Characterization of depot-specific adipokine release in diet-induced obese mice reveals a distinct secretion profile in paracardial adipose tissue

Abstract

Obesity is characterized by excessive adipose tissue deposition across multiple depots, accompanied by profound alterations in systemic metabolism. Beyond their role in lipid storage, adipose tissues are active endocrine organs that secrete a range of bioactive proteins, i.e., adipokines, that regulate energy homeostasis, inflammation, and metabolic health. Alterations in adipokine secretion are increasingly recognized as key contributors to the development of obesity-related comorbidities. Here, we investigate depot-specific patterns of adipokine release from paracardial (PAT), inguinal (ING), and epididymal (EPI) adipose tissues in healthy lean mice and mice fed a high-fat diet. Using untargeted proteomics, we identify distinct secretory profiles across depots, revealing marked heterogeneity in the composition and magnitude of secreted proteins. In healthy lean mice, PAT exhibited the highest overall protein secretion rates, with enrichment in proteins linked to cellular and oxidative stress responses. High fat feeding induced extensive remodelling of the adipose secretome across all depots. EPI adipose tissue displayed increased secretion of proteins involved in cellular stress, injury and inflammation, whereas both ING and PAT depots demonstrated substantial suppression of protein release, including reductions in proteins associated with carbohydrate homeostasis and insulin signalling. These findings highlight that obesity drives depot-specific adaptations in adipokine secretion, potentially contributing to regional differences in metabolic risk. Together, this study provides an in-depth proteomic characterization of adipose tissue secretory remodeling in diet-induced obesity, offering new insights into how distinct fat depots influence systemic metabolic regulation.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1038/s41598-026-64158-8.

Introduction

The maintenance of systemic metabolic homeostasis depends on the integrated endocrine capacities of multiple tissues, each of which responds to neural and circulating signals to produce and secrete peptides and hormones that act locally and systemically. Classically, the endocrine glands, including the pancreas, adrenal, thyroid and pituitary, have been the focus of hormone biology^1–3. However, over the past three decades, major metabolic organs and tissues, including liver, skeletal muscle, gut and adipose tissue, have been recognized as important endocrine regulators. These tissues form a network of endocrine crosstalk to regulate a myriad of processes, including energy metabolism, inflammation, and nutrient partitioning that is altered in conditions of overnutrition and obesity.

Adipose tissue occupies a central role as an active, versatile endocrine organ⁴ that secretes proteins into the circulation which regulate food intake⁵, local/systemic inflammation, and lipid and glucose metabolism in liver and skeletal muscle⁶. Both insufficiency of adipose tissue (i.e., lipodystrophies) and adipose tissue excess (i.e., obesity) can lead to metabolic dysfunction^{7, 8}, highlighting that the endocrine roles of adipose tissue are indispensable for systemic metabolic health.

In obesity, adipose tissue undergoes marked morphological and functional adaptations, including adipocyte hypertrophy, immune cell infiltration, as well as alterations in vascularization and extracellular matrix remodeling. Together, these changes contribute to adipose tissue dysfunction and systemic impairments in metabolic health^9–11. In addition, adipose tissue dysfunction in the setting of obesity is closely associated with changes in protein secretion⁴, which can exert autocrine, paracrine and endocrine actions to exacerbate metabolic pathologies. Importantly, adipose depots are not homogeneous. Visceral adipose tissue (VAT) is often more strongly associated with metabolic risk than subcutaneous adipose tissue (SAT)¹². Adipose depots differ in adipocyte size, vascularization and innervation, immune cell content, metabolic activity, and basal secretory profile^12,13. Regional differences in adipokine expression and secretion partly determine the capacity of specific adipose depots in the development of insulin resistance, dyslipidemia, and cardiovascular disease^14,15. For example, VAT tends to secrete more pro-inflammatory cytokines, has a higher basal lipolytic rate, and contributes more free fatty acids to the portal circulation, thereby influencing hepatic metabolism^12,14. In contrast, SAT tends to be more insulin sensitive and less inflammatory^12,16 Adipose tissues are localized to many discrete regions of the body, often in close contact with other tissues/organs and sharing direct portal communication, thereby allowing for direct, high-concentration transport of substances from adipose tissues to the respective organ.

We and others have recently shown that pericardial (epicardial and paracardial) adipose tissues exhibit unique properties that distinguish them from other adipose depots, having important implications for cardiovascular health and disease^17–19. Epicardial adipose tissue shares an unobstructed microcirculation with the underlying myocardium and has been shown to physically infiltrate into the epicardium, allowing for direct paracrine communication²⁰. In contrast, paracardial adipose tissue sits adjacent to the pericardium and has the capacity to exert both local cardiac and systemic cardiovascular performance^17,21.

Despite growing research in this field, it remains unclear how different adipose depots contribute differentially to the circulating adipokine pool, especially in the setting of obesity, and how protein secretion is remodelled in a depot-specific manner in response to overnutrition. In this study, we aimed to increase our understanding of how obesity reprograms adipose endocrine function in a location-dependent manner. Using untargeted label-free proteomics, we assessed and compared adipokine secretion from three distinct adipose tissue depots (epididymal, inguinal, paracardial) in lean mice and mice rendered obese by high-fat feeding, and show depot-specific hierarchy of adipokine secretion that is dynamically reshaped by nutritional state.

Materials and methods

Animals

All experimental procedures were approved by the University of Melbourne Anatomy and Neuroscience, Pathology, Pharmacology, and Physiology Animal Ethics Committee (ethics ID 1814403), and conformed to the National Health and Medical Research Council of Australia guidelines regarding the care and use of experimental animals. Male 8–10 week old C57BL/6J mice were obtained from the Animal Resources Centre (Canning Vale, Australia), and group-housed in a temperature-controlled room (22 °C ± 1 °C, 40–60% humidity) with a 12-hour light-dark cycle. Mice had ad libitum access to food and water. Mice were fed either a chow control diet (12% kcal from fat; SF13-081, Specialty Feeds, Australia) or a diet enriched in fat (207 g/kg diet) and sucrose (201 g/kg diet; SF04–001; Specialty Feeds, Australia) for a total of 12 weeks. Following the feeding regime, mice were anaesthetised with an intraperitoneal injection of pentobarbitone sodium (50 mg/kg). This was followed by cervical dislocation to assure death of the animals, and adipose tissues were carefully dissected.

Collection of conditioned media containing secreted factors

Epididymal, inguinal and paracardial adipose tissues were excised and adipose tissue explants (~ 20–30 mg) incubated in oxygenated (95% O₂, 5% CO₂) EXCELL^® 325 PF CHO Serum-Free medium (Sigma-Aldrich, Castle Hill, Australia) in a shaking water bath at 37 °C for 4 h. Previous studies have shown that tissues remain stable for over 12 h under similar conditions^18,22,23. The conditioned media containing all secreted factors was collected, centrifuged at 2000 × g for 10 min and the supernatant frozen and stored at −80 °C for subsequent proteomics analysis.

Proteomics processing

Conditioned media was processed, as described in detail²⁴. The media was concentrated using Amicon Ultra-4 Centrifugal Filters, total protein content determined by Bradford Assay (Sigma Aldrich), disulfide bonds reduced through the addition of 10 mM TCEP (Tris(2-carboxyethyl)phosphine) at 65 °C for 20 min and the sample mixed with 6.6× volumes of 5 M urea. Samples were added to Microcon Centrifugal Filters (Sigma-Aldrich), spun at 14,000 × g and 10 °C for 15 min, washed with urea, and chloroacetamide was added to a final concentration of 10 mM to alkylate proteins. Samples were incubated for 20 min in the dark, and washed 3× with 5 M urea and 3× with 50 mM ammonia bicarbonate (pH 8.5). After the last wash step, a digestion solution was added to the spin filter (1 µg trypsin/100 µg protein in 75 µL of 50 mM ammonia bicarbonate) and samples were incubated at 37 °C overnight. The following day, peptides were eluted by addition of 2 × 40 µL of 50 mM ammonia bicarbonate, and 1 × 40 µL 0.5 M NaCl. Each step was followed by centrifugation at 14,000 × g and 10 °C for 15 min. Eluted peptides were acidified with formic acid to a final pH of 3.0, and purified with Millipore^® C18 Ziptips (Sigma-Aldrich). Subsequently, samples were dried in a SpeedVac concentrator (Eppendorf Concentrator Plus), resuspended in 20 µL 0.1% formic acid, sonicated for 10 min at room temperature, centrifuged at 16,000 × g for 10 min, and transferred into HPLC vials for analysis.

Mass spectrometry proteomics analysis

All samples/peptides were analyzed by LC-MS/MS using a Q Exactive Plus mass spectrometer (Thermo Fisher Scientific) coupled online to a RSLC nano HPLC (Ultimate 3000, UHPLC Thermo Fisher Scientific). Samples were loaded onto a 100 μm, 2-cm nanoviper Pepmap100 trap column, eluted and separated on a RSLC nano column 75 μm × 50 cm, Pepmap100 C18 analytical column (Thermo Fisher Scientific). The detailed gradients used for protein identification have been previously published²⁴. The eluent was nebulized and ionized using a nanoelectrospray source (Thermo Fisher Scientific) with a distal coated fused silica emitter (New Objective). The capillary voltage was set at 1.7 kV. The Q Exactive mass spectrometer was operated in the data-dependent acquisition mode to automatically switch between full MS scans and subsequent MS/MS acquisitions. Survey full-scan MS spectra (m/z 375–1575) were acquired in the Orbitrap with 70,000 resolution (at m/z 200) after accumulation of ions to a 3 × 10⁶ target value with a maximum injection time of 54 ms. Dynamic exclusion was set to 15 s. The 12 most intense multiply charged ions (z ≥ 2) were sequentially isolated and fragmented in the collision cell by higher-energy collisional dissociation (HCD) with a fixed injection time of 54 ms, 17,500 resolution, and automatic gain control (AGC) target of 2 × 10⁵.

Proteomics data analysis and statistics

Data were analyzed with MaxQuant (version 1.5.3.30)²⁵ and its implemented Andromeda search engine to obtain protein identifications as well as their label-free quantitation (LFQ) intensities. Database searching was performed, as described²⁴. The MaxQuant result output was further processed with Perseus (Version 1.5.0.40)²⁶, a module from the MaxQuant suite. After removing reversed and known contaminant proteins, the LFQ values were log2 transformed and the reproducibility across the biological replicates was evaluated by a Pearson’s correlation analysis. The replicates were grouped accordingly, and all proteins were removed that had less than 70% “valid value” in at least one of the groups. The missing values were replaced by imputation and a two-sample t test (FDR < 5%) was performed to obtain a list of significantly regulated proteins. All proteomics data are provided in Supplementary Data.

Proteomics data were further analysed using Qiagen Ingenuity Pathway Analysis (IPA) (QIAGEN Inc., https://digitalinsights.qiagen.com/IPA), including canonical pathway analysis (comprehensive activation prediction using human-curated signalling and metabolic pathways), analysis of diseases, functions and toxicity (identification of perturbed biological functions, diseases and toxicological functions) and upstream regulator analysis, which allows for the identification of transcription factors and kinases (among others) that may regulate the expression profiles observed in the proteomics dataset.

For assessment of total protein secretion, data are presented as mean ± standard error of the mean (SEM). A parametric two-tailed unpaired student t-tests or a one-way ANOVA followed by a Bonferroni post-hoc analysis was used, as appropriate. Statistical significance was determined as P < 0.05.

Results

Paracardial adipose tissue exhibits a distinct adipokine signature in lean mice

We initially compared protein secretion across the three adipose tissue depots in lean mice. Total protein secretion was the greatest in paracardial adipose tissue (PAT) and 3-fold lower in epididymal adipose tissue (EPI), with inguinal adipose tissue (ING) showing an intermediate phenotype (Fig. 1A). Untargeted proteomics of the secreted proteins identified a total of 722 proteins across all groups, with principal component analysis (PCA) revealing distinct clustering of the proteomes by depot type (Fig. 1B). In each of the groups, ~ 25% of the secreted proteins contained an N-terminal signal peptide and were therefore most likely classically secreted (Fig. 1C). The low percentage of classically secreted proteins is common across the majority of metabolic tissues, including liver/hepatocytes (~ 20–31% ^24–28), skeletal muscle (10–17%^29–31 and adipose tissue (21–39% ^29,32,33. Of the 722 identified proteins, 403 proteins (55.8%) were present in the conditioned media of each of the three depots, while 137 proteins (19%) were secreted only from PAT (Fig. 1D) – far greater than the relatively few proteins secreted only by EPI (11 proteins) or ING (2 proteins). Common adipose-secreted factors found in conditioned media of all adipose depots included adiponectin, retinol-binding protein 4 (Rbp4), extracellular matrix/space constituents lumican (Lum), mimecan (Ogn) and vitronectin (Vtn), the extracellular chaperone Clusterin (Clu), as well as nidogen-1 and nidogen-2, sulfated glycoproteins that enhance glucose metabolism³⁴ and play a role in extracellular matrix (ECM) deposition in adipose tissue³⁵(Fig. 1E). Other factors involved in adipose tissue ECM remodelling included several laminins (Lama4, Lamb1 and Lamc1)³⁵, while we further show secretion of various cathepsins (Ctsb, Ctsd and Ctss) and pro-saposin (Psap), that regulate tissue-specific and systemic insulin sensitivity^36,37, inflammation³⁸ and/or autophagy in adipose tissue³⁹. Lastly, conditioned media of all adipose depots included various protein disulfide-isomerases (Pdia3, Pdia4, Pdia6), which function as protein folding chaperones (Fig. 1E; of note, leptin and resistin were not identified in this analysis). The proteins secreted uniquely by PAT were associated with several processes, including ‘amino acid metabolism’ and ‘cellular stress response’ (Fig. 1F). In addition, the most abundant proteins secreted only by PAT included Sec14L3, a phophatidylinositol transfer protein essential for the biogenesis of Golgi-derived transport vesicles⁴⁰, SCGB1A1 (Secretoglobin Family 1 A Member 1), implicated in inflammation and immune response⁴¹, as well as TPD52L2, a member of the tumor protein D52-like family, that may play a role in tumor development and immune cell infiltration^42,43.

Fig. 1.

Comparison of adipokine secretion across three adipose tissue depots in chow-fed, lean mice. (A) Rates of protein secretion in epididymal (EPI), inguinal (ING) and paracardial (PAT) adipose tissue. (B) Principal component analysis (PCA) and (C) percentage of proteins containing a N-terminal signal peptide in the secreted proteome between EPI, ING and PAT. (D) Venn diagram highlighting commonly and uniquely secreted proteins across the three adipose depots. (E) Protein secretion of common adipokines across all three depots and (F) Qiagen IPA of the pathways associated with proteins secreted uniquely from PAT. Volcanoplot analysis showing significantly regulated proteins and corresponding IPA comparing (G-H) PAT and EPI, (I-J) PAT and ING, and (L-M) EPI and ING. (K) Heatmap analysis of components of the 20S core proteasome complex across all three adipose depots. Data is presented as mean ± SEM, *P < 0.05, with n = 8 mice/group. Significance was assessed using one-way ANOVA and Bonferroni post-hoc analysis.

In addition to the unique depot-specific secretion profiles, volcanoplot analysis further identified differential abundance of hundreds of proteins across groups (Fig. 1G, I, L). When compared to the other two depots, EPI showed the highest secretion of adiponectin, FAPB4 (fatty acid-binding protein 4), CA3 (carbonic anhydrase 3) and PTRF/CAVIN1 (caveolae associated protein 1), essential for caveolae formation and organization. In contrast, ING showed the highest secretion of the serine/cysteine peptidase inhibitor Serpina3K, the actin-modulating protein gelsolin and the proteoglycan lumican, which plays an important role in the extracellular matrix (ECM), including collagen organization. Lastly, PAT exhibited the highest secretion of fatty acid synthase, a key regulator of de novo lipogenesis, as well as alpha-enolase, malate dehydrogenase and pyruvate kinase, all crucial enzymes involved in glycolysis and the TCA cycle.

Comparing PAT with EPI, Qiagen Ingenuity Pathway Analysis (IPA) analysis highlighted increased secretion of proteins involved in processes including ‘IL1 signalling’, ‘NF-kB signalling’ and the ‘KEAP1-NFE2L2 antioxidant pathway’, pathways crucial in the regulation of cellular immune response, inflammation and oxidative stress. Proteins reduced in secretion from PAT (i.e., increased in EPI) were associated with ‘LXR/RXR activation’ and ‘DHCR24 signalling’ (Fig. 1H). In contrast, comparison of PAT with ING highlighted increased release of proteins involved in ‘RAR activation’, ‘mitosis’ and ‘neutrophil degranulation’, and lower secretion of proteins associated with ‘mTOR signalling’, ‘Hippo signalling’ and ‘JAK-STAT signalling’ (Fig. 1J), pathways important in regulating cell growth, metabolism, and proliferation. Furthermore, we found that PAT showed a pronounced release of multiple components of the 20S core proteasome complex (Fig. 1K), which plays a crucial role in the proteolytic degradation of most intracellular proteins. These proteins are likely to be released within extracellular vesicles and may influence protein degradation in other tissues, although their endocrine role is still unclear⁴⁴.

Finally, a comparison of ING and EPI conditioned media pointed to enrichment of proteins associated with ‘RAF/MAP kinase signalling’ and ‘LXR/RXR activation’ in ING (i.e., reduced in EPI), important pathways in cancer development, lipid metabolism and cholesterol homeostasis (Fig. 1M). These included proteins such as hydroxyacyl-coenzyme A dehydrogenase (HADH), which is essential for mitochondrial fatty acid oxidation, lipoprotein lipase (LPL), an important enzyme in triglyceride hydrolysis, and retinol binding protein 4 (RBP4), a crucial transporter protein for retinol. On the other hand, proteins associated with ‘NOTCH4 signalling’, ‘signalling through ROBO receptors’ and ‘hypoxia response’ were enriched in EPI when compared to ING (Fig. 1M). Given that these pathways are associated with cancer progression^45–47, adipokine release from EPI is a potential contributor to various cancer pathologies. Overall, these data highlight depot-specific differences in adipokine secretion, that are likely to contribute to regional and/or systemic differences in (patho)physiology.

PAT and EPI exhibit similarities in adipokine secretion in the setting of diet-induced obesity

A similar comparison of protein secretion across the three adipose tissue depots was performed in tissues from obese mice. Similar to our observations in lean mice, total protein secretion was the greatest in PAT being 2-fold greater than in EPI and ING, which exhibited similar rates of protein secretion (Fig. 2A). Untargeted proteomics analysis identified a total of 739 secreted proteins in the conditioned medium, with the PCA revealing distinct clustering of ING, but not PAT or EPI (Fig. 2B). Similar to the findings in lean mice, 19–29% of the identified proteins were predicted to be classically secreted (Fig. 2C). Of the 739 proteins that were identified across all groups, 460 proteins (62.2%) were present in the conditioned media of each of the three depots (Fig. 2D). Surprisingly, only 1.9% of proteins (14 proteins) were uniquely secreted from PAT (compared to the 137 proteins uniquely secreted in lean mice) and only 2 proteins were uniquely secreted from ING. In ING, these included fibronectin type III domain containing 1 (Fndc1), a close family member of FNDC5/irisin shown to promote muscle regeneration⁴⁸, which to our knowledge has not been characterized in adipose tissue, and tetranectin/Clec3b, a previously described adipokine that has been shown to be negatively regulated by SREBP-1c and FoxO1⁴⁹.

Fig. 2.

Comparison of adipokine secretion across three adipose tissue depots in HFD-fed mice. (A) Rates of protein secretion in epididymal (EPI), inguinal (ING) and paracardial (PAT) adipose tissue. (B) Principal component analysis (PCA) and (C) percentage of proteins containing a N-terminal signal peptide in the secreted proteome between EPI, ING and PAT. (D) Venn diagram highlighting commonly and uniquely secreted proteins across the three adipose depots, and (E) pathway analysis of proteins uniquely secreted from EPI. (F) Protein secretion of common adipokines across all three depots. (G-H) Volcanoplot analysis showing significantly regulated proteins and corresponding IPA comparing PAT and EPI. (I-J) Heat maps highlighting greater abundance of proteins related to glucose metabolism, fatty acid metabolism and TCA cycle in PAT compared with EPI and ING. (K-L) Analysis of upstream regulators and associated disease & biological functions in PAT compared to EPI. Volcano plot analysis showing significantly regulated proteins and corresponding IPA comparing (M-N) PAT and ING and (O-P) EPI and ING. Data is presented as mean ± SEM, *P < 0.05, with n = 8 mice/group. Significance was assessed using one-way ANOVA and Bonferroni post-hoc analysis.

Among others, PAT-specific proteins included Thy-1/CD90, which plays a role in the fate of mesenchymal stem cells (MSC) between osteogenic and adipogenic differentiation, promoting bone formation and protecting against obesity⁵⁰, as well as CBR2/AP27, which also suppresses adipocyte differentiation⁵¹, overall suggesting that PAT (or MSC within PAT) might be characterized by a less pro-adipogenic signature. We further observed PAT-specific secretion of dipeptidyl peptidase 3 (DPP3), which regulates blood pressure by interacting with the renin-angiotensin system, and further participates in pain signaling, inflammation, and oxidative stress⁵², glycerol-3-phosphate phosphatase (PGP) which hydrolyzes glycerol-3-phosphate into glycerol, and the Sec14-like phosphatidylinositol transfer protein Sec14l3, which may play a tumour suppressive role⁵³, although its role in adipose tissue is not well defined. Lastly, EPI exhibited the largest number of uniquely secreted proteins (61 proteins; 8% of all identified proteins; Fig. 2D). These EPI-specific proteins were associated with ‘membrane trafficking’ and ‘vesicle-mediated transport’, among other functions (Fig. 2E), and included Copb2 (coatomer subunit beta), essential for Golgi budding and vesicular trafficking, CopS4, and important component of the COP9 signalosome complex, which regulates the ubiquitin conjugation pathway, and the GTPases Rab1b, Rab5c, Rab10 and Rac1, key regulators of intracellular membrane trafficking and secretory processes. As in lean mice (Fig. 1E), we observed abundant secretion of classical adipokines that varied between adipose depots (Fig. 2F).

Direct comparison of the secreted proteins from PAT and EPI of obese mice identified only very minor depot-specific differences (Figs. 2G and 6% of all proteins differently expressed) compared to the findings in lean mice (Figs. 1G and 55% of all proteins differentially expressed). In PAT, IPA and heatmap analysis pointed to enrichment of secreted proteins associated with metabolic pathways, including the ‘TCA cycle’, ‘glucose metabolism’ and ‘fatty acid oxidation’ (Fig. 2H-J). In contrast, proteins associated with oxidative phosphorylation/electron transport were not actively secreted from either of the adipose depots. Furthermore, upstream regulator analysis, which allows for the identification of transcription factors and kinases (among others) that may regulate the expression profiles (or in this case, secretion profiles) observed in the proteomics dataset, predicted activation of peroxisome proliferator-activated receptor γ (PPARG; lipid sensor transcription factor) and adipose triglyceride lipase (PNPLA; crucial to initial steps of lipolysis) in PAT when compared to EPI (Fig. 2K). Supporting these observations, additional IPA analysis of biological functions suggested that proteins secreted from PAT were associated with ‘ATP synthesis and metabolism’ (Fig. 2L). Comparison of PAT and ING in obese mice showed a similar pattern to lean mice (Fig. 1I), with many proteins showing greater secretion from PAT (Figs. 2M and 52% of proteins differentially expressed). These enriched proteins were associated with ‘triglyceride metabolism’ and ‘inositol metabolism’ (Fig. 2N). Lastly, comparison of ING and EPI conditioned media identified a significantly increased abundance of > 100 proteins in EPI (Figs. 2O and 51% of proteins differentially expressed), with these proteins being associated with the ‘TCA cycle’, the ‘KEAP1-NFE2L2 pathway’ and ‘glutaryl-CoA degradation’ (Fig. 2P), indicating enrichment in proteins involved in oxidative metabolism, oxidative stress and amino acid degradation.

Obese mice show reduced protein secretion from PAT – changes that are associated with suppression of ChREBP and oxidative glucose metabolism

Given the distinct secretion profiles in EPI, ING and PAT, we next assessed the impact of diet-induced obesity on protein secretion from the three adipose depots. Surprisingly, the rate of protein secretion from PAT was significantly reduced in obese mice (Fig. 3A). PCA showed clear segregation between the PAT secretome of lean and obese mice (Fig. 3B) and volcano plot analysis highlighted a marked suppression of the majority of secreted proteins in PAT from obese mice (Fig. 3C). Specifically, secretion of 244 proteins was decreased in PAT from obese mice compared with lean mice, while only 9 proteins were increased (Fig. 3C). These changes were associated with activation of ‘GAIT translation signalling’, with the interferon (IFN)-γ‐activated inhibitor of translation (GAIT) system playing an important role in the resolution of chronic inflammation⁵⁴ and increased ’JAK-STAT signalling’, a pathway that regulates various biological processes, including adipogenesis and adipose tissue development⁵⁵(Fig. 3D). Upstream regulators predicted to drive these changes included activation of LARP1, which is involved in mTOR signaling, particularly in various cancer settings⁵⁶, and RICTOR, also a key component of mTORC2 complex, that regulates cell growth and survival (Fig. 3E). In contrast, ChREBP (i.e., MLXIPL) was predicted to be inhibited in PAT of obese mice (Fig. 3E), with numerous ChREBP target genes (e.g., LIPE, PLIN1, FABP4, PCK1) reduced in the conditioned media of obese mice (Fig. 3F). Given that ChREBP is an important regulator of glucose metabolism, we further observed reduced secretion of proteins associated with glycolysis (Fig. 3G-H), but not fatty acid oxidation or the TCA cycle (Fig. 3I-J).

Fig. 3.

Regulation of adipokine secretion from paracardial adipose tissue (PAT) in diet-induced obese mice. C57BL/6 mice were fed a control chow or a high-fat diet (HFD) for 12 weeks, followed by assessment of protein secretion from PAT. (A) Rate of protein secretion in PAT, (B) Principal component analysis and (C) volcanoplot analysis of the PAT secreted proteome between control and HFD mice. Qiagen Ingenuity Pathway Analysis (IPA) highlighting (D) pathways and (E) upstream regulators in PAT of control and HFD mice. (F) Heatmap analysis of carbohydrate-responsive element-binding protein (ChREBP, i.e., MXLIPL) target genes in control and HFD mice. IPA analysis of the (G) diseases and biological functions and (H-J) heatmap analysis of glycolysis, fatty acid oxidation and TCA cycle-related genes identified in the PAT secreted proteome in control and HFD mice. (K) IPA analysis of the tox functions associated with the secreted proteome in control chow and HFD mice. (L) Uniquely secreted proteins only identified in PAT of chow control or HFD mice, with pathway analysis of the uniquely secreted proteins in (M) control and (N) HFD mice. Data is presented as mean ± SEM, *P < 0.05, with n = 7–8 mice/group. Significance was assessed using unpaired t-tests (A-C).

When assessing tox functions, we found that the diet-induced remodelling of protein secretion from PAT was associated with exacerbation of liver cancer and hepatocellular carcinoma, but surprisingly, suppression of ‘heart damage’ and ‘heart dysfunction’ (Fig. 3K). Lastly, we identified 88 proteins that were uniquely secreted from PAT of lean mice and 66 proteins that were uniquely secreted from PAT of obese mice (Fig. 3L). The unique proteins released from PAT of lean mice were primarily associated with ‘cellular stress response’ (Fig. 3M), including interleukin enhancer-binding factor 2 (ILF2), which has been shown to promote tumor development and cell proliferation⁵⁷, the molecular chaperone HSP70, and beta-catenin (CTNB1), a gatekeeper of adipose tissue identity and plasticity⁵⁸. In contrast, the proteins uniquely secreted from EPI of obese mice were related to ‘protein metabolism’ and ‘vesicle-mediated transport’ (Fig. 3N), including the ECM binding protein SPARC (secreted protein acidic and rich in cysteine; also known as osteonectin or BM-40), which promotes adipose tissue inflammation in obesity⁵⁹, Sec23a, a component of COPII transport vesicles, as well as the GTPases Sar1b and Rab7a and the coatamer subunit COPA, which are involved in vesicle-mediated ER to Golgi transport, overall highlighting that the adipokine profile from PAT is uniquely remodelled in obesity.

Diet-induced obesity is associated with marked remodelling in adipokine release from epididymal adipose tissue

Diet-induced obesity did not impact the rate of protein secretion from EPI (Fig. 4A). PCA analysis showed clear segregation of the EPI secretome between lean and obese mice (Fig. 4B), with a marked increase in secretion of 354 proteins in EPI from obese mice, while only 20 proteins were reduced when compared to lean mice (Fig. 4C). These changes were associated with activation of pathways related to ‘neutrophil degranulation’ and ‘signalling by ROBO receptors’ and suppression of pathways associated with ‘IL12 signalling’ and ‘AMPK signalling’, among others (Fig. 4D). With ROBO receptor signalling being implicated in adipocyte thermogenesis⁶⁰, the HFD-induced adipokine release from EPI may promote thermogenic reprogramming via autocrine/endocrine signalling, while the reduction in IL12-associated signalling may suggest diet-induced suppression of T cell differentiation and natural killer (NK) cell activity in EPI or other tissues⁶¹. The top upstream regulators predicted to drive these HFD-induced changes in EPI secretion included NRF2 (i.e., NFE2L2), a key transcription factor that regulates the expression of antioxidant genes, and ChREBP (i.e., MLXIPL) (Fig. 4E). In this respect, ChREBP-regulated proteins that were increased in protein secretion from EPI included fatty acid synthase (FASN), acetyl-CoA carboxylase (ACACA) and ATP citrate lyase (ACLY), the major regulators of de novo lipogenesis, as well as key regulators of lipolysis, including hormone-sensitive lipase (LIPE) and perilipin 1 (PLIN1) (Fig. 4F). Supporting these findings, the major biological function associated with the obesity-induced increase in EPI-secreted proteins was ‘lipid synthesis’ (Fig. 4G), suggesting that proteins secreted from EPI of obese mice may regulate/induce de novo lipogenesis on other tissues. Interestingly, processes/diseases predicted to be suppressed included apoptosis, cardiomyopathy, cardiac lesions, myocardial infarctions and heart fibrosis (Fig. 4G-H), pointing to a potential visceral adipose tissue-heart endocrine axis that regulates physiological processes leading to cardiac dysfunction. We further identified 133 proteins that were released from EPI in obese mice but were absent in the conditioned media from lean mice and 14 proteins that were uniquely secreted from EPI of lean mice (Fig. 4I). Among others, proteins secreted only from lean EPI (i.e., absent in EPI from obese mice) included epidermal growth factor receptor (EGFR), which may induce adipose tissue macrophage activation and promote obesity and insulin resistance⁶², the laminin subunit Lamb2, an extracellular matrix protein important in adipose tissue structure 63, ABHD5 which functions in phosphatidic acid synthesis, and leukemia inhibitory factor receptor (LIFR), which has been shown to drive lipolysis, adipose tissue expansion and obesity⁶⁴. In contrast, proteins uniquely secreted from EPI of obese mice were associated with cellular response to stress/survival (Fig. 4J), including superoxide dismutase SOD2, thioredoxin reductase 1 (TXNRD1), involved in the regulation of cellular redox reactions, thioredoxin-dependent peroxide reductase (PRDX3), that catalyzes the reduction of hydrogen peroxide and organic hydroperoxides, and heat shock protein A12A (HSPA12A), which has been shown to impede M2 macrophage polarization and inhibit adipose tissue browning⁶⁵. Overall, these data suggest that the obesity-induced changes in EPI secretion are likely to contribute to the observed low-grade chronic inflammation and oxidative stress observed in the setting of cardiometabolic disease.

Fig. 4.

Regulation of adipokine secretion from epidydimal adipose tissue (EPI) in diet-induced obese mice. C57BL/6 mice were fed a control chow or a high-fat diet (HFD) for 12 weeks, followed by assessment of protein secretion from EPI. (A) Rate of protein secretion in EPI, (B) Principal component analysis and (C) volcanoplot analysis of the EPI secreted proteome between control and HFD mice. Qiagen Ingenuity Pathway Analysis (IPA) highlighting (D) pathways and (E) upstream regulators in EPI of control and HFD mice. (F) Heatmap analysis of carbohydrate-responsive element-binding protein (ChREBP, i.e., MXLIPL) target genes in control and HFD mice. IPA analysis of the (G) diseases and biological functions and (H) tox functions of the EPI secreted proteome in control and HFD mice. (I) Uniquely secreted proteins only identified in chow control or HFD mice, with (J) pathway analysis of the uniquely secreted proteins in HFD mice. Data is presented as mean ± SEM, *P < 0.05, with n = 7–8 mice/group. Significance was assessed using unpaired t-tests (A-C).

Diet-induced remodelling of adipokine release from inguinal adipose tissue is associated with defects in glucose metabolism and mitochondrial function

Similar to that in PAT, the extent of protein secretion from ING was less in obese mice compared with lean control (Fig. 5A). The PCA showed clear segregation of the ING secretome of lean and obese mice (Fig. 5B) and volcano plot analysis pointed to a pronounced reduction in numerous proteins (80 proteins decreased, 3 proteins increased) in obese vs. lean mice (Fig. 5C), contrary to the remodelling observed in EPI (Fig. 4C). Pathway analysis pointed to activation of pathways associated with ‘mitochondrial dysfunction’ and suppression of pathways associated with ‘gluconeogenesis’, ‘PI3K/Akt signalling’ and ‘GLUT4 translocation’, insulin-dependent signalling pathways essential for glucose uptake into adipose tissue⁶⁶(Fig. 5D). Supporting this finding, one of the major upstream regulators predicted to be suppressed in ING from obese mice, and which may regulate the expression profiles observed in the proteomics dataset, was the insulin receptor (INSR) (Fig. 5E), with heatmap analysis highlighting reduced secretion of numerous INSR target genes (Fig. 5F). In addition, analysis of biological functions associated with the HFD-induced changes in protein secretion from ING included suppression of oxidative glucose metabolism, specifically glycolysis, and activation of carbohydrate synthesis (Fig. 5G). Together, these data suggest that the pronounced remodelling of protein secretion from ING in diet-induced obese mice is indicative of suppressed insulin signalling, glucose uptake and carbohydrate metabolism in adipose tissue⁶⁶. Furthermore, the marked suppression of protein secretion in ING was associated with pathways related to liver steatosis and liver cancer (Fig. 5G-H), pointing to a potential subcutaneous adipose-liver endocrine axis that contributes to metabolic liver disease. Additionally, we identified 64 proteins that were only secreted from ING in lean mice (Fig. 5I), with these unique proteins being associated with ‘interleukin signalling’, ‘cellular stress response’ and the ‘KEAP1/NFE2L2 (NRF2) pathway’ (Fig. 5J). Among others, proteins secreted only from lean ING (i.e., not secreted from ING of obese mice) included epidermal growth factor receptor (EGFR), as also observed in EPI, interleukin-1 receptor accessory protein (IL1RAP), involved in interleukin-1-dependent NFκB activation, and macrophage migration inhibitory factor (MIF), a pro-inflammatory cytokine involved in the innate immune response. We further identified 40 proteins that were only released from ING of obese mice (i.e., that were absent in conditioned media from lean mice) (Fig. 5I), including calumenin, which has recently been suggested to contribute to ECM production, inflammation and various cancers⁶⁷, the fermitin family homologue FERMT2, which enhances integrin-mediated cell adhesion onto the extracellular matrix, as well as several proteins involved in lipid metabolism, such as perilipin 4 (PLIN4), long-chain-fatty-acid-CoA ligase 1 (ACSL1) and the Ganglioside GM2 activator (GM2A).

Fig. 5.

Regulation of adipokine secretion from inguinal adipose tissue (ING) in diet-induced obese mice. C57BL/6 mice were fed a control chow or a high-fat diet (HFD) for 12 weeks, followed by assessment of protein secretion from ING. (A) Rate of protein secretion in ING, (B) Principal component analysis and (C) volcanoplot analysis of the ING secreted proteome between control and HFD mice. Qiagen Ingenuity Pathway Analysis (IPA) highlighting (D) pathways and (E) upstream regulators in ING of control and HFD mice. (F) Heatmap analysis of insulin receptor (INSR) target genes in control and HFD mice. IPA analysis of the (G) diseases and biological functions and (H) tox functions of the ING secreted proteome in control and HFD mice. (I) Uniquely secreted proteins only identified in chow control or HFD mice, with (J) pathway analysis of the uniquely secreted proteins in HFD mice. Data is presented as mean ± SEM, *P < 0.05, with n = 7–8 mice/group. Significance was assessed using unpaired t-tests (A-C).

To assess aspects of tissue integrity between the adipose depots that may impact release of proteins into the conditioned media, we further assessed presence/abundance of structural (i.e., cytoskeletal) and extracellular matrix (ECM) proteins in conditioned media of all six groups (Fig. S1A-C). For all three adipose depots, we found that the majority of both ECM and structural proteins were not different between lean and obese mice (Figure S1A-B), with ACTB (beta-actin) showing low abundance or even complete absence in each group (Figure S1C). Given the marginal differences in most markers identified, it is likely that differences in tissue integrity alone cannot explain the observed patterns in protein release between adipose depots and/or with high-fat feeding.

The ’obese’ PAT adipokine signature is closely associated with changes observed in ING but not EPI

We next compared obesity-induced changes in adipokine secretion between the three adipose depots. Comparing PAT and EPI, we found that the majority of proteins that were increased in secretion from EPI following the dietary intervention were decreased in PAT (blue box, Fig. 6A), showing opposing effects of the high-fat diet across these two depots. This was similarly observed when comparing ING and EPI (Fig. 6B). In contrast, the diet-induced changes in protein secretion from PAT and ING were similar between these two depots (Fig. 6C), highlighting that PAT and ING are similar in endocrine remodelling in the setting of obesity. Confirming this finding, we found that numerous common pathways were suppressed to a similar extent in both PAT and ING, including ‘ribosomal quality control’ and ‘cell cycle’ (Fig. 6D). Likewise, we observed similar predictions of common upstream regulators in PAT and ING, including activation of RICTOR, and suppression of ChREBP (MXLIPL) and NRF2 (NFE2L2) (Fig. 6E), as well as similar disease associations, such as positive associations with liver cancer and hepatic steatosis, and negative associations with cardiac inflammation and heart damage (Fig. 6F). Taken together, these data highlight that epididymal adipose tissue has a distinct adipokine signature in the setting of diet-induced obesity, contrasting with that evident in both paracardial and inguinal adipose tissues.

Fig. 6.

Comparison of diet-induced remodelling in adipokine secretion across the three adipose depots. C57BL/6 mice were fed a control chow or a high-fat diet (HFD) for 12 weeks, followed by comparison of diet-induced changes in protein secretion from epididymal (EPI), inguinal (ING) and paracardial (PAT) adipose tissue. Comparison of diet-induced changes in protein secretion between (A) EPI and PAT, (B) EPI and ING, (C) PAT and ING. (D-F) Qiagen Ingenuity Pathway Analysis (IPA) comparing diet-induced changes associated with (C) pathways, (D) upstream regulators and (E) diseases and biological functions across EPI, ING and PAT.

Discussion

Here, we provide a comprehensive comparative analysis of protein secretion from three white adipose tissue depots obtained from distinct anatomical locations in lean and obese mice. Our findings reveal profound depot-specific differences in both the magnitude and composition of secreted proteins, underscoring the likelihood of diverse endocrine roles of different adipose tissues in regulating an array of functions. Collectively, the data highlight that (i) under healthy conditions, PAT exhibits the highest secretory activity, (ii) diet-induced obesity elicits a robust suppression in protein secretion from subcutaneous ING and PAT depots, and (iii) the secretory pathways and associated signaling networks are depot- and disease-specific, suggesting distinct physiological roles and adaptive mechanisms. Specifically, we show marked depot-specific regulation of proteins associated with ChREBP transcriptional activity, oxidative glucose metabolism and insulin signalling (Fig. 7).

Fig. 7.

Summary diagram highlighting obesity-associated depot-specific remodelling in protein secretion from paracardial, epididymal and inguinal adipose tissue. ChREBP, Carbohydrate-responsive element-binding protein.

Under physiological conditions (i.e., in lean healthy mice), PAT adipose tissue displayed the greatest overall rates of protein release, followed by ING and lastly EPI. These differences likely reflect intrinsic variations in cellular composition, vascularization, and metabolic activity (lipolysis, lipogenesis, glycolysis, mitochondrial function) between these depots, as recently reported^68–70. PAT is located outside the parietal pericardium, and in contrast to epicardial adipose tissue, does not have direct contact with the myocardium⁷¹. Its influence on the heart is therefore more likely to be through endocrine regulation. In this respect, the overwhelming majority of studies assessing pericardial adipose-cardiac crosstalk focus on the influence of epicardial adipose tissue, due to its anatomical contiguity with the epicardium and clinical link with electromechanical dysfunction and associated disorders^20,71. In contrast, while studies assessing protein secretion from PAT and the capacity of PAT to modulate heart function are lacking, our team and others have previously shown that PAT conditioned media disrupts cardiomyocyte electrophysiology in vitro in a manner that could contribute to atrial fibrillation¹⁹, and that PAT-derived leptin induces myocardial apoptosis in obese rats through activation of JAK/STAT signalling and inhibition of downstream Na⁺/K⁺-ATPase activity⁷².

The proteomic analyses within the current study show that PAT displayed an adipokine signature that was distinct from the other two adipose depots, with enrichment in proteins associated with NRF2 transcriptional response, as well as enhanced interleukin 1 and NFκB signalling, implicating PAT as a source of stress-responsive and inflammatory mediators. The PAT secreted proteome was further enriched in components of the 20S proteasome core complex, which plays a crucial role in the degradation of damaged, oxidized, or misfolded proteins⁷³. In adipose tissue, this process helps to prevent accumulation of toxic protein aggregates, which can impair adipocyte metabolism and insulin sensitivity⁷⁴. In this respect, the increase in components of the 20S proteasome core complex detected in PAT may play a role in the resolution of oxidative stress⁷⁴ and contribute to cell survival under stress associated with damaged proteins⁷⁵. Interestingly, active 20S proteasome complexes have been shown to circulate in plasma, and increased circulating levels have been found in patients with various cancers, autoimmune disorders and/or sepsis⁷⁶. While these proteins are likely to be released within extracellular vesicles and may influence protein degradation in other tissues, their endocrine role is still unclear⁴⁴. Here we show that components of this complex are released from all adipose depots with the greatest secretion observed from PAT, suggesting that adipose tissues are likely to contribute to the observed differences in circulating levels in various (patho)physiological conditions.

Strikingly, in lean mice PAT released 137 unique proteins that were not detected in EPI or ING, many of which were associated with cellular stress and signalling by ROBO receptors. Within adipose tissue, ROBO receptors are important in regulating norepinephrine secretion and contributing to both adipose tissue lipolysis and adipocyte thermogenesis⁷⁷. While some evidence exists on depot-specific differences in adipose tissue lipolysis⁷⁸, including higher lipolysis rates in adipocytes located in subcutaneous compared with visceral regions⁷⁹, to our knowledge little is known about basal and norepinephrine-stimulated lipolysis in PAT, and how PAT lipolysis compares to lipolytic rates across other adipose tissue depots. Having said that, pathway analysis pointed to activation of the lipid sensor PPARγ and adipose triglyceride lipase, an enzyme essential for lipolysis⁸⁰ in PAT of obese mice, which might be associated with higher rates of lipolysis. In addition to their potential for autocrine regulation, proteins associated with ROBO receptor signalling are most likely to be released within EV and may therefore also influence (patho)physiological processes in other tissues, including bone formation⁸¹, axon guidance and neuronal migration in the nervous system⁸², as well as liver fibrosis and cancer development⁸³.

Comparing the adipokine signatures between the three depots in obese mice, we further observed a PAT-specific enrichment of secreted proteins that are involved in fatty acid oxidation, the TCA cycle and glycolysis (including hexokinase 1/2, aldolase a/b, enolase 1, phosphoglycerate mutase 1). The increased release of glycolysis-associated enzymes would suggest an ChREBP-driven response. ChREBP is highly expressed in lipogenic tissues, such as liver and adipose tissue, where it regulates the production of acetyl CoA from glucose (i.e., glycolysis) and contributes to fatty acid synthesis, elongation, and desaturation⁸⁴. This finding is surprising as, on the other hand, the PAT-specific adipokine remodelling observed in obese mice was associated with marked suppression of ChREBP transcriptional activity, together with decreased glycolytic enzyme secretion. This change in PAT may therefore suggest an obesity-induced attenuation of glucose metabolism and de novo lipogenesis. Notably, PAT also exhibited suppression of MYC, TGFβ, and NRF2-related pathways, indicating decreased responses in respect to adipose stem cell differentiation⁸⁵, extracellular matrix (ECM) remodeling and adipose tissue fibrosis⁸⁶ and adaptive antioxidant response and detoxification⁸⁷, respectively. These findings suggest that PAT adapts to chronic nutrient excess by downregulating stress- and fibrosis-related pathways, perhaps to preserve local cardiac homeostasis and prevent inflammatory spillover into the myocardium. Conversely, predicted activation of LARP1 and RICTOR point to enhanced mTORC1-⁸⁸ and mTORC2-specific⁸⁹ regulation of cell growth and metabolism, respectively. Given that the phosphorylation response to insulin is largely mTOR dependent⁹⁰, the predicted increase in mTOR signalling may suggest positive effects on insulin sensitivity and glucose uptake in PAT of obese mice. However, mTORC2 has also been shown to control expression of ChREBP in adipose tissue and, therefore, to regulate ChREBP-driven de novo lipogenesis and glucose metabolism⁹¹, contrary to the observed suppression of ChREBP in PAT of obese mice. In this respect, future studies are required to fully elucidate the link between mTOR signalling and ChREBP in the regulation of lipid and glucose metabolism in paracardial adipose tissue.

In contrast to the lack of research assessing protein secretion from PAT, various studies have previously assessed (and compared) protein secretion from visceral and subcutaneous adipose depots in both rodents^29,92,93 and humans^14,94, including proteins specifically contained with adipose tissue EV⁹⁵, overall highlighting that proteins released from visceral adipose tissue exacerbate inflammation and suppress peripheral insulin signalling to a greater extent than proteins released from subcutaneous adipose tissue. Here we show that ING adipose tissue displayed a secretory response to diet-induced obesity that closely mirrored the changes observed in PAT, including marked reductions in overall protein release and ChREBP transcriptional activity, as well as enhancement of RICTOR-associated mTOR signalling. The ING proteomic signature further indicated suppression of PI3K/Akt signaling and GLUT4 translocation, consistent with reduced insulin sensitivity and glucose uptake⁹⁶, as commonly observed in the context of obesity⁹⁷. We further observed a robust increase in many proteins released from EPI in obese mice, as previously reported (reviewed in⁹⁸), contrary to the marked suppression in PAT and ING.

The EPI secreted proteome was enriched for proteins linked to NRF2 and oxidative stress, and notably for ChREBP transcriptional activity. In this context, the contrasting regulation of ChREBP between EPI (activation), PAT (suppression) and ING (suppression) underscores the potential for depot-specific transcriptional control of lipid and glucose metabolism. Upregulation of canonical ChREBP targets in EPI, including increased secretion of acetyl-CoA carboxylase (ACC), fatty acid synthase (FAS), and ATP-citrate lyase (ACLY)⁸⁴, as well as increased secretion of hormone-sensitive lipase (LIPE) and perilipin 1 (PLIN1), markers of lipogenesis and lipolysis⁹⁹, respectively, suggests a potential EPI-driven autocrine/endocrine modulation of lipid metabolism in both adipose tissue and peripheral target tissues. In this context, chronic overnutrition has been shown to promote lipogenesis, inhibit lipolysis and/or increase lipid storage in adipocytes¹⁰⁰ liver¹⁰¹ and heart¹⁰², among other tissues. These findings suggest that obesity reprograms visceral adipocytes towards a secretory profile implicated in increased lipid synthesis and lipid mobilization. In this respect, visceral adipose tissue is associated with increased risk for developing cardiometabolic diseases, whereas subcutaneous adipose tissue is considered protective¹⁰³. This is likely related to higher rates of, and sensitivity to, lipolysis¹², in addition to the distinct adipokine secretion profile observed here.

Given the absence of a N-terminal signal peptide in these lipogenic and lipolytic proteins, as well as the glycolytic proteins discussed above, they are most likely released within extracellular vesicles (EV). Supporting this notion, we found that only 18–29% of all identified proteins were predicted to be classically secreted (i.e., contain a N-terminal signal peptide), which is similar to recent studies^29,32,33, suggesting that the majority of proteins secreted from adipose tissue are contained within EV. Similarly, various lipolysis- and lipogenesis-related proteins have been previously shown to be present within adipose tissue-derived EV^104,105 as well as in plasma EV^104,106. It is unclear why these proteins are released from adipose tissue. We hypothesize that they may impact lipid and glucose metabolism in peripheral tissues¹⁰⁷. Supporting this notion, a recent study has shown that adipose tissue-derived EV transfer cargo to the brain, and specifically to neurons within the hippocampus, which induces synaptic loss and cognitive impairment¹⁰⁸. Similarly, previous research highlighted that adipose EV can induce systemic glucose intolerance and insulin resistance¹⁰⁹, integrate into liver cells and promote dysregulation of TGF-β signalling¹¹⁰, aggravate temporomandibular joint osteoarthritis¹¹¹, and may contribute to cancer progression¹¹². Together, these studies highlight that EV released from adipose tissue of obese rodents (or individuals) are likely to contribute to pathologies across diverse clinical settings.

Taken together, the present data highlight depot-specific differences of adipokine secretion and their regulation in the setting of obesity. In lean mice, PAT adipose tissue exhibits the highest rates of protein secretion, particularly secretion of proteins associated with cellular stress, interleukin and NFκB signalling, potentially conferring local cardiometabolic regulation. In obesity, however, the endocrine landscape shifts substantially, with both subcutaneous (ING) and paracardial (PAT) depots downregulating secretion, while EPI shows increased secretion of a myriad of proteins. This asymmetric remodelling suggests that visceral adipose tissue assumes a dominant endocrine role in obesity, possibly at the expense of other depots’ regulatory capacity. The identification of unique depot-specific and obesity-responsive secreted proteins offers new insights into how regional adipose tissues coordinate systemic energy homeostasis and organ communication.

Limitations and future directions

Although this study provides a comprehensive comparative analysis of protein secretion from three white adipose tissue depots in lean and obese mice, we acknowledge limitations of the experimental approaches used. Firstly, protein secretion was assessed ex vivo in adipose tissue explants, and proteins detected in the conditioned media could have been released due to tissue injury and cell death following excision, and may not be bona fide adipokines. In addition, a potential limitation of this ex vivo approach is the absence of hormonal and neuronal input, which could affect protein secretion. In this study, all protein secretion has been normalized to tissue weight. However, given the large increase in adipose tissue mass in obesity, the observed changes in protein secretion may not be fully representative of the total rates of protein release from each adipose depot in the setting of obesity. We further acknowledge that the impact of adipokine release from the different adipose tissue depots on peripheral (patho)physiology and their potential role in inter-tissue communication was not assessed in this study. In this respect, future work is required to identify any functional properties of proteins released from the different adipose depots, particularly from PAT which showed a unique secretion profile. Lastly, the amount of epicardial adipose tissue in mice is much lower than that of humans, potentially limiting insights into PAT as a collective regulator of the cardiometabolic environment.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.R.B. and M.K.M. conceived and designed research;.J.R.B., P.M.M, H.W., and M.K.M. performed experiments; J.R.B. and M.K.M. analyzed data; J.R.B. and M.K.M. interpreted results of experiments and prepared figures; J.R.B. and M.K.M. drafted manuscript; J.R.B., P.M.M., H.W., L.M.D.D., M.J.W., and M.K.M. edited and revised manuscript; J.R.B, P.M.M., H.W., L.M.D.D., M.J.W., and M.K.M. approved final version of manuscript.

Funding

These studies were supported by funding from the National Health and Medical Research Foundation of Australia (NHMRC; Grant Nos. APP1162511 and APP1156508). P.M.M. is supported by an NHMRC Investigator Grant (No. APP2018187), and previously supported by a Canadian Natural Sciences and Engineering Research Council of Canada (NSERC) postdoctoral fellowship (No. PDF-516731-2018) and a Canadian Institutes of Health Research (CIHR) postdoctoral fellowship. M.J.W. and M.K.M. were supported by Research Fellowships from the NHMRC (Nos. APP1077703, APP1143224).

Data availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE¹⁰⁷ partner repository with the dataset identifier PXD076502. The submission details are as follows: Project Name: Interrogation of depot-specific adipokine release in diet-induced obese mice reveals suppression of oxidative glucose metabolism in paracardial adipose tissue. Project accession: PXD076502. .

Declarations

Competing interests

The authors declare no competing interests.