Integrative Proteomic and Bioenergetic Profiling Reveals Diet- and Strain-Specific Mitochondrial Dysfunction in Cohen Diabetic Rat Hearts

TL;DR

This study shows how diet and genetics together affect heart mitochondria in diabetic rats, leading to heart disease.

Contribution

The study reveals diet- and strain-specific mitochondrial changes in diabetic rat hearts using proteomic and bioenergetic profiling.

Findings

Cytochrome c oxidase subunits are downregulated in diabetic rat hearts.

Mitochondrial respiration and ATP levels are impaired in sensitive strain rats.

Diet-specific tyrosine phosphorylation of CcO subunit I indicates inflammation-driven inhibition.

Abstract

Mitochondrial dysfunction contributes to diabetic cardiomyopathy, yet how genetic predisposition and diet interact to reshape cardiac metabolism in diabetic and prediabetic states remains unclear. The Cohen diabetic rat model, comprising diabetes-resistant (CDr) and diabetes-sensitive (CDs) strains, provides a unique platform to dissect this interplay. Here, we present an integrative global proteomic and bioenergetic characterization of cardiac tissue from CDr and CDs rats fed either a regular or a diabetogenic diet. Proteomic pathway mapping revealed downregulation of cytochrome c oxidase (CcO) subunits, strain-dependent rewiring of fatty-acid oxidation pathways, and CcO subunits switch from “heart-type” to “liver-type” isoforms in the sensitive strain. These changes were accompanied by impaired mitochondrial respiration, ATP depletion, and disruption of mitochondrial quality-control mechanisms, together with increased accumulation of tyrosine 304 phosphorylation of cytochrome c oxidase subunit I, indicative of inflammation-driven regulatory inhibition in a diet-specific manner. These findings establish an understanding of how genetic susceptibility and diet contribute to cardiac mitochondrial dysfunction in the Cohen diabetic rat model.

Figures

Full text

1. Introduction

Mitochondria are essential organelles responsible for producing over 90% of cellular ATP through the oxidative phosphorylation (OXPHOS) system located in the inner mitochondrial membrane. OXPHOS is particularly vital in metabolically active tissues such as the heart, brain, liver, and skeletal muscle, which require substantial energy to sustain physiological processes and adapt to fluctuating energy demands. A decline in ATP production results in an energy deficit, compromising the ability of these tissues to effectively metabolize glucose and fatty acids [1]. The OXPHOS system comprises five multi-subunit enzymes and two small electron carriers, coenzyme Q and cytochrome c. Complexes I, III, and IV (cytochrome c oxidase, CcO) assemble into mitochondrial supercomplexes (SCs), forming higher molecular weight multi-enzyme structures. SCs exhibit various stoichiometric compositions, such as I_1-2_ + III_2_, III_2_ + IV_1,_ and I + III_2_ + IV_1-4_ [2,3], which differ significantly across species and tissues, and their physiological importance remains a topic of debate [4,5].

Growing evidence suggests that SCs can act as independent entities capable of performing respiration [6,7,8]. SCs are generally proposed to be associated with enhanced respiration efficiency and reduced reactive oxygen species (ROS) production [9]. The regulation of mitochondrial ROS generation is crucial, as excess ROS contributes to cellular damage and aging [10]. Consequently, SCs are often proposed to be associated with improved mitochondrial function and cellular health [7,11]. Molecular changes that determine SC composition and distribution, such as the possibility of their regulation through posttranslational modifications (PTMs), in particular protein phosphorylation, have not been considered.

Protein PTMs are crucial and tightly controlled regulatory mechanisms, fine-tuning protein function and activity in response to cellular signaling and environmental changes. For example, CcO is the terminal and proposed rate-limiting enzyme of the electron transport chain (ETC) in mammals [12] and acts as a crucial regulatory checkpoint for mitochondrial respiration by catalyzing the reduction in molecular oxygen to water. Together with Complexes I and III, CcO generates the proton gradient across the inner mitochondrial membrane necessary for ATP synthesis. CcO activity is modulated through various mechanisms, including tissue-specific isoforms [13], allosteric regulation by the ATP/ADP ratio and thyroid hormones [14,15], and PTMs [12]. Previously, we showed that tyrosine 304 (Y304) of the catalytic subunit I of CcO can be phosphorylated via a cAMP-dependent mechanism and by inflammatory signaling in liver and lung tissue [16,17]. Y304 is located near the catalytic oxygen binding site of CcO subunit I and functions as a respiratory switch for enzyme activity. Phosphorylation of this tyrosine results in CcO inhibition and pronounced sigmoidal kinetics in the reaction with cytochrome c [16], leading to disruption of mitochondrial respiration and energy production [16,17]. Changes in the phosphorylation state of CcO within SCs may play an important role in regulating SC composition and the overall activity of SCs, and a mechanistic understanding of how SC activity is fine-tuned may help to further decipher their physiological role and their contribution to mitochondrial dysfunction. It is well established that impairment of mitochondrial function is associated with numerous metabolic diseases [18,19,20,21,22], contributing to disorders such as insulin resistance, heart failure, and diabetic cardiomyopathy. Among the most severe consequences of type 2 diabetes (T2D) are cardiovascular complications, which arise from the interplay of metabolic dysregulation, insulin resistance, and mitochondrial impairment [23]. Cardiomyocytes, which possess a high demand for ATP, are especially susceptible to the ramifications of mitochondrial dysfunction [24].

The healthy heart relies heavily on fatty acids to meet its high energetic demands, with 50–70% of ATP generated through fatty acid oxidation [25,26]. This dependence requires precise regulation, as cardiomyocytes possess only a limited capacity for lipid storage. Under normal conditions, fatty acid uptake and oxidation are therefore carefully balanced to ensure a steady energy supply without accumulating excess lipids. It has been established that in diabetic hearts, this balance becomes disrupted, leading to an increase in fatty acid oxidation rates, lipid deposition within the heart, resulting in lipotoxicity, and thus causing mitochondrial dysfunction, diabetic cardiomyopathy, and inflammation [27,28,29].

Here, we investigated the effect of a high-sucrose low copper diet (HSD) on cardiac mitochondrial function using the Cohen rat model, developed through over 100 generations of selective inbreeding. This model includes two distinct strains: the Cohen diabetes-sensitive (CDs) rats, which develop T2D when exposed to an HSD but maintain normoglycemia on a regular diet (RD), and the Cohen diabetes-resistant (CDr) rats, which maintain normal blood glucose levels on both diets [30,31]. These strains provide a unique platform for exploring the interplay between dietary and genetic factors within metabolic disease [32]. Cardiac dysfunction is well established for the CDs rats, which have increased heart mass and develop cardiac hypertrophy with left-ventricular wall thickening, diffuse interstitial connective-tissue accumulation, and perivascular fibrosis in the left ventricle, contributing to stiffening and diastolic dysfunction [33,34]. The animals also present with ischemic necrosis foci and hyperplastic vascular changes in the myocardium [35]. In addition, when switched to the HSD, the hearts of hyperglycemic CDs rats failed to exhibit ischemic preconditioning-associated protection, a critical adaptive mechanism that preserves cardiac resilience during metabolic stress [36].

In previous studies, we showed that CDs rats demonstrated reduced CcO activity in several tissues [31,37,38,39,40]. In the present study, we show that HSD induces pathological remodeling of cardiac mitochondria, characterized by impaired bioenergetic function, decreased CcO activity, CcO isoform switching, disrupted mitochondrial quality control, proteomic changes, and activation of inflammatory signaling. We identified a diet-specific increase in the phosphorylation of CcO within cardiac respiratory SCs, and defective fatty-acid metabolism arising from genetic susceptibility in the diabetes-sensitive strain, collectively leading to mitochondrial dysfunction and energy depletion under metabolic stress.

2. Materials and Methods

2.1. Animal Information

Animal information was described in detail in [38] and heart tissue was derived from the same animals. Briefly, the Cohen rat diabetes model was developed by selective inbreeding of rat [30,31,38]. At 6 weeks of age, male Cohen diabetes-sensitive (CDs) and Cohen diabetes-resistant (CDr) were switched to a high-sucrose diet, which is also low in copper (HSD) for 6 weeks, or maintained on a regular chow-diet (RD; Teklad, Harlan Laboratories, Placentia, CA, USA) [40]. While systemic copper levels in diabetic patients can be elevated, studies indicate a paradoxical decrease in intracellular myocardial copper content or a functional deficiency within cardiomyocytes caused by defective cellular uptake and/or impaired intracellular trafficking of copper [41,42], which is mimicked by the diet. The HSD consisted of a custom-prepared diabetogenic diet comprising 72% sucrose, 18% vitamin-free casein, 5% salt mixture no. II USP (MP Biomedicals, Solon, OH, USA), 4.5% butter, 0.5% corn oil, vitamins, and low copper (1.2 ppm) [32]. Blood glucose levels 2 h post-oral glucose tolerance test and body weights were previously reported [38]. Rats were euthanized using an overdose of a combined ketamine/xylazine mixture, and hearts were immediately removed, flash-frozen in liquid nitrogen, and stored at −80 °C until processing. Animal experiments were approved under protocol number MD-22-16924-3 by the Institutional Animal Experiments Committee (IACUC).

2.2. Proteomic Analysis

Whole heart tissue lysates were processed for proteomic analysis using the EasyPep™ MS Sample Prep Kit (Thermo Fisher Scientific, Waltham, MA, USA, cat. no. A40006). A Thermo Scientific Orbitrap Eclipse Tribrid mass spectrometer, operated in data-dependent acquisition (DDA) mode, was used to identify peptides. Raw data files were processed using the MaxQuant platform (v2.6.8.0) and searched against the UniProt Rattus norvegicus reference proteome as described (https://www.uniprot.org/proteomes/UP000002494 (accessed on 25 February 2025)) in [43] for a simple data set. Subsequent data processing and statistical analysis were performed using the Perseus framework (v2.1.4.0) [44]. Briefly, protein groups annotated as “contaminants” or “reverse” were excluded. Label-free quantification (LFQ) intensity values were grouped into experimental conditions consisting of five biological replicates per (except CDr-RD, n = 4; each from a separate animal). Protein identifications were filtered to retain only those detected with ≥2 unique peptides and present in ≥70% of samples within at least one group. Proteins were filtered to include only those identified with ≥2 unique peptides and present in at least 70% of samples within one group. LFQ intensities were log_2_-transformed, and missing values were imputed using a Gaussian distribution-based method. Differential protein expression was assessed using a permutation-based ANOVA (multiple-sample t-test) with a false discovery rate (FDR) < 0.05. Data visualization included unsupervised principal component analysis (PCA) and hierarchical clustering heatmaps. Pathway enrichment analysis was performed using ShinyGO v0.82, focusing on proteins significantly affected by strain and/or diet, with significance defined as FDR < 0.05. Results were visualized as dot plots ranked by fold enrichment.

2.3. Heart Tissue Homogenization and Mitochondria Isolation

Homogenization was carried out using gentleMACs Dissociator (Miltenyi Biotec, Bergisch Gladbach, Germany, cat. No 130-093-235) with the protocol m_lung_02.01. Heart tissue (70 mg) was added to 2 mL of mitochondrial isolation buffer (MIB: 200 mM sucrose, 10 mM Tris, 1 mM EGTA, pH 7.4 adjusted with 1 M HEPES solution, Gibco, Waltham, MA, USA, cat. no. 15630080) and placed in gentleMACs C Tubes (Miltenyi Biotec, cat. no. 130-093-237). The entire process was performed on ice to help preserve protein integrity and prevent dephosphorylation. To further breakdown the tissue, 0.4 mg of Subtilisin A (Protease) (Sigma, Burlington, MA, USA, cat. no. P5380) was added to the tissue homogenate and incubated at 4 °C for 10 min with gentle shaking. A series of filtration steps was done to enrich for mitochondrial fractions. A total of 2 mM phenylmethylsulphonylfluoride, 5 mM heat-activated sodium vanadate, and Protease Inhibitor Cocktail (Sigma, Cat. no. P8849) were added to the MIB. The filters were pre-wet with MIB. First, large cellular debris and nuclei were removed with a 40 M filter (pluriSelect, El Cajon, CA, USA, cat. no. 43-50040-50). The flowthrough containing mitochondria was filtered a second time through a 10 µM filter (pluriSelect cat. no. 43-50010-03) to further remove contaminants. To precipitate the crude mitochondrial fraction and further remove contaminants, the second flowthrough was then subjected to two medium speed centrifugations (9000× g, 15 min, 4 °C). Protein concentration was measured using the DC protein assay kit (Bio-Rad, Hercules, CA, USA, cat. no. 5000111) according to the manufacturer’s protocol.

2.4. Cytochrome c Oxidase Activity Measurements Using the Polarographic Method

To measure respiration of heart mitochondria in the Cohen CDr and CDs rat strains, we measured CcO-specific activity using a Clark-type oxygen electrode (Oxygraph system, Hansatech, Norfolk, UK) as described in [38]. Briefly, for each measurement, 30 μg of total mitochondrial protein was used and resuspended in solubilization buffer (10 mM K-HEPES, pH 7.4; 40 mM KCl; 1% Tween 20). Next, mitochondrial membranes were disrupted using a Fisher Scientific (Waltham, MA, USA) Sonic Dismembrator Model 100 ultrasonicator (2 rounds of 3 s pulses) at 4 °C. Oxygen consumption was assessed in the presence of 20 mM ascorbate and 30 μM cytochrome c (Sigma, cat no. C3131). Oxygen consumption rates were continuously monitored, and the data were recorded in nmol/min/mg mitochondrial protein and analyzed with the OxyTrace+ software (version 1.0 Build 48, 2016, Hansatech Instruments Ltd., Amesbury, MA, USA). The remaining sample was run on an SDS-PAGE to normalize the measurements to mitochondrial amount and CcO amount.

2.5. Mitochondrial Membrane Solubilization for BN-PAGE

The BN-PAGE protocol was adapted from [38,45]. All components were used from the NativePAGE™ Sample Prep Kit (ThermoFisher Scientific, cat. no. BN2008), including NativePAGE™ 4 × sample buffer, NativePAGE™ 5% G-250 sample additive, except the digitonin (ThermoFisher Scientific, cat. no. 407565000), which was diluted to 5%. Briefly, NativePAGE™ 4 × sample buffer supplemented with digitonin (5%) was used to reach a final concentration of 1.5% (digitonin/protein ratio of 6 g/g) to solubilize the mitochondria. To ensure complete solubilization of the mitochondrial membranes, the mixture was incubated on ice for 15 min and resuspended every 5 min. The mixture was then centrifuged at 4 °C at 17,000× g for 30 min to remove the membranes. After centrifugation, 0.5% G-250 sample additive was added to the supernatant. Next, 40 μg or 20 μg of these samples were loaded onto 3–12% native gradient gels (Invitrogen, Carlsbad, CA, USA, cat. no. BN1001BOX); 40 μg was used for in-gel activity assay, and 20 μg protein was used for immunoblotting. Electrophoresis was conducted at 4 °C, applying 150 V with light blue cathode buffer until the proteins entered 1/3 into the gel (~30 min), when the voltage was increased to 200 V and the clear cathode buffer (50 mM Tricine, 15 mM Bis-Tris, pH 7.0) was used. The gel ran for ~1 h to achieve the desired separation. One of the gels used for Western blot was stained with GEL Blue (Biosciences, St. Louis, MO, USA, cat. no. 786-35G) for 30 min. Before imaging, the gel was washed with water twice for 5 min. The other gel was left unstained. The mitochondrial proteins on the native gel were then transferred to a PVDF membrane (Bio-Rad, cat. no. 1620177) via a wet transfer at 10 V overnight. To help transfer large proteins, the transfer buffer (25 mM Tris base 192 mM glycine, 20% methanol, 0.05% SDS, pH adjusted to 8.3) contained 0.05% SDS.

2.6. Mitochondrial Native Protein In-Gel Activity Assay (IGA)

In-gel activity (IGA) assay was conducted after the BN-PAGE as described in [38]. Briefly, gels were immersed in the respective complex assay buffer (10 mL) at room temperature. CcO and complex I (CI) activity assays were conducted on separate gels. Both reactions were carried out at room temperature for ~1 h with gentle shaking and images were recorded every 15 min. Before imaging again, the gels were washed twice for 5 min in water. The gels were then imaged again the following day after being washed in water overnight at 4 °C. Pictures obtained after 1 h of incubation were used for quantification to avoid oversaturated signals.

2.7. Preparation of Tissue Lysates for SDS-PAGE

Twenty mg of heart tissue was homogenized with manual grinding with 200 µL of RIPA Buffer (150 mM NaCl, 1% Nonidet P 40, 0.5% sodium deoxycholate, 0.1% SDS, 50 mM Tris-HCl, pH 8.0) with 1:100 protease inhibitor cocktail (PIC) for 5 min on ice. The samples were then sonicated using a Fisher Scientific Sonic Dismembrator Model 100 ultrasonicator (4 rounds of 5 s pulses) at 4 °C. Further, dilution in RIPA buffer was done if needed. Protein concentration was determined using the DC protein assay kit (Bio-Rad, cat. no. 5000111) according to the manufacturer’s protocol.

2.8. SDS-PAGE

For CcO activity normalization, leftover isolated mitochondria from oxygen consumption readings (4 µg) were added to SDS-loading buffer (4% SDS, 10% 2-mercaptoethanol, 20% glycerol, 0.004% bromophenol blue, 0.125 M Tris-HCl, adjusted pH to 6.8) and incubated for 20 min at 55 °C. Next, we loaded the samples onto mini-protean TGX precast gradient SDS-PAGE gels (4–15%) (Bio-Rad, cat. no. 4561085). In order to load all samples on the same gel, the first set was loaded into the gel, and the gel was run for 30 min at 100 V. The wells were then washed with fresh cathode buffer, and the next set of samples was loaded. The gel was then run for an additional 30 min at 100 V. This was followed by a wet transfer to a PVDF membrane (transfer buffer: 25 mM Tris base, 192 mM glycine, 20% methanol, pH adjusted to 8.3) at 10 V or rapid transfer using ready-to-assemble transfer kit (Bio-Rad, cat. no. 1704273) with Trans-Blot^®^ Turbo™ Transfer System (Bio-Rad, cat. no. 1704150) at 2.5 A for 7 min. For total protein SDS-PAGE, ~30 ug were added to SDS-loading buffer (4% SDS, 10% 2-mercaptoethanol, 20% glycerol, 0.004% bromophenol blue, 0.125 M Tris-HCl, adjusted pH to 6.8) and gently shaken at room temperature for 30 min. Samples were then boiled at 95 °C for 5 min. Next, the samples were loaded onto mini-protean TGX precast gradient SDS-PAGE gels (4–15%) (Bio-Rad, cat. no. 4561096) and run for 1 h at 120 V. This was followed by a wet transfer overnight to a PVDF membrane at 10 V.

2.9. Western Blot Analysis

A 15 min methanol wash step was performed to remove Coomassie dye in the cases where the membranes were stained. This was done to improve antibody binding. Membranes were washed 2 × 5 min with water and 2 × 5 min with TBST to remove methanol. Membranes were blocked with a blocking reagent (5% bovine serum albumin (BSA), 1% Tris-buffered saline with Tween-20: TBST, 0.1% Tween-20) for 45 min at room temperature. The primary antibody (Table 1) was added with gentle shaking for 90 min at room temperature or overnight at 4 °C. Membranes were washed twice for 5 min using TBST as the wash solution. Finally, membranes were incubated with the secondary antibody for 1 h at room temperature (Table 1). The membranes were imaged using Pierce™ ECL Western blotting Substrate (ThermoFisher, cat. no. 32106), and signals were recorded on ChemiDoc MP imaging system (Bio-Rad, Cat. no.12003154) or on an X-ray film. Production of our anti-pY304 CcO subunit I specific antibody is described in detail [16,38]

2.10. ATP Measurements

ATP levels were assessed using the bioluminescent method in combination with the boiling method as described [38]. To normalize the ATP concentrations, samples were added to RIPA buffer and subjected to further sonication. The samples were allowed to self-pellet for 1–2 min, and the supernatant was then combined with SDS-loading buffer (4% SDS, 10% 2-mercaptoethanol, 20% glycerol, 0.004% bromophenol blue, 0.125 M Tris-HCl, adjusted pH to 6.8). The samples were further incubated at room temperature while shaking for 15 min to ensure all the protein was denatured. Samples were then heated to 95 °C for 5 min and loaded onto an SDS-PAGE (10%) followed by an overnight wet transfer at 10 V. Western blotting was performed with β-actin.

2.11. RT-qPCR

RNA was isolated from ~50 mg of heart tissue using TRIzol^TM^ Reagent (Invitrogen, cat. no. 15596026) and PureLink^®^ RNA Mini Kit (Ambion, Waltham, MA, USA, cat. no. 12183018A), according to the manufacturer’s protocol. RNA concentration was determined using NanoDrop One (ThermoFisher, cat. no. ND-ONE-W) with high purity (A_280_/A_260_~2.0 for all samples). Total cDNA was generated from RNA using iScript cDNA Synthesis kit (Bio-Rad, cat. no. 1708891) according to the manufacturer’s protocol. qPCR reactions were conducted using PowerUp™ SYBR™ Green Master Mix (Applied Biosystems, Waltham, MA, USA, cat. No. A25776) based on the manufacturer’s protocol. Plates (Applied Biosystems, cat. no. 4346906) were run using the 7500 Fast Real-Time PCR System (Applied Biosystems) based on the Life Technologies system set up with 45 cycles.

Primer sequences for rat

GAPDH Forward 5′-CCCCAATGTATCCGTTGTG-3′; Reverse 5′-CTCAGTGTAGCCCAGGATGC-3′ [46]

TNFa Forward 5′-AAATGGGCTCCCTCTCATCAGTTC 3′ Reverse 5′ TCTGCTTGGTGGTTTGCTACGAC-3′ [47]

CXCL2 (IL-8 analogue) Forward 5′-CATTAATATTTAACGATGTGGATGCGTTTCA-3′ Reverse 5′-GCCTACCATCTTTAAACTGCACAAT-3′ [48]

IL-1β Forward 5′-CCAAGCACCTTCTTTTCCTTCA-3′

Reverse 5′-AGCCTGCAGTGCAGCTGTCTAA-3′ [49]

CCL2 Forward 5′-TAGCATCCACGTGCTGTCTC-3′

Reverse 5′-CCGACTCATTGGGATCATCT-3′ [50]

IL-10 Forward 5′-TTGAACCACCCGGCATCTAC-3′

Reverse 5′-CCAAGGA GTTGCTCCCGTTA-3′ [51]

2.12. Dephosphorylation Experiment

To perform the dephosphorylation experiment, the complexes were freed from the inner mitochondrial membrane using our BN-PAGE mitochondrial solubilization protocol (see Section 2.5). After centrifugation, the supernatant was removed and diluted with water to 1 mg/mL. The reaction was carried out in a Clark-type oxygen electrode (Oxygraph system, Hansatech, Norfolk, UK) set at 37 °C with FastAP Thermosensitive Alkaline Phosphatase (Thermofisher, cat. No. EF0652) according to the manufacturer’s protocol. Oxygen consumption was assessed in the presence of 20 mM ascorbate and 30 μM cytochrome c (Sigma, cat no. C3131) at the start of the reaction and 60 min into the reaction. The data were recorded in nmol/min/mg mitochondrial protein and analyzed with the OxyTrace+ software (version 1.0 Build 48, 2016, Hansatech Instruments Ltd., Amesbury, MA, USA). The remaining sample was run on an SDS-PAGE to normalize the measurements to mitochondrial amount and CcO amount.

2.13. Statistical Analysis

All data are presented as means ± SEM (standard error of the mean) values. To assess the differences between HSD and RD in both rat strains, i.e., Cohen diabetes-sensitive (CDs) and Cohen diabetes-resistant (CDr) rats, we conducted a two-way ANOVA followed by Tukey’s multiple comparisons test. These statistical analyses were carried out using GraphPad Prism Version 10.3.0 for macOS (GraphPad Software, Boston, MA, USA) with significance level set at p ≤ 0.05.

3. Results

3.1. Global Proteomic Remodeling and Inflammatory Signaling in the Hearts of Cohen Diabetic Rats

Although heart pathology is well characterized in the Cohen diabetic sensitive model, the underlying mechanisms are less clear. To identify pathways that may be altered by genetic background and diet and that lead to cardiac mitochondrial adaptations in the context of diabetic stress, we conducted unbiased data-dependent acquisition proteomic (DDA) analysis using cardiac tissue from CDr and CDs rat strains. While the Cohen diabetic rat model has been studied metabolically, a comprehensive comparison of their global cardiac proteomes under dietary stress has not previously been reported. Cardiac tissues were collected from both strains after feeding them a regular diet (RD) or a high-sucrose low-copper diet (HSD), a regimen known to promote type 2 diabetes (T2D) [40].

Figure 1A presents a heatmap of significantly altered proteins identified using permutation-based false discovery rate (FDR) correction. The most pronounced proteomic differences were observed between the HSD and RD groups in the CDs strain and between the strains, supporting their heightened susceptibility to dietary challenge in the CDs rats. Among the proteins associated with OXPHOS, components of cytochrome c oxidase (CcO, CIV) were, in particular, downregulated under HSD in both strains, such as CcO4I1 (Figure 1D; CcO4I1 plot) corresponding to ~49%, ~23%, and ~69% reductions in CDr-HSD, CDs-RD, and CDs-HSD, respectively, compared to CDr-RD. In addition, we confirmed previous studies reporting that the Ndufa4 gene is deleted in the CDs strain [52,53] (Figure 1D; NDUFA4 plot). Although NDUFA4 was originally annotated as a Complex I (CI) subunit, it is now recognized as an accessory subunit of CcO [54]. NDUFA4 is not necessary for CcO monomer assembly into supercomplexes. However, its deficiency has been associated with Leigh syndrome and metabolic disorders, and it could lead to lower CcO stability and activity [55,56]. In contrast, MT-ND1, a core mitochondrial-encoded subunit of CI located in the proton-pumping P-module of CI, was upregulated under HSD (Figure 1D; MT-ND1 plot), showing an opposite tendency to CcO4I1. It is important to note that, ATP synthase (complex V) subunits did not show a change in their expression pattern, as shown in Figure 1D; ATP5F1A plot.

Interestingly, we identified switching of tissue-specific CcO isoforms by comparing heart-type (CcO7A1, CcO6A2) and liver-type (CcO7A2, CcO6A1) subunits (Figure 1D; CcO7A2/CcO7A1 and CcO6A1/CcO6A2 plot). The ratio of liver-type/heart-type revealed similar regulation patterns for CcO7A and CcO6A. For both isoform pairs, the ratio showed a shift toward the liver-type isoforms in the CDs-HSD group. The liver-type isoforms represent the fetal versions of the subunits, and the isoform switching would be consistent with the pattern seen for other genes during heart failure when fetal isoform genes are induced [57].

Several key enzymes involved in mitochondrial β-oxidation were consistently upregulated in CDs hearts, including ACADM (medium-chain acyl-CoA dehydrogenase, mitochondrial), ACAA2 (acetyl-CoA acyltransferase 2), HSD17B4 (17β-hydroxysteroid dehydrogenase type 4, peroxisomal), and ECI1 (enoyl-CoA delta isomerase 1, mitochondrial) (Figure 1D, ACADM, ACAA2, HSD17B4, and EC11 plot). This pattern was further amplified under HSD, suggesting a strain-dependent enhancement of lipid catabolism in response to dietary challenge. These results indicate a substrate shift toward medium and longer-chain fatty acids, a sign of increased fatty acid accumulation in the cardiac tissue, which potentially causes lipotoxicity and diabetic cardiomyopathy [27,28,29].

Notably, the upregulation of AK1 (adenylate kinase isoenzyme) (Figure 1D; AK1 plot) is an indicator of altered cellular energy state and increased adenylate turnover in the CDs strain, as the upregulation occurred in a strain-dependent manner, possibly caused by a lower ATP/ADP ratio and increased ATP demand in the CDs cardiac tissue.

Additionally, Venn diagram analysis (Figure 2C) comparing proteins significantly altered by diet alone versus strain alone uncovered distinct, non-overlapping sets of proteins associated with each factor, suggesting unique molecular responses to genetic background and dietary stress. The corresponding protein lists are provided in Table S1.

Pathway enrichment analysis of the biological processes predominantly enriched includes fatty acid metabolism, OXPHOS, and mitochondrial redox buffering systems (Figure 1B). Additionally, pathways associated with cardiac dysfunction, such as diabetic cardiomyopathy and cardiac muscle contraction, were significantly enriched when considering both diet and strain as combined factors (Figure 1B,C). To disentangle the individual contributions of diet and strain, we next performed pathway enrichment analyses using only proteins significantly altered by diet or strain alone. The results are summarized in Figure 2A–C and Supplementary Figure S1. Diet-driven enrichment revealed strong activation of pathways related to cardiac muscle contraction, diabetic cardiomyopathy, and OXPHOS dysregulation. Notably, acute-phase response protein haptoglobin (Hp), copper-binding protein ceruloplasmin (Cp), copper-binding glycoprotein, and heme-binding protein hemopexin (Hpx) were altered under HSD conditions, suggesting increased oxidative stress and disturbances in iron, copper, and heme homeostasis. Additionally, the upregulation of thiol protease inhibitor kininogen-1 (Kng1), involved in coagulation and inflammation, along with serpina3n, a serine protease inhibitor A3N linked to immune modulation in acute-phase response and response to cytokines, indicates a diet-induced pro-inflammatory response (heatmaps specific to diet effect vs. strain effect are shown in Figure 2A,B).

When focusing on strain-specific effects, we observe distinct enrichment in pathways related to fatty acid degradation, core metabolic processes (e.g., carbon and propionate metabolism), and modest enrichment of diabetic cardiomyopathy-related pathways, though to a lesser extent than the diet effect alone (Supplementary Figure S1A,B).

Principal Component Analysis (PCA) reveals that each animal group forms a distinct cluster, indicating that both diet and strain significantly influence the proteomic profile. Notably, animals on the HSD cluster toward the top of the plot, reflecting higher PC2 values, whereas their RD counterparts are positioned lower on the PC2 axis. Additionally, the CDs-HSD group shows a more pronounced shift in proteomic profile compared to its RD counterpart, suggesting that the CDs strain exhibits a greater response to dietary changes than the CDr strain (Supplementary Figure S1C).

Together, our findings suggest that the CDs strain exhibits an inherent genetic susceptibility in protein pathways related to dietary stress, marked by disrupted OXPHOS function and enhanced fatty acid oxidation. When isolating the effects of diet, we observed a consistent molecular signature of mitochondrial dysfunction, alterations in cardiac muscle-specific pathways, and activation of inflammatory and stress response pathways.

Based on the proteomic data and to further investigate the presence of inflammation in our animal model, we conducted RT-qPCR using total cDNA generated from heart RNA of CDr and CDs animals. Our results show that HSD increases TNF⍺ endogenous mRNA levels in both CDr and CDs strains’ heart tissue about 4-fold on HSD compared to their respective RD groups (Figure 2D; TNF⍺ plot), with diet being the only significant contributing factor. CXCL2 (IL-8 homolog) mRNA levels showed a 3-fold increase in CDr-HSD compared to RD and no statistically significant increase in CDs-RD or HSD, indicating both diet and strain as contributing factors (Figure 2D; CXCL2 plot). IL-1β mRNA levels increased approximately 6-fold in CDs-HSD compared to RD (Figure 2D; IL-1β plot). The results show that both factors are significant, with diet being the most influential. CCL2 exhibited a 5-fold increase in CDr-HSD versus RD and a 3-fold increase in CDs-HSD compared to RD (Figure 2D; CCL2 plot). Diet was the highest contributor for CCL2. In addition to pro-inflammatory cytokines, we also tested the anti-inflammatory cytokine IL-10 and found no change in its mRNA levels (Figure 2D; IL-10 plot). These findings indicate that the HSD increases pro-inflammatory signaling in both CDr and CDs animals. However, it appears that the strain is a contributing factor in determining which specific cytokine or interleukin is increased.

3.2. Complex IV Activity and ATP Decline in a Diet- and Strain-Dependent Manner

Due to its central role in mitochondrial respiration, CcO is tightly regulated via tissue-specific isoforms, allosteric control, and PTMs [12,16]. We previously demonstrated that CcO activity is significantly reduced in pancreatic β-cells of normoglycemic CDs rats fed a RD. When exposed to HSD, CcO function was further diminished, leading to reduced insulin secretion and diabetic onset [31,39]. Similarly, in the liver of both CDr and CDs rats, HSD resulted in reduced CcO activity, increased inhibitory phosphorylation of the dimeric form of CcO, and decreased ATP production [38]. However, the potential impact of HSD-driven diabetic maladaptation of mitochondrial function in other vital tissues, such as the heart of CDr and CDs rats, remained unknown.

To examine the impact of diet and strain variation on mitochondrial heart function, we first measured the oxygen consumption rate (OCR) of CcO after the purification and solubilization of mitochondria using a Clark-type electrode (Figure 3A,B). When CcO activity was normalized to total mitochondrial content (TOM20), CDr-HSD and CDs-HSD animals showed ~65% and ~54% lower OCR, respectively, compared to their RD-fed controls (Figure 3A). To determine whether the observed decrease in OCR was also due to functional inhibition of CcO enzyme activity rather than merely a reduction in the overall amount of CcO, we normalized CcO activity to the CcO enzyme amount (Figure 3B). OCR reduction persisted and we observed a ~65% decrease in CDr rats fed an HSD compared to CDr-RD rats. Interestingly, CDs rats showed strongly decreased CcO activity independent of diet, similar to that seen in CDr rats on an HSD. Strain was the highest contributor, but diet and their interaction were also statistically significant contributors to changes in CcO-specific activity.

Immunoblotting revealed substantial reductions in CcO content in both HSD-fed groups (Figure 3C). Specifically, CDr-HSD and CDs-HSD hearts showed 55% and 45% less CcO protein, respectively, compared to their RD counterparts. Notably, even under RD conditions, CDs rats had ~30% lower CcO content than CDr, indicating a strain-specific difference independent of diet (Figure 3E). Both loss of CcO protein and enzymatic inhibition contributed to the observed respiratory decline. Consistent with these findings, cardiac ATP levels were markedly reduced. CDr-HSD hearts exhibited an ~85% reduction compared to RD, while both CDs-RD and CDs-HSD showed ~90% ATP depletion relative to CDr-RD (Figure 3F). Levels of ATP synthase remained unchanged (Figure 1D and Figure 3D ATP5F1A plot from proteomic analysis), suggesting impaired proton flux, rather than changes in ATP synthase expression, as the primary driver of ATP depletion. These data support the concept that both genetic predisposition and dietary challenge contribute to compromised mitochondrial respiratory capacity. The measured CcO activity reduction cannot only be explained through CcO content reduction but also through a decrease in CcO enzymatic turnover rate.

3.3. Tyrosine Phosphorylation Accumulates on CIV Within Dimers and Supercomplexes and Suppresses Regulation and Reorganization

Consistent with our RT-qPCR data, it has been shown that T2D in rats leads to a chronic low-grade inflammatory state [58]. We previously found that treating liver tissue with proinflammatory cytokine TNF⍺ led to decreased CcO activity that correlated with an increase in Y304 phosphorylation on CcO subunit I [17]. CcO-I Y304 phosphorylation inhibits enzyme function and induces sigmoidal kinetics typically associated with dimerization [16,17]. We also previously reported elevated Y304 phosphorylation and reduced CcO activity in the livers of diabetic Cohen rats [38]. In this study, we extended our investigation to the heart to determine whether Y304 phosphorylation is specific to the liver or if it is also present in the heart, suggesting it may play a more global role in mitochondrial dysfunction in more than one tissue.

Using BN-PAGE followed by immunoblotting (Figure 4A,C) and in-gel activity (IGA) assays (Figure 4B,D; note that brown color indicates CIV activity and violet color indicates CI activity and not protein amount), we assessed CcO phosphorylation and function in cardiac mitochondria. A custom phospho-specific antibody targeting CcO-I-pY304 revealed significant increases in phosphorylation across multiple conditions (Figure 4A,C). Compared to CDr-RD, dimeric CcO showed a 120% increase in CDr-HSD, a 150% increase in CDs-RD, and a ~400% increase in Y304 phosphorylation in CDs-HSD (Figure 4E).

We also detected inhibitory phosphorylation of CcO in high-molecular-weight supercomplexes (SC_pY304_) (Figure 4F). Notably, SC_pY304_ was already present in CDs-RD, suggesting a “primed” phosphorylation state even in the absence of dietary stress. Functionally, Y304 phosphorylation was strongly associated with enzymatic suppression (Figure 4B,D). CcO activity in SC_pY304_ complexes was decreased in CDs-HSD hearts (50%) (Figure 4H), while dimeric CcO activity declined by 25% in CDr-HSD and 42% in CDs-HSD compared to RD-fed counterparts (Figure 4I). Monomeric activity was also reduced in both HSD-fed groups (Supplementary Figure S3A). These findings support the hypothesis that Y304 phosphorylation inactivates CcO, particularly within supercomplexes in heart mitochondria (Figure 4A,C,H). In contrast, Complex I (CI) activity increased significantly under the same conditions. CDs-HSD rats exhibited a 54% increase in monomeric CI activity relative to CDr-RD (Supplementary Figure S3C), along with elevated CI activity in SC_pY304_-containing bands (Supplementary Figure S3G). Additionally, CI activity in I_2_ + III_2_ supercomplexes rose by 25% and 30% in CDs-RD and CDs-HSD, respectively, compared to their CDr counterparts (Supplementary Figure S3H). These results suggest that increased CI activity may act as a compensatory response to CIV dysfunction.

To further explore the relationship between CcO function and total pY304 status, we performed correlation analysis and unsupervised k-means clustering using total CcO-I pY304 and CcO activity across all groups (Figure 4G). Three distinct metabolic clusters emerged: (1) CDr-RD, (2) CDr-HSD + CDs-RD, and (3) CDs-HSD. These clusters reflected a graded loss of mitochondrial function associated with increased phosphorylation. Importantly, in CDs-HSD, total phosphorylation predicted CcO inhibition (R^2^ = 0.735), supporting a mechanistic link between PTM level and respiratory dysfunction. A moderate correlation was observed in CDr-HSD (R^2^ = 0.441). Conversely, in CDs-RD (R^2^ = 0.010), phosphorylation did not correlate with CcO activity, suggesting that genetic predisposition alone is insufficient to drive respiratory impairment through inhibitory phosphorylation. Together, these findings support a model in which both strain susceptibility and dietary stress converge to promote phosphorylation-dependent CcO inhibition.

To confirm that phosphorylation of Y304 was leading to the decrease in CcO activity, we performed a dephosphorylation experiment using alkaline phosphatase and isolated mitochondria. To directly access oxygen consumption and CcO activity, the dephosphorylation experiment was performed in the Clark-type electrode. CcO maximum turnover was assessed at time point zero (T0) and 60 min (T60) into the reaction. When normalized to CcOIV levels, we observed a significant increase in CcO activity at T60 in CDr-HSD (~80%), CDs-RD (~50%), and CDs-HSD (~105%) compared to T0. This suggests that the elevated phosphorylation in the animals fed a HSD was leading to a decrease in CcO activity, and by removing the phosphorylations, some of the CcO activity could be restored (Figure 4J–M). However, we were unable to recover the decrease due to the genetic predisposition (with decreased CcO levels) present in the CDs animals. Finally, the dephosphorylation reaction and alignment of pY304 immunoblots with CcO IGA demonstrated that elevated phosphorylation coincided with gel regions of CcO activity loss (Figure 4A,C,J,K).

To determine the molecular composition of the novel SC_pY304_ band observed in Figure 3B,D,F and Supplementary Figure S3A,B, we performed BN-PAGE followed by immunoblotting using a combination of phospho-specific and fluorescently multiplexed antibodies. Although technical limitations prevented multiplexing of our anti-CcO-I-pY304 antibody, due to the high hydrophobicity of CcOI, we successfully aligned the pY304 signal with the total CcO-I on the same blot, enabling accurate positional mapping (Supplementary Figure S5A).

To identify co-migrating complexes, we performed dual-channel fluorescent immunoblotting for CcO-I together with either CORE1 (Complex III) or NDUFS3 (Complex I) (Supplementary Figure S5B,C). Notably, SC_pY304_ co-migrated with NDUFS3-positive bands, confirming the presence of Complex I. In contrast, no CORE1 signal was detected in this region, indicating the absence of Complex III. The SC_pY304_ band partially overlapped with the migration zone of the canonical I + III_2_ supercomplex but lacked CIII content, suggesting that this structure represents a potential non-canonical I + IV_1_/2 supercomplex and/or IV_n_ aggregates that may be formed under stress conditions such as phosphorylation or high-sucrose exposure (Supplementary Figure S5A–C).

We next examined how strain and diet affected CI assembly and its distribution across SCs. Immunoblot analysis of NDUFS3 revealed an increase in monomeric CI as well as in the putative I + IV_1_/2 supercomplexes (Supplementary Figures S5D and S6C quantification is summarized in the visual map image in Supplementary Figure S5G). The I_2_ + III_2_ supercomplex also increased by over 200% in both CDs-RD and CDs-HSD compared to CDr-RD, while no significant change was observed between CDr-RD and CDr-HSD (Supplementary Figure S5G, quantified in Supplementary Figure S3F). Total CI levels were elevated in the CDs strain relative to CDr-RD (quantified in Supplementary Figure S6G; the two-way ANOVA results are summarized in Supplementary Table S2).

For CIII, CORE1 immunoblotting showed increased levels of the CIII_2_ dimer in CDs rats compared to CDr (Figure 4E, quantification in Supplementary Figure S6H). Interestingly, the CIII_2_ + IV complex was reduced in CDs-RD animals (Supplementary Figure S6I), despite no significant change in total CIII levels (Supplementary Figure S6N). Additionally, CIII content within the I_2_ + III_2_ supercomplex increased by ~85% in CDs-HSD compared to CDr-RD (Supplementary Figure S6M). These findings suggest that strain is the predominant factor influencing CI and CIII organization.

Together, these results reveal that both genetic background and dietary environment contribute to the remodeling of mitochondrial respiratory complexes. Specifically, they support the emergence of a stress-induced, non-canonical I + IV_1/2_ supercomplex and/or IV_n_ aggregates (SC_pY304_) lacking Complex III and altered SC stoichiometry involving Complex I and III, particularly in the diabetes-prone CDs strain.

3.4. High Sucrose Diet Induces Changes in Mitochondrial Biogenesis, Antioxidants, and Quality Control Mechanisms

Based on our proteomic data and qPCR, it appears that there is an increased susceptibility to genetic stress and OXPHOS dysfunction. Stress, including cardiovascular disease and T2D, can contribute to dysfunctional mitochondrial quality control as seen by changes in biogenesis, mitophagy, and autophagy [59]. To further investigate this, we used total protein lysate from heart tissue for immunoblotting with different quality control markers to see if there was an effect caused by differences in strain and/or diet (Figure 5A,H). We first investigated markers of quality control linked to mitochondrial dynamics (fission and fusion) and autophagy/mitophagy. We found an increase in fusion (Mitofusin 2, MFN2, and OPA1) and a decrease in fission markers (active Dynamin-related protein 1, pS616 DRP1) in the CDs strain, indicating impaired mitochondrial dynamics (Figure 5A). For MFN2, there was a 60% increase in CDs-RD compared to CDr-RD, and a 90% increase in CDs-HSD compared to CDr-RD (Figure 5C). The increases in MFN2 are statistically significant when normalized to VDAC (Supplementary Figure S7A), indicating a mitochondrial mass-independent increase. For OPA1, although not statistically significant, there was a trend of increased protein levels in the CDs strain (Figure 4D). For active DRP1 (p-S616 DRP1), there was a ~40% decrease when comparing CDs-RD and CDs-HSD to CDr-RD (Figure 5E). The results show that only the strain contributes to these alterations.

We next examined markers associated with autophagy and mitophagy. There was a decrease in autophagy markers (microtubule-associated protein light chain 3, LC3) and an increase in mitophagy pathways (sequestosome 1, SQSTM1/p62), suggesting impaired mitochondrial turnover. We found a decrease in the ratio of LC3 II to LC3 I (Figure 5F), with the only statistically significant decrease being a 75% drop in CDs-HSD compared to CDr-RD. There was also an increase in SQSTM1/p62 in the CDs strain (Figure 5G), with an 82% increase in the CDs-RD compared to the CDr-RD group and a 78% increase in CDs-HSD compared to CDs-RD group. In summary, there is an increase in mitophagy signaling but a decrease in autophagy in the CDs group under RD and HSD with the CDs-HSD group showing the most significant changes. This could indicate that mitochondria are labeled for degradation, but degradation is not executed, leading to the accumulation of damaged mitochondria.

To study mitochondrial biogenesis, we analyzed mitochondrial transcription markers and observed an increase in mitochondrial transcription factor A (TFAM) and mitochondrial biogenesis marker (peroxisome proliferator-activated receptor gamma coactivator 1-alpha, PGC1⍺), but no change in mitochondrial mass (voltage-dependent anion channel, VDAC) (Figure 5H). Surprisingly, TFAM was increased 60% and 25% in the CDs-HSD group compared to CDr-RD and CDs-RD, respectively (Figure 5I). No significant change was observed in the CDr rats on RD and HSD (Figure 5H,I). However, when TFAM protein levels were normalized to VDAC, there was no increase compared to CDr-RD, suggesting that the increase in TFAM was due to an increase in mitochondrial mass. There was a trend for increased VDAC amounts in CDs-HSD compared to CDr-HSD hearts (p = 0.097) (Figure 5J). Additionally, we found a strong 385% and 60% increase in PGC1⍺ protein levels in CDs-HSD compared to CDr-RD and CDs-RD, respectively, and a 145% increase in CDs-RD compared to CDr-RD (Figure 5K). No differences were noted between the CDr-HSD and CDr-RD groups.

Lastly, we found changes in antioxidant levels. Superoxide dismutase 1 (SOD1), a cytoplasmic ROS scavenger, showed a trend (45% decrease; not significant) in CDs-HSD compared to CDr-RD, and both the CDr-HSD and CDs-RD showed a trend (20% decrease, not significant; Figure 5L). In contrast, superoxide dismutase 2 (SOD2), which is a ROS scavenger localized to the mitochondrial matrix, was increased by 70% in CDs-HSD compared to CDs-RD, whereas the CDr-RD, CDr-HSD, and CDs-RD groups showed no statistically significant differences (Figure 5M). 4-HNE is a marker for lipid peroxidation, which reflects the levels of oxidative damage. 4-HNE levels were only significantly increased in CDs-RD compared to CDr-RD (150%) (Figure 5N). The higher levels of SOD2 in CDs-HSD may account for the lower levels of 4-HNE signal. Interestingly, the strain factor was the primary contributor to the observed differences in mitochondrial biogenesis and antioxidant markers.

These findings suggest that CDs animals have increased fusion, decreased fission, and defective autophagy (summarized in Figure 5B). Altogether, there are strain-dependent alterations in mitochondrial biogenesis, antioxidant levels, mitophagy, and autophagy that may contribute to diabetic cardiovascular disease.

4. Discussion

Diabetes is among the top risk factors for cardiovascular disease. This vulnerability is evident in diabetic cardiomyopathy, a condition characterized by left ventricular dysfunction, cardiac remodeling, and an elevated risk of heart failure independent of conventional cardiovascular risk factors [60,61]. Diabetic cardiomyopathy can occur in both type 1 and type 2 diabetes, with type 2 diabetes accounting for 90–95% of cases [62]. Furthermore, mitochondrial dynamics, including biogenesis, fragmentation, and mitophagy, are altered in diabetic cardiomyopathy, exacerbating impairments in cardiac efficiency and function [63,64,65,66,67]. It is therefore essential to reveal diabetes-induced molecular changes to identify enzymes and pathways that can be targeted to improve cardiac function.

Fatty acids are critical oxidation substrates for the heart [26]. Our proteomics analysis revealed that genetic background profoundly shapes how cardiac mitochondria respond to dietary stress in the Cohen model. In the diabetes-sensitive CDs strain, core enzymes of mitochondrial and peroxisomal β-oxidation, ACADM, ACAA2, HSD17B4, and ECI1, were consistently upregulated, which may suggest increased utilization of fatty acids; however, this can also occur as an adaptive response to impaired fatty acid oxidation. This remodeling was further amplified under HSD, suggesting a diet-responsive enhancement of fatty-acid catabolism. This pattern is consistent with prior reports showing enhanced cardiac fatty-acid oxidation in diabetes and insulin resistance [68,69,70]. Such upregulation may represent a compensatory mechanism to handle elevated lipid availability and prevent the accumulation of toxic lipid intermediates. This adaptation in CDr hearts could explain their increased resilience in comparison to CDr-HSD liver mitochondria. However, in addition to changes in protein levels, regulation of fatty acid oxidation also depends on post-translational modifications, such as acetylation of important enzymes such as long-chain acyl-CoA dehydrogenase, which can inhibit long-chain fatty acid metabolism [71,72]. Further studies are needed to define the specific alterations of the fatty acid metabolic pathways in the Cohen model. We also validated previous studies that found a deletion of the Ndufa4 gene in the CDs strain [52,53] and found further diet and strain-specific maladaptations as contributing factors explaining the pathologies found in the Cohen diabetic rat model. The loss of NDUFA4 (also referred to as COXFA4 [73]) was found to reduce CcO activity [54].

Proteins involved in the acute-phase response and redox homeostasis were dysregulated under HSD, including Hp (haptoglobin), Cp (ceruloplasmin), and Hpx (ceruloplasmin). These molecules are critical for iron and heme scavenging (Hpx) [74], copper transport (Cp) [75], and oxidative damage mitigation (Hp) [76], suggesting that the HSD induces systemic and cardiac oxidative stress and perturbs metal homeostasis. Moreover, the upregulation of Kng1 (kininogen-1) and serpina3n (serine protease inhibitor A3N), two proteins linked to coagulation, inflammation, and cytokine signaling, further supports a diet-induced pro-inflammatory state. Together, these findings indicate that HSD imposes a combined oxidative, inflammatory, and metal-dyshomeostasis burden on the cardiac tissue.

We showed previously that in vitro treatment of liver and pancreatic tissue with TNFα and IL-1β leads to CcO inhibition [17,31]. We propose that HSD induces an inflammatory state, triggering a signaling cascade that activates an unknown mitochondrial tyrosine kinase to phosphorylate CcO, resulting in its dimerization and subsequent inactivation in heart tissue. TNFα and IL-1β are implicated in the development of atherosclerotic plaques and heart failure [77,78], suggesting that these signaling pathways play an important role in the pathology of diabetic hearts. We also explored the role of cytokines such as CCL2 and CXCL2 that are involved in early and late inflammatory responses. CCL2 is associated with chronic low-grade inflammation frequently observed in T2D, while CXCL2 is linked to acute inflammation [79]. Our analysis indicates that within the Cohen rat model, the diabetes-sensitive CDs strain on an HSD shows patterns corresponding to later stages of heart failure and the manifestation of a type 2 diabetic phenotype, whereas the diabetes-resistant CDr strain on the same diet exhibits traits indicative of early-stage heart disease caused by mitochondrial dysfunction despite still being normoglycemic after six weeks on an HSD. Accumulations of p62 and decreased levels of LC3-II in unstable atherosclerotic plaques signify impaired mitophagy, which is also evident in diabetic cardiomyopathy, where changes in cardiac composition contribute to heart failure [59]. Other studies have found that PGC1⍺ levels are increased in T2D, as it can directly increase cardiac fatty acid oxidation [28,69].

In previous studies, we demonstrated that Y304 in the catalytic subunit I of CcO acts as a regulatory switch through phosphorylation [16,17]. Phosphorylation at this site inhibits CcO and produces pronounced sigmoidal kinetics [16] as a means to regulate and fine-tune enzymatic activity. We recently reported that the livers of Cohen diabetic rats exhibited decreased CcO activity alongside increased Y304-CcO-I phosphorylation, particularly in the dimeric form of CcO within the diabetic liver tissue [38]. Interestingly, comparing those results with this study shows a distinct impact of diet and strain when contrasting mitochondria from the liver and heart. This underscores the different susceptibility of various tissues to diet and strain as disease modifiers [31]. Notably, CDr livers [38] seem more negatively affected than CDr hearts. In liver mitochondria, there was an 80% reduction in OCR for CDr rats on HSD compared to RD [38], while heart mitochondria in this study showed a 65% decline when comparing CDr-RD with CDr-HSD. Both liver and heart are affected by insulin resistance; however, heart tissue appears to have a somewhat protective mechanism against the inactivation of CcO in the context of T2D and inflammation. The liver is more severely affected in its metabolic function, while cardiac tissue is primarily structurally impacted and found to be more resistant to metabolic changes under hyperglycemic conditions [80]. These differences between the heart and liver may be explained by their distinct energy metabolism. Mitochondrial fatty acid oxidation constitutes about 60–80% of the mitochondrial substrate for cardiac ATP production [81], whereas liver, in the fed state, relies predominantly on glucose metabolism [82].

pY304-CcO-I accumulated not only in the dimeric CcO, but also in a subset of SCs—specifically, IV_n_ or I + IV_1/2_. IV_n_ has previously been reported [83]; however, they run much lower (between IV_2_ and III_2_ + IV) compared to our gel region. The aggregates reported in our study could contain more copies of CIV (5–6 copies based on molecular weight). The latter SC was recently reported in two independent studies [84,85] and discussed by Cogliati et al. [6]; however, the function of this specific supercomplex remains unknown. Similar to our findings, I + IV was reported to co-migrate alongside free Complex I and I + III_2_, and I + IV_2_ runs close to I + III_2_. In this study, we show that both dimeric CcO (IV_2_) and I + IV_1-2_ undergo increased phosphorylation in diabetic conditions (IV_2 pY304_ and I + IV_1/2 pY304_ or IV_n_), resulting in an inhibition of CcO activity. This is paralleled by increased activity and protein levels of I_2_ + III_2_ and monomeric complex I, predominantly observed in diabetes-sensitive strain CDs on an HSD. Recent reports suggest that the formation of I_2_ + III_2_ provides protection against ischemic heart failure by reducing ROS production and stabilizing mitochondrial function [2]. Additionally, the structural integrity of SCs, such as complex I + IV and I + IV_2_ raises questions regarding potential structural in addition to functional roles, highlighting the intricacies of the mitochondrial ETC architecture [83]. pY304-CcO-I is known to correlate with reduced enzymatic activity, and its accumulation within mitochondrial supercomplexes suggests that a subset of SCs may become selectively inhibited through this PTM under inflammatory conditions, such as that which is seen in diabetes.

In addition to PTMs and SCs, tissue-specific subunit isoforms of CcO finetune CcO activity. In human heart and skeletal muscle, there are two known tissue-specific isoforms, CcO6A2 and CcO7A1 (the “heart-type” isoform, also called COX6AH and COX7AH), which are distinct from their ubiquitously expressed counterparts, CcO6A1 and CcO7A2 (the “liver-type” isoform, also known as COX6AL and COX7AL; note that the numbering of the heart and liver isoforms is the opposite for CcO subunits 6A and 7A) [12]. In rodents and other mammals but not in humans, there is a third heart/liver isoform pair of the smallest CcO subunit 8 [13,86]. These subunits experience a developmental transition from the ubiquitous liver and fetal isoforms to the heart-type isoforms during early postnatal skeletal muscle maturation in humans and mice. The heart/muscle isoforms become the main forms by around three months in human infants and by four weeks in mice [87]. Heart-type isoforms of CcO have been linked to increased mitochondrial capacity compared to liver-type [12,13]. CcO6a2 knockout mice have cardiac dysfunction [21], and, interestingly, skeletal muscle is protected from insulin resistance and glucose intolerance [88], but no further in-depth mitochondrial studies were carried out for this heart isoform. We previously generated CcO7A1-deficient mice and reported an ~50% reduction in total CcO activity, and the mice presented with dilated cardiomyopathy [20]. Interestingly, Cox7a1 knockout mice compensate for loss of CcO7A1 by upregulating the liver-type CcO7A2 isoform, indicating that an isoform switch is possible. We also reported that loss of CcO7A1 specifically impairs dimeric and higher-order SCs-associated CcO activity in mouse hearts, while monomeric CcO activity remained preserved [89]. The compensatory increase in CcO7A2 incorporation in Cox7a1 knockout mice led to rearrangements that favor CI-containing SCs, which seems to help reduce respiratory dysfunction [89].

Interestingly, in CDs-HSD hearts, we observed a switch from the heart-type to the liver-type isoforms for subunits 6A and 7A (our MS analysis did not show fragments of subunit 8, likely because it is very short and hydrophobic), following the pattern of isoform transitioning to the fetal versions seen in heart failure [57], likely driven by inflammatory signaling causing fibrosis and structural and functional changes in the heart [90]. CDs rats lack NDUFA4 and simultaneously exhibit a marked reduction in CcO7A1, therefore, CcO dimers are expected to be structurally unstable, and the increase in phosphorylation detected on CcO dimers in the sensitive HSD animals suggests an adaptive response to ETC stress. Dimer-selective phosphorylation has been reported as a stress-responsive regulatory mechanism that slows electron flux and may stabilize residual dimeric complexes [16,17,38]. This aligns with the model proposed by Ramzan et al., in which cardiac CcO transitions between monomeric and dimeric states via phosphorylation, and inhibitory dimer phosphorylation is reversed upon CcO monomerization [91]. Thus, the decrease in CcO7A1 may represent an attempt to shift CcO towards the more active monomeric form. The increased phosphorylation observed here may reflect a compensatory regulation of structurally compromised CcO dimers under the diabetic condition. In summary, we propose that isoform switching and lack of NDUFA4 may compromise the structural integrity required to form functional CcO dimers. As a result, increased expression of the liver isoforms cannot rescue CcO dimer dysfunction, leading instead to an accumulation of hyperphosphorylated dimers. This state is promoted by inflammatory signaling, resulting in synergistic pathological effects caused by genetic predisposition and diabetic mitochondrial remodeling.

We also found a reduction in CcO protein content when the Cohen rats were given the HSD. It should be noted that the HSD is also low in copper, which may in part explain lower CcO levels because copper plays an important role in CcO biogenesis and function [40]. However, SOD1 is also a copper-dependent protein, and its levels remained unchanged under the HSD, which suggests that factors other than the low copper diet may contribute to a decrease in CcO amount, such as changes in the activity or amount of mitochondrial proteases, which should be explored in future studies.

Given the need of cardiac tissue for increased ATP production, quality control mechanisms become critical. A balance between mitochondrial fission and fusion and mitophagy are essential for maintaining cellular health, particularly in the heart [63,92], where high ATP demand prevails. We measured an increase in TFAM levels in CDs-HSD compared to CDr-RD and HSD. However, we did not see an increase in mitochondrial mass or replication. It has been reported that strong increases in TFAM levels can be deleterious, specifically in heart tissue, where high TFAM-to-mtDNA ratios can lead to decreased mtDNA expression, resulting in dysfunctional OXPHOS and mitochondria [93]. The increase in TFAM in CDs-HSD could further exacerbate mitochondrial dysfunction already present, amplifying the diabetic phenotype.

5. Conclusions

Collectively, our findings describe a multi-layered mechanism by which genetic predisposition and dietary stress converge to impair mitochondrial metabolism (Figure 6). In the diabetes-sensitive strain, the high-sucrose diet exacerbated mitochondrial dysfunction by amplifying inhibitory phosphorylation of CcO and altering respiratory supercomplex organization. These defects coincided with impaired fatty-acid oxidation and compromised mitochondrial quality control and isoform switching, together contributing to energy depletion. Further studies are needed to elucidate the upstream signaling pathways that mediate CcO phosphorylation and supercomplex remodeling that may uncover new therapeutic targets to restore mitochondrial efficiency and prevent diabetic cardiomyopathy.

6. Limitations

A limitation of our study is the use of only male Cohen rats. We previously investigated the sex differences in Cohen rats [30], the male rats exhibited more severe expression of the metabolic diabetic phenotype on the HSD, and that is why they were the focus of our current study. Future work should be done to include the use of female rats and investigate potential sex-specific differences in mitochondrial signaling and regulatory pathways. Moreover, extending these analyses to additional type 2 diabetes models would help establish the generality of these findings beyond the Cohen strain model. Another limitation was the use of qPCR only for markers of inflammation that oftentimes, but not always, correlate with changes in protein levels [94,95]. Finally, future studies should involve measurements with intact mitochondria such as respiratory control ratio and mitochondrial substrate oxidation rates to better understand mitochondrial coupling and bioenergetics.