If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Copenhagen, DenmarkDepartment of Physiology and Pharmacology, Karolinska Institutet, Stockholm, SwedenDepartment of Molecular Medicine and Surgery, Karolinska Institutet, Stockholm, Sweden
Novo Nordisk Foundation Center for Basic Metabolic Research, University of Copenhagen, Copenhagen, DenmarkNovo Nordisk Foundation Center for Protein Research, University of Copenhagen, Copenhagen, Denmark
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.
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) (
). Furthermore, dysregulated hepatic triglyceride metabolism contributes to insulin resistance and the development of type 2 diabetes alongside obesity (
). 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 (
) 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 (
). 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 (
). 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 (
). 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 (
) 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 (
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) (
). 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 (
) 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 (
) (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 (
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) (
). 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 (
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 (
) 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 (
); 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 (
). 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 (
) in significantly regulated proteins compared with the whole proteome. Additional enrichment analyses were performed using a one-dimensional enrichment analysis (
), 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 (
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 (
) 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 (
). 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.
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.
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.
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 (
) 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.
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.
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 (
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 (
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, A–F). We identified 7684 proteins from two fractions using the MED-FASP protocol (
) (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 (
). 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 (
). 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 (
). Peroxisomal biogenesis factor (PEX) 19, which binds newly synthesized peroxisomal membrane proteins and shuttles them to the membrane protein import machinery (
) 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 (
), 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 (
), 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) (
). 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 (
). 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 (
). 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 (
) 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 (
). 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 (
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 (
). 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 (
). 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 (
). 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).
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 (
). 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.
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 (
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 (
). 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 (
). 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 (
). 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 (
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 (
). 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 (
). 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 (
). 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 (
), 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 (
). 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 (
). Endoplasmic reticulum stress is a common feature of obesity and NAFLD; however, this is typically characterized by an upregulation of chaperone genes (
). However, consistent with a reduction in protein content the mRNA expression of endoplasmic reticulum chaperon complex genes typically decreases in ob/ob mice (
). 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 (
). 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 (
). 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 (
). 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 (
), 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 (
). Enriched within the upregulated proteins were GOBP terms relating to lipid and cholesterol homeostasis, which is in line with obesity-induced dyslipidemia (
). 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 (
). 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 (
), 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 (
). 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 (
), 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 (
), 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 (
Galectin-3 ablation enhances liver steatosis, but attenuates inflammation and IL-33-dependent fibrosis in obesogenic mouse model of nonalcoholic steatohepatitis.
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 (
). 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 (
). 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 (
). Furthermore, individuals with mutations in the human leptin gene have been identified, which is also associated with severe early onset of obesity (
). 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 (
) (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 (
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.
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.
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).
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.
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).
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).
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).
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. (
Proteomics analysis of skeletal muscle from leptin-deficient ob/ob mice reveals adaptive remodeling of metabolic characteristics and fiber type composition.
Extending the limits of quantitative proteome profiling with data-independent acquisition and application to acetaminophen-treated three-dimensional liver microtissues.
Galectin-3 ablation enhances liver steatosis, but attenuates inflammation and IL-33-dependent fibrosis in obesogenic mouse model of nonalcoholic steatohepatitis.
00015-9&id=fx1.jpg){kind=link}
00015-9&id=gr1.jpg){kind=link}
00015-9&id=gr2.jpg){kind=link}
00015-9&id=gr3.jpg){kind=link}
00015-9&id=gr4.jpg){kind=link}
00015-9&id=gr5.jpg){kind=link}
00015-9&id=gr6.jpg){kind=link}
00015-9&id=gr7.jpg){kind=link}