Clinical studies
Anthropometric and metabolic examinations
All examinations were conducted in the morning after an overnight fast. Anthropometric measures were determined; obesity was defined as body mass index (BMI) ≥ 30 kg m−2, and waist and hip circumferences were used to calculate waist-to-hip ratio. Body fat percentage was determined by bioimpedance (Bodystat). Energy expenditure was measured by indirect calorimetry using an open-system ventilated hood (Deltatrac II; Datex-Ohmeda). Blood samples were obtained to measure circulating levels of leptin and insulin. Homeostasis Model Assessment of insulin resistance (HOMA-IR) was calculated as (fP-glucose in mmol l−1 × fS-insulin in mU l−1)/22.5 (ref. 49). Participants in cohort 1 (which includes people living with obesity scheduled for bariatric surgery and non-obese controls, NCT01727245) underwent a hyperinsulinaemic euglycemic clamp as described50. The mean glucose infusion rate (glucose disposal) between 60 and 120 min was determined (M value is mg of glucose uptake per kg of body weight per min) and was expressed corrected for mean plasma insulin during steady state (M/I). Clinical data for cohorts 1 and 2 are presented in Supplementary Table 1.
Investigations of WAT biopsies
Subcutaneous WAT needle biopsies were obtained under local anaesthesia from the periumbilical area51. These isolated fat cells were prepared by collagenase digestion52. Mean fat cell volume was determined and lipogenesis experiments were performed as described53. For the latter, the medium was supplemented with a low concentration of unlabelled and tritiated glucose and insulin-stimulated incorporation of 3H into total lipids was determined. Results were expressed as amount of glucose incorporated into lipids (nmol glucose 2 h−1 (107 fat cells)−1).
Omics in clinical samples
Plasma and WAT metabolomics data from cohort 1 and WAT transcriptomic data from cohort 3 were generated and presented in ref. 7, ref. 54 and ref. 15, respectively. Targeted gene expression analyses in men and women were performed in previously published data from ref. 14. For correlations between metabolites, gene expression and clinical parameters, the Hmisc package (rcorr function) in R Studio (v.4.2.1) was used.
Ethical approvals
All studies were approved by the regional ethics boards in Stockholm (cohort 1 and 3) and Paris (cohort 2), and informed written consent was obtained from all study participants.
Mouse studies
For all animal studies, mice were handled following the European Union laws and guidelines for animal care, health inventories were performed according to the guidelines of the Federation of European Laboratory Animal Science Associations and special care was taken to minimize animal suffering and to reduce the number of mice used.
HFD intervention in wild-type mice
Male C57BL/6J mice, sourced from Charles River Laboratories (France) at 7 weeks of age, were group-housed under a 12-h light-dark cycle with ad libitum access to food and water. Following 1 week of acclimatization, the mice were assigned to either a standard chow diet or a HFD for an additional 15 weeks. For the CB-839 study, C57BL/6J male mice from Envigo (France) were acclimatized to a standard chow diet for 2 weeks. Housed in groups of five, they were subjected to a 12-h light/12-h dark cycle with ad libitum access to food and water. On reaching 10 weeks of age, mice were switched to either a standard chow diet or an HFD for 5 weeks. After 3 weeks on the HFD, mice were randomly allocated to treatment groups on the basis of body weight. Those on the HFD received daily gavage administration of CB-839 (200 mg kg−1 body weight) or vehicle for 19 days (25% (w/v) hydroxypropyl-β-cyclodextrin (ThermoFisher) in 10 mmol l−1 citrate buffer (pH 2)), while the standard chow group received vehicle only. Blood glucose and plasma insulin levels were monitored 12 days posttreatment initiation. After 19 days, mice were euthanized and tissue weights were recorded. Samples were snap-frozen in liquid nitrogen. All mice were housed in the animal facility of Pitié-Salpetrière, in conformity with EU regulations, and studied according to a protocol that received ethics approval from the French ministry for research.
Generation, interventions and phenotyping of GlsAdipoq-Cre mice
Adiponectin-Cre mice were bred with Glsfl/fl mice (Glstm2.1Sray/J, stock 017894) to generate adipocyte-specific Gls-depleted mice (GlsAdipoqCre). Glsfl/fl littermates were used as controls. Mice were housed in groups at the KM-B animal facility in ventilated cages (with a 12-h light/12-h dark cycle (lights between 6:00 and 18:00) in a temperature-controlled (20–24 °C, 50% humidity) facility with ad libitum access to food and water. All experimental procedures were approved by the Stockholm North Animal Ethical Committee. Five to seven-week-old male mice were fed with an HFD (D12492i, 60% kcal fat, Research Diets) for a duration of 7 weeks. Five weeks after the start of the HFD, mice were fasted for 4 h and glucose tolerance was assessed by an intraperitoneal glucose injection (1.5 g kg−1). Blood glucose concentration was monitored from the tail tip using a glucometer (Contour XT, Bayer) before and at the indicated time intervals following glucose injection. Indirect calorimetry was performed 1 week later in the Phenomaster Home cage system (TSE Systems). Ambulatory and locomotor activity was automatically assessed by counting the number of photo beam breaks in the x and y axis. A feeding sensor monitored food intake without disturbances by the experimenters. Lean and fat mass were measured before the start of the HFD and the indirect calorimetry protocols using the EchoMRI-100 (EchoMRI). Seven weeks after the start of the HFD protocol, at 13–15 weeks of age, mice were euthanized under general anaesthesia by avertin injection, and the wet weight of each dissected tissue was measured. Subsequently, one part of the samples was snap-frozen in liquid nitrogen and one part was fixed in in 4% paraformaldehyde (PFA). The same procedures were conducted in another cohort of mice kept on a standard chow diet (CRM (P), 801722, Special Diets Services). Chow mice were individually housed in metabolic cages at 14–18 weeks of age and dissected 4 weeks later.
Plasma, media and tissue analyses of glutamine, glutamate and lactate
The glutamine-to-glutamate ratio in plasma, media and in iWAT was measured using the Glutamine/Glutamate-Glo Assay (Promega). Lactate measurements in snap-frozen iWAT samples were performed at The Swedish Metabolic Center (detailed in the targeted metabolic analysis below). Data were normalized by tissue weight.
Immunofluorescence analyses in murine WAT
As described in ref. 54, WAT samples were fixed in 4% PFA for 1 day, embedded in paraffin, cut into 5-μm sections and stained with hematoxylin and eosin (Sigma-Aldrich). For immunostaining, sections were rehydrated by successive baths in xylene, ethanol and PBS followed by blocking with 10% goat serum. The tissue sections were subsequently incubated overnight with antibodies directed against UCP1 (1:100), COX4 (1:100) and TOM20 (1:100). Goat anti-Rabbit Rhodamine Red-X (1:500) was used a as a secondary antibody and Hoechst 34580 (1:500) was applied for 20 min to counterstain nuclei. Images were acquired using a Axio Observer Z1 inverted fluorescence microscope (Zeiss) and the AxioVision software.
Determination of fat cell size by image analysis
Qupath (v.0.3.4)55 was used to export histology images (tiles 2,048 × 2,048 pixels in .tiff format, 5× downsample). Fiji56 with the Adiposoft57 plugin (v.1.16) was used to quantify adipocyte size of whole tissue sections, pixel size was set to 1.2 μm, with expected diameter of 5–150 μm. An unpaired Student’s t-test was performed in R v.4.2.3 (R Core Team, v.2021) using the rstatix package.
Single-nucleus RNA sequencing
Nuclei isolation and library preparation
BATs and WATs were collected from chow diet-fed Glsfl/fl mice and GlsAdipoqCre mice (n = 5 for each genotype) and promptly frozen in liquid nitrogen. Samples from each depot and genotype were pooled for nuclei isolation. Briefly, frozen tissue samples were minced and homogenized in cold lysis buffer using a gentleMACS Dissociator (Miltenyi Biotec). After addition of lysis buffer with Triton X-100 (X100, Sigma-Aldrich), the lysates were filtered and washed before centrifugation to collect the nuclei pellet. The isolated nuclei were stained with DAPI (D9542, Sigma-Aldrich), sorted, and counted before loading onto a 10X Chip G (10X Genomics). Libraries were prepared using the Chromium Single-Cell v.3.1 kit and sequenced on a Nextseq 2000 platform (Illumina).
Single-nucleus RNA sequencing data preprocessing
Raw sequencing files were processed using Cell Ranger v.7.0.1 based on the default parameters to demultiplex cell barcodes and generate cell-by-gene expression matrices. The mm10 mouse genome (refdata-gex-mm10-2020-A) from 10X genomics was used. We next applied SoupX v.1.6.2 (ref. 58) and DoubletFinder v.2.0.3 (ref. 59) to remove ambient RNA contamination and doublets, respectively. We filtered cells with more than 5% of mitochondrial RNA and excluded haemoglobin and mitochondrial genes, ribosomal protein families, MTRNR and MALAT1 for downstream analyses.
Integration and cell type identification
We integrated the libraries on experimental conditions to remove batch effects using scVI v.0.16.2 (ref. 60). Briefly, all matrices were merged by Seurat v.4.1.3 (ref. 61), and a subset of the top 2,000 highly variable features was identified. From scVI, get_latent_representation was used for generating the latent embedding. By leveraging this latent embedding, we built a shared nearest neighbour graph, clustered cells and visualized all cells in a two-dimensional embedding by using Seurat’s FindNeighbors, FindClusters and RunUMAP. FindMarker was used for determination of differentially expressed genes. We annotated clusters through comparison of highly expressed genes of each cluster with well-established adipose cell type specific markers62,63.
High-resolution respirometry in adipose tissues
Dissection and preparation of iWAT pieces
Inguinal WAT mitochondrial respiration was measured ex vivo using high-resolution respirometry following previously established methods (Oxygraph 2k, Oroboros)64. From Glsfl/fl and GlsAdipoqCre mice, iWAT was dissected, cleaned and cut in approximately 20 mg pieces. In total, 40 to 55 mg of the tissue was placed in 2 ml of ice-cold BIOPS buffer (2.77 mmol l−1 CaK2EGTA anhydrous, 7.23 mmol l−1 K2EGTA anhydrous, 5.77 mmol l−1 Na2ATP, 6.56 mmol l−1 MgCl2-6H2O, 20 mmol l−1 Taurine, 15 mmol l−1 Na2 phosphocreatine, 20 mmol l−1 Imidazole, 0.5 mmol l−1 eithiothreitol, 50 mmol l−1 MES) until the start of the assay. Excess of BIOPS buffer was then removed by blotting the tissue on filter paper before placing the samples in a respirometry chamber containing 2.1 ml of MIR05 buffer (0.5 mmol l−1 EGTA, 3 mmol l−1 MgCl2-6H2O, 60 mmol l−1 lactobionic acid, 20 mmol l−1 taurine, 10 mmol l−1 KH2PO4, 20 mmol l−1 HEPES, 110 mmol l−1 d-sucrose, 1 g l−1 fatty acid free bovine serum albumin).
High-resolution respirometry for inguinal white adipose
After adding the tissue into the respiration chamber, the chamber was closed and allowed to equilibrate before adding any substrates. Leak respiration associated with complex I was assessed by adding 2 mmol l−1 malate, 10 mmol l−1 pyruvate and 10 mmol l−1 glutamate. After O2 consumption rates were stabilized, ADP was added to a final concentration of 5 mmol l−1, triggering complex I-driven coupled respiration. After stabilization, complex II respiration was stimulated by adding 10 mmol l−1 succinate, obtaining the respiration rate corresponding to complex I + complex II-driven respiration. The maximal capacity of the electron transfer system (ETS) was assessed by titrating carbonyl cyanide-4-(trifluoromethoxy) phenylhydrazone in 0.5 μmol l−1 steps. Complex II-linked ETS capacity was evaluated by inhibiting complex I respiration with 0.5 µmol l−1 rotenone. Finally, 2.5 µmol l−1 antimycin A was added into the chamber, inhibiting complex III and therefore blocking mitochondrial electron circulation. The subsequent residual oxygen consumption—corresponding to non-mitochondrial respiration—was subtracted from the previous measured states. O2 consumption was normalized to the amount of tissue mass added in each assay. Mitochondrial membrane integrity and damage was assessed by adding cytochrome C after measuring complex I respiration. In our measurements, we did not observe differences in O2 consumption following this control step.
Dissection and preparation of BAT pieces
Interscapular BAT was dissected and separated from the surrounding white fat. Approximately 20 mg of the tissue was placed in a microcentrifuge tube containing 200 µl of MIR05 buffer and gently homogenized with a handheld pestle homogenizer. Then 20 µl (roughly 2 mg) of homogenate was added to the respiratory chamber containing 2.1 ml of MIR05 buffer.
High-resolution respirometry protocol for BAT
Uncoupled complex I- and complex I + complex II-driven mitochondrial respiration and maximal ETS capacity were evaluated by following the protocol described above for iWAT, minus the addition of ADP into the respirometry chamber.
Cell studies
Differentiation and perturbations of human adipocytes
Isolation, proliferation and differentiation of human adipocyte progenitor cells (from an anonymous male donor, ethical permit no. 2009/764-32, regional ethics board of Stockholm) were performed as described7. Short interfering oligonucleotides (siRNAs) were transfected by electroporation using a Neon Transfection System (1,300 V, 20 ms, two pulses) using the 100 µl of Kit (Invitrogen) at day eight of adipocyte differentiation. All transfections were performed using a final concentration of 20 nmol l−1 siRNA oligonucleotides. Results were compared with non-silencing control siRNAs. A full list of RNA interference (RNAi) oligonucleotides is provided in Supplementary Table 2. In addition to the inhibitors used in the Seahorse assays, the cells were treated with the following chemicals (final concentrations and incubation times): dimethylsulfoxide (DMSO), BPTES (10 μmol l−1 for 3 h), CB-839 (10 μmol l−1 for 6 h), eta-ketoglutarate (2 mmol l−1 for 1 day), GPR81 agonist (500 nmol l−1 for 1 day), insulin (50 nmol l−1 for 15 min or 4 h), rapamycin (50 nm l−1 for 4 h), TNF (2 ng ml−1 for 4 h), STAT3 inhibitor VII (5 μmol l−1 for 5 h), NF-κB inhibitor (trifluoroacetate, 50 μg ml−1 for 5 h), JNK inhibitor SP600125 (10 μmol l−1 for 5 h), isoprenaline (10 μmol l−1 for 6 h), SB203580 (10 μmol l−1 for 24 h in RNAi experiments and a total of 3 h in lactate treatments including 2.5 h of preincubation), 2-deoxyglucose (100 μmol l−1 for 4 h), UK5099 (10 μmol l−1 for 4 h), pyruvate (5 mmol l−1 for 4 h), PF-739 (5 μmol l−1 for 1 day), sodium lactate (20 mmol l−1 for 30 min) and HIF Inhibitor VI (100 μmol l−1 for 4 h). For the glutamine depletion experiment, DMEM/F12 without glutamine was used either supplemented or not with 2.5 or 10 mmol l−1 of glutamine. All catalogue numbers and suppliers are detailed in Supplementary Table 2.
To generate cells with doxycycline-inducible GLS expression, the pCW-Cas9 plasmid (Addgene, no. 50661) was digested using BamHI-HF and NheI-HF and dephosphorylated using Calf Intestinal Alkaline Phosphatase according to the instructions from New England Biolabs (NEB). The digested plasmid was gel-purified using NucleoSpin Gel and PCR Clean-Up (Macherey-Nagel) and ligated overnight at 16 °C using T4 DNA ligase (NEB) with a codon-optimized sequence of GLS. The insert was generated by PCR from two distinct gBlocks as the sequence complexity was too high for synthesis (primers listed in Supplementary Table 2). Stable competent cells (E. coli, High efficiency, NEB, cat. no. C3040H) were transformed with the ligated plasmid according to the manufacturer’s instructions and single colonies were selected and cultured. From these, plasmids were extracted using QIAprep Spin Miniprep Kit (Qiagen) and verified by Sanger sequencing. Lentiviruses were created by transfecting human embryonic kidney 293 cells with this plasmid together with two additional packaging vectors (Addgene nos. 12259, 12260) using Opti-MEM and Lipofectamine 3000 (ThermoFisher) according to the manufacturer’s instructions. Viruses were collected from the conditioned media 2 days following transfection and filtered using Puradisc 25-mm 0.45-μm PES syringe filters (VWR). Around 200,000 proliferating adipocyte progenitor cells per well were seeded and cultured without antibiotics, then spinfected with 50,000 ng of virus for 1 h, 800g at 37 °C. Three days later, cells were incubated with 1 ng ml−1 puromycin. The selection was stopped when control cells in parallel wells were no longer viable. Induction of GLS was carried out by adding 2 ng ml−1 doxycycline to the cell culture media for at least 1 day.
To reintroduce GLS in siRNA-transfected adipocytes, GLS mRNA was in vitro transcribed using a DNA template encoding a codon-optimized coding sequence of GLS driven by a T7 promoter according to the HiScribe T7 ARCA mRNA Kit (with tailing, NEB no. E2060S). N1-Methylpseudo-UTP (cat. no. NU-890L, Saveen Werner) was added to the reaction to increase mRNA stability and thereby extend the intracellular expression. A total of 37 pmol of GLS mRNA was added in each Neon transfection (1,700 V, 20 ms, one pulse) with or without GLS siRNA as described above. Primers used to generate the DNA template are listed in Supplementary Table 2.
Mature adipocyte isolation and culture
Following dissection, WAT was subjected to careful washing and fine mincing. The minced tissue was thereafter placed in vials containing a Krebs–Ringer phosphate buffer (composed of 0.9% NaCl, 0.1 M NaH2PO4, 0.11 M CaCl2 and 0.154 M KCl, with a pH of 7.4) supplemented with 10% FBS, 5 mmol l−1 d-glucose, 1% penicillin-streptomycin (Thermo Fisher Scientific) and 0.05% collagenase type 1 (Sigma-Aldrich). The samples were placed in a shaking water bath set at 37 °C for a duration of 45 min. Subsequently, any connective tissue or substantial undigested fragments were effectively eliminated by filtration through a 250-µm nylon mesh. The resultant floating adipocytes underwent a series of four washes using PBS supplemented with 10% FBS, 1% penicillin-streptomycin and 5 mmol l−1 glucose (all from Sigma-Aldrich). The washed adipocytes were then poised for downstream analyses, either in their fresh state or following freezing at a temperature of −80 °C.
Human mature adipocytes were cultured within 6.5-mm Transwells (Costar-3413 and 3397), following a previously published protocol65. In brief, 30 µl of densely packed human adipocytes (approximately 60,000 cells) were carefully dispensed into each well. The Transwells were then inverted and positioned over a well of 24-well plates containing 1 ml of medium comprising DMEM/F12, 10% FBS and 1% penicillin-streptomycin. After a 4-h stabilization period, the cells were incubated with the following chemicals (final concentrations and incubation times): DMSO, CB-839 (10 μmol l−1 for 2 days) and TNF (10 ng ml−1 for 1 day).
RNA isolation, cDNA synthesis and qPCR
Total RNA was extracted from cells or intact human/murine WAT as described7. The RNA concentration and purity was measured using Varioskan Lux (ThermoFisher) and samples were reverse transcribed using iScript complementary DNA (cDNA) synthesis kits (BioRad). Messenger RNA levels were determined using TaqMan (Applied Biosystems) or SYBR-green (BioRad) assays and relative expression levels were calculated with the comparative Ct-method, that is, 2ΔCt-target gene/2ΔCt-reference gene. All primers are listed in Supplementary Table 2.
Western blot
Western blots were performed as described previously7 except that the membranes were scanned using a ChemiDoc MP system (BioRad). For experiments where the proteins of interest run at the same size, lysates were subdivided in equal amounts and loaded on separate gels as detailed in the respective figure legends. All antibodies used are listed in Supplementary Table 2.
Immunofluorescence analyses in human adipocytes
Cells were seeded on milli EZSlide-4 well slides and fixed in 4% PFA for 15 min at room temperature. Cells were washed twice with PBS, permeabilized using PBS supplemented with 0.2% Triton-X100 for 10 min and blocked for 30 min in PBS containing 2% bovine serum albumin (Sigma-Aldrich). The cells were subsequently incubated overnight with primary antibodies targeting UCP1 (1:100) or COX4 (1:100). The antibodies were diluted in PBS supplemented with 5% normal goat serum. Cells were washed three times with PBS containing 0.05% Tween-20 and incubated with goat anti-Rabbit Rhodamine Red-X secondary antibody (1:500) for 60 min. Subsequently PBS supplemented with Hoechst 34580 (1:500) and BODIPY 493/503 (1:2,500, ThermoFisher) was applied for 15 min to counterstain nuclei and lipid droplets, respectively. Finally, cells were washed in PBS and then mounted in fluorescence mounting medium (Fluoromount Aqueous Mounting Medium, refractive index 1.4). Images were obtained using an Axio Observer.Z1 inverted fluorescence microscope (Zeiss) and the AxioVision software.
Human adipocyte adiponectin secretion
For analyses of conditioned media from human adipocytes, samples were collected at day 14 of differentiation. Secretion of adiponectin was determined by ELISA (R&D Systems) according to the manufacturer’s instructions.
Redox balance measurements
Cultured siC or siGLS adipocytes were collected at day 14 of differentiation. For analyses of NAD+ and NADH, the NAD+/NADH Assay Kit (cat. no. ab65348) was used according to the manufacturer’s instructions. Data were normalized by protein concentrations per sample.
Seahorse assays and CyQUANT analyses
Real-time measurements of oxygen consumption and ECARs were performed using a Seahorse XF96 Extracellular Flux Analyzer (Agilent Technologies) as previously described54. In brief, adipocytes were incubated in Seahorse DMEM medium (pH 7.4) supplemented with 1 mmol l−1 pyruvate, 2 mmol l−1 glutamine and 10 mmol l−1 glucose. The assays were performed by sequential addition of 1.5 μmol l−1 oligomycin (inhibitor of ATP synthesis), 1.5 μmol l−1 carbonyl cyanide-p-trifluoromethoxyphenylhydrazone and 0.5 μmol l−1 rotenone/antimycin A (inhibitors of complex I and complex III of the respiratory chain, respectively). To assess responses to glucose addition during glycostress tests, cells were incubated in medium with 2 mmol l−1 glutamine but without glucose and pyruvate. The assays were performed by sequential addition of 10 mmol l−1 glucose, 1 μmol l−1 oligomycin and 50 mmol l−1 2-deoxyglucose. Mitochondrial fuel stress tests were performed using the XF96 fuel kit after acute injection of UK5099 (2 μmol l−1 per well) or Etomoxir (4 μmol l−1 per well). Seahorse data were normalized using the CyQUANT Kit (ThermoFisher) according to the manufacturer’s instructions. Immediately after Seahorse analysis, the cells were incubated with the CyQUANT reagent and fluorescence was measured. For estimation of basal and maximal respiration, the mean non-mitochondrial respiration was subtracted from the mean values of basal and maximal respiration. For glycostress tests, the mean non-glycolytic acidification was subtracted from all data points.
RNA sequencing and pathway analyses
Total RNA from siC and siGLS cells was isolated as described above and used for library preparation. The yield and quality of the amplified libraries were analysed using Qubit (Thermo Fisher) and Tapestation (Agilent) and subsequently normalized and combined. These pools were sequenced on the Illumina Nextseq 2000 P2 100 cycle sequencing run, generating 59 base paired end reads with dual index. Basecalling and demultiplexing was performed using CASAVA software with default settings generating Fastq files aligned to GRCh38 for further downstream mapping and analysis. Raw counts were normalized and analysed using DESeq2 in R studio (v.4.1.1). Pathway analyses were performed using GSEA (gene set enrichment analysis) and ProFat18.
GLS activity assay
In human and mouse adipocytes, GLS activity was determined using the Glutaminase Assay Kit (Abcam). In brief, samples were homogenized on ice in 1 ml of assay buffer and subjected to 8,000g centrifugation for 10 min at 4 °C. Protein concentrations were determined by Pierce BCA assay (Thermo Scientific). Samples containing equal amounts of protein were incubated with kit reagents at 37 °C, and fluorescence was detected over time using a Varioskan Lux. Control samples exposed to the kit reagent mixture lacking GLS substrate were used to measure endogenous basal glutamate levels. GLS activity was calculated on the basis of the increase in glutamate over time. Data were normalized by protein concentration for each sample.
Targeted glutamine and glutamate measurements
Glutamine and glutamate levels in adipocytes were determined using the Glutamine/Glutamate-Glo Assay (Promega). One half of the volume of the lysed cells was used to measure glutamate and the other half was used to measure glutamine. Protein concentrations were determined by Pierce BCA assay (Thermo Scientific) and samples were normalized to total protein content.
Chromatin immunoprecipitation
Chromatin immnuprecipiation against c-Jun was performed using Magna ChIP HiSense Chromatin Immunoprecipitation Kit (cat. no. 17-10460, Sigma), following the manufacturer’s instructions. The antibody targeting c-Jun (cat. no. 9165, Cell Signalling), was added at the concentration of 1:50, as suggested by the manufacturer. A negative control was generated using IgG Ab (cat. no. 7074S, Cell Signalling). Pull down enrichment quantification was performed using normal qPCR with reverse transcription as above described with the primer set in Supplementary Table 2. Relative enrichment was calculated using input enrichment ((Ct Input − log2100) → 100 × 2(Adjusted input−Ct IP)) and fold change against IgG signal (Ct IP/Ct IgG).
Adipocyte U-13C glucose labelling
Cells were washed twice with PBS and thereafter incubated with DMEM without glucose (cat. no. 11966025, Thermo Fisher) supplemented with an addition of 5.5 mM glucose (cat. no. G8270, Sigma) or U-13C6 d-Glucose (cat. no. 389374,Sigma) for 4 h. Cells were washed with cold PBS twice before being lysed in 90% ice-cold methanol (cat. no. 34860-1L-R, Sigma). Before cell collection, 1 ml of the media was extracted and centrifuged at 21,000g for 10 min at 4 °C then frozen. One millilitre of ice-cold 90% methanol was added to facilitate subsequent metabolic tracing analysis.
Lactate measurements
Lactate was measured in cell media using the Lactate-Glo Assay (Promega). Data were normalized by protein concentration per well.
Targeted metabolite analyses
Metabolites were profiled at the Swedish Metabolomics Centre. To trace U-13C-glutamine conversion, human adipocytes were treated with 10 μmol l−1 of BPTES or vehicle (DMSO) and 2.5 mmol l−1 U-13C-glutamine (Sigma-Aldrich) for 1 h. The cell, tissue and media samples were extracted with 1 ml of 90% methanol (diluted in water). Two aliquots of the extracts were analysed by (1) gas-chromatography-mass spectrometry (GC–MS) and (2) liquid chromatography–mass spectrometry (LC–MS). For GC–MS measurements, derivatization and GC–MS analysis were performed as described previously66,67. The extracted mass spectra were annotated (putatively/tentatively identified) by library comparisons of their retention index and mass spectra68. Mass spectra and retention index comparison was performed using NIST MS v.2.2 software, and annotation of mass spectra was based on reverse and forward searches. Both the Swedish Metabolomics Centre’s in-house standards libraries and public libraries as NIST ( MoNA ( and MS-DIAL ( was used. Peak detection and peak area calculations of both labelled and unlabelled fragments was done as described in ref. 69
The LC–MS analysis of amino acids was conducted using ultrahigh-performance liquid chromatography wtih electrospray ionizaton and quadrupole time of flight after the amino acids were derivatized with AccQ-Tag (Waters). The derivatized amino acids were identified on the basis of their retention time and exact mass. The acquired LC–MS data were converted to XML and eventually to NetCDF format, then the peaks integration and labelling calculation were conducted as described41.
Statistical analyses
Unless otherwise stated, results are reported as mean ± standard error of the mean (s.e.m.) with individual data points shown for experiments with fewer than ten replicates per sample group. The number of independent experiments and relevant statistical methods (all two-sided) for each panel are detailed in the corresponding figure legends. Statistical analyses were performed using Prism (v.9.2.0, GraphPad Software), JMP (v.15.1, SAS) or R v.4.1.1./4.2.1.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
link