA from different tissues of mice, and the products were sequenced. Results from sequencing determined precisely in all cases that the fusion transcript generated by the proper joining of exon 1 to the SA sequence had caused activation of a cryptic splice donor site located 115 nt downstream in the En2 sequence, which spliced to the SA of Sucnr1-exon 2 (Fig. 2E). This splicing resulted in a frameshift with premature stop codons within exon 2, predicting termination of a truncated Sucnr1 translation product. Curiously, this cryptic splice element in the En2 gene has been previously described in the literature (31,32), and it seems rather puzzling that a large-scale genetic KO platform would use the En2-SA unit with a documented alternative splicing phenomenon. Because of this incongruity, only KO mice were characterized further.

In accord with mRNA expression, strong and homogenous β-galactosidase staining was observed throughout WAT from KO mice but not WT littermates (Fig. 2F, top panels), indicating robust β-galactosidase activity in this tissue driven by the endogenous Sucnr1 promoter. Subsequently, immunofluorescence analysis of paraffin-embedded WAT sections revealed that Sucnr1 is present in the plasma membrane of WT adipocytes but is not detected in equivalent samples of KO mice (Fig. 2F, bottom panels). Collectively, these results identify the white adipocyte as a significant reservoir of Sucnr1 expression and suggest that it might have a functional role in adipocyte metabolism.

Hydrolysis of fat stored as TG in adipose tissue, termed lipolysis, is tightly controlled by hormones and metabolites that are secreted according to nutritional status (33). Enzymatic breakdown of TG occurs through a series of steps (34) to release one molecule of glycerol and three molecules of fatty acid. Released fatty acids enter the circulation and can be taken up and oxidized by peripheral tissues. Interestingly, succinate has been shown to inhibit lipolysis in isolated adipose tissue, presumably through engagement with Sucnr1. To investigate whether lipolysis is perturbed in Sucnr1 KO mice, we dissected WAT from epididymal fat regions and measured lipolytic activity in explant tissue. Glycerol release, as a metric of lipolysis, was similar in WT and Sucnr1 KO mice, both under basal and β-adrenergic agonist (isoproterenol)-stimulated conditions (Fig. 3A). In accord with previous findings (22), addition of succinate (100 μmol/L) induced an inhibition of isoproterenol-stimulated glycerol release in WAT explants from WT mice, and pretreatment of adipocytes with pertussis toxin blocked the antilipolytic effect of succinate (Fig. 3B). As anticipated, inhibition of lipolysis by succinate was absent in equivalent WAT from Sucnr1 KO mice (Fig. 3C). Moreover, intraperitoneal injection of succinate (1.2 g/kg) provoked a significant decrease in serum levels of NEFA in WT mice, whereas the decrease in Sucnr1 KO mice was not significant (Fig. 3D). Further, intraperitoneal injection of the adenosine A₁ receptor agonist PIA could inhibit lipolysis both in WT and Sucnr1 KO mice (Fig. 3E), demonstrating that the response to general GPCR-mediated antilipolysis was not perturbed in Sucnr1 KO mice. Taken together, these results establish that succinate exerts its antilipolytic actions through Sucnr1 coupled to Gi-type G proteins.

Given that expression of Sucnr1 in WAT alluded to an involvement in energy homeostasis, we assessed weight gain in WT and Sucnr1 KO mice on an SD. Total body weight did not differ at weaning (13.1 ± 1.1 g for male WT vs. 13.2 ± 0.6 g for male KO; n = 8), and both genotypes displayed similar weight gains, although a discernible trend for decreased weight gain in Sucnr1 KO mice was apparent from 10 weeks of diet (Fig. 3F), which was significant when measured at 50 weeks (41.6 ± 1.5 g for WT vs. 35.5 ± 1.9 g for Sucnr1 KO mice; P < 0.05) (Fig. 3G). Analysis of body composition was performed by dissection of organs and fat pads, nuclear MRI (NMRI), and MRS. After 18–20 weeks on the diet, mice were killed and organ weights were measured. No differences between genotypes were found in the weights of heart, spleen, liver, and BAT at the end of the study (not shown); however, evident differences were found in the WAT depots. Accordingly, retroperitoneal, subcutaneous, and epididymal fat mass was significantly reduced in Sucnr1 KO mice relative to WT littermates (Fig. 3H). Further, compared with total body weight, cumulative fat content was significantly reduced in Sucnr1 KO animals (Fig. 3I). Visual inspection of NMRI images of WT and Sucnr1 KO animals suggested a greater percentage of abdominal fat mass in WT animals relative to Sucnr1 KO littermates (Fig. 3J), and quantification analysis of axial slices demonstrated a trend for reduction (Fig. 3J). In the same experimental setting, ¹³C-NMR spectra of glycerides were recorded over the entire abdominal region (Supplementary Fig. 1A). Quantification of the spectra (24) showed that the fractional contribution of saturated fatty acids to the total pool of glycerides was significantly increased in Sucnr1 KO animals (0.583 ± 0.292 vs. 0.481 ± 0.189; P < 0.05 for Sucnr1 KO vs. WT, respectively), whereas monounsaturated and polyunsaturated fatty acids were comparable between genotypes. Finally, ¹H-MRS were also acquired from the abdominal region to assess lipid content (25). As depicted in representative spectra of mice, the ratio of water to fat content was altered in Sucnr1 KO mice relative to control littermates (Supplementary Fig. 1B), and area under the curve measurements revealed a significant decrease in fat content in Sucnr1 KO mice (Fig. 3K). Collectively, these results are consistent with the fat weight analysis and strongly suggest that Sucnr1 KO mice have a reduced WAT content.

In agreement with the decreased size of the different anatomical WAT compartments in SD-fed Sucnr1 KO mice, histological analysis of hematoxylin and eosin–stained WAT established a significant decrease in adipocyte size relative to WT mice (Fig. 3L). Further, flow cytometry analysis of disaggregated adipocytes from epididymal WAT revealed a significant decrease in adipocyte size, measured as forward scatter (Fig. 3M). These changes were mirrored by a reduction in the concentration of adipose triglycerides (Fig. 3N). As the decrease in adipose mass in Sucnr1 KO mice can result from a reduction in adipocyte size and/or decreased differentiation, we examined the expression of genes related to adipocyte differentiation. Although expression levels of the adipogenic transcription factors, peroxisome proliferator-activated receptor γ (Pparγ) and CCAAT/enhancer-binding protein α (Cebpα), together with the late differentiation marker fatty acid binding protein 4 (Fabp4), tended to be slightly lower in WAT taken from Sucnr1 KO mice, this decrease was not significant (Fig. 3O). Moreover, the differentiation ability of preadipocytes extracted from the SVF of adult mice was comparable between both groups, as measured by Oil Red O staining of neutral lipids (Fig. 3P) and by expression of the adipocyte marker genes Pparγ, protein delta homolog-1 (Pref1), and SREBP-1c (Srebp1c), although a nonsignificant increase in Pparγ and Srebp1c expression was observed in adipocytes from Sucnr1 KO mice (Fig. 3Q). These results, taken together with the findings of low levels of Sucnr1 expression in SVF, a reduction in adipose tissue mass and adipocyte size, and decreased lipid content, collectively suggest that Sucnr1 deletion does not alter adipogenesis but rather results in decreased lipid accumulation and smaller adipocytes.

To further evaluate the metabolic changes caused by Sucnr1 deletion, we measured energy expenditure by indirect calorimetry. Interestingly, the rate of VO₂ consumption in the Sucnr1 KO group was consistently higher than WT counterparts (Fig. 4A). The respiratory exchange ratio (RER; VO₂/VCO₂) of Sucnr1 KO mice was also significantly higher than that of WT mice during the dark cycle (Fig. 4B and E), indicating that perhaps less tissue fat is available for oxidation in these mice. Alternatively, the increased RER values may suggest an increase in whole-body utilization of carbohydrates and protein over lipids. Surprisingly, food intake was increased in Sucnr1 KO mice compared with WT (Fig. 4C and F), and energy expenditure was greater during the dark cycle (Fig. 4D and G), with no change in locomotor activity between the two genotypes (Fig. 4H). This increase in food intake together with the reduction in fat mass could be due to evident increases in energy expenditure. Though we were unable to find differences in BAT expansion between WT and Sucnr1 KO mice, dysregulated energy expenditure often results in overexpression of uncoupling proteins (UCPs) to dissipate energy. However, analysis of UCP expression revealed comparable levels of UCP2 and UCP3 in skeletal muscle and liver (Fig. 4I). Interestingly, although changes did not reach significance, there was a trend for lower UCP1 expression in BAT tissue of Sucnr1 KO mice, together with elevated levels of UCP1 and UCP2 expression in WAT tissue (Fig. 4I), suggesting perhaps compensatory mechanisms. Collectively, these results suggest that deletion of Sucnr1 results in substantial alterations in energy expenditure and disposition.

We next challenged mice with HFD to provoke obesity. HFD feeding led to increased weight in both genotypes; however, Sucnr1 KO mice gained weight at a slower rate than WT littermates, which was significant after only ∼3 weeks of feeding (25.7 ± 0.3 vs. 28.6 ± 0.5 g; P = 0.0006). Variance in body weight widened until week 14 when differences became less pronounced and weight curves were superimposable (Fig. 5A). Predictably, animals on HFD were substantially heavier than SD-fed animals at the end of the study (compare Fig. 5A and Fig. 3F), and this correlated with an increase in the weight of WAT and BAT compartments and also liver. Surprisingly, although Sucnr1 KO mice on HFD were generally leaner than WT counterparts, the weight of WAT stores at the end of the study was significantly increased (Fig. 5B), whereas adipocyte size was comparable (Supplementary Fig. 2). This finding was in contrast to that observed under SD, where Sucnr1 KO mice had significantly less WAT (Fig. 3H and I). In a second cohort of HFD-fed mice used for NMRI analysis, body weights of Sucnr1 KO mice after 20 weeks on diet were found to be significantly lower than equivalent WT mice (Fig. 5C), in keeping with the overall observations. Although no differences were found in the weight of different anatomical WAT compartments between groups in this cohort (Fig. 5D), the proportion of fat as a percentage of total body weight was significantly higher in Sucnr1 KO animals, as determined by measurement of combined fat depots (Fig. 5E). Inspection of NMRI images indicated comparable abdominal fat volume in both groups (Fig. 5F), and quantitative assessment of fat from axial slices revealed no differences (Fig. 5F). A more accurate quantification was carried out using ¹H-MRS to determine lipid content. In this case, area under the curve measurements revealed a significant increase in fat content in Sucnr1 KO mice (Fig. 5G), in keeping with the findings from the first cohort.

Results from indirect calorimetry at 8 weeks on HFD revealed no significant differences in VO₂ consumption or RER (Fig. 5H) in Sucnr1 KO mice compared with WT control, although ingestion of the HFD reduced RER values in both groups relative to SD. Again, total food consumption was modestly increased in Sucnr1 KO mice, although this did not reach significance (Fig. 5H). Furthermore, locomotor activity was not significantly different between groups (data not shown). These results indicated that the initial slower rate of weight gain in Sucnr1 KO mice under HFD was not attributable to alterations in energy expenditure, food intake, or activity.

Changes in adipose mass were reflected in alterations of adipokine levels in serum of animals. Compared with WT littermates, the reduced quantity of adipose tissue in Sucnr1 KO mice under SD was paralleled by decreased circulating levels of the obesity indicator leptin (Fig. 5I). This finding is in keeping with the increased feeding of Sucnr1 KO animals on SD (Fig. 4C and F) and the satiety effect of this hormone (35). In contrast, HFD feeding induced a >10-fold increase in serum leptin, which was not significantly different between Sucnr1 KO and WT mice (Fig. 5I). In addition, a small but significant decrease in serum levels of the insulin-sensitizing hormone adiponectin was observed for Sucnr1 KO mice under both dietary regimes (Fig. 5I). However, serum levels of the adipokines resistin and plasminogen activator inhibitor-1, although increased in HFD, were not statistically different between genotypes (Fig. 5I). Obesity in humans and animal models is often associated with increased circulating lipids and ectopic lipid accumulation in nonadipose tissues including liver (36). Consistent with this phenotype, both fasted serum levels of NEFAs and cholesterol were elevated in WT mice and Sucnr1 KO mice fed HFD for 18 weeks, whereas TG concentrations remained in the normal range (Fig. 5J). In accord with the measured lipolytic rates in explants (Fig. 3A), no differences in plasma parameters were observed between genotypes. Although no morphologic alterations were found in liver between groups fed an SD, hepatic steatosis was induced in both genotypes under HFD, evidenced by hematoxylin and eosin staining of abundant vacuolar vesicles in liver sections, which were confirmed as lipid droplets by Oil Red O staining (Fig. 5K). Additionally, hepatic triglyceride content in both genotypes was significantly increased under HFD, consistent with the fatty changes to liver (Fig. 5L); however, no significant differences in liver TG content were found between Sucnr1 KO and WT mice under either diet. Further, HFD levels of AST and ALT were found to be significantly increased in Sucnr1 KO mice compared with WT mice, suggesting increased hepatocyte damage (Fig. 5M); the origin of these differences is not known. Consistent with the increased lipid load in liver, expression analysis revealed changes in relevant genes (Supplementary Fig. 3), including the lipogeneic transcription factor sterol regulatory binding protein 1c (Srebp1c), and its target stearyol CoA desaturase-1 (Scd1), which were induced under HFD in both genotypes; however, other genes involved in triglyceride synthesis, fatty acid synthase (Fas) and stearyol CoA desaturase (Acc), were unaltered in HFD-fed mice. Similarly, expression of the fatty acid transporter CD36 was upregulated under HFD in both genotypes, although expression was significantly reduced in Sucnr1 KO mice compared with WT littermates on SD. Additionally, a modest but significant increase in the expression of a second fatty acid transporter, Fabp4, was found in Sucnr1 KO but not WT mice under HFD diet, whereas expression of the fatty acid transport protein 2 (Fatp2) was downregulated in Sucnr1 KO mice under HFD. In contrast to these changes, no differences were found between genotypes for the expression of genes related to fatty acid oxidation, although carnitine palmitoyltransferase 1a and 1b (Cpt1a and Cpt1b) were downregulated under HFD in both genotypes. Collectively, these data indicate that although the absence of Sucnr1 results in subtle changes to liver function metrics compared with WT mice, HFD appeared to cause comparable steatosis.

Although the metabolic phenotype of Sucnr1 KO mice appeared modest, the evident differences in adiposity and weight gain prompted us to evaluate glucose and insulin homeostasis. These parameters were measured in mice after 16 weeks on diet. Compared with SD, HFD led to an increase in both fasted and fed levels of serum glucose in Sucnr1 KO mice and WT littermates (Fig. 6A). Serum glucose was comparable between genotypes in SD; however, despite their similar final body weights in HFD, Sucnr1 KO mice had significantly higher levels of fasting blood glucose, whereas fed glucose was unaltered (Fig. 6A, right panel). Nevertheless, Sucnr1 KO mice showed similar levels of fasting insulin as those from WT mice in HFD, which were elevated compared with those in SD (Fig. 6B). To directly assess whether glucose homeostasis was dysregulated in Sucnr1 KO mice, we performed glucose tolerance tests and insulin tolerance tests. Interestingly, in the SD group, intraperitoneal challenge with 1.5 g/kg glucose resulted in faster clearance in Sucnr1 KO mice, as glucose concentrations remained significantly lower compared with WT mice (Fig. 6C, left panel). This is in keeping with the general leanness of KO animals and higher RER values (Fig. 4B and E), and suggested improved insulin sensitivity that promotes carbohydrate utilization. As anticipated, glucose clearance deteriorated in the HFD group (Fig. 6D, left panel). However, in marked contrast to the findings under SD, Sucnr1 KO mice fed an HFD developed a clear impairment in glucose elimination compared with WT mice, with manifest hyperglycemia 120 min after glucose injection. Notably, this decreased ability to lower blood glucose was not apparent in Sucnr1 KO mice at 11 weeks on HFD (Supplementary Fig. 4), suggesting a progressive glucose intolerance concurring with the increased WAT weight. Abnormal glucose homeostasis could result from either a defect in insulin sensitivity of the peripheral tissues or a defect in insulin secretion. Insulin tolerance tests showed that although both genotypes exhibited decreased insulin sensitivity under HFD, Sucnr1 KO mice exhibited comparable insulin sensitivity to WT littermates, as demonstrated by a quick decline in plasma glucose after an acute insulin injection (Fig. 6C and D, right panels). Furthermore, insulin was shown to induce phosphorylation of the major insulin-signaling protein Akt (on Ser473) in liver, WAT, and skeletal muscle to similar intensities in Sucnr1 KO and WT mice on HFD (Fig. 6E). These results pointed to a defect in insulin secretion as the origin of the hyperglycemia in HFD-fed KO mice. To consider this, we repeated the glucose tolerance test and measured circulating insulin levels at baseline and 10 min after a glucose load (1.5 g/kg). Whereas WT mice showed a characteristic increase in insulin levels at t = 10 min, Sucnr1 KO mice failed to secrete insulin, although GLP-1 levels were similar in both genotypes (Fig. 6F). Collectively, these results imply that while Sucnr1 KO mice have significantly enhanced glucose tolerance under SD, the HFD causes a sufficient deficiency in insulin secretion to cause glucose intolerance.

Guided by these findings, we examined islet anatomy and endocrine cell size in pancreatic tissue sections of HFD-fed mice (Fig. 6G). As expected, morphometric analysis revealed that high-fat feeding increased islet cell size, and this increase was comparable in both genotypes (Fig. 6H). Moreover, β-cell number was also similarly increased in Sucnr1 KO and WT mice under HFD, compared with SD (Fig. 6I). Also, total (acid-extracted) pancreatic insulin concentration was similar between genotypes on HFD (Fig. 6J). Immunohistochemistry staining with an antibody to insulin showed that the morphologic features of WT and KO islets were comparable (Fig. 6K, left panels), and detailed analysis of the islets revealed a normal cell architecture shared between genotypes, with positive immunostaining for glucagon (α-cells) in the periphery, while the area of anti-insulin staining was observed mostly in the central region (Fig. 6K, right panels). Examination of β-cell proliferation with an antibody to Ki67 indicated a similar and low (<1%) frequency of Ki67-positive cells in both WT and KO mice, and very few apoptotic (caspase-3 positive) β-cells were detected in either genotype (Fig. 6L). Thus our results are consistent with an insulin secretory defect in Sucnr1 KO mice.

Impaired cellular metabolism is a defining hallmark of many disease states and can affect homeostatic signaling processes. As the Krebs cycle is the central metabolic pathway in aerobic organisms, it is particularly significant that succinate, an intermediate metabolite of this pathway, has an alternative occupation as a signaling molecule for Sucnr1 coupling.

Information regarding the function of Sucnr1 is limited. Previous studies have shown that Sucnr1 is distributed in organs with high metabolic demand, including the kidney and liver (19). Signaling of Sucnr1 appears to be an early response to hyperglycemic conditions in the kidney and leads to activation of the intrarenal renin-angiotensin system, resulting in hypertension and potential tissue injury (37). Here we show that Sucnr1 is expressed in high levels in the adipocyte and that loss of Sucnr1 leads to dichotomous effects on body weight and metabolism. Although not conspicuously different during early growth, after 16–20 weeks on SD (at 21–25 weeks of age) Sucnr1 KO mice exhibited a modestly lean phenotype compared with WT mice, with significantly reduced levels of plasma leptin and reduced WAT mass containing smaller adipocytes and decreased triglyceride content. This decrease in adiposity was not associated with ectopic triglyceride storage in liver or plasma (Fig. 5) or skeletal muscle (not shown) but correlated with a significant increase in glucose tolerance and elevated metabolic activity, including increased energy expenditure, which might partly explain why Sucnr1 KO mice were lean despite hyperphagia. However, this phenotype was seen in the setting of normal levels of BAT content and tissue expression of UCPs, which are often elevated in models of increased energy expenditure (38). Decreased weight gain in Sucnr1 KO mice was also observed in the HFD study, particularly during the initial 10 weeks. This finding supports our observations on SD and strongly suggests that leanness is a primary outcome of Sucnr1 deletion. Since adipogenesis was not affected by deletion of Sucnr1 and an antilipolytic effect of succinate had been previously described, this metabolically favorable phenotype might be explained, in part, by the release from succinate-inhibited lipolysis. Notably, experiments both in vitro and in vivo verified that succinate inhibited the breakdown of triglycerides, although basal and isoproterenol-induced lipolysis was unchanged in Sucnr1 KO animals. This phenotype, in part, is consistent with other mouse models of “unrestrained” lipolysis and shows that release from Gi-coupled inhibition of fat mobilization results in leanness (39–41) and increased energy expenditure (41). After prolonged high-fat feeding, Sucnr1 KO mice displayed a rather more complex phenotype, presumably a result of nutritional stress induced by the HFD. Under these conditions, end-of-study body weights of Sucnr1 KO mice were, in the main, comparable with WT controls, but WAT was significantly heavier. Furthermore, after 16 weeks on HFD, Sucnr1 KO mice had significantly greater levels of blood glucose and failed to secrete insulin in glucose tolerance tests; however, pancreas architecture was not different than in WT mice. Ostensibly, this hyperglycemic state may have resulted from the failure to buffer serum glucose, although this inference is tempered by the result that insulin sensitivity was unchanged in Sucnr1 KO mice. Still, compensatory mechanisms such as neuronal or endocrine modulation may account for this discrepancy between insulin secretion and resistance.

Although our original hypothesis presupposed that a lean phenotype would result from augmented WAT lipolysis, our findings of increased metabolic activity, under basal conditions, in Sucnr1 KO mice points to a more systemic effect of Sucnr1 deletion. Whether these phenotypes are triggered directly by differences in adipose mass in Sucnr1 KO mice or through a more circuitous route awaits further study. Nevertheless, adiponectin levels, which are typically inversely correlated with obesity, were decreased in Sucnr1 KO mice and might reflect specific changes in adipocytokine signaling (42), leading us to hazard that a hyperglycemic phenotype might occur when fat content reaches a critical mass.

Many mechanisms can account for leanness, but they all generally result from deregulation of energy balance, as noted here. We can speculate on several potential mechanisms that may promote this phenotype in Sucnr1 KO mice (43): hypothalamic signal transduction (44), reduced intestinal absorption of foods leading to decreased energy availability (45), enhanced energy metabolism in peripheral tissues, or a defect in adipose tissue storage capability. It remains to be determined whether one or more of these mechanisms are operating in our model.

Finally, it should be noted that an underlying concept in Sucnr1 signaling is hypoxia-induced modulation of receptor engagement through local increases in succinate levels. This is particularly manifest in ischemic retinopathy, where microvascular degeneration results in an abnormal hypoxia-driven succinate-induced neovascularization through Sucnr1-dependent upregulation of proangiogenic factors (13,46). Interestingly, hyperglycemia can also result in regional increases in succinate (16,46), most likely through overactivation of the Krebs cycle (47). Indeed, adipose tissue succinate has been shown to increase (twofold) in mice under HFD (48), and hypoxia has been detected during expansion of adipose tissue in obesity (49,50), doubtless occurring when vascularization is insufficient to maintain oxygen levels. Thus increases in succinate, hypoxia, and/or hyperglycemia might trigger Sucnr1-induced signaling at the level of the adipocyte. Clearly, because Sucnr1 is expressed in organs with high metabolic demands, tissue-selective KOs will be required to identify the specific cell type(s) in which Sucnr1 plays a direct role; nevertheless, our results indicate that activation of Sucnr1 may have therapeutic implications in the treatment of obesity-related disorders.

Acknowledgments. The authors thank Guadalupe Sabio and Tamas Rözer for helpful discussions and Roisin Brid Doohan, Ana Belén Ricote, and Jesús María Ruiz-Cabello for technical assistance (all from CNIC).

Funding. This work was supported in part by a research grant from the Ministerio de Ciencia e Innovación (SAF2009-07965 to K.J.M. and SAF2010-15239 to B.G.G.). The CNIC is supported by Ministerio de Economía y Competitividad and the Pro-CNIC Foundation.

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

Author Contributions. K.J.M. and A.M.C. conceived the project, performed experiments, interpreted data, and wrote the paper. S.E., B.G.G., M.B., A.d.M., and P.S. performed experiments and interpreted data. A.M.C. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.