Multi-omics analysis reveals metabolic disruptions in cardiac tissues of aging nonhuman primates with spontaneous type 2 diabetes

Alteration of proteomics profile in the OB and TM group compared to the CON

To explore the metabolic consequences of aging and TM on cardiac function, we conducted a comprehensive proteomic analysis. We screened a large cohort of aging monkeys from a monkey farm and identified three groups: aging with obesity (OB), aging with TM, and healthy aging controls (CON) (Figure 1A). Body weight, body mass index, and fasting blood glucose data for these groups are detailed in Supplementary Table 4. Through label-free mass spectrometry, we profiled the left ventricular proteomes of these groups and performed a principal component analysis (PCA), which revealed significant proteomic differences (Figure 1B).

A total of 2555 proteins were detected across all groups (Supplementary Table 1), with 90 and 41 proteins uniquely present in the OB and TM groups, respectively, and 56 proteins exclusive to the CON group (Figure 1C). The heatmaps in Figure 1D to F present the comprehensive differential protein profiles of the three groups. Proteins with more than a 1.2-fold change in expression and a p-value less than 0.05 were considered as candidates for further analysis, providing a robust dataset to investigate the molecular signatures of cardiac remodeling in the context of aging and OB or TM.

Comparative analysis revealed that, relative to the CON group, 8 and 29 proteins were upregulated and downregulated in the OB group, respectively, while 1 and 38 proteins were upregulated and downregulated in the TM group. When the TM group was compared to the OB group, 8 and 181 proteins were found to be upregulated and downregulated, respectively (Figure 1G–1I).

Multiple extracellular matrix (ECM) related proteins were affected by obesity or diabetes

To explore the effects of differential protein expression on cardiac function, we undertook GO and KEGG analyses of our proteomic dataset. GO analysis revealed significant alterations in extracellular matrix and mitochondrial component pathways in the obesity (OB) group. In contrast, the triglyceride metabolism (TM) group exhibited broader disruptions in energy metabolism and myocardial structural pathways compared to the healthy aging control (CON) group (Figure 2A, 2B). Additionally, the TM group displayed unique changes in ribosome complex, nucleic acid binding, and peptide biosynthetic pathways relative to the OB group (Figure 2C).

Building on these findings, KEGG analysis pinpointed the complement and coagulation cascade, along with cytoskeletal pathways, as being most affected in the OB group. In contrast, the TM group showed more pronounced disruptions in focal adhesion and muscle cytoskeletal pathways (Supplementary Figure 1).

Our proteomic analysis further revealed significant downregulation of key extracellular matrix proteins Dermatopontin (DPT), Lumican (LUM), and Haptoglobin (EGK_12958) in both OB and TM groups relative to controls. This was accompanied by substantial downregulation of Fibrinogen gamma chain protein (FGG) and beta chain protein (EGK_16175), indicating a potential disruption in the coagulation cascade (Figure 2D).

The serpin family members SERPINA1 and SERPINA3 were significantly downregulated in both OB and TM groups, consistent with alterations in the complement system. Another serpin family member, SERPING1, which regulates the complement cascade, was also downregulated, along with a decrease in the major complement component C3 and the immune-related protein Alpha-1-acid glycoprotein (ORM1) (Figure 2E, 2F). This downregulation across OB and TM groups points to a common impact on the immune and complement systems.

Proteins associated with metal ion metabolism were also affected, with Nucleobindin-1 (NUCB1) increased and Calumenin (CALU) decreased in the TM group. Vitamin D binding protein, GC, showed a consistent decrease in both groups. Additionally, iron metabolic regulation was altered, as Ap2 associate kinase (AAK-1) was elevated in the TM group, while transferrin (TF) and serotransferrin (EGK_12095) were decreased in both groups (Figure 2G).

Lastly, protein homeostasis factors, including Histone H3 (HIST2H3A), Aspartyl-tRNA synthetase (DARS1), 39S ribosomal protein L39 (MRPL39), and Ubiquitin thioesterase (OTUB1), exhibited differential regulation between the OB and TM groups (Figure 2H).

To validate the proteomic findings, we assessed the expression of key proteins—TF, ORM1, C3, NUCB1, and serpin family A member 1 (SERPINA1)—via Western blot analysis. The results corroborated the proteomic data, exhibiting consistent trends across the OB, TM, and CON groups (Supplementary Figure 2). These findings confirm the reliability and accuracy of the label-free mass spectrometry approach, reinforcing the observed metabolic and structural perturbations in the DCM myocardium.

Metabolic profiling reveals distinct cardiac metabolite alterations in obesity and type 2 diabetes

To determine whether altered protein expression corresponds to changes in myocardial metabolites, we conducted a comprehensive metabolomic analysis on left ventricular heart tissue from the three study groups (Supplementary Table 2). This analysis was designed to explore the downstream metabolic effects of the protein expression changes observed in our proteomic study. PCA revealed significant metabolic distinctions among the groups, highlighting the divergent metabolite compositions (Figure 3A). In total, we identified 194 measurable and reproducible metabolites, encompassing 54 fatty acids, 37 amino acids, 27 organic acids, and 19 carnitines (Figure 3B).

The metabolic profile of the TM group notably deviated from that of the CON and OB groups, exhibiting a reduction in fatty acids and carnitine levels alongside an increase in amino acids and carbohydrate content (Figure 3C). These observations suggest that the metabolic shifts in the TM group may be indicative of an adaptive response or a pathological change in energy metabolism. Heatmaps visually depicted the profiles of differential metabolites, providing a clear overview of metabolic variations (Figure 3D).

To determine which pathways were affected by obesity or diabetes, the Metabolic Pathways Enrichment Analysis (MPEA) was applied. MEPA revealed that lysine biosynthesis, fatty acid biosynthesis, and biotin metabolism were the most significantly affected pathways in the OB group (Figure 3E). In contrast, the TM group showed pronounced disruptions in amino acids metabolism, taurine metabolism, and fatty acid biosynthesis when compared to the CON group (Figure 3F). A comparative analysis from OB to TM revealed that amino acids metabolism and fatty acid biosynthesis were the primary metabolic pathways that changed, indicating a shift in metabolic flux associated with the progression from obesity to TM (Figure 3G).

Comprehensive metabolomic profiling unveils diverse metabolic disruptions in obesity and type 2 diabetes

In our metabolomic analysis, we utilized univariate statistical analysis to discern 24 metabolites that were significantly differentially expressed, providing insights into the metabolic repercussions of obesity and TM within an aging context. The metabolism of amino acids was particularly perturbed; L-Alanine was significantly downregulated by 36% in the TM group relative to the control group, with no significant change observed in the obesity group. This pattern was echoed by N-Acetyl-L-alanine, another alanine derivative, which also saw a 36% decrease in the TM group. Additionally, Gamma-Aminobutyric-acid (GABA) was specifically and significantly downregulated in the TM group, with a pronounced 63% reduction. Other amino acids including L-Alpha-aminobutyric acid, Dimethylglycine, Glycine, and Methylcysteine were significantly downregulated in TM groups, albeit from a low baseline. In contrast, N-Acetylglutamine, and L-Aspartic acid were selectively affected by obesity, indicating distinct metabolic sensitivities to these conditions (Figure 4A).

The metabolism of fatty acids was predominantly affected in the TM group. Palmitoleic acid, a saturated fatty acid, showed a paradoxical 25% increase in the OB group but a substantial 50% decrease in the TM group when compared to controls. This dichotomy was also observed in Dodecanoic acid, with the latter decreasing by 47% in the TM group. The mono-unsaturated fatty acids, oleic acid and 11-cis-Eicosenoic acid, exhibited a non-significant trend towards increase in the obesity group but a significant 47% and 38% decrease in the TM group respectively. This pattern was mirrored by poly-unsaturated fatty acids such as eicosadienoic acid, adrenic acid, and 8,11,14-Eicosatrienoic acid, which showed a mild increase in the obesity group followed by a sharp decrease in the TM group. The long-chain fatty acids, 13,16,19-docosatrienoic acid, demonstrated a stepwise reduction from both obesity to TM groups, with the TM group exhibiting the most significant decreases. Arachidonic acid significantly downregulated in the TM group (Figure 4B–4D).

In concordance with the observed decline in fatty acid content in the TM group, short-chain carnitines were significantly downregulated. Hydroxybutyrylcarnitine and methylbutyroylcarnitine were the primary contributors to this decrease, with substantial reductions observed in both the obesity and TM groups compared to controls. Even the low-abundance Glutarylcarnitine was significantly downregulated in both groups, indicating a broad impact on fatty acid metabolism (Figure 4E).

Furthermore, threonic acid, a sugar acid, was significantly downregulated by 17.4% in the TM group. Other organic acids including Citramalic acid and Indoleacetic acid also showed significant changes, albeit from very low baseline levels, suggesting a complex metabolic reprogramming in the TM group (Figure 4F).

Lipidomic profiling reveals distinct lipid alterations in obesity and type 2 diabetes

Following our metabolomic findings that indicated suppressed short-chain acylcarnitines in both OB and TM groups, we conducted a lipidomic study to further elucidate the lipid changes (Supplementary Table 3). We measured acylcarnitines, glycerophospholipids, and sphingolipids and observed significant differences among the CON, OB, and TM groups using PCA (Figure 5A). A total of 87 glycerophospholipids, 40 acylcarnitines, and 14 sphingolipids were detected (Figure 5B). Notably, the TM group exhibited lower levels of glycerophospholipids and higher levels of sphingolipids compared to the CON and OB groups (Figure 5C).

Through pairwise T-tests and one-dimensional statistical analysis, we identified 13 significantly differential lipid species, including 3 acylcarnitines, 8 lysoPCs, and 2 PCs. (Figure 5D–5F). In both the OB and TM groups, we observed a decrease in lysoPC species C16:0, C17:0, C18:0, C18:2, and C28:1, with the most substantial reductions in the OB group. Notably, lysoPC C16:0 levels were significantly decreased by 66% and 48% in the OB and TM groups, respectively, relative to controls. The levels of lysoPC C18:0 were also reduced by 62% and 41% in the OB and TM groups, respectively, though the change in the TM group was not significant compared to controls. LysoPC species C16:1, C18:1, and C20:4 showed a stepwise downregulation from the control group to the OB and TM groups, suggesting a progressive impact of obesity and diabetes on lipid metabolism (Figure 5D).

Particularly, lysoPC C16:1 and C18:1, which are delta 9 unsaturated fatty acids, were significantly decreased by 43% and 39% in the OB group, and by 63% and 53% in the TM group, respectively, compared to controls. The decrease in lysoPC C20:4 was also significant, showing a 61% reduction in the TM group (Figure 5D).

Surprisingly, though we detected downregulation of short chain acylcarnitine, C18:2 acylcarnitine, a long chain acylcarnitine significantly upregulated 4.4 times in the TM group when compared with the CON group. C3-OH, another short chain acylcarnitine, which is negligible in the CON group, was upregulated 47 times in the OB group but not changed in the TM group when compared with the CON group. C9 acylcarnitine content mildly but significantly decreased by 26% in the OB group but unchanged in the TM group when compared with the CON group (Figure 5E).

PC C32:3 is the only one significantly changed Phosphatidylcholine, it is decreased and increased by 20% and 13% in the OB and the TM group respectively when compared with the CON group (Figure 5F).

Generation of spontaneous obese and diabetic rhesus monkeys (Macaca mulatta)

Adrenal glands were obtained from rhesus macaques (Macaca mulatta) that were raised at the Kunming Institute of Zoology, Chinese Academy of Sciences. The criteria used for screening of spontaneously obese and diabetic rhesus monkeys were as previously described [40]. In brief, body mass index (BMI) ≥ 40 kg/m2 was defined as obesity and plasma levels of fasting blood glucose ≥ 7.0 mmol/L was defined as type 2 diabetes (TM). In total 9 monkeys were included in this study with 3 in each group: healthy male monkeys (CON), obese male monkeys (OB), and TM male monkeys (TM). All the nine monkeys had libitum access to water and food including 21.6% of calories as protein, 5.4% as fat, and 56.6% as carbohydrates. The monkeys were euthanized using 20 mg/kg ketamine, and cardiac muscle was collected (100 mg/animal).

Metabolomics

Metabolomics analysis was performed using a Q300 Kit (Metabo-Profile, Shanghai, China). Each harvest tissue was stored in an Eppendorf Safelock microcentrifuge tube (Eppendorf, Hamburg, Germany) and mixed with 10 pre-chilled zirconium oxide beads and 20 μL deionized water. Metabolites were extracted using 150 μL methanol containing the internal standard after homogenization of 3 min. Then another 3 mini homogenization was performed and followed by centrifuged at 18,000 × g for 20 min. The supernatants were transferred to 96-well plates and a Biomek 4000 workstation (Biomek 4000; Beckman Coulter Inc., Brea, CA, USA) was used for the following procedures. 20 μL freshly prepared derivative reagent was added into each well. The plate was sealed, and derivatization was conducted at 30° C for 60 min. After derivatization, the samples were evaporated for 2 h and 330 μL ice-cold 50% (v/v) methanol was added to each well to reconstitute the samples. The plate was stored at − 20° C for 20 min and the samples were centrifuged at 4000 × g and 4° C for 30 min. Then, 135 μL supernatant was transferred to a new 96-well plate with 15 μL internal standards in each well. Serial dilutions of derivatized stock standards were added to the wells on the left side and the plate was sealed for liquid chromatography-mass spectrometry (LC–MS) analysis. All metabolite targets were quantitated using an ultraperformance liquid chromatography coupled with tandem mass spectrometry (UPLC-MS/MS) system (ACQUITY UPLC-Xevo TQ-S; Waters Corp., Milford, MA, USA) by Metabo-Profile Biotechnology (Shanghai) Co. Ltd.

Raw data files generated by UPLC-MS/MS were processed using the QuanMET software (v2.0; Metabo-Profile, Shanghai, China) to obtain peak integration, calibration and quantitation for each metabolite.

Lipidomic

We performed the lipidomics analysis with primary targeted metabolic profiling platform BAP Ultra (Metabo-Profile, Shanghai, China). Harvested samples were stored in an Eppendorf Safelock microcentrifuge tube (Eppendorf, Hamburg, Germany) and centrifuged at 2000 rcf for 10 min. After that, the supernatant was collected and transferred to a 96-well plate. All samples were qualified after quality control. Then the lipidomics profile was detected using Ultra Performance Liquid Chromatography Tandem Mass Spectrometer (UPLC-MS/MS) with normalization based on mass following the manufacturer's instructions by Metabo-Profile Inc. (Shanghai, China).

Raw data were processed by an automation assay workflow using LipidMET software streamlines and lipids in the samples were identified and quantitated for further analysis.

Proteome

Protein was extracted using SDT method (4% sodium dodecyl sulfate [SDS], 100 mmol/L Tris/HCl, 0.1 mol/L DTT, pH 7.6). After trypsin digested with FASP (Filter aided proteome preparation), the peptides were desalted and reconstituted with 40 μL of 0.1% formic acid solution. Then the peptides were labeled using the TMTsixplex™ Isobaric Label Reagent Set (#90111, Thermo Fisher Scientific, USA). TMT-labeled peptides were fractionated using high pH Reversed-Phase Peptide Fractionation Kit (#84868, Thermo Fisher Scientific). After desalting and vacuum drying, the peptide was reconstituted using 12 μL of 0.1% fatty acid (FA) and concentration was based on ultraviolet (UV) absorbance at 280 nm. The TMT-labeled peptide set was blended using AKTA Purifier 100 (GE, Uppsala, Sweden) and subjected to chromatographic column for fractionation. The prepared samples were then used for liquid chromatography with tandem mass spectrometry (LC-MS/MS) and mass spectrometric analysis using a Q-Exactive mass spectrometer.

Raw peptides were identified and quantitated using Maxquant software (https://www.maxquant.org/, version 1.5.1.0).

Western blots

Proteins were extracted from left ventricle samples using RIPA buffer supplemented with protease inhibitors. Protein concentration was determined by BCA assay. Samples (20–50 μg) were denatured in Laemmli buffer, resolved by 10% SDS-PAGE, and transferred to a PVDF membrane. The membrane was blocked with 5% non-fat milk in TBST and incubated overnight at 4° C with primary antibodies from Proteintech (USA): rabbit anti-transferrin (1:1000), rabbit anti-ORM1 (1:1000), rabbit anti-C3 (1:1000), rabbit anti-NUCB1 (1:1000), rabbit anti-SERPINA1 (1:1000), or mouse anti-GAPDH (1:5000). After washing, the membrane was incubated with HRP-conjugated secondary antibody (1:5000) for 1 hour at room temperature. Bands were visualized using ECL and imaged with a ChemiDoc system.

Data availability

The datasets generated and analyzed during the current study are not publicly available due to restrictions, such as containing information that could compromise the privacy of research participants or proprietary considerations. However, these datasets are available from the corresponding author upon reasonable requests and for researchers who meet the criteria for access to confidential data. The mass spectrometry proteomics data have been deposited in Dryad with the accession code DOI: 10.5061/dryad.c2fqz61ks. Metabolomics and lipidomics data are also deposited in Dryad under the same identifier, facilitating integrated access to related datasets.

Data analysis

All data analyses were performed in language R. The significance was assessed using Student’s t-test with a p-value of < 0.05. The heatmap, boxplot and volcano were conducted using package “pheatmap” and “ggplot” respectively.