Advertisement

Integrated Liver and Plasma Proteomics in Obese Mice Reveals Complex Metabolic Regulation

Open AccessPublished:January 27, 2022DOI:https://doi.org/10.1016/j.mcpro.2022.100207

      Highlights

      • Proteomics reveals liver-derived proteins systemically dysregulated in obesity.
      • Obesity increases hepatic lipid metabolism via peroxisomal biogenesis.
      • Obesity dysregulates secretory machinery and secreted proteins within the liver.
      • Metabolic and immune proteins dysregulated in plasma of obese mice.
      • Comparative proteomics of high-fat diet and monogenic (ob/ob) models of obesity.

      Abstract

      Obesity leads to the development of nonalcoholic fatty liver disease (NAFLD) and associated alterations to the plasma proteome. To elucidate the underlying changes associated with obesity, we performed liquid chromatography–tandem mass spectrometry in the liver and plasma of obese leptin-deficient ob/ob mice and integrated these data with publicly available transcriptomic and proteomic datasets of obesity and metabolic diseases in preclinical and clinical cohorts. We quantified 7173 and 555 proteins in the liver and plasma proteomes, respectively. The abundance of proteins related to fatty acid metabolism were increased, alongside peroxisomal proliferation in ob/ob liver. Putatively secreted proteins and the secretory machinery were also dysregulated in the liver, which was mirrored by a substantial alteration of the plasma proteome. Greater than 50% of the plasma proteins were differentially regulated, including NAFLD biomarkers, lipoproteins, the 20S proteasome, and the complement and coagulation cascades of the immune system. Integration of the liver and plasma proteomes identified proteins that were concomitantly regulated in the liver and plasma in obesity, suggesting that the systemic abundance of these plasma proteins is regulated by secretion from the liver. Many of these proteins are systemically regulated during type 2 diabetes and/or NAFLD in humans, indicating the clinical importance of liver–plasma cross talk and the relevance of our investigations in ob/ob mice. Together, these analyses yield a comprehensive insight into obesity and provide an extensive resource for obesity research in a prevailing model organism.

      Graphical Abstract

      Keywords

      Abbreviations:

      DIA (data-independent acquisition), FDR (false discovery rate), GO (gene ontology), GOBP (gene ontology biological processes), GOCC (gene ontology cellular compartment), HFD (high-fat diet), KEGG (Kyoto encyclopedia of genes and genomes), LC-MS/MS (liquid chromatography–tandem mass spectrometry), MED-FASP (multienzyme digestion with filter-aided sample preparation), NAFLD (nonalcoholic fatty liver disease), PTS (peroxisomal targeting sequences), TPA (total protein abundance)
      Obesity can lead to a multitude of chronic liver conditions, collectively known as nonalcoholic fatty liver disease (NAFLD). NAFLD encompasses increased hepatic triglyceride storage (steatosis) with or without liver inflammation (steatohepatitis) (
      • Kawano Y.
      • Cohen D.E.
      Mechanisms of hepatic triglyceride accumulation in non-alcoholic fatty liver disease.
      ) and the possible progression to cirrhosis and/or hepatocellular carcinoma (
      • Cohen J.C.
      • Horton J.D.
      • Hobbs H.H.
      Human fatty liver disease: Old questions and new insights.
      ). Furthermore, dysregulated hepatic triglyceride metabolism contributes to insulin resistance and the development of type 2 diabetes alongside obesity (
      • Gruben N.
      • Shiri-Sverdlov R.
      • Koonen D.P.Y.
      • Hofker M.H.
      Nonalcoholic fatty liver disease: A main driver of insulin resistance or a dangerous liaison?.
      ,
      • Knobler H.
      • Schattner A.
      • Zhornicki T.
      • Malnick S.D.
      • Keter D.
      • Sokolovskaya N.
      • Lurie Y.
      • Bass D.D.
      Fatty liver–an additional and treatable feature of the insulin resistance syndrome.
      ). In turn, insulin resistance may also contribute to the pathogenesis and development of fatty liver by leading to increased adipose tissue lipolysis and de novo lipogenesis in the liver (
      • Utzschneider K.M.
      • Kahn S.E.
      The role of insulin resistance in nonalcoholic fatty liver disease.
      ). Obesity is considered an epidemic (
      • Organization W.H.
      Global Status Report on Noncommunicable Diseases 2014.
      ) with 13% of the world’s adult population classified as obese (Global Health Observatory, 2017). At the same time, NAFLD has a 25% prevalence in Western and Asian populations (
      • Bellentani S.
      • Scaglioni F.
      • Marino M.
      • Bedogni G.
      Epidemiology of non-alcoholic fatty liver disease.
      • Fan J.G.
      • Kim S.U.
      • Wong V.W.
      New trends on obesity and NAFLD in Asia.
      ), raising to 80% to 90% in obese individuals (
      • Bellentani S.
      • Scaglioni F.
      • Marino M.
      • Bedogni G.
      Epidemiology of non-alcoholic fatty liver disease.
      ). Hence, obesity and NAFLD are substantial global health challenges and understanding the molecular drivers and implications of obesity-related liver dysfunction is of increasing importance.
      Obesity is characterized by chronic systemic inflammation and impaired immune function (
      • Hotamisligil G.S.
      Inflammation and metabolic disorders.
      ,
      • Shoelson S.E.
      • Lee J.
      • Goldfine A.B.
      Inflammation and insulin resistance.
      ). Although white adipose tissue is purported to be a major source of obesity-induced inflammation (
      • Hotamisligil G.S.
      • Shargill N.S.
      • Spiegelman B.M.
      Adipose expression of tumor necrosis factor-alpha: Direct role in obesity-linked insulin resistance.
      ), the liver secretes the majority of highly abundant systemic proteins, including those involved in immune function (
      • Shoelson S.E.
      • Lee J.
      • Goldfine A.B.
      Inflammation and insulin resistance.
      ). Thus, the development of liver dysfunction alongside obesity exerts considerable influence upon the plasma proteome (
      • Niu L.
      • Geyer P.E.
      • Wewer Albrechtsen N.J.
      • Gluud L.L.
      • Santos A.
      • Doll S.
      • Treit P.V.
      • Holst J.J.
      • Knop F.K.
      • Vilsboll T.
      • Junker A.
      • Sachs S.
      • Stemmer K.
      • Muller T.D.
      • Tschop M.H.
      • et al.
      Plasma proteome profiling discovers novel proteins associated with non-alcoholic fatty liver disease.
      ) and, likely, the immune system. Obesity is also associated with the development of cardiovascular disease, Alzheimer’s disease, and cancer (
      • Anstey K.J.
      • Cherbuin N.
      • Budge M.
      • Young J.
      Body mass index in midlife and late-life as a risk factor for dementia: A meta-analysis of prospective studies.
      ,
      • Calle E.E.
      • Thun M.J.
      • Petrelli J.M.
      • Rodriguez C.
      • Heath C.W.
      Body-mass index and mortality in a prospective cohort of U.S. adults.
      ,
      • Calle E.E.
      • Rodriguez C.
      • Walker-Thurmond K.
      • Thun M.J.
      Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults.
      ). Indeed, liver dysfunction is emerging as a contributing factor to several diseases whereby the primary pathogenesis occurs in nonhepatic tissues (
      • Nho K.
      • Kueider-Paisley A.
      • Ahmad S.
      • MahmoudianDehkordi S.
      • Arnold M.
      • Risacher S.L.
      • Louie G.
      • Blach C.
      • Baillie R.
      • Han X.
      • Kastenmüller G.
      • Trojanowski J.Q.
      • Shaw L.M.
      • Weiner M.W.
      • Doraiswamy P.M.
      • et al.
      Association of altered liver enzymes with Alzheimer disease diagnosis, cognition, neuroimaging measures, and cerebrospinal fluid biomarkers.
      ,
      • Bassendine M.F.
      • Taylor-Robinson S.D.
      • Fertleman M.
      • Khan M.
      • Neely D.
      Is Alzheimer's disease a liver disease of the brain?.
      ,
      • Sorensen H.T.
      • Friis S.
      • Olsen J.H.
      • Thulstrup A.M.
      • Mellemkjaer L.
      • Linet M.
      • Trichopoulos D.
      • Vilstrup H.
      • Olsen J.
      Risk of liver and other types of cancer in patients with cirrhosis: A nationwide cohort study in Denmark.
      ,
      • Targher G.
      • Bertolini L.
      • Rodella S.
      • Tessari R.
      • Zenari L.
      • Lippi G.
      • Arcaro G.
      Nonalcoholic fatty liver disease is independently associated with an increased incidence of cardiovascular events in type 2 diabetic patients.
      ).
      The ob/ob mouse is a widespread model organism for obesity research (
      • Barrett P.
      • Mercer J.G.
      • Morgan P.J.
      Preclinical models for obesity research.
      ), which harbors a mutation in the leptin gene (
      • Zhang Y.
      • Proenca R.
      • Maffei M.
      • Barone M.
      • Leopold L.
      • Friedman J.M.
      Positional cloning of the mouse obese gene and its human homologue.
      ,
      • Mayer J.
      • Bates M.W.
      • Dickie M.M.
      Hereditary diabetes in genetically obese mice.
      ,
      • Mayer J.
      • Dickie M.M.
      • Bates M.W.
      • Vitale J.J.
      Free selection of nutrients by hereditarily obese mice.
      ). Leptin is a circulating adipose tissue-derived hormone that, among other things, signals to the hypothalamus, ultimately conveying information regarding energy storage and results in the modulation of feeding behaviors (
      • Jéquier E.
      Leptin signaling, adiposity, and energy balance.
      ). Within 4 weeks of age, ob/ob mice develop obesity, insulin resistance, and hepatic steatosis (
      • Dubuc P.U.
      The development of obesity, hyperinsulinemia, and hyperglycemia in ob/ob mice.
      ,
      • Trak-Smayra V.
      • Paradis V.
      • Massart J.
      • Nasser S.
      • Jebara V.
      • Fromenty B.
      Pathology of the liver in obese and diabetic ob/ob and db/db mice fed a standard or high-calorie diet.
      ). Thus, ob/ob mice are a conventional model organism for studying obesity, type 2 diabetes, and liver dysfunction.
      Mass spectrometry–based proteomics provides a powerful tool to study (mal)adaptations across multiple tissues and biological fluids in pathophysiological conditions, such as obesity. Nonetheless, proteomics remains challenging in many tissues and fluids, such as liver and plasma, owing to their large dynamic range of protein abundance, in which highly abundant proteins hinder the detection of proteins with a low abundance. Here, we quantified the abundances of 7173 and 555 proteins within the liver and plasma of ob/ob mice and C57BL/6J lean control mice, providing biological insight into obesity and the hepatic alterations associated with NAFLD development, as well as a resource for the obesity research community. The ob/ob mice displayed liver and plasma proteomes that indicate impaired metabolism and immune function. Furthermore, we integrated the liver and plasma proteomes and identified circulating proteins putatively influenced by altered hepatic secretion during obesity and mapped their regulation in human metabolic diseases (e.g., type 2 diabetes and NAFLD) and interventions (e.g., diet and gastric bypass).

      Experimental Procedures

      Animal Experiments

      All animal experiments were approved by the Regional Animal Ethical Committee (Stockholm, Sweden). Four-month-old male ob/ob mice (
      • Zhang Y.
      • Proenca R.
      • Maffei M.
      • Barone M.
      • Leopold L.
      • Friedman J.M.
      Positional cloning of the mouse obese gene and its human homologue.
      ) and wild-type (WT) lean controls on a C57BL/6J background (liver samples: Charles River, Italy; plasma samples: Janvier Labs, France) had free access to water and standard rodent chow and were maintained in a temperature- and light-controlled (22–24 °C, 12-h light/dark cycle) environment. For the analysis of the liver proteome, mice were fasted for 4 h prior to anesthetization with Avertin (2,2,2-tribromoethanol 99% and tertiary amyl alcohol [1:1 w/v], 500 mg⋅kg−1 body weight). Livers were removed, frozen in liquid nitrogen, and stored at −80 °C until processing. For study of the plasma proteome, blood samples were collected into EDTA-coated tubes via tail blood sampling in the fed state. Blood samples were subjected to centrifugation at 12,000g for 5 min at 4 °C, and the plasma supernatant was collected and stored at −80 °C until processing.

      Mass Spectrometry Sample Preparation and Measurement

      Liver

      Liver samples from ob/ob mice were prepared for mass spectrometry as described for skeletal muscle (
      • Schonke M.
      • Bjornholm M.
      • Chibalin A.V.
      • Zierath J.R.
      • Deshmukh A.S.
      Proteomics analysis of skeletal muscle from leptin-deficient ob/ob mice reveals adaptive remodeling of metabolic characteristics and fiber type composition.
      ). Briefly, approximately 50 mg of liver (n = 4 per group) was homogenized in 750 μl of SDS buffer (0.1 M Tris-HCl, pH 7.5, 0.1 M DTT and 4% SDS) with an Ultra-Turrax homogenizer (IKA) and boiled at 95 °C for 5 min. Lysates were sonicated using a Branson tip-sonicator and subjected to centrifugation at 16,000g for 10 min. The supernatant was collected and 200 μg proteins was processed according to the multienzyme digestion with filter-aided sample preparation (MED-FASP) protocol using LysC (1:100 enzyme to protein ratio) and trypsin (1:100 enzyme to protein ratio) (
      • Wisniewski J.R.
      • Mann M.
      Consecutive proteolytic digestion in an enzyme reactor increases depth of proteomic and phosphoproteomic analysis.
      ). Peptides (10 μg) originating from trypsin and LysC digests were purified on 2 × C18 StageTips (
      • Rappsilber J.
      • Ishihama Y.
      • Mann M.
      Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics.
      ). Peptides were separated on a 50-cm C18 column (inner diameter 75 μm, 1.8 μm beads, Dr. Maisch GmbH) and measured via liquid chromatography–tandem mass spectrometry (LC-MS/MS) using an Easy nano-flow HPLC coupled via a nanoelectrospray ion source to a Q Exactive mass spectrometer (Thermo Fisher Scientific) using data-dependent acquisition. Peptides obtained from the LysC and Trypsin fractions were loaded onto the column with buffer A (0.5% formic acid) and eluted with a 270-min linear gradient increasing from 2% to 60% buffer B (80% acetonitrile, 0.5% formic acid). After the gradient, the column was washed with 95% buffer B and re-equilibrated with buffer A. The mass spectra were acquired in positive ion mode with automatic switching between mass spectrometry (MS) and MS/MS using a top 5 method. MS spectra were acquired in the Orbitrap analyzer with a mass range of 300 to 1750 m/z at a resolution of 70,000 with a target of 3 x 106 ions and a maximum injection time of 20 ms. Higher-energy C-trap dissociation peptide fragments acquired at 25 normalized collision energy were analyzed at a resolution of 17,500 in the Orbitrap analyzer with a target of 1 × 105 ions and a maximum injection time of 120 ms. Peptides were measured in singleton.

      Plasma

      Plasma samples from ob/ob mice were prepared for mass spectrometry as described (
      • Geyer P.E.
      • Kulak N.A.
      • Pichler G.
      • Holdt L.M.
      • Teupser D.
      • Mann M.
      Plasma proteome profiling to assess human health and disease.
      ) with minor modifications. Briefly, 1 μl of plasma was mixed with 24 μl of PreOmics lysis buffer (P.O. 00001, PreOmics GmbH) and boiled at 95 °C for 10 min. Proteins were digested with LysC (1:100 enzyme to protein ratio) for 1 h at 37 °C prior to the addition of trypsin (1:100 enzyme to protein ratio) for overnight digestion at 37 °C. About 10 to 15 μg of peptides were purified on 3 × styrenedivinylbenzene–reverse phase sulfonate StageTips. Peptides were separated on a 15-cm C18 column (inner diameter 75 μm, 1.8 μm beads, Dr. Maisch GmbH) and measured via LC-MS/MS using an Easy nano-flow HPLC coupled via a nanoelectrospray ion source to a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific) using data-independent acquisition (DIA). Peptides were loaded onto the column with buffer A (0.5% formic acid) and eluted with a 45-min linear gradient increasing from 3% to 35% buffer B (80% acetonitrile, 0.5% formic acid). After the gradient, the column was washed with 98% buffer B. The DIA method used for the plasma proteome measurements contained one full scan (350–1650 m/z, resolution = 120,000) at a target of 3 × 106 ions, followed by 22 windows with a resolution of 30,000 where precursor ions were fragmented with higher-energy collisional dissociation (stepped collision energy 25%, 27.5%, 30%) and analyzed with an automatic gain control target of 3 × 106 ions and maximum injection time at 54 ms in profile mode using positive polarity. Different proteomic methods were employed to measure the liver and plasma proteomes, reflecting the different characteristics of these biological matrixes (e.g., tissue versus biofluid, sample complexity, and expected proteome depth) and the technological and methodological advances in proteomic analyses that occurred over the course of this investigation.

      Mass Spectrometry Data Analyses

      Liver Proteome

      Raw MS files were analyzed using MaxQuant version 1.6.7.0 (
      • Tyanova S.
      • Temu T.
      • Cox J.
      The MaxQuant computational platform for mass spectrometry-based shotgun proteomics.
      ) (http://www.maxquant.org). MS/MS spectra were searched by the Andromeda search engine against the mouse FASTA database (version 2017_11, containing 69,720 entries from the SwissProt and TrEMBL databases) supplemented with 262 frequently observed contaminants and forward and reverse sequences. In the main Andromeda search precursor, mass and fragment mass were matched with an initial mass tolerance of 6 and 20 ppm, respectively. Minimum score for modified and unmodified peptides were set at 0 and 40, respectively. The search included variable modifications of methionine oxidation and N-terminal acetylation and fixed modification of carbamidomethyl cysteine. The specific digestion method was set as “LysC” for the first Lys-C-derived fraction and “Trypsin/P” for the second trypsin-derived fraction. Minimal peptide length was set to seven amino acids, and a maximum of two miscleavages were allowed. The false discovery rate (FDR) was set to 0.01 for peptide and protein identifications. MS runs were analyzed with the “match between runs” option, with a retention time window of 30 s. When all identified peptides were shared between two proteins, results were combined and reported as one protein group. Matches to the reverse database were excluded.

      Plasma Proteome

      Plasma proteome files acquired in DIA mode were processed using Biognosys Spectronaut software version 12 (
      • Bruderer R.
      • Bernhardt O.M.
      • Gandhi T.
      • Miladinović S.M.
      • Cheng L.Y.
      • Messner S.
      • Ehrenberger T.
      • Zanotelli V.
      • Butscheid Y.
      • Escher C.
      • Vitek O.
      • Rinner O.
      • Reiter L.
      Extending the limits of quantitative proteome profiling with data-independent acquisition and application to acetaminophen-treated three-dimensional liver microtissues.
      ). DIA raw data were analyzed using an in-house generated spectral library of mouse plasma that contains 11,566 peptides and 1461 protein groups (supplemental Table S1; mouse UniProt FASTA database, version 2018_06, containing 92,096 entries) (
      • Dall M.
      • Hassing A.S.
      • Niu L.
      • Nielsen T.S.
      • Ingerslev L.R.
      • Sulek K.
      • Trammell S.A.J.
      • Gillum M.P.
      • Barrès R.
      • Larsen S.
      • Poulsen S.S.
      • Mann M.
      • Ørskov C.
      • Treebak J.T.
      Hepatocyte-specific perturbation of NAD+ biosynthetic pathways in mice induces reversible nonalcoholic steatohepatitis–like phenotypes.
      ). The experimental DIA runs were analyzed in Spectronaut using default settings. The mass tolerance for the precursor ion envelope in MS1 was determined from the data using a calibration search, with a starting width of 20 ppm. For peptide identification, the six most intense fragment ions are selected (minimally three when less than six fragment ions are apparent). The mass tolerance for the main library search is determined from the data using a calibration search, with a starting width of 20 ppm (for MS1 and MS2 separately). FDR was calculated run-wise at the precursor level and experiment-wide on the protein group level using the Storey method (
      • Storey J.D.
      • Tibshirani R.
      Statistical significance for genomewide studies.
      ).

      Bioinformatics Analyses

      Bioinformatics analyses were performed on MaxQuant raw intensities for liver proteomes and on Spectronaut intensities for plasma proteomes. Bioinformatics analysis was performed in the Perseus software (
      • Tyanova S.
      • Temu T.
      • Sinitcyn P.
      • Carlson A.
      • Hein M.Y.
      • Geiger T.
      • Mann M.
      • Cox J.
      The Perseus computational platform for comprehensive analysis of (prote)omics data.
      ) unless otherwise stated. For the liver proteome, relative protein quantification was performed via the total protein abundance (TPA) methodology, whereby the raw protein intensity values were divided by the total protein intensity in a sample (
      • Wisniewski J.R.
      Label-free and standard-free absolute quantitative proteomics using the “total protein” and “proteomic ruler” approaches.
      ,
      • Wiśniewski J.R.
      • Hein M.Y.
      • Cox J.
      • Mann M.
      A “proteomic ruler” for protein copy number and concentration estimation without spike-in standards.
      ); all analyses in the liver proteome were performed on log2-transformed TPA values. TPA values were multiplied by 108 prior to log2-transformation to ensure positive log2 values were displayed in figures. Categorical annotation was supplied in the form of gene ontology (GO) biological process (GOBP), gene ontology cellular component (GOCC), Kyoto encyclopedia of genes and genomes (KEGG) pathway and UniProt keyword terms. All annotations were extracted from the UniProt database. Protein data were filtered to remove proteins identified with only a single modified peptide “only identified by site,” reverse sequences, and potential contaminants and to have >60% coverage in at least one group (Liver proteome: three of four, plasma proteome: four of seven). Data were imputed to fill missing abundance values by drawing random numbers from a Gaussian distribution with a standard deviation of 30% and a downshift of 1.8 standard deviations from the mean relative to that of the proteome abundance distribution. Two sample t tests were performed on WT and ob/ob groups with a permutation-based FDR and an S0 of 0.1 applied (
      • Tusher V.G.
      • Tibshirani R.
      • Chu G.
      Significance analysis of microarrays applied to the ionizing radiation response.
      ). Summed protein abundances were calculated by summing the total protein abundance values for each protein annotated to a particular GO term . Overrepresentation enrichment analyses were performed using the ClueGo App on Cytoscape (
      • Bindea G.
      • Mlecnik B.
      • Hackl H.
      • Charoentong P.
      • Tosolini M.
      • Kirilovsky A.
      • Fridman W.-H.
      • Pagès F.
      • Trajanoski Z.
      • Galon J.
      ClueGO: A cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks.
      ) using a right-sided hypergeometric test (Benjamini–Hochberg FDR <0.02 ) (
      • Cox J.
      • Mann M.
      1D and 2D annotation enrichment: A statistical method integrating quantitative proteomics with complementary high-throughput data.
      ) in significantly regulated proteins compared with the whole proteome. Additional enrichment analyses were performed using a one-dimensional enrichment analysis (
      • Cox J.
      • Mann M.
      1D and 2D annotation enrichment: A statistical method integrating quantitative proteomics with complementary high-throughput data.
      ), in which protein log2-fold change values were ranked and a test was performed for each annotation term to determine whether the log2-fold change for the proteins belonging to the annotation were systemically larger (positive enrichment) or smaller (negative enrichment) than the global distribution of all the log2-fold change for all the proteins. Proteomic data from the liver of ob/ob mice were compared with publicly available datasets, including the ob/ob liver transcriptome (
      • Egami R.
      • Kokaji T.
      • Hatano A.
      • Yugi K.
      • Eto M.
      • Morita K.
      • Ohno S.
      • Fujii M.
      • Hironaka K.I.
      • Uematsu S.
      • Terakawa A.
      • Bai Y.
      • Pan Y.
      • Tsuchiya T.
      • Ozaki H.
      • et al.
      Trans-omic analysis reveals obesity-associated dysregulation of inter-organ metabolic cycles between the liver and skeletal muscle.
      ) and the liver proteome of high-fat diet (HFD)-fed mice (
      • Krahmer N.
      • Najafi B.
      • Schueder F.
      • Quagliarini F.
      • Steger M.
      • Seitz S.
      • Kasper R.
      • Salinas F.
      • Cox J.
      • Uhlenhaut N.H.
      • Walther T.C.
      • Jungmann R.
      • Zeigerer A.
      • Borner G.H.H.
      • Mann M.
      Organellar proteomics and phospho-proteomics reveal subcellular reorganization in diet-induced hepatic steatosis.
      ). Enrichment of GO terms and KEGG pathways were compared in each dataset using a two-dimensional enrichment analysis (
      • Cox J.
      • Mann M.
      1D and 2D annotation enrichment: A statistical method integrating quantitative proteomics with complementary high-throughput data.
      ). Hierarchical clustering analysis was performed using Euclidean distance (
      • Egami R.
      • Kokaji T.
      • Hatano A.
      • Yugi K.
      • Eto M.
      • Morita K.
      • Ohno S.
      • Fujii M.
      • Hironaka K.I.
      • Uematsu S.
      • Terakawa A.
      • Bai Y.
      • Pan Y.
      • Tsuchiya T.
      • Ozaki H.
      • et al.
      Trans-omic analysis reveals obesity-associated dysregulation of inter-organ metabolic cycles between the liver and skeletal muscle.
      ,
      • Krahmer N.
      • Najafi B.
      • Schueder F.
      • Quagliarini F.
      • Steger M.
      • Seitz S.
      • Kasper R.
      • Salinas F.
      • Cox J.
      • Uhlenhaut N.H.
      • Walther T.C.
      • Jungmann R.
      • Zeigerer A.
      • Borner G.H.H.
      • Mann M.
      Organellar proteomics and phospho-proteomics reveal subcellular reorganization in diet-induced hepatic steatosis.
      ).
      Figure thumbnail gr1
      Fig. 1Lipid metabolic processes are increased in the liver of ob/ob mice. A, ob/ob mice and wild-type (WT) controls were sacrificed for liver (n = 4) and plasma (n = 7) proteomics. Liver samples were prepared using the multienzyme digestion with filter-aided sample preparation (MED-FASP) protocol (
      • Wisniewski J.R.
      • Mann M.
      Consecutive proteolytic digestion in an enzyme reactor increases depth of proteomic and phosphoproteomic analysis.
      ) and measured with a data-dependent acquisition (DDA) method. The resulting spectra were analyzed by MaxQuant software and were quantified using the total protein abundance (TPA) approach (
      • Wisniewski J.R.
      Label-free and standard-free absolute quantitative proteomics using the “total protein” and “proteomic ruler” approaches.
      ). Plasma samples were measured with a data-independent acquisition (DIA) method, and the resulting raw data were analyzed using Spectronaut (version 12) and quantified by protein intensities (INT). B, volcano plot of the liver proteome showing 264 significantly upregulated proteins in red and 223 significantly downregulated proteins in blue (FDR < 0.05, S0 = 0.1). C, summed protein abundance of gene ontology biological processes (GOBP) terms related to metabolism. D, ClueGO enrichment analysis of GOBP terms within significantly upregulated proteins. Right-sided hypergeometric test, Benjamini–Hochberg FDR < 0.02, functionally related terms are grouped and color coded based on overlapping proteins. The most significantly regulated terms in each group are labeled. Data are represented as mean ± standard deviation (SD) (n = 4), Student’s t test (summed protein abundance): ∗∗p < 0.01, ∗∗∗ p < 0.001. LC-MS/MS, liquid chromatography–tandem mass spectrometry.
      Figure thumbnail gr2
      Fig. 2Peroxisomal proliferation in the liver of ob/ob mice. A, one-dimensional enrichment analysis based on Kyoto encyclopedia of genes and genomes (KEGG) pathways (FDR < 0.02). B, summed protein abundance of the gene ontology cellular compartment (GOCC) term peroxisome. C, partial reproduction of the KEGG Peroxisome pathway, showing proteins involved in peroxisomal biogenesis with proteins denoted as upregulated or downregulated at an FDR < 0.05 or 0.10. D, protein abundance of PEX11 isoforms. Data are represented as mean ± SD (n = 4), permutation-based FDR-corrected Student’s t test (proteome): $FDR < 0.05, Student’s t test (summed protein abundance): ∗∗∗p < 0.001.
      Figure thumbnail gr3
      Fig. 3The secretory machinery and secreted proteins are dysregulated within the liver of ob/ob mice. A, one-dimensional enrichment analysis based on GOCC terms (FDR < 0.02). B, regulation of proteins annotated to contain a signal peptide (UniProt Keyword). The ten proteins with the largest positive and negative fold-changes, respectively, are listed. C, protein abundance of the ten proteins annotated to the endoplasmic reticulum chaperone complex (GOCC) filtered from the proteome, of which five were significantly downregulated. D, summed protein abundance of the GOCC term Golgi apparatus. E, regulation of proteins annotated to extracellular exosome (GOCC). The ten proteins with the largest positive and negative fold-changes, respectively, are listed. F, protein abundance of the 13 significantly regulated proteins annotated to extracellular matrix filtered from the proteome. Data are represented as mean ± SD (n = 4), permutation-based FDR corrected Student’s t test (proteome): $FDR < 0.05, Student’s t test (summed protein abundance): ∗∗∗p < 0.001. GOCC, gene ontology cellular compartment.
      Figure thumbnail gr4
      Fig. 4Transcriptional regulation of the liver proteome of ob/ob mice. A, Venn diagram displaying the overlap of proteins and transcripts quantified in the liver proteome and a previously published transcriptome (
      • Egami R.
      • Kokaji T.
      • Hatano A.
      • Yugi K.
      • Eto M.
      • Morita K.
      • Ohno S.
      • Fujii M.
      • Hironaka K.I.
      • Uematsu S.
      • Terakawa A.
      • Bai Y.
      • Pan Y.
      • Tsuchiya T.
      • Ozaki H.
      • et al.
      Trans-omic analysis reveals obesity-associated dysregulation of inter-organ metabolic cycles between the liver and skeletal muscle.
      ) of ob/ob mice. B, scatter plot of the log2-fold change in the transcriptome and proteome. Proteins/transcripts significantly regulated only in the transcriptome or proteome are displayed in green and purple, respectively; proteins/transcripts significantly regulated in both the transcriptome and proteome are highlighted in orange. C and D, two-dimensional enrichment analysis based on GOBP (C) and GOCC (D) terms for the proteome and transcriptome (FDR < 0.02). GOBP, gene ontology biological processes; GOCC, gene ontology cellular compartment.
      Figure thumbnail gr5
      Fig. 5Proteasomal proteins and lipid metabolic processes are increased in the plasma proteome of ob/ob mice. A, volcano plot of the plasma proteome showing 146 significantly upregulated proteins in red and 119 significantly downregulated proteins in blue (FDR < 0.05, S0 = 0.1). B and C, functional enrichment analysis by ClueGO for the significantly downregulated (B) and upregulated (C) proteins. Right-sided hypergeometric test, Benjamini–Hochberg FDR < 0.02, functionally related terms are grouped and color coded based on overlapping proteins. The most significantly regulated terms in each group are labeled in color. D, protein abundance of the ten proteins annotated to the proteasome complex (GOCC). E, hierarchical cluster of the proteins within the purple enrichment term groups (whose most significantly regulated terms were carboxylic acid metabolic process and organic hydroxyl compound metabolic process) in (C). Data are represented as mean ± SD (n = 7), permutation-based FDR-corrected Student’s t test (proteome): $FDR < 0.05. GOCC, gene ontology cellular compartment.

      Physiological Phenotyping

      Mice were single housed in individual Plexiglas cages of the Comprehensive Lab Animal Monitoring System (CLAMS; Columbus Instruments), and food intake, locomotor activity, and oxygen consumption were continuously monitored. During the CLAMS session, mice were maintained on a 12:12-h light–dark cycle and had free access to standard rodent chow and water. Glucose tolerance was assessed using an intraperitoneal glucose tolerance test after a 4-h fast, as described (
      • Gonzalez-Franquesa A.
      • Stocks B.
      • Borg M.L.
      • Kuefner M.
      • Dalbram E.
      • Nielsen T.S.
      • Agrawal A.
      • Pankratova S.
      • Chibalin A.V.
      • Karlsson H.K.R.
      • Gheibi S.
      • Björnholm M.
      • Jørgensen N.R.
      • Clemmensen C.
      • Hostrup M.
      • et al.
      Discovery of thymosin β4 as a human exerkine and growth factor.
      ).

      Experimental Design and Statistical Rationale

      Liver samples were analyzed from four mice per genotype (biological replicates). Plasma samples were analyzed from seven mice per genotype (biological replicates). Statistical power was deemed to be sufficient based on our previous observations in skeletal muscle of ob/ob mice (
      • Schonke M.
      • Bjornholm M.
      • Chibalin A.V.
      • Zierath J.R.
      • Deshmukh A.S.
      Proteomics analysis of skeletal muscle from leptin-deficient ob/ob mice reveals adaptive remodeling of metabolic characteristics and fiber type composition.
      ). Student’s t tests were used to compare means after log-transformation and visual inspection of normal distribution. For physiological data, independent Student’s t test or two-way analysis of variance, with Ryan–Holm–Bonferroni corrections where appropriate, were used to assess differences between genotypes and interactions with time.

      Results

      High Coverage and Reproducibility of Liver and Plasma Proteomes

      We performed LC-MS/MS on liver and plasma samples from leptin-deficient ob/ob mice and wild-type C57BL/6J controls (Fig. 1A). ob/ob mice were obese and glucose intolerant, displaying elevated body mass, food intake, blood glucose, and plasma insulin, as well as reduced activity and maximal oxygen uptake (supplemental Fig. S1, AF). We identified 7684 proteins from two fractions using the MED-FASP protocol (
      • Wisniewski J.R.
      • Mann M.
      Consecutive proteolytic digestion in an enzyme reactor increases depth of proteomic and phosphoproteomic analysis.
      ) (supplemental Table S2). Filtering for three valid values (out of four) in at least one genotype resulted in the final quantification of 7173 proteins (supplemental Table S3). The number of quantified proteins per sample ranged between 6427 and 6911 (supplemental Fig. S2A). Good reproducibility between biological replicates was apparent (Pearson’s r range: 0.89–0.98), which was higher than between samples from different genotypes (Pearson’s r range: 0.86–0.93) (supplemental Fig. S2B). Proteins were quantified across a wide range of TPA values, with six orders of magnitude between the highest and lowest abundant quantified proteins (supplemental Fig. S2C). The quantified proteins included several transcription factors (e.g., ARNTL (BMAL1), CLOCK, CREB1, FOXO1, HNF1A, HNF4A, and PPARA), which are challenging to quantify in tissue owing to their low protein abundance.
      Plasma samples were prepared for proteomics in a single fraction (Fig. 1A), as described (
      • Geyer P.E.
      • Kulak N.A.
      • Pichler G.
      • Holdt L.M.
      • Teupser D.
      • Mann M.
      Plasma proteome profiling to assess human health and disease.
      ). We identified 592 proteins in plasma (supplemental Table S4). Intensities were filtered for four valid values (out of seven) in at least one genotype, resulting in the final quantification of 555 proteins (supplemental Table S5). The number of quantified proteins per sample ranged between 426 and 516 (supplemental Fig. S2D). Good reproducibility between biological replicates was also apparent in the plasma proteome (Pearson’s r range: 0.87–0.95), which was substantially higher than between samples from different genotypes (Pearson’s r range: 0.78–0.85) (supplemental Fig. S2E). Similarly to the liver proteome, proteins were quantified across approximately five orders of magnitude (supplemental Fig. S2F), demonstrating the large dynamic range of the liver and plasma proteomes.

      The Liver Proteome of ob/ob Mice Displays Elevated Lipid Metabolic Processes

      Of the 7173 quantified liver proteins, 487 proteins were differentially regulated (FDR <0.05, S0 = 0.1), with 264 proteins upregulated and 223 proteins downregulated within the liver of ob/ob mice (Fig. 1B and supplemental Table S3). These included fructose-biphosphate aldolase B (ALDOB), an enzyme involved in glycolysis and gluconeogenesis, as well the apolipoproteins E (APOE) and M (APOM). Furthermore, fatty acid synthase (FASN) and the lipid droplet protein perilipin-5 (PLIN5) were among the upregulated proteins. To assess the effect of obesity on aspects of metabolism, the abundance of proteins annotated to various GOBP terms were summed and compared between groups (Fig. 1C). Increased summed protein abundance of processes involved in fat metabolism (e.g., β-oxidation, fatty acid metabolic process, glycerol metabolic process), carbohydrate metabolism (e.g., glycolysis, gluconeogenesis, and carbohydrate metabolic process), and alcohol metabolism were observed in liver from ob/ob mice (Fig. 1C). Conversely, the summed protein abundance for proteins annotated to amino acid metabolic process was decreased (Fig. 1C). Despite the increase in summed protein abundance related to substrate metabolism, neither oxidative phosphorylation (OXPHOS) nor the tricarboxylic acid cycle was affected (Fig. 1C).
      To systematically identify proteomic changes to biological processes, we performed enrichment analyses for GOBP terms in the upregulated and downregulated proteins using ClueGO (
      • Bindea G.
      • Mlecnik B.
      • Hackl H.
      • Charoentong P.
      • Tosolini M.
      • Kirilovsky A.
      • Fridman W.-H.
      • Pagès F.
      • Trajanoski Z.
      • Galon J.
      ClueGO: A cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks.
      ). No GOBP terms were significantly enriched in the downregulated proteins; however, a multitude of GOBP terms were enriched within the upregulated proteins (Fig. 1D and supplemental Table S6). Groups of enriched terms and their interactions are displayed in Figure 1D. The largest cluster of enriched terms are primarily related to fatty acid metabolism, including fatty acid metabolic process, cellular lipid metabolic process, lipid biosynthetic process, and monocarboxylic acid metabolic process. Additional biological processes enriched within the upregulated proteins included terms related to glucose-6-phosphate metabolism, double-strand break repair via break-induced replication, and extracellular matrix organization (Fig. 1D and supplemental Table S6). Thus, changes to the ob/ob liver proteome are primarily related to the metabolism of fatty acids.

      Peroxisomal Biogenesis in the Liver of ob/ob Mice

      In line with the enriched GOBP terms related to fatty acid metabolism, the KEGG terms biosynthesis of unsaturated fatty acids and peroxisome were enriched within the upregulated proteins (Fig. 2A and supplemental Table S7). In addition, summed protein abundance of the GOCC term peroxisome was increased (Fig. 2B). This is complementary to elevated fatty acid metabolic processes in the absence of changes to mitochondrial processes in both the summed protein abundance (Fig 1C and supplemental Fig. S3) and GOBP enrichment (Fig. 1D) analyses. Together, these results implicate the peroxisome as the site of elevated fatty acid metabolism within ob/ob liver; thus, we investigated the mediators of peroxisomal biogenesis in further detail (Fig. 2C and supplemental Fig. S4A). Peroxisomal biogenesis encompasses the import of peroxisomal membrane and matrix proteins and the elongation and division of existing peroxisomes (
      • Fujiki Y.
      • Okumoto K.
      • Mukai S.
      • Honsho M.
      • Tamura S.
      Peroxisome biogenesis in mammalian cells.
      ). Peroxisomal biogenesis factor (PEX) 19, which binds newly synthesized peroxisomal membrane proteins and shuttles them to the membrane protein import machinery (
      • Jones J.M.
      • Morrell J.C.
      • Gould S.J.
      PEX19 is a predominantly cytosolic chaperone and import receptor for class 1 peroxisomal membrane proteins.
      ), was increased in liver from ob/ob mice. PEX3 and PEX16, which are also required for de novo peroxisomal biogenesis (
      • Fujiki Y.
      • Okumoto K.
      • Mukai S.
      • Honsho M.
      • Tamura S.
      Peroxisome biogenesis in mammalian cells.
      ) remained unchanged (Fig. 2C). Of the proteins involved in the import of peroxisomal matrix proteins, PEX7 increased at a less stringent false discovery rate (FDR < 0.10), whereas PEX5, the other cytosolic peroxisomal matrix protein receptor (
      • Fujiki Y.
      • Okumoto K.
      • Mukai S.
      • Honsho M.
      • Tamura S.
      Peroxisome biogenesis in mammalian cells.
      ), remained unchanged. PEX14, which PEX5 and PEX7 dock to during peroxisomal matrix protein import, increased in the absence of changes to its docking complex partner PEX13. PEX1, PEX6, and PEX26, which mediate the recycling of peroxisomal matrix protein receptors (
      • Fujiki Y.
      • Okumoto K.
      • Mukai S.
      • Honsho M.
      • Tamura S.
      Peroxisome biogenesis in mammalian cells.
      ), were all increased at an FDR < 0.10 (Fig. 2C). Peroxisomal elongation and division is initiated via PEX11 isoforms, which recruit the fission proteins dynamin-1-like protein (DNM1L), mitochondrial fission 1 protein (FIS1), and mitochondrial fission factor (MFF) (
      • Koch J.
      • Brocard C.
      PEX11 proteins attract Mff and human Fis1 to coordinate peroxisomal fission.
      ). PEX11A and PEX11G were increased in ob/ob liver, although PEX11B was unaltered (Fig. 2D, isoforms quantified with unique peptides). Changes to PEX11 isoforms occurred in the absence of significant changes to the peroxisomal fission proteins DNM1L, FIS1, and MFF, although MFF displayed a nonsignificant decrease in protein abundance (q = 0.103; supplemental Fig. S4B).

      Obesity Dysregulates Secretory Machinery and Secreted Proteins Within the Liver

      A one-dimensional enrichment analysis of GOCC terms identified systematic regulation of cellular compartments involved in protein secretion (supplemental Table S8). Endoplasmic reticulum chaperone complex was enriched within the downregulated proteins (Fig. 3A), whereas extracellular vesicle exosome and extracellular matrix were enriched within the upregulated proteins (Fig. 3A). Protein secretion is mediated via different secretory pathways (i.e., classical secretion or nonclassical secretion). Thus, we investigated these secretory pathways in further detail. Classically secreted proteins contain an endoplasmic reticulum–targeting signal peptide (
      • Walter P.
      • Gilmore R.
      • Blobel G.
      Protein translocation across the endoplasmic reticulum.
      ). The majority of signal peptide–containing proteins are transported via vesicles through the endoplasmic reticulum and the Golgi apparatus and are released into the extracellular space by vesicular fusion with the plasma membrane (
      • Palade G.
      Intracellular aspects of the process of protein synthesis.
      ). Within the liver proteome, we quantified 741 proteins annotated to contain a signal peptide (UniProt Keyword: signal peptide) (Fig. 3B). Of these, 79 were differentially regulated with 38 upregulated and 41 downregulated in the liver of ob/ob mice (Fig. 3B). The signal peptide–containing proteins with the highest positive and negative fold-changes are identified in Figure 3B; these include the upregulation of Interferon gamma receptor 1 (INGR1) and Interferon alpha-inducible protein 27-like protein 2B (IFI27L2B), indicating liver damage–associated interferon signaling (
      • Horras C.J.
      • Lamb C.L.
      • Mitchell K.A.
      Regulation of hepatocyte fate by interferon-γ.
      ,
      • Volpes R.
      • van den Oord J.J.
      • De Vos R.
      • Depla E.
      • De Ley M.
      • Desmet V.J.
      Expression of interferon-γ receptor in normal and pathological human liver tissue.
      ) in ob/ob mice. In line with the enrichment of the endoplasmic reticulum chaperone complex within the downregulated proteins, five of the ten quantified proteins within the endoplasmic reticulum chaperone complex were significantly downregulated (Fig. 3C), whereas the summed protein abundance of the endoplasmic reticulum was also decreased (supplemental Fig. S3). Furthermore, the total abundance of proteins annotated to the Golgi apparatus was increased in ob/ob mice (Fig. 3D). During nonclassical secretion, proteins are secreted via several mechanisms independently of the endoplasmic reticulum and the Golgi apparatus. One such mechanism is the secretion of exosomes, a form of extracellular vesicle (
      • Hessvik N.P.
      • Llorente A.
      Current knowledge on exosome biogenesis and release.
      ). The GOCC term extracellular vesicle exosome was enriched within the upregulated proteins (Fig. 3A). Filtering for proteins annotated to extracellular vesicle exosome (GOCC) resulted in the quantification of 1757 proteins within the liver that are putatively associated with exosomes (Fig. 3E). Of these, 88 and 51 proteins were significantly upregulated and downregulated, respectively. Among the most highly upregulated was galectin-3 (LGALS3) (Fig. 3E), a profibrotic extracellular matrix protein (
      • Henderson N.C.
      • Mackinnon A.C.
      • Farnworth S.L.
      • Poirier F.
      • Russo F.P.
      • Iredale J.P.
      • Haslett C.
      • Simpson K.J.
      • Sethi T.
      Galectin-3 regulates myofibroblast activation and hepatic fibrosis.
      ).
      The extracellular matrix and related protein groups (e.g., collagen and laminin-1 complex) were also enriched within the upregulated proteins (Fig. 3A). The significantly upregulated proteins within the GOCC term extracellular matrix are displayed in Figure 3F. The majority of these are structural proteins of the extracellular matrix (e.g., collagen proteins COL1A1, COL3A1, and COL14A1; fibrillin 1 [FBN1]; fibronectin [FN1]; and laminin [LAMC1]), although a number of secreted proteins annotated to the extracellular matrix were also upregulated (e.g., APOE, galectin-1 [LGALS1], LGALS3, and periostin precursor [POSTN]). All together, we demonstrate a substantial regulation of the secretory machinery and the extracellular matrix within the liver of ob/ob mice.

      Transcriptional Regulation of the Liver Proteome of ob/ob Mice

      To investigate the degree to which the observed changes in the liver proteome are transcriptionally regulated, we took advantage of our deep liver proteome and integrated it with a published transcriptomic dataset from the liver of ob/ob mice (
      • Egami R.
      • Kokaji T.
      • Hatano A.
      • Yugi K.
      • Eto M.
      • Morita K.
      • Ohno S.
      • Fujii M.
      • Hironaka K.I.
      • Uematsu S.
      • Terakawa A.
      • Bai Y.
      • Pan Y.
      • Tsuchiya T.
      • Ozaki H.
      • et al.
      Trans-omic analysis reveals obesity-associated dysregulation of inter-organ metabolic cycles between the liver and skeletal muscle.
      ). Of the 7173 proteins we quantified in the liver, 6935 had corresponding transcripts quantified in the published transcriptome dataset (Fig. 4A). Correlation of the transcriptomic and proteomic log2-fold changes compared with wild-type controls displayed only a moderate correlation (Fig. 4B; Pearson’s r = 0.419). A total of 377 proteins/transcripts were differentially regulated in both transcriptome and the proteome. Indeed, the proteins with the largest changes in the proteome (both up- and downregulated) were also the most regulated at the transcriptional level (e.g., SERPINA1E, MUP7, LY6D, and CYP42B9). However, there were 4018 differentially regulated transcripts that did not change at the protein level, whereas 97 proteins were regulated without changes in transcription. To identify concomitantly and divergently regulated cellular compartments and biological processes between the proteome and transcriptome, we performed a two-dimensional enrichment analysis and displayed how enriched terms were regulated in each dataset. The metabolic processes carbohydrate catabolic process, lipid biosynthetic process, and alcohol catabolic processes (GOBP) were positively enriched in both the transcriptome and the proteome, indicating substantial transcriptional control of these processes in obesity (Fig. 4C and supplemental Table S9). Similarly, GOCC terms relating to the extracellular matrix and fibrillar collagen displayed concordant regulation in the proteome and transcriptome, as was also the case with the peroxisome (Fig. 4D and supplemental Table S10). Conversely, terms related to the mitochondria, including mitochondrial membrane part (Fig. 4D), were upregulated at the transcriptional level but not in the proteome, whereas the ribosome is decreased in the proteome and moderately upregulated in the transcriptome (Figs. 2D and 4D). Thus, although some of the major metabolic pathways and cellular compartments that are regulated in the liver of ob/ob mice are driven by transcriptional regulation, posttranscriptional mechanisms (e.g., translation and/or degradation) likely also regulate the liver proteome during obesity.
      To validate our results and identify common obesity phenotypes between models, we compared the liver proteome of ob/ob mice (fed a chow diet) to a liver proteome derived from HFD-fed wild-type mice of the same age and genetic background (
      • Krahmer N.
      • Najafi B.
      • Schueder F.
      • Quagliarini F.
      • Steger M.
      • Seitz S.
      • Kasper R.
      • Salinas F.
      • Cox J.
      • Uhlenhaut N.H.
      • Walther T.C.
      • Jungmann R.
      • Zeigerer A.
      • Borner G.H.H.
      • Mann M.
      Organellar proteomics and phospho-proteomics reveal subcellular reorganization in diet-induced hepatic steatosis.
      ). We observed only a very small correlation (Pearson’s r = 0.158) between the log2-fold changes observed in the liver of ob/ob and HFD-fed mice (supplemental Fig. S5A), indicating that changes at the individual protein level are markedly different between ob/ob mice and HFD-fed mice. However, when comparing systematic changes via two-dimensional enrichment analyses, we can identify cellular compartments, processes, and pathways that are similarly regulated in ob/ob and HFD-fed mice (supplemental Fig. S5B and supplemental Table S11). For example, terms related to fatty acid metabolism, glycolysis/gluconeogenesis, and peroxisomal proliferation were upregulated in both models of obesity. Conversely, terms related to the extracellular matrix (e.g., extracellular matrix and fibrillar collagen) were divergently regulated, displaying substantial upregulation in the liver of ob/ob mice and downregulation in HFD-fed mice (supplemental Fig. S4B). These concordant and divergent phenotypes are important to consider when choosing a model organism in obesity and/or NAFLD research.

      ob/ob Mice Exhibit a Dysregulated Plasma Proteome Enriched for Proteasomal and Metabolic Processes

      Of the 555 quantified plasma proteins, 265 proteins were significantly regulated (FDR <0.05, S0 = 0.1), with 146 proteins upregulated and 119 proteins downregulated in the plasma of ob/ob mice (Fig. 5A and supplemental Table S5). These included the upregulation of the clinical NAFLD markers ALDOB, APOE, dipeptidyl peptidase 4 (DPP4), and polymeric immunoglobulin receptor (PIGR) and the downregulation of APOM (
      • Niu L.
      • Geyer P.E.
      • Wewer Albrechtsen N.J.
      • Gluud L.L.
      • Santos A.
      • Doll S.
      • Treit P.V.
      • Holst J.J.
      • Knop F.K.
      • Vilsboll T.
      • Junker A.
      • Sachs S.
      • Stemmer K.
      • Muller T.D.
      • Tschop M.H.
      • et al.
      Plasma proteome profiling discovers novel proteins associated with non-alcoholic fatty liver disease.
      ). We performed enrichment analyses based on GOBP terms enriched within the upregulated and downregulated proteins using ClueGO (
      • Bindea G.
      • Mlecnik B.
      • Hackl H.
      • Charoentong P.
      • Tosolini M.
      • Kirilovsky A.
      • Fridman W.-H.
      • Pagès F.
      • Trajanoski Z.
      • Galon J.
      ClueGO: A cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks.
      ). GOBP terms enriched within the downregulated proteins were related to protein secretion, including regulation of hormone secretion and regulation of glucose metabolic process. These enriched terms were driven by the regulation of proteins such as adiponectin (ADIPOQ), annexin-a1 (ANXA1), epidermal growth factor receptor (EGFR), glycosylphosphatidylinositol-specific phospholipase D (GPLD1), and major urinary proteins (MUP1, MUP2, and MUP3). Terms related to negative regulation of macromolecule metabolic process and negative regulation of cell adhesion were also enriched in the downregulated proteins (Fig. 5B and supplemental Table S12). Enriched within the upregulated proteins were terms related to cellular catabolic process (Fig. 5C and supplemental Table S13), which was particularly driven by the coordinated upregulation of the 20S proteasome complex proteins in plasma (Fig. 5D). Carboxylic acid metabolic process, lipid homeostasis, cholesterol homeostasis, and plasma lipoprotein organization were all enriched within the enrichment term cluster depicted in the two shades of purple (Fig. 5C). Proteins driving the enrichment of these two clusters are shown in Figure 5E and are predominantly apolipoproteins, as well as proteins involved in glycolysis and the tricarboxylic acid cycle. Terms related to cellular aldehyde metabolic process, cofactor metabolic process, and oxidoreductase activity were also enriched in the upregulated proteins (Fig. 5C).

      Impaired Coagulation and Complement Cascades in the Plasma of Obese Mice

      KEGG enrichment analysis revealed that complement and coagulation cascades were enriched in the downregulated proteins (Fig. 6A and supplemental Table S14). Plasma proteomics allowed for the quantification of the majority of the proteins within the complement and coagulation cascades KEGG pathway (Fig. 6B and supplemental Fig. S6), which were predominantly downregulated in obesity. Hence, within the coagulation cascade, the fibrinogen subunits (FGA, FGB, and FGC) were downregulated alongside the coagulation factors F13A1 and F13B. Plasminogen, which degrades fibrin clots, was also decreased in the plasma of ob/ob mice. Furthermore, the protease inhibitors alpha-1-antitrypsin 1-1, 1-4, and 1-5 (SERPINA1A, SERPINA1D, and SERPINA1E, respectively) and alpha-1-antiplasmin SERPINF2 were also decreased. In the case of the complement cascade, plasma abundance of the terminal complement complex (complement components 5–9; C5–C9) was decreased. This was accompanied by a decrease in various protein precursors from the alternate pathway (e.g., complement component 3; C3), the classical pathway (e.g., complement subcomponent C1r; C1RA), and the lectin pathway (e.g., mannan-binding lectin serine protease 1; MASP1) (Fig. 6B).
      Figure thumbnail gr6
      Fig. 6Downregulation of the complement and coagulation cascades in the plasma of ob/ob mice. A, one-dimensional enrichment analysis based on KEGG terms (FDR < 0.02). B, reproduction of the KEGG complement and coagulation cascades pathway with proteins denoted as upregulated or downregulated at an FDR < 0.05 or 0.10. KEGG, Kyoto encyclopedia of genes and genomes.

      Integrated Liver and Plasma Proteomics Reveal Putative Liver-Derived Proteins Dysregulated in the Plasma of ob/ob Mice

      To identify proteins whose plasma abundance is likely to be regulated by altered secretion from the liver, we integrated the liver and plasma proteomes by looking for common features among the two datasets. Of the 7173 and 555 proteins quantified in liver and plasma, respectively, 224 proteins were quantified in both proteomes, of which 31 proteins were differentially regulated (FDR <0.05) in both liver and plasma (Fig. 7A). For proteins quantified in both liver and plasma, we plotted the log2-fold change in each dataset on a scatter plot and highlighted the proteins that were significantly regulated in both datasets (Fig. 7B and supplemental Table S15). This allowed us to identify proteins whose abundance changed in the same direction in liver and plasma (Fig. 7B), suggesting a coordinated regulation during obesity. This analysis identified the concomitant downregulation of the liver-secreted membrane attack complex proteins C8A and C9 (Fig. 7B). Proteins involved in glucose (glucose-6-phosphate isomerase; GPI) and glycerol (GPD1) metabolism were upregulated in both datasets. Vanin 1 (VNN1), an enzyme involved in oxidative stress, coenzyme A metabolism, and lipid metabolism was also upregulated in both datasets. Proteins changing in the same direction in liver and plasma also included apolipoproteins (APOA4, APOE, and APOM), serpin proteins (SERPINA7 and SERPINA1E), and EGFR (Fig. 7B). To investigate whether these concomitantly regulated proteins are also regulated at the transcriptional level in the liver of ob/ob mice, we took advantage of the published transcriptomic dataset from ob/ob liver used in Figure 5 (
      • Egami R.
      • Kokaji T.
      • Hatano A.
      • Yugi K.
      • Eto M.
      • Morita K.
      • Ohno S.
      • Fujii M.
      • Hironaka K.I.
      • Uematsu S.
      • Terakawa A.
      • Bai Y.
      • Pan Y.
      • Tsuchiya T.
      • Ozaki H.
      • et al.
      Trans-omic analysis reveals obesity-associated dysregulation of inter-organ metabolic cycles between the liver and skeletal muscle.
      ). The majority of the concomitantly regulated proteins displayed significant changes in transcription (supplemental Fig. S7) that mirrored changes in the liver and plasma proteomes (Fig. 7B). However, ALDOB and APOA5 did not change at the transcriptional level (supplemental Fig. S7), reflecting that the plasma concentration of ALDOB is indicative of hepatocyte damage, rather than a regulated process of transcription and secretion.
      Figure thumbnail gr7
      Fig. 7Integrated liver and plasma proteomics reveal concomitantly regulated proteins in the liver and plasma of ob/ob mice. A, Venn diagram displaying the overlap of quantified and significantly regulated proteins in the liver and plasma proteomes. B, scatter plot of the log2-fold-change in the liver and plasma of proteins quantified in both proteomes. Proteins significantly regulated in both proteomes are highlighted in red. C, heat map displaying the systemic (plasma or serum) regulation of proteins found to be concomitantly regulated in liver and plasma of ob/ob mice in human investigations of prediabetes and type diabetes (
      • Diamanti K.
      • Cavalli M.
      • Pereira M.J.
      • Pan G.
      • Castillejo-Lopez C.
      • Kumar C.
      • Mundt F.
      • Komorowski J.
      • Deshmukh A.S.
      • Mann M.
      • Korsgren O.
      • Eriksson J.W.
      • Wadelius C.
      Organ-specific metabolic pathways distinguish prediabetes, type 2 diabetes and normal tissues.
      ), nonalcoholic fatty liver disease (NAFLD) and cirrhosis (
      • Niu L.
      • Geyer P.E.
      • Wewer Albrechtsen N.J.
      • Gluud L.L.
      • Santos A.
      • Doll S.
      • Treit P.V.
      • Holst J.J.
      • Knop F.K.
      • Vilsboll T.
      • Junker A.
      • Sachs S.
      • Stemmer K.
      • Muller T.D.
      • Tschop M.H.
      • et al.
      Plasma proteome profiling discovers novel proteins associated with non-alcoholic fatty liver disease.
      ), high-fat diet (HFD) (
      • Lundsgaard A.-M.
      • Holm J.B.
      • Sjøberg K.A.
      • Bojsen-Møller K.N.
      • Myrmel L.S.
      • Fjære E.
      • Jensen B.A.H.
      • Nicolaisen T.S.
      • Hingst J.R.
      • Hansen S.L.
      • Doll S.
      • Geyer P.E.
      • Deshmukh A.S.
      • Holst J.J.
      • Madsen L.
      • et al.
      Mechanisms preserving insulin action during high dietary fat intake.
      ), low-calorie diet (LCD) (
      • Geyer P.E.
      • Wewer Albrechtsen N.J.
      • Tyanova S.
      • Grassl N.
      • Iepsen E.W.
      • Lundgren J.
      • Madsbad S.
      • Holst J.J.
      • Torekov S.S.
      • Mann M.
      Proteomics reveals the effects of sustained weight loss on the human plasma proteome.
      ), and Roux-en-Y gastric bypass (RYGB) (
      • Wewer Albrechtsen N.J.
      • Geyer P.E.
      • Doll S.
      • Treit P.V.
      • Bojsen-Møller K.N.
      • Martinussen C.
      • Jørgensen N.B.
      • Torekov S.S.
      • Meier F.
      • Niu L.
      • Santos A.
      • Keilhauer E.C.
      • Holst J.J.
      • Madsbad S.
      • Mann M.
      Plasma proteome profiling reveals dynamics of inflammatory and lipid homeostasis markers after roux-en-Y gastric bypass surgery.
      ).
      To identify whether these putatively liver-derived plasma proteins are relevant to human metabolic disease we investigated how these proteins were regulated in plasma and serum proteomes from human cohorts (Fig. 7C). Similarly to in the plasma of ob/ob mice, APOA4 increased in the serum of individuals with type 2 diabetes but not prediabetes (
      • Diamanti K.
      • Cavalli M.
      • Pereira M.J.
      • Pan G.
      • Castillejo-Lopez C.
      • Kumar C.
      • Mundt F.
      • Komorowski J.
      • Deshmukh A.S.
      • Mann M.
      • Korsgren O.
      • Eriksson J.W.
      • Wadelius C.
      Organ-specific metabolic pathways distinguish prediabetes, type 2 diabetes and normal tissues.
      ); furthermore, APOA4 displayed a trend to increase after 6 weeks of HFD (
      • Lundsgaard A.-M.
      • Holm J.B.
      • Sjøberg K.A.
      • Bojsen-Møller K.N.
      • Myrmel L.S.
      • Fjære E.
      • Jensen B.A.H.
      • Nicolaisen T.S.
      • Hingst J.R.
      • Hansen S.L.
      • Doll S.
      • Geyer P.E.
      • Deshmukh A.S.
      • Holst J.J.
      • Madsen L.
      • et al.
      Mechanisms preserving insulin action during high dietary fat intake.
      ). ALDOB and VNN1 are increased in the plasma of individuals with NAFLD (
      • Niu L.
      • Geyer P.E.
      • Wewer Albrechtsen N.J.
      • Gluud L.L.
      • Santos A.
      • Doll S.
      • Treit P.V.
      • Holst J.J.
      • Knop F.K.
      • Vilsboll T.
      • Junker A.
      • Sachs S.
      • Stemmer K.
      • Muller T.D.
      • Tschop M.H.
      • et al.
      Plasma proteome profiling discovers novel proteins associated with non-alcoholic fatty liver disease.
      ), whereas ALDOB is reduced after 2 years of Roux-en-Y gastric bypass-induced weight loss (
      • Wewer Albrechtsen N.J.
      • Geyer P.E.
      • Doll S.
      • Treit P.V.
      • Bojsen-Møller K.N.
      • Martinussen C.
      • Jørgensen N.B.
      • Torekov S.S.
      • Meier F.
      • Niu L.
      • Santos A.
      • Keilhauer E.C.
      • Holst J.J.
      • Madsbad S.
      • Mann M.
      Plasma proteome profiling reveals dynamics of inflammatory and lipid homeostasis markers after roux-en-Y gastric bypass surgery.
      ). These data indicate the clinical relevance of liver–plasma cross talk and our investigations in ob/ob mice.

      Discussion

      Obesity is one of the world’s largest healthcare challenges. Obesity leads to a dysregulated hepatic metabolism and to an excess deposition of lipids in the liver, known as steatosis (
      • Kawano Y.
      • Cohen D.E.
      Mechanisms of hepatic triglyceride accumulation in non-alcoholic fatty liver disease.
      ). Here, we integrated liver and plasma proteomics in leptin-deficient obese (ob/ob) mice. Our analyses illuminated the complexity of hepatic peroxisomal proliferation and protein secretion during obesity. Furthermore, we identified substantial dysregulation of the plasma proteome in ob/ob mice, particularly of proteins involved in lipid and cholesterol metabolism, protein catabolism, and the immune system. Integration of the liver and plasma proteomes identified proteins whose differential plasma abundance is putatively regulated by altered liver expression and secretion, of which APOA4, ALDOB, and VNN1 are systemically regulated in the development and/or treatment of human metabolic disease. Therefore, these integrated analyses provide biological insight into the alterations associated with obesity, as well as provide a comprehensive resource for obesity research in a prevailing model organism.
      Obesity produced a predominant increase in proteins involved in hepatic lipid metabolism. Hyperphagia and excess energy intake lead to obesity and hepatic steatosis in ob/ob mice (
      • Jéquier E.
      Leptin signaling, adiposity, and energy balance.
      ,
      • Trak-Smayra V.
      • Paradis V.
      • Massart J.
      • Nasser S.
      • Jebara V.
      • Fromenty B.
      Pathology of the liver in obese and diabetic ob/ob and db/db mice fed a standard or high-calorie diet.
      ). Our results suggest that this increase in lipid metabolism primarily results from peroxisomal proliferation with little effect on the mitochondrial proteome, which is in contrast to the skeletal muscle of these ob/ob mice in which increased mitochondrial protein content was apparent (
      • Schonke M.
      • Bjornholm M.
      • Chibalin A.V.
      • Zierath J.R.
      • Deshmukh A.S.
      Proteomics analysis of skeletal muscle from leptin-deficient ob/ob mice reveals adaptive remodeling of metabolic characteristics and fiber type composition.
      ). We capitalized on the apparent peroxisomal proliferation and the deep coverage of the liver proteome to get an insight into peroxisomal biogenesis in the liver. Increased peroxisomal content occurs via de novo synthesis, and expansion and/or division of existing peroxisomes, which requires the import of peroxisomal membrane and matrix proteins. Import of newly synthesized peroxisomal membrane proteins occurs via two classes of mechanisms. In the class I mechanism, peroxisomal membrane proteins are targeted directly to the peroxisome from the cytosol in a PEX19-dependent manner, whereas the class II mechanism indirectly imports peroxisomal membrane proteins via a PEX16-mediated recruitment to the endoplasmic reticulum (
      • Jones J.M.
      • Morrell J.C.
      • Gould S.J.
      PEX19 is a predominantly cytosolic chaperone and import receptor for class 1 peroxisomal membrane proteins.
      ,
      • Aranovich A.
      • Hua R.
      • Rutenberg A.D.
      • Kim P.K.
      PEX16 contributes to peroxisome maintenance by constantly trafficking PEX3 via the ER.
      ). PEX19 increased in liver of ob/ob mice in the absence of changes to other proteins in the peroxisomal membrane import machinery (e.g., PEX3 or PEX16). Thus, the increase in PEX19, without an increase in PEX16, may indicate a predominance of class I peroxisomal membrane protein import in liver of obese mice. Initiation of peroxisomal matrix protein import occurs through recognition and binding of proteins containing peroxisomal targeting sequences (PTS) 1 and 2 by the cytosolic proteins PEX5 and PEX7, respectively (
      • Fujiki Y.
      • Okumoto K.
      • Mukai S.
      • Honsho M.
      • Tamura S.
      Peroxisome biogenesis in mammalian cells.
      ). PEX5 and PEX7 then translocate PTS1- and PTS2-containing proteins to the PEX13-PEX14 docking complex, which transiently form an import pore (
      • Fujiki Y.
      • Okumoto K.
      • Mukai S.
      • Honsho M.
      • Tamura S.
      Peroxisome biogenesis in mammalian cells.
      ). Here we report a significant increase in PEX14 in the liver of ob/ob mice, which raises the hypothesis that PEX14 is a rate-limiting protein in peroxisomal matrix protein import. Peroxisomal elongation and division is regulated by PEX11 proteins. Three PEX11 isoforms exist in mammalian cells, namely, the alpha (PEX11A), beta (PEX11B), and gamma (PEX11G) isoforms. Knockout of PEX11B is embryonically lethal, although it only moderately affects peroxisomal fatty acid oxidation (
      • Li X.
      • Baumgart E.
      • Morrell J.C.
      • Jimenez-Sanchez G.
      • Valle D.
      • Gould S.J.
      PEX11 beta deficiency is lethal and impairs neuronal migration but does not abrogate peroxisome function.
      ). Conversely, Pex11a-/- mice are viable and they display impaired hepatic fatty acid oxidation and develop NAFLD (
      • Weng H.
      • Ji X.
      • Naito Y.
      • Endo K.
      • Ma X.
      • Takahashi R.
      • Shen C.
      • Hirokawa G.
      • Fukushima Y.
      • Iwai N.
      Pex11α deficiency impairs peroxisome elongation and division and contributes to nonalcoholic fatty liver in mice.
      ). In the liver of ob/ob mice, PEX11A and PEX11G increased in the absence of changes to PEX11B. This corroborates previous evidence that PEX11A is an inducible gene, whereas PEX11B is constitutively expressed (
      • Abe I.
      • Fujiki Y.
      cDNA cloning and characterization of a constitutively expressed isoform of the human peroxin Pex11p.
      ,
      • Abe I.
      • Okumoto K.
      • Tamura S.
      • Fujiki Y.
      Clofibrate-inducible, 28-kDa peroxisomal integral membrane protein is encoded by PEX11.
      ). However, PEX11G, which is putatively constitutively expressed (
      • Tanaka A.
      • Okumoto K.
      • Fujiki Y.
      cDNA cloning and characterization of the third isoform of human peroxin Pex11p.
      ), is inducible during severe obesity. These data provide further evidence that PEX11 isoforms have distinct regulation and function, implicating PEX11A and PEX11G in the metabolic adaptation to excess fatty acid availability.
      Various components of secretory pathways were also dysregulated within ob/ob liver. In the classical secretion pathway, proteins with a signal peptide are targeted to the endoplasmic reticulum and transported via a multistep vesicular-mediated mechanism through the Golgi apparatus and released into the extracellular space through vesicular fusion with the plasma membrane (
      • Palade G.
      Intracellular aspects of the process of protein synthesis.
      ). Greater than 10% of proteins containing a signal peptide were differentially regulated in liver of ob/ob mice. Furthermore, the summed protein abundance of the endoplasmic reticulum was decreased alongside the enrichment of the GOCC term endoplasmic reticulum chaperone complex among the downregulated proteins. The lumen of the endoplasmic reticulum, and particularly the chaperone proteins, provides a specialized environment for folding and maturation of secreted proteins (
      • Braakman I.
      • Hebert D.N.
      Protein folding in the endoplasmic reticulum.
      ). Endoplasmic reticulum stress is a common feature of obesity and NAFLD; however, this is typically characterized by an upregulation of chaperone genes (
      • Pagliassotti M.J.
      • Kim P.Y.
      • Estrada A.L.
      • Stewart C.M.
      • Gentile C.L.
      Endoplasmic reticulum stress in obesity and obesity-related disorders: An expanded view.
      ). However, consistent with a reduction in protein content the mRNA expression of endoplasmic reticulum chaperon complex genes typically decreases in ob/ob mice (
      • Egami R.
      • Kokaji T.
      • Hatano A.
      • Yugi K.
      • Eto M.
      • Morita K.
      • Ohno S.
      • Fujii M.
      • Hironaka K.I.
      • Uematsu S.
      • Terakawa A.
      • Bai Y.
      • Pan Y.
      • Tsuchiya T.
      • Ozaki H.
      • et al.
      Trans-omic analysis reveals obesity-associated dysregulation of inter-organ metabolic cycles between the liver and skeletal muscle.
      ), although the mRNA expression of BiP (HSPA5) has also been shown to increase (
      • Kammoun H.L.
      • Chabanon H.
      • Hainault I.
      • Luquet S.
      • Magnan C.
      • Koike T.
      • Ferré P.
      • Foufelle F.
      GRP78 expression inhibits insulin and ER stress-induced SREBP-1c activation and reduces hepatic steatosis in mice.
      ,
      • Özcan U.
      • Cao Q.
      • Yilmaz E.
      • Lee A.-H.
      • Iwakoshi N.N.
      • Özdelen E.
      • Tuncman G.
      • Görgün C.
      • Glimcher L.H.
      • Hotamisligil G.S.
      Endoplasmic reticulum stress links obesity, insulin action, and type 2 diabetes.
      ). It could be hypothesized that a reduction in protein abundance of endoplasmic reticulum chaperone proteins contributes to altered hepatic protein secretion, through increased misfolding of proteins that were destined for secretion. Proteins can also be secreted via endoplasmic reticulum–independent, nonclassical mechanisms, including via the release of exosomes (
      • Hessvik N.P.
      • Llorente A.
      Current knowledge on exosome biogenesis and release.
      ). Proteins annotated to the extracellular exosome (GOCC) were enriched within the upregulated proteins. Nonetheless, several exosome-annotated proteins were downregulated, including proteins forming the membrane attack complex of the immune system. Extracellular matrix (GOCC) was also enriched within the upregulated proteins and transcripts of ob/ob mice, which contrasted with a downregulation of extracellular matrix terms in the proteome of HFD-fed mice (
      • Krahmer N.
      • Najafi B.
      • Schueder F.
      • Quagliarini F.
      • Steger M.
      • Seitz S.
      • Kasper R.
      • Salinas F.
      • Cox J.
      • Uhlenhaut N.H.
      • Walther T.C.
      • Jungmann R.
      • Zeigerer A.
      • Borner G.H.H.
      • Mann M.
      Organellar proteomics and phospho-proteomics reveal subcellular reorganization in diet-induced hepatic steatosis.
      ). However, extracellular matrix deposition has been described in the liver of HFD-fed mice elsewhere, whereby increased hepatic staining of alpha-smooth muscle actin is apparent, as well as increased gene expression of the type I and type II collagen molecules (
      • Wada T.
      • Miyashita Y.
      • Sasaki M.
      • Aruga Y.
      • Nakamura Y.
      • Ishii Y.
      • Sasahara M.
      • Kanasaki K.
      • Kitada M.
      • Koya D.
      • Shimano H.
      • Tsuneki H.
      • Sasaoka T.
      Eplerenone ameliorates the phenotypes of metabolic syndrome with NASH in liver-specific SREBP-1c Tg mice fed high-fat and high-fructose diet.
      ,
      • Dixon L.J.
      • Flask C.A.
      • Papouchado B.G.
      • Feldstein A.E.
      • Nagy L.E.
      Caspase-1 as a central regulator of high fat diet-induced non-alcoholic steatohepatitis.
      ,
      • Williams A.S.
      • Kang L.
      • Zheng J.
      • Grueter C.
      • Bracy D.P.
      • James F.D.
      • Pozzi A.
      • Wasserman D.H.
      Integrin α1-null mice exhibit improved fatty liver when fed a high fat diet despite severe hepatic insulin resistance.
      ). Furthermore, both portal and perisinusoidal fibrosis are apparent in ob/ob mice fed a standard chow diet (
      • Trak-Smayra V.
      • Paradis V.
      • Massart J.
      • Nasser S.
      • Jebara V.
      • Fromenty B.
      Pathology of the liver in obese and diabetic ob/ob and db/db mice fed a standard or high-calorie diet.
      ,
      • Sutter A.G.
      • Palanisamy A.P.
      • Lench J.H.
      • Jessmore A.P.
      • Chavin K.D.
      Development of steatohepatitis in ob/ob mice is dependent on Toll-like receptor 4.
      ). In addition to collagen deposition, we identified the upregulation of fibrillin 1, fibronectin, and laminin, as well as several secreted proteins that are annotated to the extracellular matrix (e.g., APOE, LGALS1, LGALS3, and POSTN). Together, these data indicated a dysregulation of secreted proteins and the secretory machinery within the liver and led us to investigate whether these derangements altered the plasma proteome of ob/ob mice.
      The plasma proteome of ob/ob mice displayed substantial regulation, whereby over half of the quantified proteins were differentially regulated. This is a somewhat larger proportion of the proteome than for HFD-fed mice, in which 30% of the plasma proteome is differentially regulated (
      • Martinez-Huenchullan S.F.
      • Shipsey I.
      • Hatchwell L.
      • Min D.
      • Twigg S.M.
      • Larance M.
      Blockade of high-fat diet proteomic phenotypes using exercise as prevention or treatment.
      ), consistent with a greater degree of obesity in ob/ob mice. Differentially regulated proteins included the potential NAFLD biomarkers ALDOB, APOE, DPP4, PIGR, and APOM identified in humans (
      • Niu L.
      • Geyer P.E.
      • Wewer Albrechtsen N.J.
      • Gluud L.L.
      • Santos A.
      • Doll S.
      • Treit P.V.
      • Holst J.J.
      • Knop F.K.
      • Vilsboll T.
      • Junker A.
      • Sachs S.
      • Stemmer K.
      • Muller T.D.
      • Tschop M.H.
      • et al.
      Plasma proteome profiling discovers novel proteins associated with non-alcoholic fatty liver disease.
      ). Enriched within the upregulated proteins were GOBP terms relating to lipid and cholesterol homeostasis, which is in line with obesity-induced dyslipidemia (
      • Klop B.
      • Elte J.W.F.
      • Cabezas M.C.
      Dyslipidemia in obesity: Mechanisms and potential targets.
      ). Furthermore, GOBP terms related to catabolism and the proteasome were also enriched among the upregulated proteins, whereas each protein of the 20S proteasome was increased in ob/ob plasma. The 26S proteasome, of which the 20S proteasome is a subunit, degrades the majority of cellular proteins. Although the role and source of the extracellular proteasome is less clear (
      • Dwivedi V.
      • Yaniv K.
      • Sharon M.
      Beyond cells: The extracellular circulating 20S proteasomes.
      ), the circulating proteasome is enzymatically functional (
      • Zoeger A.
      • Blau M.
      • Egerer K.
      • Feist E.
      • Dahlmann B.
      Circulating proteasomes are functional and have a subtype pattern distinct from 20S proteasomes in major blood cells.
      ) and has been identified in the human plasma proteome (
      • Wewer Albrechtsen N.J.
      • Geyer P.E.
      • Doll S.
      • Treit P.V.
      • Bojsen-Møller K.N.
      • Martinussen C.
      • Jørgensen N.B.
      • Torekov S.S.
      • Meier F.
      • Niu L.
      • Santos A.
      • Keilhauer E.C.
      • Holst J.J.
      • Madsbad S.
      • Mann M.
      Plasma proteome profiling reveals dynamics of inflammatory and lipid homeostasis markers after roux-en-Y gastric bypass surgery.
      ). Indeed, the extracellular proteasome may still be related to the degradation of short-lived and misfolded proteins, or alternatively it simply may represent a product of damaged cells. Nonetheless, the extracellular accumulation of damaged proteins has been associated with various disease states, including in a mouse model of type 2 diabetes (
      • Wyatt A.R.
      • Yerbury J.J.
      • Poon S.
      • Wilson M.R.
      Therapeutic targets in extracellular protein deposition diseases.
      ). Furthermore, emerging evidence indicates circulating proteasome activity is negatively correlated with physical activity in individuals with chronic obstructive pulmonary disease (
      • Lee K.Y.
      • Chen T.T.
      • Chiang L.L.
      • Chuang H.C.
      • Feng P.H.
      • Liu W.T.
      • Chen K.Y.
      • Ho S.C.
      Proteasome activity related with the daily physical activity of COPD patients.
      ) and positively correlated with squamous cell carcinoma (skin cancer) (
      • Kakurina G.V.
      • Cheremisina O.V.
      • Choinzonov E.L.
      • Kondakova I.V.
      Circulating proteasomes in the pathogenesis of head and neck squamous cell carcinoma.
      ), indicating its clinical relevance. Future studies should investigate the relevance of elevated plasma proteasome abundance in the context of metabolic dysfunction and human obesity.
      Enriched within the downregulated plasma proteins was the KEGG term coagulation and complement cascade, implicating hemostatic and immune system dysfunction in obesity. The reduction in plasma abundance of thrombotic proteins (e.g., the fibrinogens FGA, FGB, and FGG and factor VII [F7]) is at odds with elevations in circulating prothrombotic factors and hypercoagulation in human obesity and type 2 diabetes (
      • Carr M.E.
      Diabetes mellitus: A hypercoagulable state.
      ,
      • Kaye S.M.
      • Pietiläinen K.H.
      • Kotronen A.
      • Joutsi-Korhonen L.
      • Kaprio J.
      • Yki-Järvinen H.
      • Silveira A.
      • Hamsten A.
      • Lassila R.
      • Rissanen A.
      Obesity-related derangements of coagulation and fibrinolysis: A study of obesity-discordant monozygotic twin pairs.
      ,
      • De Pergola G.
      • De Mitrio V.
      • Giorgino F.
      • Sciaraffia M.
      • Minenna A.
      • Di Bari L.
      • Pannacciulli N.
      • Giorgino R.
      Increase in both pro-thrombotic and anti-thrombotic factors in obese premenopausal women: Relationship with body fat distribution.
      ). Nonetheless, this is in agreement with data identifying impaired platelet function and coagulation in leptin-deficient ob/ob mice (
      • Henry M.L.
      • Davidson L.B.
      • Wilson J.E.
      • McKenna B.K.
      • Scott S.A.
      • McDonagh P.F.
      • Ritter L.S.
      Whole blood aggregation and coagulation in db/db and ob/ob mouse models of type 2 diabetes.
      ). These data highlight the need to carefully choose a model organism dependent upon the specific aspect or stage of obesity that is to be studied. Nonetheless, we were interested by the observation that proteins from the membrane attack complex, the terminal complex of the complement cascade, were downregulated in both the liver and the plasma. Indeed, C8A and C9 were downregulated to a similar degree in both liver and plasma. Given the predominant expression of these proteins in the liver (
      • Uhlén M.
      • Fagerberg L.
      • Hallström B.M.
      • Lindskog C.
      • Oksvold P.
      • Mardinoglu A.
      • Sivertsson Å.
      • Kampf C.
      • Sjöstedt E.
      • Asplund A.
      • Olsson I.
      • Edlund K.
      • Lundberg E.
      • Navani S.
      • Szigyarto C.A.
      • et al.
      Proteomics. Tissue-based map of the human proteome.
      ), plasma levels appear to be governed by altered secretion from the liver.
      To identify additional proteins whose plasma regulation is likely influenced by altered abundance and secretion from the liver we integrated the liver and plasma proteomes. We identified 31 proteins that were significantly regulated in both liver and plasma, of which 25 were regulated in the same direction, including proteins of the complement and coagulation cascades: C8A, C9, SERPINA1E, and SERPINA7. The NAFLD marker ALDOB was upregulated in both datasets, whereas DPP4 was upregulated in the plasma but downregulated in the liver. Furthermore, the apolipoproteins APOA4, APOA5, APOE, and APOM were also regulated in the same direction in liver and plasma. Interesting, the obesity-induced alteration in plasma abundance of a number of the proteins identified here as concomitantly regulated in the liver and plasma by obesity (e.g., ALDOB, APOE, APOM, EGFR, SERPINA1E, and SERPIN7) can be reversed by exercise training (
      • Martinez-Huenchullan S.F.
      • Shipsey I.
      • Hatchwell L.
      • Min D.
      • Twigg S.M.
      • Larance M.
      Blockade of high-fat diet proteomic phenotypes using exercise as prevention or treatment.
      ). Plasma ALDOB and VNN1 also increase in humans with NAFLD (
      • Niu L.
      • Geyer P.E.
      • Wewer Albrechtsen N.J.
      • Gluud L.L.
      • Santos A.
      • Doll S.
      • Treit P.V.
      • Holst J.J.
      • Knop F.K.
      • Vilsboll T.
      • Junker A.
      • Sachs S.
      • Stemmer K.
      • Muller T.D.
      • Tschop M.H.
      • et al.
      Plasma proteome profiling discovers novel proteins associated with non-alcoholic fatty liver disease.
      ), whereas elevated plasma ALDOB can be reversed by sustained weight loss induced by Roux-en-Y gastric bypass (
      • Wewer Albrechtsen N.J.
      • Geyer P.E.
      • Doll S.
      • Treit P.V.
      • Bojsen-Møller K.N.
      • Martinussen C.
      • Jørgensen N.B.
      • Torekov S.S.
      • Meier F.
      • Niu L.
      • Santos A.
      • Keilhauer E.C.
      • Holst J.J.
      • Madsbad S.
      • Mann M.
      Plasma proteome profiling reveals dynamics of inflammatory and lipid homeostasis markers after roux-en-Y gastric bypass surgery.
      ). Furthermore, serum APOA4 is increased during type 2 diabetes and the consumption of a HFD in humans (
      • Diamanti K.
      • Cavalli M.
      • Pereira M.J.
      • Pan G.
      • Castillejo-Lopez C.
      • Kumar C.
      • Mundt F.
      • Komorowski J.
      • Deshmukh A.S.
      • Mann M.
      • Korsgren O.
      • Eriksson J.W.
      • Wadelius C.
      Organ-specific metabolic pathways distinguish prediabetes, type 2 diabetes and normal tissues.
      ,
      • Lundsgaard A.-M.
      • Holm J.B.
      • Sjøberg K.A.
      • Bojsen-Møller K.N.
      • Myrmel L.S.
      • Fjære E.
      • Jensen B.A.H.
      • Nicolaisen T.S.
      • Hingst J.R.
      • Hansen S.L.
      • Doll S.
      • Geyer P.E.
      • Deshmukh A.S.
      • Holst J.J.
      • Madsen L.
      • et al.
      Mechanisms preserving insulin action during high dietary fat intake.
      ), suggesting it is a sensitive systemic marker of altered liver metabolism. LGALS3, which lacks a signal peptide (
      • Menon R.P.
      • Hughes R.C.
      Determinants in the N-terminal domains of galectin-3 for secretion by a novel pathway circumventing the endoplasmic reticulum-Golgi complex.
      ), was also identified as one of the most upregulated exosome-annotated proteins in the liver and was also upregulated in the plasma. In humans, plasma LGALS3 correlates with disease states such as type 2 diabetes and cardiovascular disease (
      • van Kimmenade R.R.
      • Januzzi Jr., J.L.
      • Ellinor P.T.
      • Sharma U.C.
      • Bakker J.A.
      • Low A.F.
      • Martinez A.
      • Crijns H.J.
      • MacRae C.A.
      • Menheere P.P.
      • Pinto Y.M.
      Utility of amino-terminal pro-brain natriuretic peptide, galectin-3, and apelin for the evaluation of patients with acute heart failure.
      ,
      • van der Velde A.R.
      • Gullestad L.
      • Ueland T.
      • Aukrust P.
      • Guo Y.
      • Adourian A.
      • Muntendam P.
      • van Veldhuisen D.J.
      • de Boer R.A.
      Prognostic value of changes in galectin-3 levels over time in patients with heart failure: Data from CORONA and COACH.
      ,
      • Weigert J.
      • Neumeier M.
      • Wanninger J.
      • Bauer S.
      • Farkas S.
      • Scherer M.N.
      • Schnitzbauer A.
      • Schäffler A.
      • Aslanidis C.
      • Schölmerich J.
      • Buechler C.
      Serum galectin-3 is elevated in obesity and negatively correlates with glycosylated hemoglobin in type 2 diabetes.
      ) and is an emerging therapeutic target for treating NAFLD (
      • Al Attar A.
      • Antaramian A.
      • Noureddin M.
      Review of galectin-3 inhibitors in the treatment of nonalcoholic steatohepatitis.
      ). LGALS3 is profibrotic (
      • Jeftic I.
      • Jovicic N.
      • Pantic J.
      • Arsenijevic N.
      • Lukic M.L.
      • Pejnovic N.
      Galectin-3 ablation enhances liver steatosis, but attenuates inflammation and IL-33-dependent fibrosis in obesogenic mouse model of nonalcoholic steatohepatitis.
      ,
      • Yu L.
      • Ruifrok Willem P.T.
      • Meissner M.
      • Bos Eelke M.
      • van Goor H.
      • Sanjabi B.
      • van der Harst P.
      • Pitt B.
      • Goldstein Irwin J.
      • Koerts Jasper A.
      • van Veldhuisen Dirk J.
      • Bank Ruud A.
      • van Gilst Wiek H.
      • Silljé Herman H.W.
      • de Boer Rudolf A.
      Genetic and pharmacological inhibition of galectin-3 prevents cardiac remodeling by interfering with myocardial fibrogenesis.
      ) and immune modulatory (
      • Dumic J.
      • Dabelic S.
      • Flögel M.
      Galectin-3: An open-ended story.
      ), regulating the extracellular matrix of numerous tissues, including the cardiac tissue of patients with heart failure (
      • Lin Y.H.
      • Lin L.Y.
      • Wu Y.W.
      • Chien K.L.
      • Lee C.M.
      • Hsu R.B.
      • Chao C.L.
      • Wang S.S.
      • Hsein Y.C.
      • Liao L.C.
      • Ho Y.L.
      • Chen M.F.
      The relationship between serum galectin-3 and serum markers of cardiac extracellular matrix turnover in heart failure patients.
      ). Although increased LGALS3 expression is not exclusive to the liver during obesity (
      • Papaspyridonos M.
      • McNeill E.
      • de Bono Joe P.
      • Smith A.
      • Burnand Kevin G.
      • Channon Keith M.
      • Greaves David R.
      Galectin-3 is an amplifier of inflammation in atherosclerotic plaque progression through macrophage activation and monocyte chemoattraction.
      ), our data does indicate that the liver is a source, or potentially a target (
      • Jeftic I.
      • Jovicic N.
      • Pantic J.
      • Arsenijevic N.
      • Lukic M.L.
      • Pejnovic N.
      Galectin-3 ablation enhances liver steatosis, but attenuates inflammation and IL-33-dependent fibrosis in obesogenic mouse model of nonalcoholic steatohepatitis.
      ), of altered systemic LGALS3 abundance in obesity. Together, these data identify clinically relevant liver-derived plasma proteins in obesity and indicate that the ob/ob mouse is a useful model of liver–plasma cross talk in hepatic dysfunction.
      One limitation of this investigation is that ob/ob mice are a monogenic model of obesity, harboring a mutation in the leptin gene resulting in a truncated nonfunctional protein (
      • Zhang Y.
      • Proenca R.
      • Maffei M.
      • Barone M.
      • Leopold L.
      • Friedman J.M.
      Positional cloning of the mouse obese gene and its human homologue.
      ), which is in contrast to the vast majority of cases of human obesity, in which circulating levels of leptin increase (
      • Considine R.V.
      • Sinha M.K.
      • Heiman M.L.
      • Kriauciunas A.
      • Stephens T.W.
      • Nyce M.R.
      • Ohannesian J.P.
      • Marco C.C.
      • McKee L.J.
      • Bauer T.L.
      • Caro J.F.
      Serum immunoreactive-leptin concentrations in normal-weight and obese humans.
      ). Some phenotypic and proteomic aspects of obesity in ob/ob mice will be directly due to the loss of functional leptin, rather than an outcome of obesity. For example, leptin reduces the expression of SCD1, the rate-limiting enzyme in the biosynthesis of monounsaturated fatty acids (
      • Biddinger S.B.
      • Miyazaki M.
      • Boucher J.
      • Ntambi J.M.
      • Kahn C.R.
      Leptin suppresses stearoyl-CoA desaturase 1 by mechanisms independent of insulin and sterol regulatory element-binding protein-1c.
      ), whereas leptin per se is immunoregulatory (
      • Naylor C.
      • Petri Jr., W.A.
      Leptin regulation of immune responses.
      ). Indeed, the poor correlation between the liver proteomes of ob/ob and HFD-fed mice further highlights the issue of translating findings from a single model of obesity, albeit many processes were commonly regulated between ob/ob and HFD-fed mice when viewed at a systems level (e.g., fatty acid metabolism, glycolysis and gluconeogenesis, and peroxisomal biogenesis). Nonetheless, the ob/ob mouse remains a useful model for obesity research (
      • Barrett P.
      • Mercer J.G.
      • Morgan P.J.
      Preclinical models for obesity research.
      ). Furthermore, individuals with mutations in the human leptin gene have been identified, which is also associated with severe early onset of obesity (
      • Montague C.T.
      • Farooqi I.S.
      • Whitehead J.P.
      • Soos M.A.
      • Rau H.
      • Wareham N.J.
      • Sewter C.P.
      • Digby J.E.
      • Mohammed S.N.
      • Hurst J.A.
      • Cheetham C.H.
      • Earley A.R.
      • Barnett A.H.
      • Prins J.B.
      • O'Rahilly S.
      Congenital leptin deficiency is associated with severe early-onset obesity in humans.
      ). Thus, our proteomic investigation of the ob/ob mouse may also have relevance to a rare congenital form of leptin-deficient obesity.
      By quantifying the abundance of 7173 and 555 proteins within the liver and plasma of obese leptin-deficient ob/ob mice, respectively, we illuminated the complexity of adaptations during severe obesity. Within the liver, fatty acid metabolism, peroxisomal biogenesis, and the secretory pathways were dysregulated. Furthermore, we identified the substantial dysregulation of the plasma proteome in ob/ob mice, whereby more than 50% of the quantified plasma proteins were differentially regulated, identifying metabolic, proteasomal, and immune system dysregulation. Furthermore, by integrating the liver and plasma proteomes, we identified concomitantly regulated proteins with clinical relevance to human metabolic diseases. In addition to these biological insights, these analyses provide a comprehensive resource in the study of obesity.

      Data Availability

      The mass spectrometry proteomics data produced in this investigation, including peptide and protein level identifications, have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository (
      • Perez-Riverol Y.
      • Csordas A.
      • Bai J.
      • Bernal-Llinares M.
      • Hewapathirana S.
      • Kundu D.J.
      • Inuganti A.
      • Griss J.
      • Mayer G.
      • Eisenacher M.
      • Perez E.
      • Uszkoreit J.
      • Pfeuffer J.
      • Sachsenberg T.
      • Yilmaz S.
      • et al.
      The PRIDE database and related tools and resources in 2019: Improving support for quantification data.
      ) (http://www.proteomexchange.org/) with the accession number PXD026821. The in-house library used for the quantification of the plasma proteome has previously been deposited to PRIDE with the identifier PXD017170 (
      • Dall M.
      • Hassing A.S.
      • Niu L.
      • Nielsen T.S.
      • Ingerslev L.R.
      • Sulek K.
      • Trammell S.A.J.
      • Gillum M.P.
      • Barrès R.
      • Larsen S.
      • Poulsen S.S.
      • Mann M.
      • Ørskov C.
      • Treebak J.T.
      Hepatocyte-specific perturbation of NAD+ biosynthetic pathways in mice induces reversible nonalcoholic steatohepatitis–like phenotypes.
      ). Annotated spectra of data-dependent acquisition data can be visualized using the MS-Viewer in ProteinProspector with the search key t2hr4mzs7d: http://msviewer.ucsf.edu:443/prospector/cgi-bin/mssearch.cgi?report_title=MS-Viewer&search_key=t2hr4mzs7d&search_name=msviewer.

      Supplemental data

      This article contains supplemental data.

      Conflict of interest

      The authors declare no competing interests.

      Acknowledgments

      This work was supported by unconditional donations from the Novo Nordisk Foundation (NNF) to the NNF Center for Basic Metabolic Research (http://www.cbmr.ku.dk) (Grant number NNF18CC0034900) and the NNF Center for Protein Research (https://www.cpr.ku.dk/) (Grant number NNF14CC001). We would like to acknowledge Matthias Mann and the mass spectrometry platform at the NNF Center for Protein Research for technical assistance and access to mass spectrometers.

      Funding and additional information

      This work was supported by grants from the Swedish Research Council (2015-00165) and the Knut and Alice Wallenberg Foundation (2018-0094) to J. R. Z., an EFSD Rising Star Fellowship Award to B. S., and an EFSD/NNF Future Leaders Award to A. S. D. (Grant number: NNF19SA058976).

      Author contributions

      A. S. D. conceptualization; B. S., A. G.-F., M. L. B., and M. B. methodology; B. S. formal analysis; B. S., A. G.-F., M. L. B., and M. B. investigation; A. S. D. and J. R. Z. resources; B. S. data curation; B. S. writing – original draft; B. S., A. G.-F., M. L. B., M. B., L. N., J. R. Z., and A. S. D. writing – review & editing; B. S. visualization; A. S. D. and J. R. Z. supervision; A. S. D. project administration; A. S. D. and J. R. Z. funding acquisition.

      Supplemental Data

      • Supplemental Figures S1

        Physiological phenotyping of ob/ob mice. A, body mass. B, food intake in the light and dark phases. C, total cage activity in the light and dark phases. D, oxygen consumption in the light and dark phases. E, blood glucose concentration in response to an intraperitoneal glucose tolerance test. F, plasma insulin in the fasted state and 15 min after an intraperitoneal glucose bolus. Data are represented as mean ± SD (n = 7–8); difference between genotypes within a time point: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001; difference between time points within a genotype: #p < 0.05, ##p < 0.01, ###p < 0.001.

      • Supplemental Figures S2

        Quality control of proteomics data. A, the number of quantified proteins ranged between 6427 and 6911 for each liver sample. B, inter-sample Pearson’s correlation of log2 TPA values in the liver. C, dynamic range of the liver proteome (median protein abundance). D, the number of quantified proteins ranged between 426 and 516 for each plasma sample. E, inter-sample Pearson’s correlations of protein intensities in the plasma. F, dynamic range of the plasma proteome (median intensity).

      • Supplemental Figures S3

        Summed protein abundance of cellular compartments. Summed protein abundance of cellular compartments (GOCC) within liver. Data are represented as mean ± SD (n = 4), student’s t-test: ∗p < 0.05, ∗∗p < 0.01.

      • Supplemental Figures S4

        Abundance of proteins regulating peroxisomal biogenesis in the liver of ob/ob mice. Liver protein abundance of proteins involved in peroxisomal biogenesis (A) and peroxisomal fission (B). Data are represented as mean ± SD (n = 4).

      • Supplemental Figures S5

        Comparison of the liver proteomes of ob/ob and HFD-fed mice. A, scatter plot of the log2-fold change in the liver of ob/ob and HFD-fed mice. B, two-dimensional enrichment analysis based on GOBP (purple), GOCC (blue), and KEGG (orange) terms for ob/ob and HFD-fed mice (FDR < 0.02).

      • Supplemental Figures S6

        Abundance of proteins in the complement and coagulation cascades in the plasma of ob/ob mice. Plasma protein abundance of proteins from the complement (A) and coagulation (B) cascades (KEGG pathway). Data are represented as mean ± SD (n = 7).

      • Supplemental Figures S7

        mRNA expression of proteins concomitantly regulated in the liver and plasma of ob/ob mice. mRNA expression of proteins concomitantly regulated in the liver and plasma of ob/ob mice. Data are represented as mean ± SD (n = 11–12). Data is reproduced from the transcriptomic study of Egami et al. (

        • Egami R.
        • Kokaji T.
        • Hatano A.
        • Yugi K.
        • Eto M.
        • Morita K.
        • Ohno S.
        • Fujii M.
        • Hironaka K.I.
        • Uematsu S.
        • Terakawa A.
        • Bai Y.
        • Pan Y.
        • Tsuchiya T.
        • Ozaki H.
        • et al.
        Trans-omic analysis reveals obesity-associated dysregulation of inter-organ metabolic cycles between the liver and skeletal muscle.
        ).

      References

        • Kawano Y.
        • Cohen D.E.
        Mechanisms of hepatic triglyceride accumulation in non-alcoholic fatty liver disease.
        J. Gastroenterol. 2013; 48: 434-441
        • Cohen J.C.
        • Horton J.D.
        • Hobbs H.H.
        Human fatty liver disease: Old questions and new insights.
        Science. 2011; 332: 1519-1523
        • Gruben N.
        • Shiri-Sverdlov R.
        • Koonen D.P.Y.
        • Hofker M.H.
        Nonalcoholic fatty liver disease: A main driver of insulin resistance or a dangerous liaison?.
        Biochim. Biophys. Acta. 2014; 1842: 2329-2343
        • Knobler H.
        • Schattner A.
        • Zhornicki T.
        • Malnick S.D.
        • Keter D.
        • Sokolovskaya N.
        • Lurie Y.
        • Bass D.D.
        Fatty liver–an additional and treatable feature of the insulin resistance syndrome.
        QJM. 1999; 92: 73-79
        • Utzschneider K.M.
        • Kahn S.E.
        The role of insulin resistance in nonalcoholic fatty liver disease.
        J. Clin. Endocrinol. Metab. 2006; 91: 4753-4761
        • Organization W.H.
        Global Status Report on Noncommunicable Diseases 2014.
        World Health Organization, Geneva2014
        • Bellentani S.
        • Scaglioni F.
        • Marino M.
        • Bedogni G.
        Epidemiology of non-alcoholic fatty liver disease.
        Dig. Dis. 2010; 28: 155-161
        • Fan J.G.
        • Kim S.U.
        • Wong V.W.
        New trends on obesity and NAFLD in Asia.
        J. Hepatol. 2017; 67: 862-873
        • Hotamisligil G.S.
        Inflammation and metabolic disorders.
        Nature. 2006; 444: 860-867
        • Shoelson S.E.
        • Lee J.
        • Goldfine A.B.
        Inflammation and insulin resistance.
        J. Clin. Invest. 2006; 116: 1793-1801
        • Hotamisligil G.S.
        • Shargill N.S.
        • Spiegelman B.M.
        Adipose expression of tumor necrosis factor-alpha: Direct role in obesity-linked insulin resistance.
        Science. 1993; 259: 87
        • Niu L.
        • Geyer P.E.
        • Wewer Albrechtsen N.J.
        • Gluud L.L.
        • Santos A.
        • Doll S.
        • Treit P.V.
        • Holst J.J.
        • Knop F.K.
        • Vilsboll T.
        • Junker A.
        • Sachs S.
        • Stemmer K.
        • Muller T.D.
        • Tschop M.H.
        • et al.
        Plasma proteome profiling discovers novel proteins associated with non-alcoholic fatty liver disease.
        Mol. Syst. Biol. 2019; 15e8793
        • Anstey K.J.
        • Cherbuin N.
        • Budge M.
        • Young J.
        Body mass index in midlife and late-life as a risk factor for dementia: A meta-analysis of prospective studies.
        Obes. Rev. 2011; 12: e426-e437
        • Calle E.E.
        • Thun M.J.
        • Petrelli J.M.
        • Rodriguez C.
        • Heath C.W.
        Body-mass index and mortality in a prospective cohort of U.S. adults.
        N. Engl. J. Med. 1999; 341: 1097-1105
        • Calle E.E.
        • Rodriguez C.
        • Walker-Thurmond K.
        • Thun M.J.
        Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults.
        N. Engl. J. Med. 2003; 348: 1625-1638
        • Nho K.
        • Kueider-Paisley A.
        • Ahmad S.
        • MahmoudianDehkordi S.
        • Arnold M.
        • Risacher S.L.
        • Louie G.
        • Blach C.
        • Baillie R.
        • Han X.
        • Kastenmüller G.
        • Trojanowski J.Q.
        • Shaw L.M.
        • Weiner M.W.
        • Doraiswamy P.M.
        • et al.
        Association of altered liver enzymes with Alzheimer disease diagnosis, cognition, neuroimaging measures, and cerebrospinal fluid biomarkers.
        JAMA Netw. Open. 2019; 2e197978
        • Bassendine M.F.
        • Taylor-Robinson S.D.
        • Fertleman M.
        • Khan M.
        • Neely D.
        Is Alzheimer's disease a liver disease of the brain?.
        J. Alzheimers Dis. 2020; 75: 1-14
        • Sorensen H.T.
        • Friis S.
        • Olsen J.H.
        • Thulstrup A.M.
        • Mellemkjaer L.
        • Linet M.
        • Trichopoulos D.
        • Vilstrup H.
        • Olsen J.
        Risk of liver and other types of cancer in patients with cirrhosis: A nationwide cohort study in Denmark.
        Hepatology. 1998; 28: 921-925
        • Targher G.
        • Bertolini L.
        • Rodella S.
        • Tessari R.
        • Zenari L.
        • Lippi G.
        • Arcaro G.
        Nonalcoholic fatty liver disease is independently associated with an increased incidence of cardiovascular events in type 2 diabetic patients.
        Diabetes Care. 2007; 30: 2119
        • Barrett P.
        • Mercer J.G.
        • Morgan P.J.
        Preclinical models for obesity research.
        Dis. Model. Mech. 2016; 9: 1245
        • Zhang Y.
        • Proenca R.
        • Maffei M.
        • Barone M.
        • Leopold L.
        • Friedman J.M.
        Positional cloning of the mouse obese gene and its human homologue.
        Nature. 1994; 372: 425-432
        • Mayer J.
        • Bates M.W.
        • Dickie M.M.
        Hereditary diabetes in genetically obese mice.
        Science. 1951; 113: 746-747
        • Mayer J.
        • Dickie M.M.
        • Bates M.W.
        • Vitale J.J.
        Free selection of nutrients by hereditarily obese mice.
        Science. 1951; 113: 745-746
        • Jéquier E.
        Leptin signaling, adiposity, and energy balance.
        Ann. N. Y. Acad. Sci. 2002; 967: 379-388
        • Dubuc P.U.
        The development of obesity, hyperinsulinemia, and hyperglycemia in ob/ob mice.
        Metabolism. 1976; 25: 1567-1574
        • Trak-Smayra V.
        • Paradis V.
        • Massart J.
        • Nasser S.
        • Jebara V.
        • Fromenty B.
        Pathology of the liver in obese and diabetic ob/ob and db/db mice fed a standard or high-calorie diet.
        Int. J. Exp. Pathol. 2011; 92: 413-421
        • Schonke M.
        • Bjornholm M.
        • Chibalin A.V.
        • Zierath J.R.
        • Deshmukh A.S.
        Proteomics analysis of skeletal muscle from leptin-deficient ob/ob mice reveals adaptive remodeling of metabolic characteristics and fiber type composition.
        Proteomics. 2018; 18e1700375
        • Wisniewski J.R.
        • Mann M.
        Consecutive proteolytic digestion in an enzyme reactor increases depth of proteomic and phosphoproteomic analysis.
        Anal. Chem. 2012; 84: 2631-2637
        • Rappsilber J.
        • Ishihama Y.
        • Mann M.
        Stop and go extraction tips for matrix-assisted laser desorption/ionization, nanoelectrospray, and LC/MS sample pretreatment in proteomics.
        Anal. Chem. 2003; 75: 663-670
        • Geyer P.E.
        • Kulak N.A.
        • Pichler G.
        • Holdt L.M.
        • Teupser D.
        • Mann M.
        Plasma proteome profiling to assess human health and disease.
        Cell Syst. 2016; 2: 185-195
        • Tyanova S.
        • Temu T.
        • Cox J.
        The MaxQuant computational platform for mass spectrometry-based shotgun proteomics.
        Nat. Protoc. 2016; 11: 2301-2319
        • Bruderer R.
        • Bernhardt O.M.
        • Gandhi T.
        • Miladinović S.M.
        • Cheng L.Y.
        • Messner S.
        • Ehrenberger T.
        • Zanotelli V.
        • Butscheid Y.
        • Escher C.
        • Vitek O.
        • Rinner O.
        • Reiter L.
        Extending the limits of quantitative proteome profiling with data-independent acquisition and application to acetaminophen-treated three-dimensional liver microtissues.
        Mol. Cell. Proteomics. 2015; 14: 1400-1410
        • Dall M.
        • Hassing A.S.
        • Niu L.
        • Nielsen T.S.
        • Ingerslev L.R.
        • Sulek K.
        • Trammell S.A.J.
        • Gillum M.P.
        • Barrès R.
        • Larsen S.
        • Poulsen S.S.
        • Mann M.
        • Ørskov C.
        • Treebak J.T.
        Hepatocyte-specific perturbation of NAD+ biosynthetic pathways in mice induces reversible nonalcoholic steatohepatitis–like phenotypes.
        J. Biol. Chem. 2021; 297: 101388
        • Storey J.D.
        • Tibshirani R.
        Statistical significance for genomewide studies.
        Proc. Natl. Acad. Sci. U. S. A. 2003; 100: 9440
        • Tyanova S.
        • Temu T.
        • Sinitcyn P.
        • Carlson A.
        • Hein M.Y.
        • Geiger T.
        • Mann M.
        • Cox J.
        The Perseus computational platform for comprehensive analysis of (prote)omics data.
        Nat. Methods. 2016; 13: 731
        • Wisniewski J.R.
        Label-free and standard-free absolute quantitative proteomics using the “total protein” and “proteomic ruler” approaches.
        Methods Enzymol. 2017; 585: 49-60
        • Wiśniewski J.R.
        • Hein M.Y.
        • Cox J.
        • Mann M.
        A “proteomic ruler” for protein copy number and concentration estimation without spike-in standards.
        Mol. Cell. Proteomics. 2014; 13: 3497-3506
        • Tusher V.G.
        • Tibshirani R.
        • Chu G.
        Significance analysis of microarrays applied to the ionizing radiation response.
        Proc. Natl. Acad. Sci. U. S. A. 2001; 98: 5116-5121
        • Bindea G.
        • Mlecnik B.
        • Hackl H.
        • Charoentong P.
        • Tosolini M.
        • Kirilovsky A.
        • Fridman W.-H.
        • Pagès F.
        • Trajanoski Z.
        • Galon J.
        ClueGO: A cytoscape plug-in to decipher functionally grouped gene ontology and pathway annotation networks.
        Bioinformatics. 2009; 25: 1091-1093
        • Cox J.
        • Mann M.
        1D and 2D annotation enrichment: A statistical method integrating quantitative proteomics with complementary high-throughput data.
        BMC Bioinformatics. 2012; 13S12
        • Egami R.
        • Kokaji T.
        • Hatano A.
        • Yugi K.
        • Eto M.
        • Morita K.
        • Ohno S.
        • Fujii M.
        • Hironaka K.I.
        • Uematsu S.
        • Terakawa A.
        • Bai Y.
        • Pan Y.
        • Tsuchiya T.
        • Ozaki H.
        • et al.
        Trans-omic analysis reveals obesity-associated dysregulation of inter-organ metabolic cycles between the liver and skeletal muscle.
        iScience. 2021; 24: 102217
        • Krahmer N.
        • Najafi B.
        • Schueder F.
        • Quagliarini F.
        • Steger M.
        • Seitz S.
        • Kasper R.
        • Salinas F.
        • Cox J.
        • Uhlenhaut N.H.
        • Walther T.C.
        • Jungmann R.
        • Zeigerer A.
        • Borner G.H.H.
        • Mann M.
        Organellar proteomics and phospho-proteomics reveal subcellular reorganization in diet-induced hepatic steatosis.
        Dev. Cell. 2018; 47: 205-221.e7
        • Gonzalez-Franquesa A.
        • Stocks B.
        • Borg M.L.
        • Kuefner M.
        • Dalbram E.
        • Nielsen T.S.
        • Agrawal A.
        • Pankratova S.
        • Chibalin A.V.
        • Karlsson H.K.R.
        • Gheibi S.
        • Björnholm M.
        • Jørgensen N.R.
        • Clemmensen C.
        • Hostrup M.
        • et al.
        Discovery of thymosin β4 as a human exerkine and growth factor.
        Am. J. Physiol. Cell Physiol. 2021; 321: C770-C778
        • Fujiki Y.
        • Okumoto K.
        • Mukai S.
        • Honsho M.
        • Tamura S.
        Peroxisome biogenesis in mammalian cells.
        Front. Physiol. 2014; 5: 307
        • Jones J.M.
        • Morrell J.C.
        • Gould S.J.
        PEX19 is a predominantly cytosolic chaperone and import receptor for class 1 peroxisomal membrane proteins.
        J. Cell Biol. 2004; 164: 57-67
        • Koch J.
        • Brocard C.
        PEX11 proteins attract Mff and human Fis1 to coordinate peroxisomal fission.
        J. Cell Sci. 2012; 125: 3813-3826
        • Walter P.
        • Gilmore R.
        • Blobel G.
        Protein translocation across the endoplasmic reticulum.
        Cell. 1984; 38: 5-8
        • Palade G.
        Intracellular aspects of the process of protein synthesis.
        Science. 1975; 189: 347
        • Horras C.J.
        • Lamb C.L.
        • Mitchell K.A.
        Regulation of hepatocyte fate by interferon-γ.
        Cytokine Growth Factor Rev. 2011; 22: 35-43
        • Volpes R.
        • van den Oord J.J.
        • De Vos R.
        • Depla E.
        • De Ley M.
        • Desmet V.J.
        Expression of interferon-γ receptor in normal and pathological human liver tissue.
        J. Hepatol. 1991; 12: 195-202
        • Hessvik N.P.
        • Llorente A.
        Current knowledge on exosome biogenesis and release.
        Cell. Mol. Life Sci. 2018; 75: 193-208
        • Henderson N.C.
        • Mackinnon A.C.
        • Farnworth S.L.
        • Poirier F.
        • Russo F.P.
        • Iredale J.P.
        • Haslett C.
        • Simpson K.J.
        • Sethi T.
        Galectin-3 regulates myofibroblast activation and hepatic fibrosis.
        Proc. Natl. Acad. Sci. U. S. A. 2006; 103: 5060
        • Diamanti K.
        • Cavalli M.
        • Pereira M.J.
        • Pan G.
        • Castillejo-Lopez C.
        • Kumar C.
        • Mundt F.
        • Komorowski J.
        • Deshmukh A.S.
        • Mann M.
        • Korsgren O.
        • Eriksson J.W.
        • Wadelius C.
        Organ-specific metabolic pathways distinguish prediabetes, type 2 diabetes and normal tissues.
        bioRxiv. 2021; ([preprint])https://doi.org/10.1101/2021.05.09.443296
        • Lundsgaard A.-M.
        • Holm J.B.
        • Sjøberg K.A.
        • Bojsen-Møller K.N.
        • Myrmel L.S.
        • Fjære E.
        • Jensen B.A.H.
        • Nicolaisen T.S.
        • Hingst J.R.
        • Hansen S.L.
        • Doll S.
        • Geyer P.E.
        • Deshmukh A.S.
        • Holst J.J.
        • Madsen L.
        • et al.
        Mechanisms preserving insulin action during high dietary fat intake.
        Cell Metab. 2019; 29: 50-63.e54
        • Wewer Albrechtsen N.J.
        • Geyer P.E.
        • Doll S.
        • Treit P.V.
        • Bojsen-Møller K.N.
        • Martinussen C.
        • Jørgensen N.B.
        • Torekov S.S.
        • Meier F.
        • Niu L.
        • Santos A.
        • Keilhauer E.C.
        • Holst J.J.
        • Madsbad S.
        • Mann M.
        Plasma proteome profiling reveals dynamics of inflammatory and lipid homeostasis markers after roux-en-Y gastric bypass surgery.
        Cell Syst. 2018; 7: 601-612.e3
        • Aranovich A.
        • Hua R.
        • Rutenberg A.D.
        • Kim P.K.
        PEX16 contributes to peroxisome maintenance by constantly trafficking PEX3 via the ER.
        J. Cell Sci. 2014; 127: 3675-3686
        • Li X.
        • Baumgart E.
        • Morrell J.C.
        • Jimenez-Sanchez G.
        • Valle D.
        • Gould S.J.
        PEX11 beta deficiency is lethal and impairs neuronal migration but does not abrogate peroxisome function.
        Mol. Cell. Biol. 2002; 22: 4358-4365
        • Weng H.
        • Ji X.
        • Naito Y.
        • Endo K.
        • Ma X.
        • Takahashi R.
        • Shen C.
        • Hirokawa G.
        • Fukushima Y.
        • Iwai N.
        Pex11α deficiency impairs peroxisome elongation and division and contributes to nonalcoholic fatty liver in mice.
        Am. J. Physiol. Endocrinol. Metab. 2012; 304: E187-E196
        • Abe I.
        • Fujiki Y.
        cDNA cloning and characterization of a constitutively expressed isoform of the human peroxin Pex11p.
        Biochem. Biophys. Res. Commun. 1998; 252: 529-533
        • Abe I.
        • Okumoto K.
        • Tamura S.
        • Fujiki Y.
        Clofibrate-inducible, 28-kDa peroxisomal integral membrane protein is encoded by PEX11.
        FEBS Lett. 1998; 431: 468-472
        • Tanaka A.
        • Okumoto K.
        • Fujiki Y.
        cDNA cloning and characterization of the third isoform of human peroxin Pex11p.
        Biochem. Biophys. Res. Commun. 2003; 300: 819-823
        • Braakman I.
        • Hebert D.N.
        Protein folding in the endoplasmic reticulum.
        Cold Spring Harb. Perspect. Biol. 2013; 5a013201
        • Pagliassotti M.J.
        • Kim P.Y.
        • Estrada A.L.
        • Stewart C.M.
        • Gentile C.L.
        Endoplasmic reticulum stress in obesity and obesity-related disorders: An expanded view.
        Metabolism. 2016; 65: 1238-1246
        • Kammoun H.L.
        • Chabanon H.
        • Hainault I.
        • Luquet S.
        • Magnan C.
        • Koike T.
        • Ferré P.
        • Foufelle F.
        GRP78 expression inhibits insulin and ER stress-induced SREBP-1c activation and reduces hepatic steatosis in mice.
        J. Clin. Invest. 2009; 119: 1201-1215
        • Özcan U.
        • Cao Q.
        • Yilmaz E.
        • Lee A.-H.
        • Iwakoshi N.N.
        • Özdelen E.
        • Tuncman G.
        • Görgün C.
        • Glimcher L.H.
        • Hotamisligil G.S.
        Endoplasmic reticulum stress links obesity, insulin action, and type 2 diabetes.
        Science. 2004; 306: 457
        • Wada T.
        • Miyashita Y.
        • Sasaki M.
        • Aruga Y.
        • Nakamura Y.
        • Ishii Y.
        • Sasahara M.
        • Kanasaki K.
        • Kitada M.
        • Koya D.
        • Shimano H.
        • Tsuneki H.
        • Sasaoka T.
        Eplerenone ameliorates the phenotypes of metabolic syndrome with NASH in liver-specific SREBP-1c Tg mice fed high-fat and high-fructose diet.
        Am. J. Physiol. Endocrinol. Metab. 2013; 305: E1415-E1425
        • Dixon L.J.
        • Flask C.A.
        • Papouchado B.G.
        • Feldstein A.E.
        • Nagy L.E.
        Caspase-1 as a central regulator of high fat diet-induced non-alcoholic steatohepatitis.
        PLoS One. 2013; 8: e56100
        • Williams A.S.
        • Kang L.
        • Zheng J.
        • Grueter C.
        • Bracy D.P.
        • James F.D.
        • Pozzi A.
        • Wasserman D.H.
        Integrin α1-null mice exhibit improved fatty liver when fed a high fat diet despite severe hepatic insulin resistance.
        J. Biol. Chem. 2015; 290: 6546-6557
        • Sutter A.G.
        • Palanisamy A.P.
        • Lench J.H.
        • Jessmore A.P.
        • Chavin K.D.
        Development of steatohepatitis in ob/ob mice is dependent on Toll-like receptor 4.
        Ann. Hepatol. 2015; 14: 735-743
        • Martinez-Huenchullan S.F.
        • Shipsey I.
        • Hatchwell L.
        • Min D.
        • Twigg S.M.
        • Larance M.
        Blockade of high-fat diet proteomic phenotypes using exercise as prevention or treatment.
        Mol. Cell. Proteomics. 2021; 20: 100027
        • Klop B.
        • Elte J.W.F.
        • Cabezas M.C.
        Dyslipidemia in obesity: Mechanisms and potential targets.
        Nutrients. 2013; 5: 1218-1240
        • Dwivedi V.
        • Yaniv K.
        • Sharon M.
        Beyond cells: The extracellular circulating 20S proteasomes.
        Biochim. Biophys. Acta Mol. Basis Dis. 2021; 1867: 166041
        • Zoeger A.
        • Blau M.
        • Egerer K.
        • Feist E.
        • Dahlmann B.
        Circulating proteasomes are functional and have a subtype pattern distinct from 20S proteasomes in major blood cells.
        Clin. Chem. 2006; 52: 2079-2086
        • Wyatt A.R.
        • Yerbury J.J.
        • Poon S.
        • Wilson M.R.
        Therapeutic targets in extracellular protein deposition diseases.
        Curr. Med. Chem. 2009; 16: 2855-2866
        • Lee K.Y.
        • Chen T.T.
        • Chiang L.L.
        • Chuang H.C.
        • Feng P.H.
        • Liu W.T.
        • Chen K.Y.
        • Ho S.C.
        Proteasome activity related with the daily physical activity of COPD patients.
        Int. J. Chron. Obstruct. Pulmon. Dis. 2017; 12: 1519-1525
        • Kakurina G.V.
        • Cheremisina O.V.
        • Choinzonov E.L.
        • Kondakova I.V.
        Circulating proteasomes in the pathogenesis of head and neck squamous cell carcinoma.
        Bull. Exp. Biol. Med. 2017; 163: 92-94
        • Carr M.E.
        Diabetes mellitus: A hypercoagulable state.
        J. Diabetes Complications. 2001; 15: 44-54
        • Kaye S.M.
        • Pietiläinen K.H.
        • Kotronen A.
        • Joutsi-Korhonen L.
        • Kaprio J.
        • Yki-Järvinen H.
        • Silveira A.
        • Hamsten A.
        • Lassila R.
        • Rissanen A.
        Obesity-related derangements of coagulation and fibrinolysis: A study of obesity-discordant monozygotic twin pairs.
        Obesity. 2012; 20: 88-94
        • De Pergola G.
        • De Mitrio V.
        • Giorgino F.
        • Sciaraffia M.
        • Minenna A.
        • Di Bari L.
        • Pannacciulli N.
        • Giorgino R.
        Increase in both pro-thrombotic and anti-thrombotic factors in obese premenopausal women: Relationship with body fat distribution.
        Int. J. Obes. Relat. Metab. Disord. 1997; 21: 527-535
        • Henry M.L.
        • Davidson L.B.
        • Wilson J.E.
        • McKenna B.K.
        • Scott S.A.
        • McDonagh P.F.
        • Ritter L.S.
        Whole blood aggregation and coagulation in db/db and ob/ob mouse models of type 2 diabetes.
        Blood Coagul. Fibrinolysis. 2008; 19: 124-134
        • Uhlén M.
        • Fagerberg L.
        • Hallström B.M.
        • Lindskog C.
        • Oksvold P.
        • Mardinoglu A.
        • Sivertsson Å.
        • Kampf C.
        • Sjöstedt E.
        • Asplund A.
        • Olsson I.
        • Edlund K.
        • Lundberg E.
        • Navani S.
        • Szigyarto C.A.
        • et al.
        Proteomics. Tissue-based map of the human proteome.
        Science. 2015; 347: 1260419
        • Menon R.P.
        • Hughes R.C.
        Determinants in the N-terminal domains of galectin-3 for secretion by a novel pathway circumventing the endoplasmic reticulum-Golgi complex.
        Eur. J. Biochem. 1999; 264: 569-576
        • van Kimmenade R.R.
        • Januzzi Jr., J.L.
        • Ellinor P.T.
        • Sharma U.C.
        • Bakker J.A.
        • Low A.F.
        • Martinez A.
        • Crijns H.J.
        • MacRae C.A.
        • Menheere P.P.
        • Pinto Y.M.
        Utility of amino-terminal pro-brain natriuretic peptide, galectin-3, and apelin for the evaluation of patients with acute heart failure.
        J. Am. Coll. Cardiol. 2006; 48: 1217-1224
        • van der Velde A.R.
        • Gullestad L.
        • Ueland T.
        • Aukrust P.
        • Guo Y.
        • Adourian A.
        • Muntendam P.
        • van Veldhuisen D.J.
        • de Boer R.A.
        Prognostic value of changes in galectin-3 levels over time in patients with heart failure: Data from CORONA and COACH.
        Circ. Heart Fail. 2013; 6: 219-226
        • Weigert J.
        • Neumeier M.
        • Wanninger J.
        • Bauer S.
        • Farkas S.
        • Scherer M.N.
        • Schnitzbauer A.
        • Schäffler A.
        • Aslanidis C.
        • Schölmerich J.
        • Buechler C.
        Serum galectin-3 is elevated in obesity and negatively correlates with glycosylated hemoglobin in type 2 diabetes.
        J. Clin. Endocrinol. Metab. 2010; 95: 1404-1411
        • Al Attar A.
        • Antaramian A.
        • Noureddin M.
        Review of galectin-3 inhibitors in the treatment of nonalcoholic steatohepatitis.
        Expert Rev. Clin. Pharmacol. 2021; 14: 457-464
        • Jeftic I.
        • Jovicic N.
        • Pantic J.
        • Arsenijevic N.
        • Lukic M.L.
        • Pejnovic N.
        Galectin-3 ablation enhances liver steatosis, but attenuates inflammation and IL-33-dependent fibrosis in obesogenic mouse model of nonalcoholic steatohepatitis.
        Mol. Med. 2015; 21: 453-465
        • Yu L.
        • Ruifrok Willem P.T.
        • Meissner M.
        • Bos Eelke M.
        • van Goor H.
        • Sanjabi B.
        • van der Harst P.
        • Pitt B.
        • Goldstein Irwin J.
        • Koerts Jasper A.
        • van Veldhuisen Dirk J.
        • Bank Ruud A.
        • van Gilst Wiek H.
        • Silljé Herman H.W.
        • de Boer Rudolf A.
        Genetic and pharmacological inhibition of galectin-3 prevents cardiac remodeling by interfering with myocardial fibrogenesis.
        Circ. Heart Fail. 2013; 6: 107-117
        • Dumic J.
        • Dabelic S.
        • Flögel M.
        Galectin-3: An open-ended story.
        Biochim. Biophys. Acta. 2006; 1760: 616-635
        • Lin Y.H.
        • Lin L.Y.
        • Wu Y.W.
        • Chien K.L.
        • Lee C.M.
        • Hsu R.B.
        • Chao C.L.
        • Wang S.S.
        • Hsein Y.C.
        • Liao L.C.
        • Ho Y.L.
        • Chen M.F.
        The relationship between serum galectin-3 and serum markers of cardiac extracellular matrix turnover in heart failure patients.
        Clin. Chim. Acta. 2009; 409: 96-99
        • Papaspyridonos M.
        • McNeill E.
        • de Bono Joe P.
        • Smith A.
        • Burnand Kevin G.
        • Channon Keith M.
        • Greaves David R.
        Galectin-3 is an amplifier of inflammation in atherosclerotic plaque progression through macrophage activation and monocyte chemoattraction.
        Arterioscler. Thromb. Vasc. Biol. 2008; 28: 433-440
        • Considine R.V.
        • Sinha M.K.
        • Heiman M.L.
        • Kriauciunas A.
        • Stephens T.W.
        • Nyce M.R.
        • Ohannesian J.P.
        • Marco C.C.
        • McKee L.J.
        • Bauer T.L.
        • Caro J.F.
        Serum immunoreactive-leptin concentrations in normal-weight and obese humans.
        N. Engl. J. Med. 1996; 334: 292-295
        • Biddinger S.B.
        • Miyazaki M.
        • Boucher J.
        • Ntambi J.M.
        • Kahn C.R.
        Leptin suppresses stearoyl-CoA desaturase 1 by mechanisms independent of insulin and sterol regulatory element-binding protein-1c.
        Diabetes. 2006; 55: 2032-2041
        • Naylor C.
        • Petri Jr., W.A.
        Leptin regulation of immune responses.
        Trends Mol. Med. 2016; 22: 88-98
        • Montague C.T.
        • Farooqi I.S.
        • Whitehead J.P.
        • Soos M.A.
        • Rau H.
        • Wareham N.J.
        • Sewter C.P.
        • Digby J.E.
        • Mohammed S.N.
        • Hurst J.A.
        • Cheetham C.H.
        • Earley A.R.
        • Barnett A.H.
        • Prins J.B.
        • O'Rahilly S.
        Congenital leptin deficiency is associated with severe early-onset obesity in humans.
        Nature. 1997; 387: 903-908
        • Perez-Riverol Y.
        • Csordas A.
        • Bai J.
        • Bernal-Llinares M.
        • Hewapathirana S.
        • Kundu D.J.
        • Inuganti A.
        • Griss J.
        • Mayer G.
        • Eisenacher M.
        • Perez E.
        • Uszkoreit J.
        • Pfeuffer J.
        • Sachsenberg T.
        • Yilmaz S.
        • et al.
        The PRIDE database and related tools and resources in 2019: Improving support for quantification data.
        Nucleic Acids Res. 2019; 47: D442-D450
        • Geyer P.E.
        • Wewer Albrechtsen N.J.
        • Tyanova S.
        • Grassl N.
        • Iepsen E.W.
        • Lundgren J.
        • Madsbad S.
        • Holst J.J.
        • Torekov S.S.
        • Mann M.
        Proteomics reveals the effects of sustained weight loss on the human plasma proteome.
        Mol. Syst. Biol. 2016; 12: 901