Impaired mitochondrial proteostasis underlies cardiomyocyte bioenergetic dysfunction in a murine 3-hit model of heart failure with preserved ejection fraction

Abstract

Heart failure with preserved ejection fraction (HFpEF) predominantly affects older individuals with multiple comorbidities. Although mitochondrial dysfunction appears to be a major contributor to diastolic dysfunction, the mechanisms underlying it during HFpEF remain unclear. We used a clinically relevant two-hit HFpEF model in aged mice to investigate remodeling of cardiomyocyte mitochondrial proteins and to determine whether mitochondrial dysfunction reflects impaired biogenesis or defective mitochondrial proteostasis. HFpEF was induced in aged (15-month-old) male and female mice using a high-fat diet combined with chronic nitric oxide synthase inhibition (L-NAME). This model recapitulated key features of human HFpEF, including preserved ejection fraction with diastolic dysfunction and impaired cardiac reserve in both sexes. Consistent with impaired energy reserve, mitochondrial respiratory capacity measured by high-resolution respirometry (Oroboros) was significantly reduced in left ventricular muscle fibers. To examine cardiomyocyte-intrinsic mechanisms, mitochondria were isolated from cardiomyocytes using Langendorff perfusion and analyzed by proteomics and pathway enrichment analysis. Proteomic profiling revealed broad suppression of mitochondrial proteins that support oxidative metabolism and organelle maintenance, including those involved in oxidative phosphorylation, carbohydrate metabolism, and the TCA cycle. In contrast, key regulators of mitochondrial biogenesis were preserved. Notably, mitochondrial chaperones and proteases for protein quality control were reduced, leading to increased accumulation of ubiquitinated mitochondrial proteins and impaired mitochondrial proteostasis. Together, these findings indicate that mitochondrial dysfunction in a 3-hit HFpEF model is characterized by reduced respiratory capacity and impaired mitochondrial proteostasis, rather than defective biogenesis, suggesting that impaired mitochondrial protein quality control may contribute to reduced cardiac reserve in HFpEF.

New & Noteworthy

HFpEF lacks effective therapies in part due to an incomplete understanding of cardiomyocyte-specific mitochondrial mechanisms involved in cardiac dysfunction. Using an aged, multi-hit mouse model that recapitulates key features of human HFpEF in both sexes, we demonstrate that impaired cardiac reserve is associated with reduced mitochondrial respiratory capacity and coordinated suppression of mitochondrial protein quality-control pathways. These findings identify mitochondrial proteostasis failure as a central mechanism limiting energetic reserve in HFpEF cardiomyocytes.

Introduction

Heart failure with preserved ejection fraction (HFpEF) accounts for approximately half of all heart failure cases and mainly affects older individuals with metabolic and hypertensive comorbidities (1–5). Despite its growing prevalence and substantial morbidity, effective therapies remain limited, reflecting an incomplete understanding of disease mechanisms (6–8). Clinically, HFpEF is characterized by preserved resting systolic function but impaired diastolic performance and markedly reduced exercise tolerance, consistent with a syndrome of impaired cardiac reserve rather than baseline pump failure (9). Increasing evidence suggests that limitations in cardiac energetic capacity contribute to stress intolerance in HFpEF (10); however, the cardiomyocyte-intrinsic mechanisms responsible for this energetic deficit remain poorly understood.

A major barrier to mechanistic progress is the limited availability of experimental models that capture the clinical complexity of HFpEF. The syndrome develops over time in the setting of aging and multiple "hits," including obesity, insulin resistance, hypertension, and vascular dysfunction, and sex is an important biological variable in HFpEF prevalence and phenotype (11, 12). Many preclinical HFpEF models rely on single insults or on young animals and fail to reproduce the combined effects of aging and cardiometabolic stress that characterize human HFpEF (13). Consequently, there is a need for clinically relevant, multi-hit models in aged animals that permit rigorous testing of cardiomyocyte-specific disease mechanisms.

Mitochondria are essential for cardiomyocyte function, supplying the majority of ATP required for contraction, ion homeostasis, and adaptive responses to physiological stress. Prior studies in HFpEF reported altered substrate utilization, reduced oxidative capacity, and abnormalities in mitochondrial structure and function, implicating mitochondrial dysfunction as a potential contributor to impaired cardiac reserve (14, 15). However, key mechanistic questions remain unresolved: whether mitochondrial impairment reflects decreased biogenesis and mitochondrial abundance, altered turnover, or failure of mitochondrial maintenance processes. In addition, many studies rely on whole-heart or bulk tissue analyses, where changes in cell-type composition, inflammation, fibrosis, and vascular remodeling can obscure cardiomyocyte-intrinsic mitochondrial remodeling. These limitations motivate a cardiomyocyte-focused approach that directly interrogates mitochondrial pathways within the cell type most responsible for cardiac energetic demand. Beyond mitochondrial quantity, maintenance of mitochondrial function requires robust mitochondrial proteostasis to ensure efficient import, folding, assembly, and turnover of mitochondrial proteins. Mitochondrial proteostasis is coordinated by chaperones, proteases, and stress-responsive programs, which collectively preserve mitochondrial integrity under chronic stress. Disruption of these systems has been implicated in aging-related diseases, yet their role in cardiomyocyte mitochondrial remodeling and energetic reserve in HFpEF remains largely unexplored (16). Importantly, impaired mitochondrial proteostasis can lead to the accumulation of damaged or misassembled electron transport chain proteins, thereby compromising oxidative metabolism despite preserved biogenesis signaling, a mechanism that is consistent with the strong age dependence of HFpEF (17).

In the present study, we employed a three-hit mouse model of HFpEF that combines aging, a high-fat diet, and chronic nitric oxide synthase inhibition, producing HFpEF-like functional abnormalities in both male and female mice. Using comprehensive in vivo phenotyping, cardiomyocyte-specific mitochondrial isolation, high-resolution mitochondrial respirometry, and data-independent acquisition proteomics, we sought to define cardiomyocyte-intrinsic mitochondrial remodeling in HFpEF. This approach enables pathway-level resolution of mitochondrial energy metabolism and maintenance programs and examines whether mitochondrial dysfunction in HFpEF reflects impaired biogenesis or defective mitochondrial protein quality-control mechanisms, with distinct implications for therapeutic targeting.

Methods

Animals

The experimental procedures followed the National Institutes of Health Guide for the Care and Use of Laboratory Animals and had been approved by the Institutional Animal Care and Use Committee of the University of Mississippi Medical Center. Fifteen-month-old male and female C57BL/6J-NIA mice were obtained from the National Institute on Aging (NIA), aged rodent colony maintained by Charles River Laboratories. Young (22-week-old) C57BL/6J mice were purchased from the Jackson Laboratory.

Experimental protocol

Male and female mice were housed at 23 ± 2 °C under a 12:12-h light-dark cycle with free access to food and water. Mice were divided into four groups: (1) male mice fed a regular diet (Male RD), (2) male mice fed a high-fat diet (HFD; 45% kcal from fat, TD.08811, Inotiv) and treated with L-NAME (Male HFD + L-NAME), (3) female mice fed a regular diet (Female RD), and (4) female mice fed a high-fat diet and treated with L-NAME (Female HFD+L-NAME). HFpEF was induced in groups 2 and 4 at 15 months of age by feeding with HFD for 4 weeks to induce obesity; L-NAME (0.5 g/L; N-ω-Nitro-L-arginine methyl ester hydrochloride, MP Biomedicals, Irvine, CA) was then added to the drinking water, and HFD was continued for an additional 10 weeks. The L-NAME solution was prepared fresh every 3 days, adjusted to pH 7.0 using 2N NaOH solution, and fluid intake (~2 mL/day) was monitored to confirm drug intake. Echocardiography was performed in all groups at baseline and after 10 weeks of L-NAME treatment, followed by dobutamine stress and treadmill exercise testing. At the end of the in vivo protocol, body weight, heart weight, and lung weight (wet and dry) were measured at 19 months old. Heart tissue was harvested, and freshly isolated left ventricular muscle fibers were used to measure mitochondrial respiration. This HFpEF protocol was adapted from the Schiattarella et al. HFD+L-NAME model (18), which originally used 8–12-week-old male C57BL/6N mice receiving a 60% kcal-from-fat diet and L-NAME (0.5 g/L) concurrently. In the present study, 15-month-old male and female C57BL/6JNIA mice were used and received a 45% kcal-from-fat diet (TD.08811, Inotiv) for 4 weeks, followed by 10 weeks of combined HFD and L-NAME treatment.

Transthoracic echocardiography

Cardiac function was assessed by transthoracic echocardiography using a Vevo 3100 system (VisualSonics, Toronto, Canada). Mice were anesthetized with 1–2% isoflurane, and echocardiographic measurements were obtained at baseline and at the end of HFD+ L-NAME treatment. Data from at least three consecutive cardiac cycles were averaged for each mouse. Left ventricular systolic function was evaluated by measuring ejection fraction (EF), fractional shortening (FS), stroke volume (SV), and cardiac output (CO). Left ventricular wall thickness, including anterior and posterior walls, was measured during systole and diastole. Diastolic function was assessed using pulsed-wave Doppler imaging to determine the ratio of early mitral inflow velocity to early mitral annular velocity (E/e′), isovolumetric relaxation time (IVRT), and isovolumetric contraction time (IVCT). The myocardial performance index (MPI) was calculated as (IVCT + IVRT)/ejection time. Image acquisition and analysis were performed using Vevo LAB software (v3.2.0), with frame-by-frame tracking throughout the cardiac cycle.

Dobutamine stress test

Following completion of echocardiographic measurements at week 10 of L-NAME treatment, mice underwent dobutamine stress testing to assess cardiac reserve. Dobutamine was administered intraperitoneally at a single dose of 0.75 mg/kg. Heart rate (HR), EF, FS, SV, and CO were assessed by echocardiography at baseline (time 0) and at 1, 5, and 10 minutes after dobutamine administration (19–22).

Treadmill exercise test

A treadmill exercise stress test was performed at the end of the protocol to assess cardiac reserve. Mice were acclimated to the treadmill (Columbus Instruments, Columbus, OH) for two consecutive days by running for 5 minutes at a low speed (9 m/min) with no incline. On the test day, mice were exercised on a motorized treadmill at progressively increasing speeds and inclinations until exhaustion. The protocol began at 9 m/min and gradually increased to higher speeds with a maximum inclination of 15°, continuing until mice were unable to maintain running despite encouragement. A mild electrical stimulus (25 V, 0.34 mA, 2 Hz) was applied to motivate running. Exercise exhaustion was defined as either remaining on the shock grid for more than 5 consecutive seconds or failure to respond to gentle manual prodding. Mice were removed immediately upon reaching exhaustion and allowed to recover individually for 30 minutes. Running distance was calculated as speed multiplied by running time, and vertical workload was calculated as body weight multiplied by vertical distance (23).

High-resolution respirometry (HRR) measurement of mitochondrial respiration in permeabilized heart samples

Mitochondrial respiration was measured in saponin-permeabilized cardiac fibers using high-resolution respirometry (Oxygraph-2k, Oroboros Instruments, Innsbruck, Austria), following a substrate-uncoupler-inhibitor titration (SUIT) protocol as previously described (24). Briefly, thin left ventricular muscle fibers (1.5–2.0 mg) were permeabilized with saponin (100 μg/mL) in relaxation buffer on a shaker for 40 minutes, followed by three washes in MiRO5 buffer (10 minutes each). Fibers were then weighed and transferred to Oxygraph-2k chambers containing MiRO5 buffer at 37°C, with oxygen concentration maintained above 350 μM. Mitochondrial respiration was assessed sequentially by the addition of glutamate (10 mM) and malate (2 mM) to support complex I-linked respiration, followed by ADP (5 mM) to stimulate oxidative phosphorylation (OXPHOS). Succinate (20 mM) was then added to induce maximal OXPHOS capacity via convergent electron flow through complexes I and II. Oligomycin (5 μM) was added to inhibit ATP synthase and determine ATP-linked respiration. Residual non-mitochondrial respiration was measured following the addition of rotenone (1 μM) and antimycin A (5 μM). Substrate and inhibitor concentrations were selected based on prior optimization studies (25). Data were analyzed using DatLab software (version 4.2, Oroboros Instruments), and respiration rates were expressed as oxygen flux normalized to tissue mass (pmol O₂·s⁻¹·mg⁻¹).

Mitochondria isolation from cardiomyocytes

Cardiomyocytes were isolated from mice at the end of the animal protocol. Mice were anesthetized with 2%-3% isoflurane and anticoagulated with heparin (1000 units/kg, IV). Hearts were excised, cannulated, and perfused in Langendorff mode at 37°C with a Ca²⁺-free buffer (pH 7.2) containing NaCl (135 mM), KCl (4 mM), MgCl₂ (1 mM), HEPES (10 mM), NaH₂PO₄ (0.33 mM), glucose (10 mM), 2,3-butanedione monoxime (10 mM), and taurine (5 mM), equilibrated with 95% O₂/5% CO₂. After 3–5 minutes, hearts were enzymatically digested using collagenase D (0.3 mg/g), collagenase B (0.4 mg/g), and protease type XIV (0.05 mg/g). Digested hearts were removed, minced, and cardiomyocytes were dispersed. Mitochondria were further isolated from the cardiomyocytes using the mitochondria isolation kit (Novus Bio, NBP2–29448, Centennial CO) (26).

Mass spectrometry-based proteomics and Orbitrap Exploris data-independent acquisition (DIA)

Proteomic sample preparation, LC-MS/MS acquisition, and primary data processing were performed by the IDeA National Resource for Quantitative Proteomics at the University of Arkansas for Medical Sciences. For each sample, 100 μg total protein was brought to 100 μL with HPLC-grade water. Proteins were reduced with 10 mM tris (2-carboxyethyl) phosphine (TCEP) for 30 min at 37°C and alkylated with 30 mM iodoacetamide (IAA) for 60 min at room temperature in the dark. Proteins were purified by chloroform/methanol extraction, resuspended in 100 mM triethylammonium bicarbonate (TEAB), and digested with sequencing-grade modified porcine trypsin (Promega) at a protein: trypsin ratio of 50:1 (100 μg protein:2 μg trypsin). Samples were incubated at 47°C for 1 h, vortexed, and then digested overnight at 47°C. Peptides were subsequently acidified and desalted prior to LC-MS/MS analysis.

Tryptic peptides were separated by reverse-phase nano-liquid chromatography using an XSelect CSH C18, 2.5-μm resin, 150 × 0.075-mm column coupled to an UltiMate 3000 RSLCnano system (Thermo Fisher Scientific). Mobile phase A consisted of 0.1% formic acid and 0.5% acetonitrile, and mobile phase B consisted of 0.1% formic acid and 99.9% acetonitrile. Peptides were eluted over a 60-min gradient from 98:2 to 65:35 A: B, ionized by electrospray at 2.4 kV through a heated capillary maintained at 275°C, and analyzed on an Orbitrap Exploris 480 mass spectrometer.

Precursor spectra were acquired over 385–1015 Th at a resolution of 60,000, with a normalized automatic gain control (AGC) target of 100%, a maximum injection time of 50 ms, and an RF parameter of 40%. DIA acquisition used 40 × 15-Th isolation windows spanning 394–1009 Th, with a normalized AGC target of 500% and maximum injection time of 40 ms. A second DIA acquisition used a staggered 15-Th window pattern with optimized window placement from 400–1015 Th.

DIA data processing and differential protein analysis:

DIA data were processed using Spectronaut (Biognosys, v21.0) with the directDIA workflow and searched against the UniProt Mus musculus proteome database (Proteome ID UP000000589; Taxon ID 10090; second 2025 release). Precursor- and protein-level q-value thresholds were controlled at 1%, with decoy generation enabled. Protein inference was performed using the maxLFQ workflow with the IDPicker inference algorithm, and protein quantification was performed at the MS2 level without cross-run normalization, using the median peptide and precursor quantity for protein-group quantification. Fixed modifications included carbamidomethylation of cysteine, and variable modifications included protein N-terminal acetylation and methionine oxidation. Protein MS2 intensity values were assessed for data quality using ProteiNorm and subsequently normalized using variance-stabilizing normalization (VSN). Differential protein abundance was analyzed using proteoDA with Linear Models for Microarray Data (limma) and empirical Bayes moderation of standard errors. For individual-protein differential abundance analyses, proteins with an FDR-adjusted P value <0.05 and fold change >2 were considered significantly altered. Each biological sample represented an individual mouse. All raw mass spectrometry data and Spectronaut output files are deposited in the MassIVE with accession number MSV000102938 and ProteomeXchange dataset PXD082970.

MitoCarta annotation and mitochondrial protein filtering:

To ensure mitochondrial specificity, all quantified proteins were annotated against MitoCarta 3.0 (mouse). MitoCarta 3.0 is a curated, comprehensive inventory of experimentally supported mitochondrial proteins that enables confident identification and analysis of mitochondrial-localized proteins, allowing focused interpretation of mitochondrial-specific pathways rather than confounded whole-proteome signals (17, 27). Only high-confidence mitochondrial proteins were retained for downstream analyses. When multiple protein entries were mapped to the same gene, the entry with the most significant adjusted P value was retained.

Rank-based gene set enrichment analysis (GSEA):

To identify coordinated pathway-level changes, rank-based GSEA was performed on all quantified MitoCarta-annotated mitochondrial proteins, without pre-filtering for statistical significance. Proteins were ranked using a composite metric defined as log₂ fold change multiplied by −log₁₀ (P value). Enrichment analysis was conducted using mitochondrial pathway gene sets derived from MitoCarta 3.0 MitoPathways, with gene set sizes restricted to 10–500 proteins.

Proteomics statistical analyses and data visualization:

All analyses were performed using standard bioinformatics workflows implemented in R and Python. Differential protein abundance was assessed within the MitoCarta 3.0 annotated mitochondrial protein set, and P values were adjusted for multiple testing using the Benjamini-Hochberg false discovery rate method (Supplemental Dataset S1-S2). Rank-based GSEA was performed using a pre-ranked approach without pre-filtering by statistical significance, with mitochondrial pathway gene sets derived from MitoCarta 3.0. Pathways with FDR < 0.25 were considered significantly enriched, consistent with the commonly used exploratory threshold for GSEA that balances pathway discovery with control of false-positive findings.

Fractional mitochondria and cytosol protein isolation for western blotting

Cardiac mitochondria were isolated from the cardiomyocytes of old male RD and HFD + L-NAME mice using a mitochondria isolation kit (Novus Bio, NBP2-29448). The cytosol fraction was obtained in the supernatant, and the mitochondrial pellet was lysed using a mitochondrial lysis buffer. After protein concentration was measured by BCA assay, a total of 15 μg protein from cytosol and mitochondria was separated in 10% SDS-polyacrylamide gels (Bio-Rad, CA). After transfer to nitrocellulose membranes, blots were rinsed in PBS and blocked in Odyssey blocking buffer (LI-COR, NE) for 1 h at room temperature and then incubated with LONP1 rabbit antibody (1:1,500, #28020, Cell Signaling), mtHSP70 rabbit antibody (1:1,000, #3593, Cell Signaling), mtHSP60 rabbit antibody (1:1,000, #12165, Cell Signaling), NRF1 rabbit antibody (1:2,000, #69432 Cell Signaling), NRF2 rabbit antibody (1:1,000, #20733, Cell Signaling), TFAM rabbit antibody (1:1,000, #15218, Cell Signaling), Ubiquitin (FK2) Mouse monoclonal antibody (1:1,000, #39269, Cell Signaling) at 4 °C overnight. Membranes were probed with LI-COR fluorescent dye-labeled secondary antibodies (1:5,000) for 1 h at room temperature. Antibody labeling was visualized using the Odyssey Infrared Scanner (LI-COR, NE). Total cytosol and mitochondrial protein staining was used as the internal loading control for the mitochondrial and cytosolic fractions (Supplementary Fig. S6).

ClpP protease activity assay

ClpP protease activity was measured in lysed cardiomyocyte mitochondrial preparations using a fluorogenic ClpP substrate assay (ProFoldin, Hudson, MA). Mitochondrial lysate containing 10 μg protein per well was added to a 96-well plate in a final reaction volume of 60 μL. Samples were incubated for 20 min at 37°C after the addition of the fluorogenic substrate. Fluorescence was measured at Ex 380 nm/Em 460 nm at 37°C. Background fluorescence from substrate-only wells was subtracted, and protease activity was calculated in RFU.

Data representation and statistical analysis

Data are presented as mean ± standard error of the mean (SEM). For comparisons among multiple groups, one-way ANOVA followed by Tukey’s post hoc test was performed. For comparisons between two groups, a t-test was performed. The statistical methods applied for each comparison are detailed in the figure legends. Statistical analyses were conducted using Prism (GraphPad Software, La Jolla, CA), and p-values < 0.05 were considered statistically significant.

Results

Morphometric and organ weight characteristics in old male and female mice subjected to RD or HFD+L-NAME.

In this 3-hit HFpEF model, combined aging, high-fat diet, and L-NAME treatment resulted in significant cardiac and pulmonary remodeling in male and female mice (Table 1). Relative to sex-matched regular diet controls, HFD+L-NAME increased body weight in both sexes, consistent with diet-induced metabolic stress. Absolute heart weight was significantly increased by 25% in males and 47% in females with HFD+L-NAME compared with sex-matched regular diet controls, indicating cardiac hypertrophy. Notably, while the heart-to-body weight ratio was unchanged in males, it was significantly elevated in females, suggesting a more pronounced hypertrophic response relative to body size in female mice. In addition, wet and dry lung weights were increased in HFD+L-NAME-treated animals of both sexes, consistent with pulmonary congestion and early heart failure-associated fluid accumulation. Together, these findings demonstrate that combined metabolic and hypertensive stress induces robust cardiopulmonary remodeling in both sexes, with sex-specific differences in the magnitude of cardiac hypertrophy.

Table 1.

Morphometric and organ weight characteristics in old male and female mice subjected to regular diet (RD) or high-fat diet plus L-NAME (HFD+L-NAME) at the end of the in vivo protocol (19 months old).

	Male RD	Male HFD+ L-NAME	Female RD	Female HFD+ L-NAME
Body Weight (g)	35.4± 1.2	44.6± 2.4 *	28.7± 1.5	32.3± 2.9 +
Heart Weight (mg)	184.7± 4.3	230.7± 7.5 *	158.9± 9.4	235.2± 11.2 +
Heart/Body Weight	5.3± 0.2	5.3± 0.3	5.5± 0.2	7.5± 0.6 +
Lung Weight (wet)	172.1± 4.0	201.2± 5.0 *	163.4± 7.0	204.7± 13.9 +
Lung Weight (dry)	40.5± 1.0	45.4± 1.4 *	37.8± 1.3	47.8± 3.2 +

(n=6–10 per group; *, p<0.05 when compared to Male RD; +, p<0.05 when compared to Female RD by t test)

Effects of HFD+L-NAME on cardiac systolic and diastolic functions in old male and female mice

Echocardiographic assessment demonstrated preserved systolic function across all groups following induction of the HFpEF model. Left ventricular ejection fraction and fractional shortening remained within the normal range in male and female mice and were not significantly reduced by HFD+L-NAME treatment compared with sex-matched regular diet controls (Figs 1A and C). Consistent with preserved systolic performance, stroke volume and cardiac output remained unchanged over time in both sexes, with no evidence of systolic pump failure (Figs. 1B and D). These findings indicate that combined metabolic and hypertensive stress induced cardiac remodeling and functional changes in the absence of systolic dysfunction, consistent with a HFpEF phenotype.

Fig. 1: Cardiac systolic function assessed by echocardiography in old male and female mice.

(A) ejection fraction, (B) stroke volume, (C) fractional shortening, and (D) cardiac output measurements at baseline and 10 weeks after RD or HFD+ L-NAME. Results are expressed as mean±SEM, n=8–10.

Echocardiographic measurement of diastolic function revealed significant impairment following HFD+L-NAME treatment, consistent with the development of a HFpEF phenotype. Compared with baseline measurement, aged HFD+L-NAME mice exhibited abnormal transmittal Doppler parameters, including significantly decreased e′ in males and females and elevated E/e′ ratio in females, indicating impaired myocardial relaxation and increased left ventricular filling pressures (Figs. 2A and B). In parallel, isovolumic relaxation time (IVRT) was significantly prolonged in male and female HFD+L-NAME groups compared with baseline, reflecting delayed ventricular relaxation (Fig. 2C). Myocardial performance index (MPI), an integrated measure of systolic and diastolic function that is particularly sensitive to diastolic abnormalities, was also significantly increased in HFD+L-NAME-treated mice of both sexes (Fig. 2D). Notably, these diastolic abnormalities occurred in the absence of systolic dysfunction, further supporting that combined metabolic and hypertensive stress selectively impairs diastolic performance and recapitulates key functional features of HFpEF.

Fig. 2: Cardiac diastolic function assessed by echocardiography in old male and female mice.

(A) e’, (B) E/e’, (C) isovolumic relaxation time, and (D) myocardial performance index measurements at baseline and 10 weeks after RD or HFD+ L-NAME. Results are expressed as mean±SEM, n=8–10, * p<0.05 between two groups by paired two-tailed t-test.

HFD+L-NAME induced left ventricular hypertrophy in old male and female mice

Echocardiographic assessment of left ventricular wall thickness demonstrated structural remodeling consistent with concentric hypertrophy in aged HFD+L-NAME-treated mice. Compared with sex-matched baseline measurements, HFD+L-NAME resulted in significant increases in left ventricular anterior wall thickness during systole (LVAW;s) and diastole (LVAW;d) in female mice and significant increases of LWAW;d in male mice (Figs. 3A and D). Posterior wall thickness was also increased, with significant elevations in left ventricular posterior wall thickness during systole (LVPW;s) and diastole (LVPW;d) observed in male HFD+L-NAME-treated animals, and significant increases of LVPW;d in female mice (Figs. 3B and E). These changes indicate uniform thickening of the left ventricular wall (around 25% increase) without chamber dilation, as evidenced by unchanged internal diameters in both sexes (Figs. 3C and F), consistent with the concentric remodeling phenotype. Together with preserved systolic function and impaired diastolic relaxation, these structural alterations further support the development of a HFpEF-like phenotype in old mice under combined metabolic and hypertensive stress.

Fig. 3: Left ventricle wall thickness and internal diameter in old male and female mice.

(A) End-systolic left ventricular anterior wall thickness (LVAW,s), (B) end-systolic left ventricular posterior wall thickness (LVPW,s), (C) end-systolic left ventricular internal diameter, (D) end-diastolic left ventricular anterior wall thickness (LVAW,d), (E) end-diastolic left ventricular posterior wall thickness (LVPW,d), and (F) end-diastolic left ventricular internal diameter measurements at baseline and 10 weeks after RD or HFD+ L-NAME. Results are expressed as mean±SEM, n=8–10, * p<0.05 between two groups by between two groups by paired two-tailed t-test.

Dobutamine stress test suggested reduced cardiac reserve capacity in old male and female mice after HFD+L-NAME

Dobutamine stress echocardiography revealed impaired cardiac reserve in aged HFD+L-NAME-treated mice despite preserved baseline systolic function. At baseline and in response to β-adrenergic stimulation with dobutamine injection, heart rate was similar in HFD+L-NAME and regular diet groups in both sexes (Fig. 4A). Ejection fraction was significantly increased in all groups of mice compared to baseline after dobutamine (Fig. 4B). Fractional shortening was significantly increased following dobutamine in regular diet-fed groups of both males and females. However, the magnitude of augmentation in fractional shortening was attenuated in HFD+L-NAME-treated animals (Fig. 4C). More importantly, HFD+L-NAME mice exhibited a significant decrease in stroke volume (29% reduction in males and 24% reduction in females) and cardiac output (27% reduction in males and 20% reduction in females) at 10 minutes after dobutamine injection compared with baseline levels. In contrast, regular diet mice maintained stroke volume and cardiac output throughout the stress period (Figs. 4D and E). These findings indicate that combined metabolic and hypertensive stress impairs contractile reserve and limits the heart’s ability to augment performance under stress, a hallmark feature of HFpEF.

Fig. 4: Echocardiography during dobutamine stress test.

Time-dependent changes after dobutamine injection in (A) heart rate, (B) ejection fraction, (C) fractional shortening, (D) stroke volume, and (E) cardiac output at baseline, 1, 5, and 10 minutes after injection in old male and female mice with RD or HFD+ L-NAME. Results are expressed as mean±SEM; n=8–10; *, p<0.05 when compared with baseline by post-hoc test after one-way ANOVA.

Reduced exercise capacity in old male and female mice after HFD+L-NAME

Exercise capacity was significantly impaired in aged HFD+L-NAME-treated mice, consistent with the functional limitation characteristic of HFpEF. Compared with sex-matched regular diet controls, male HFD+L-NAME mice exhibited 26% reduced treadmill run time and 36% reduced total distance achieved during exercise testing. These significant changes can also be observed in female HFD+-L-NAME mice compared to regular diet controls (33% reduction in run time and 44% reduction in total distance) (Figs. 5A and B). In addition, vertical work performed after normalization to body weight was significantly decreased in male and female HFD+L-NAME-treated animals compared to regular diet controls, indicating reduced exercise tolerance and diminished cardiac reserve (Fig. 5C). These deficits were observed in both sexes and occurred in the setting of preserved systolic function, supporting the presence of exercise intolerance driven by diastolic dysfunction and impaired cardiopulmonary performance rather than systolic failure.

Fig. 5: Treadmill running test in old male and female mice.

(A) running time, (B) running distance, and (C) vertical work in old male and female mice with RD or HFD+ L-NAME. Results are expressed as mean±SEM; n=7–10; * p<0.05 between two groups by t-test.

Aging increases susceptibility to HFD+L-NAME-induced HFpEF cardiac remodeling

To determine whether advanced age modifies the cardiac response to metabolic and hypertensive stress, we examined young male and female C57BL/6J mice subjected to the same 14-week HFD+L-NAME protocol used in the aged mice. In contrast to aged HFD+L-NAME mice, young animals maintained relatively preserved resting cardiac systolic function (Supplementary Figs. S1 A-D). Doppler-derived indices, including e′ velocity, E/e′ ratio, IVRT, and MPI, without marked deterioration (Supplementary Figs. S2 A-D). Despite this preservation of resting cardiac function, young HFD+L-NAME mice developed evidence of structural remodeling. Left ventricular anterior and posterior wall thickness increased following treatment in males and females (Supplementary Figs. S3 A and B), whereas ventricular internal diameter remained relatively unchanged (Supplementary Fig. S3C), consistent with the development of concentric hypertrophic remodeling. Importantly, as shown in the accompanying exercise studies, these young HFD+L-NAME mice also exhibited substantial reductions in treadmill running time, distance, and vertical work (Supplementary Figs. S4 A-C).

Thus, the same metabolic and hypertensive challenge as used in old mice produced measurable hypertrophic remodeling and exercise intolerance in young mice but did not reproduce the more pronounced resting cardiac phenotype observed in aged animals over the same treatment interval. These findings support advanced age as an important modifier of susceptibility to HFD+L-NAME-induced HFpEF remodeling.

Old male and female mice with HFD+L-NAME developed mitochondrial dysfunction

To evaluate the mitochondrial functional changes during the development of HFpEF, we performed detailed mitochondrial respiration measurements in isolated cardiac fibers from the left ventricle. We assessed the maximal OXPHOS capacity induced by complex I and II substrates in the presence of ADP as an indicator of mitochondrial respiration capacity. ATP-linked respiration rate was measured as an indicator of mitochondrial respiration efficiency by comparing changes in O₂ consumption after adding the synthase inhibitor oligomycin. Compared with age-matched regular diet controls, male and female HFD+L-NAME mice exhibited a marked reduction in maximal mitochondrial respiration (Fig. 6A), indicating diminished electron transport chain capacity. In parallel, ATP-linked respiration was significantly reduced in HFD+L-NAME hearts, reflecting impaired coupling of OXPHOS to ATP production (Fig. 6B). Complex I-linked OXPHOS was significantly decreased in male HFpEF mice, whereas the reduction in females did not reach statistical significance (Fig. 6C). In contrast, convergent Complex I+II-linked OXPHOS was significantly reduced in both male and female HFpEF mice (Fig. 6D). These findings demonstrate a broad impairment of mitochondrial oxidative phosphorylation in aged HFpEF hearts. These deficits were evident in both sexes, although the magnitude of impairment varied between males and females (Maximum respiration rate: 45% reduction in males vs 28% reduction in females). Together, these findings indicate that combining aging, metabolic stress, and hypertension compromise cardiac mitochondrial bioenergetic function, providing a mechanistic link between mitochondrial dysfunction and HFpEF development in the 3-hit model.

Fig. 6: Mitochondrial respiration capacity.

measured in left ventricular cardiac fibers of old male and female mice with RD or HFD+ L-NAME. (A) Maximal respiration rate, and (B) ATP-linked oxygen consumption rate calculated as succinate-supported respiration minus oligomycin-resistant respiration. (C) Complex I-linked OXPHOS, calculated as ADP-stimulated respiration minus antimycin A-insensitive respiration. (D) Convergent Complex I+II-linked OXPHOS, calculated as succinate-supported respiration minus antimycin A-insensitive respiration. Results are expressed as mean±SEM; n=5–9; * p<0.05 between two groups by t-test.

Mitochondrial proteomic remodeling is broadly conserved between male and female HFpEF mice

To determine whether mitochondrial proteomic remodeling differed by sex, cardiomyocyte mitochondrial proteomes from old male and female HFpEF mice were analyzed separately relative to their respective same-sex controls. Volcano plots demonstrated mitochondrial protein changes in both sexes, with males showing a predominantly negative shift in many mitochondrial proteins, whereas females exhibited a broader bidirectional distribution of protein changes (Figs. 7A and C)

Fig. 7: Sex-specific and shared mitochondrial proteomic remodeling in aged HFpEF mice.

Cardiomyocyte mitochondrial proteomes from old male and female control (RD) and HFD+L-NAME mice were analyzed separately by sex and filtered using MitoCarta3.0. (A) Volcano plot of MitoCarta-filtered mitochondrial proteins in old male HFpEF versus male RD controls. (B) Bidirectional MitoPathway enrichment analysis in males showing pathways with FDR < 0.25. Negative normalized enrichment scores (NES) indicate pathways suppressed in HFpEF, whereas positive NES indicate pathways enriched in HFpEF. Dot size represents gene-set size and color represents −log10(FDR). (C) Volcano plot of MitoCarta-filtered mitochondrial proteins in old female HFpEF versus female RD controls. (D) Bidirectional MitoPathway enrichment analysis in females using the same FDR < 0.25 threshold (−log10(0.25) =0.60). (E) Comparison of mitochondrial pathways significantly altered in both males and females with concordant enrichment direction. Bars show NES for males and females.

Mitochondrial pathway enrichment analysis further identified extensive remodeling of mitochondrial metabolism and OXPHOS in both sexes. In males, multiple pathways were negatively enriched, including OXPHOS, Complex I and its subunits, TCA-cycle metabolism, amino acid and carbohydrate metabolism, Fe-S-containing proteins, and Complex IV-associated pathways. Fewer pathways, including xenobiotic metabolism, mitochondrial ribosome assembly, and apoptosis, were positively enriched. Females similarly showed strong suppression of OXPHOS, OXPHOS subunits, Complex I, Complex I subunits, TCA-cycle metabolism, amino acid metabolism, and related mitochondrial metabolic pathways, while several stress- and metabolic-response pathways, including lipid metabolism, detoxification, ROS/glutathione metabolism, xenobiotic metabolism, and apoptosis, were positively enriched (Figs. 7B and D). Tables 2 and 3 identified the proteins that contributed most strongly to the four most negatively enriched and three most positively enriched mitochondrial pathways in old male and female HFpEF cardiomyocyte mitochondria. The corresponding leading-edge proteins highlight coordinated remodeling of mitochondrial energy metabolism, respiratory chain function, stress-response pathways, and mitochondrial protein synthesis in HFpEF.

Table 2.

Leading-edge proteins contributing to the strongest negatively and positively enriched mitochondrial pathways in old male HFpEF cardiomyocyte mitochondria. Pre-ranked GSEA of MitoCarta 3.0 pathways was performed comparing old male HFD+L-NAME-treated mice with age-matched RD controls. The four pathways with the most negative normalized enrichment scores (NES) and the three pathways with the most positive NES were selected for presentation. The leading-edge protein subset that contributes most to the enrichment score for each pathway is shown. Negative NES indicates relative suppression in HFpEF, whereas positive NES indicates relative enrichment in HFpEF. FDR-adjusted P values are provided for each pathway.

Carbohydrate metabolism	TCA cycle	Amino acid metabolism	Complex I subunits	Xenobiotic metabolism	Mitochondrial ribosome assembly	Apoptosis
NES = −1.838	NES = −1.794	NES = −1.663	NES = −1.649	NES = 0.960	NES = 0.879	NES = 0.651
FDR = 0.013	FDR = 0.013	FDR = 0.013	FDR = 0.013	FDR = 0.090	FDR = 0.158	FDR = 0.231
Pdk4	Sdhd	Aldh18a1	Ndufv3	Ephx2	Eral1	Aifm2
Hmgcs2	Suclg2	Sfxn3	mt-Nd3	Mgst1	Ddx28	Bnip3
Pck2	Sdhc	Comt	Ndufa2	Maoa	Dhx30	Diablo
Sfxn5	Idh3a	Maoa	Ndufs6	Cyb5r3	Mtg1	Bcl2l13
Sdhd	Idh3g	Hibadh	Ndufb2	Acsm5	Mrm3	Htra2
Abhd11		Ldhb	Ndufb5	Mtarc2	Mrm2	Bcl2
Pdp2		Slc25a15	Ndufa7	Mpst	Mettl17	Cycs
Slc25a1		Slc25a29	Ndufa3	Bphl	Fastkd2	Endog
Suclg2		Adhfe1	Ndufb3	Maob	Ptcd3	Bak1
Fahd1		Aldh9a1	Ndufab1	Gstk1	Nsun4	Chchd2
Hadhb		Maob	Ndufa5	Cyb5b	Grsf1	Ghitm
Sdhc		Ppm1k		Tst		Aifm1
Mpc2		Hadhb		Abhd10
Pdk2		Etfdh
Ldhd		Grhpr
Me3		Hadha
Idh3a		Gls
Pdk1		Hibch
Acaa2		Shmt2
Acat1		Acaa2
Idh3g		Slc25a44
D2hgdh		Bcat2
Bdh1		Acat1
Mpc1		Prodh
Clybl		Hadh
Mmut		Bckdk
Auh		Ivd
Mdh2		Etfb

Table 3.

Leading-edge proteins contributing to the strongest negatively and positively enriched mitochondrial pathways in old female HFpEF cardiomyocyte mitochondria. Pre-ranked GSEA of MitoCarta 3.0 pathways was performed comparing old female HFD+L-NAME-treated mice with age-matched RD controls. The four pathways with the most negative normalized enrichment scores (NES) and the three pathways with the most positive NES were selected for presentation. The leading-edge protein subset that contributes most to the enrichment score for each pathway is shown. Negative NES indicates relative suppression in HFpEF, whereas positive NES indicates relative enrichment in HFpEF. FDR-adjusted P values are provided for each pathway.

OXPHOS subunits	OXPHOS	Complex I	Complex I subunits	Lipid metabolism	Detoxification	ROS and glutathione metabolism
NES = −1.846	NES = −1.780	NES = −1.729	NES = −1.722	NES = 1.686	NES = 1.595	NES = 1.511
FDR = 0.018	FDR = 0.018	FDR = 0.027	FDR = 0.032	FDR = 0.018	FDR = 0.039	FDR = 0.083
Cox6a1	Fmc1	Tmem186	Ndufb2	Ephx2	Ephx2	Cat
Cox7a2l	Cox6a1	Ndufaf2	Ndufa7	Acad11	Cat	Mgst1
Cox8b	Higd1a	Ndufaf7	Ndufab1	Hsd17b4	Mgst1	Aifm2
Cox7b	Uqcc1	Tmem126b	Ndufb11	Pnpla8	Aifm2	Sod1
Hccs	Cox7a2l	Ndufb2		Ehhadh	Mtarc2
Uqcr11	Cox17	Foxred1		Scp2	Sod1
Cyct	Tmem186	Ndufaf3		Acot2
Ndufb2	Pnkd	Ecsit		Phyh
mt-Co3	Atpaf2	Tmem70		Echdc1
Cox7c	Cox10	Dmac2		Plscr3
Ndufa4	Cox11	Ndufaf5		Amacr
Ndufa7	Ndufaf2	Ndufa7		Ech1
Cyc1	Ndufaf7	Ndufab1		Acaa1a
Ndufab1	Cox8b	Ndufb11		Acadvl
Uqcr10	Higd2a			Gpam
Ndufb11	Cox7b			Crot
Cox6a2	Hccs			Mlycd
Dmac2l	Uqcr11			Dhrs1
Ndufv2	Coa3			Acsl1
	Cox15			Cpt2
	Uqcc3			Cyp27a1
	Tmem126b			Cpt1b
	Pet100			Eci2

Among 695 MitoCarta proteins (curated reference list of proteins with strong evidence for mitochondrial localization or function) quantified in both sexes, HFpEF-associated protein changes showed moderate concordance between males and females with 74.5% of proteins changing in the same direction. At the pathway level, 51 MitoPathways were shared between sexes, and enrichment patterns were also concordant. Fourteen pathways met FDR<0.25 in both sexes and changed in the same direction. Shared suppressed pathways included Complex I, Complex I subunits, OXPHOS and OXPHOS subunits, TCA cycle, amino acid metabolism, branched-chain amino acid metabolism, carbohydrate metabolism, Complex IV assembly factors, Fe-S-containing proteins, and Complex IV, whereas xenobiotic metabolism and apoptosis were enriched in both sexes (Fig. 7E).

Together, these findings indicate that the major mitochondrial response to HFpEF is largely shared between males and females, particularly the suppression of OXPHOS and mitochondrial metabolic pathways, although sex-specific differences are evident in the magnitude of pathway changes and in several positively enriched stress-response pathways.

To determine whether mitochondrial abnormalities occurred within a broader cardiac proteomic response, we analyzed whole-heart cytosolic and mitochondrial-enriched fractions separately from female mice treated with a regular diet or HFD+L-NAME. Cytosolic protein analysis identified several significantly altered pathways at FDR < 0.25. HFpEF was associated with suppression of mRNA splicing, respiratory electron transport, mitochondrial protein degradation, and the TCA cycle, whereas ubiquitin/proteasome-associated pathways were enriched. These findings indicate that the cardiac proteomic response to HFD+L-NAME extends beyond mitochondrial pathways and includes broader alterations in RNA processing, metabolism, and protein quality-control systems. Importantly, mitochondrial-related pathways remain among the most prominently altered biological programs, with coordinated changes in respiratory metabolism, the TCA cycle, and mitochondrial protein degradation, supporting mitochondrial dysfunction as a major component of the broader cardiac remodeling response (Supplementary Fig. S5A).

Mitochondrial-restricted analysis of the mitochondrial-enriched fraction revealed a more focused mitochondrial signature. Mitochondrial translation, mitochondrial protein degradation, and the TCA cycle were suppressed in HFpEF, whereas respiratory electron transport showed only a modest positive pathway-level shift (Supplementary Fig. S5B). The concordant suppression of mitochondrial protein degradation in both fractions, together with reduced mitochondrial translation and TCA-cycle pathways, supports disruption of mitochondrial proteostasis and metabolism in HFpEF.

Loss of mitochondrial proteostasis underlies mitochondrial dysfunction in HFpEF cardiomyocytes

Mitochondrial proteomics revealed a consistent reduction in proteins responsible for maintaining mitochondrial protein quality control in aged HFpEF cardiomyocytes. As shown in Table 4, key mitochondrial chaperones and co-chaperones involved in protein folding and importing, including HSP60, HSP10, mtHSP70, and DNAJA3, were significantly downregulated in cardiomyocyte mitochondria from old HFD + L-NAME mice compared with age-matched controls. In parallel, major mitochondrial proteases that support the mitochondrial unfolded protein response and matrix quality control, including ClpP, LONP1, and AFG3L2, were also reduced. Together, these changes indicate a diminished capacity for mitochondrial protein folding, import, and degradation in HFpEF.

Table 4.

Reduced abundance of mitochondrial proteostasis and unfolded protein response (UPRmt) effector proteins in old male HFpEF cardiomyocyte mitochondria. Summary of mitochondrial chaperones, co-chaperones, and proteases identified by DIA proteomics that are significantly reduced in cardiomyocyte mitochondria from old mice with HFD+ L-NAME compared with age-matched RD controls. Proteins are annotated by functional category, sub-mitochondrial localization, log₂ fold change (HFpEF vs control), abundance ratio, and false discovery rate (FDR)-adjusted P value.

Category	Gene	Protein name	Localization	log2FC (HFpEF vs Ctrl)	HFpEF / Ctrl	Adjusted P (FDR)	Interpretation
Chaperone	Hspd1	HSP60	Matrix	–0.48	0.72	<0.05	↓ protein folding capacity
Chaperone	Hspe1	HSP10	Matrix	–0.52	0.70	<0.05	↓ chaperonin cofactor
Chaperone	Hspa9	mtHSP70 / Mortalin	Matrix	–0.58	0.67	<0.05	↓ protein import & folding
Co-chaperone	Dnaja3	Tid1	Matrix/IMM	–0.70	0.62	<0.01	↓ co-chaperone activity
Protease	ClpP	Caseinolytic peptidase P	Matrix	–0.46	0.73	<0.05	↓ UPRmt core protease
Protease	Lonp1	Lon protease	Matrix	–0.48	0.72	<0.05	↓ matrix protein quality control
Protease	Afg3l2	m-AAA protease subunit	IMM (matrix-facing)	–0.54	0.69	<0.05	↓ IMM proteostasis

Western blot validation of proteomics reveals impaired mitochondrial proteostasis and increased mitochondrial protein ubiquitination in HFpEF heart

Because the pathway analysis identified coordinated alterations in mitochondrial maintenance and protein quality-control pathways, we next determined whether these proteomic changes were accompanied by biochemical and functional evidence of impaired mitochondrial proteostasis and whether they could instead be explained by reduced mitochondrial biogenesis. Immunoblot analyses were performed on isolated cardiac mitochondria from old male mice fed a regular diet or HFD + L-NAME. As shown in Fig 8A, mitochondrial matrix protease LONP1, the mitochondrial HSP70 chaperone (mtHSP70), and the chaperonin mtHSP60 all exhibited lower abundance in HFD+L-NAME heart samples compared with regular diet controls. These changes are consistent with the downregulation of mitochondrial chaperones and proteases identified by targeted proteomics. To determine whether impaired proteostasis was associated with altered protein turnover, total mitochondrial ubiquitination was examined. Fig 8B shows an accumulation of multi-ubiquitinated proteins across a broad molecular weight range in mitochondria from HFD+L-NAME mice relative to regular diet controls, suggesting increased protein damage and/or impaired clearance of misfolded proteins. Quantification of immunoblot data by mitochondria total protein (Fig. 8C) confirmed a significant reduction in mtHSP70, mtHSP60, and LONP1 protein expression in HFpEF mitochondria compared with regular diet controls. In contrast, levels of ubiquitinated proteins were significantly increased in the HFD+L-NAME group. To directly determine whether the proteomic evidence of impaired mitochondrial protein degradation was accompanied by a functional defect in mitochondrial proteolysis, we measured caseinolytic mitochondrial matrix peptidase proteolytic subunit (ClpP) protease activity in isolated cardiomyocyte mitochondria. ClpP was selected because it is a major mitochondrial matrix protease involved in the degradation of damaged or misfolded proteins and is therefore an important component of mitochondrial protein quality control. Consistent with the proteomic finding of suppressed mitochondrial protein degradation pathways, ClpP activity was significantly reduced in HFD+L-NAME mice compared with RD controls (Fig. 8D). These findings provide functional evidence of impaired mitochondrial proteolytic capacity and support disruption of mitochondrial proteostasis in HFpEF.

Fig. 8: Impaired mitochondrial proteostasis in cardiomyocytes from old HFpEF mice.

(A) Representative immunoblots of mitochondrial fractions isolated from cardiomyocytes of old male mice fed RD or HFD+ L-NAME, probed for the mitochondrial proteostasis regulators LONP1, mtHSP70, and mtHSP60. (B) Representative immunoblot showing accumulation of multi-ubiquitinated proteins in mitochondrial fractions from RD and HFD+ L-NAME groups. (C) Quantification of mitochondrial protein expression normalized to RD controls, showing reduced levels of mtHSP70, mtHSP60, and LONP1 and increased accumulation of ubiquitinated proteins in the HFD+ L-NAME group. (D) ClpP protease activity was measured in lysed cardiac mitochondrial preparations using a fluorogenic substrate assay with 10 μg mitochondrial protein per well between groups (3 males and 4 females in each group). Each data point represents an individual mouse. Data are presented as mean ± SEM. * p<0.05 versus RD by multiple unpaired t-tests.

Together, these results demonstrate that HFpEF is associated with loss of mitochondrial chaperone and protease capacity alongside increased accumulation of ubiquitinated proteins, providing direct evidence of impaired mitochondrial proteostasis in cardiomyocytes.

Preserved cytosolic regulators of mitochondrial biogenesis in HFpEF cardiomyocytes

To assess whether mitochondrial dysfunction in HFpEF was associated with altered upstream control of mitochondrial biogenesis, cytosolic levels of key transcriptional regulators were examined by immunoblotting. NRF1 and NRF2 are nuclear transcription factors that regulate the expression of genes involved in mitochondrial biogenesis, respiratory chain assembly, and antioxidant defense, while TFAM is a critical mitochondrial transcription and DNA-binding factor required for mitochondrial DNA maintenance and transcription. As shown in Fig. 9A, cytosolic protein levels of NRF1, NRF2, and TFAM were similar between regular diet and HFD + L-NAME mice. Quantitative analysis by cytosolic total protein confirmed no significant differences in the abundance of these proteins between groups (Fig. 9B). These results indicate that the reduction in mitochondrial protein content and proteostasis capacity observed in HFpEF is not accompanied by decreased expression of major transcriptional regulators of mitochondrial biogenesis, supporting the conclusion that mitochondrial dysfunction in HFpEF may reflect impaired mitochondrial maintenance rather than reduced biogenic signaling.

Fig. 9. Mitochondrial biogenesis transcription factors are preserved in cardiomyocytes from old HFpEF mice.

(A) Representative immunoblots of cytosolic fractions from cardiomyocytes of old male mice fed RD or HFD+ L-NAME, probed for the mitochondrial biogenesis-related transcription factors NRF1, NRF2, and TFAM. (B) Quantification of protein expression levels normalized to RD controls, showing no significant differences in NRF1, NRF2, or TFAM abundance between RD and HFD + L-NAME groups. Each data point represents an individual mouse (n = 5 per group). Data are presented as mean ± SEM. Comparisons were performed using multiple unpaired t-tests.

Discussion

HFpEF is increasingly recognized as a syndrome of impaired cardiac reserve that develops with aging, usually in the setting of metabolic and hypertensive comorbidities. Using an aged, HFD+L-NAME model in male and female mice, we integrated in vivo phenotyping with tissue-level respirometry and cardiomyocyte-specific mitochondrial proteomics to define cardiomyocyte-intrinsic mitochondrial remodeling. Our results support four conclusions: (1) the model provides a reproducible HFpEF phenotype in both sexes with preserved ejection fraction but impaired diastolic function and decreased cardiac reserve; (2) maximal and ATP-linked mitochondrial respiratory capacity are reduced in left ventricular fibers, consistent with diminished energetic reserve; (3) cardiomyocyte mitochondrial proteomics reveal broad suppression of oxidative metabolism and mitochondrial maintenance programs, shared between males and females, particularly the suppression of oxidative phosphorylation and mitochondrial metabolic pathways; and (4) the pattern of remodeling is most consistent with impaired mitochondrial proteostasis rather than reduced mitochondrial biogenesis.

A clinically relevant aged HFD+L-NAME model for mechanistic studies

Mechanistic progress in HFpEF has been limited by models that rely on single insults or young animals and therefore incompletely capture the age dependence and multimorbidity of the human syndrome. By combining advanced age with metabolic stress and chronic NO synthase inhibition, this model more closely approximates the clinical context in which HFpEF develops, and the consistent phenotype in male and female mice strengthens its translational relevance. It is particularly important for mitochondrial biology because aging and metabolic stress independently compromise mitochondrial efficiency and quality control.

Recent work by Pepin et al. (20) further supports the view that cardiometabolic HFpEF is closely linked to mitochondrial dysfunction and redox/nitrosative stress. It emphasizes that incorporating aging may be critical for modeling sex-related aspects of HFpEF. The young-mouse comparison in our study provides important context for interpreting the aged HFD+L-NAME model. Exposure of young mice to the same 14-week HFD+L-NAME treatment as used in aged mice induced cardiac hypertrophic remodeling and marked exercise intolerance, while conventional resting systolic and diastolic measurements remained relatively preserved. In contrast, aged mice exposed to the same challenge developed a more extensive HFpEF phenotype. These findings suggest that advanced age increases susceptibility to progression from an earlier, compensated response toward more overt HFpEF-like cardiac dysfunction.

In young mice, increased ventricular wall thickness together with preserved resting cardiac function may represent an earlier stage of remodeling. The marked exercise intolerance despite limited resting echocardiographic changes also suggests that impaired cardiovascular and peripheral reserve may precede detectable resting dysfunction. The difference between young and aged mice may therefore reflect faster disease progression in the aged heart, with greater age-dependent vulnerability to metabolic and hypertensive stress. Because both groups were studied after the same 14-week exposure, the current data cannot determine whether longer treatment in young mice would eventually reproduce the aged phenotype. Thus, young mice should not necessarily be considered resistant to HFpEF development; rather, they appear to develop a less advanced phenotype over the same treatment interval, at least in the C57BL/6J strain.

Energetic limitation explains impaired cardiac reserve and preferential diastolic vulnerability

HFpEF commonly presents with exercise intolerance and reduced cardiac reserve that appear disproportionate to resting systolic performance (28). Our results from high-resolution respirometry demonstrate impaired mitochondrial respiratory capacity in left ventricular fibers, and in vivo stress testing shows an inadequate hemodynamic response to dobutamine, with cardiac output and stroke volume failing to increase appropriately despite increases in ejection fraction and preserved fractional shortening. Together with reduced exercise performance, these findings support the interpretation that limited energetic reserve constrains cardiac performance under demand.

A possible explanation for preserved EF and FS at rest but impaired diastolic function is differential energetic sensitivity. While systolic contraction is ATP-intensive, resting systolic function can often be maintained through length-dependent activation (Frank-Starling mechanism), basal adrenergic support (sympathetic activation), metabolic buffering via phosphocreatine (ATP buffer), and structural remodeling that reduces wall stress (hypertrophy). In contrast, diastolic relaxation depends on ATP for SERCA2a-mediated Ca²⁺ reuptake and for actomyosin cross-bridge detachment; modest reductions in ATP availability or oxidative reserve can therefore delay relaxation and increase diastolic stiffness without necessarily reducing resting systolic parameters (29). This framework is consistent with the clinical concept of HFpEF as a disease of limited cardiac reserve rather than of baseline systolic failure, although systolic function may eventually decline in severe, prolonged HFpEF.

Cardiomyocyte mitochondrial proteomics indicate impaired maintenance rather than loss of mitochondrial content

A key strength of this study is the use of cardiomyocyte mitochondrial proteomics rather than bulk tissue, reducing confounding from nonmyocyte expansion and remodeling. Using MitoCarta-based annotation and pathway-level enrichment, we identified coordinated suppression of oxidative phosphorylation, the TCA cycle, and metabolic pathways, indicating a broad impairment of programs required for mitochondrial energy production and renewal. Importantly, these changes are unlikely to reflect a primary reduction in mitochondrial number. Aging does not uniformly reduce mitochondrial content in the heart, and our prior work in a young HFpEF model showed preserved mtDNA copy number and preserved PGC-1α expression (30). In the present study, key downstream biogenesis regulators (NRF1, NRF2, and TFAM) were also preserved. Together, these data argue against impaired canonical biogenesis and instead support defective mitochondrial maintenance as the dominant mechanism underlying reduced mitochondrial protein abundance.

Sex-stratified proteomic analysis showed that the major mitochondrial response to HFpEF was broadly conserved between males and females, particularly the suppression of OXPHOS, Complex I, and TCA-cycle. However, the magnitude of these changes differed between sexes, and females showed additional enrichment of several stress- and metabolic-response pathways. These findings suggest that a common mitochondrial energetic defect is present in both sexes, while sex-specific adaptive responses may contribute to differences in the physiological phenotype.

In our study comparing whole-heart cytosolic and mitochondrial proteomics, we demonstrated that HFpEF is associated with broad remodeling extending beyond mitochondrial pathways, including changes in RNA processing and ubiquitin/proteasome-associated pathways. Nevertheless, mitochondrial-related pathways remain among the most prominently altered programs. These findings indicate that mitochondrial dysfunction occurs within a broader cardiac proteomic response and remains a major component of HFpEF remodeling.

Impaired proteostasis and a mitochondrial membrane potential (ΔΨm) import proteostasis feedforward cycle

The proteomic signature and immunoblot validation point to impaired mitochondrial proteostasis, including reduced chaperones (mtHSP60, mtHSP70, HSP10) and proteases central to protein quality control (ClpP and LONP1 etc.), accompanied by accumulation of ubiquitinated mitochondrial proteins, which indicates that damaged or misfolded mitochondrial proteins accumulate faster than they can be refolded or degraded, triggering ubiquitin-dependent quality-control pathways to mark dysfunctional proteins and mitochondria for extraction or turnover. These findings suggest that mitochondrial protein damage and aggregation are not adequately resolved in HFpEF cardiomyocytes. Increased mitochondrial ubiquitination alone cannot define proteostasis flux or localize the underlying defect. However, the concurrent suppression of mitochondrial protein-degradation pathways, reduced abundance of mitochondrial chaperones and proteases, and decreased ClpP protease activity support a defect in mitochondrial protein quality-control capacity. Direct measurements of protein turnover, autophagic flux, and mitophagy will be required to further define the specific step responsible for impaired proteostasis.

Here, we propose a feed-forward mechanism linking energetic impairment to progressive proteostasis failure. Reduced oxidative phosphorylation would be expected to lower ATP production and decrease ΔΨm, weakening the driving force for mitochondrial protein import and impairing assembly and renewal of nuclear-encoded mitochondrial proteins (31). In parallel, cardiometabolic stress and hypertension promote oxidative protein damage, increasing proteotoxic burden. With reduced chaperone/protease capacity, damaged proteins accumulate, destabilizing electron transport chain complexes and further worsening ΔΨm and ATP production. This ΔΨm-import-proteostasis cycle provides a mechanistic framework that links mitochondrial maintenance failure to the loss of energetic reserve and to HFpEF functional limitation.

Translational implications

These data suggest that reduced mitochondrial protein quality control is present in cardiomyocytes under conditions of aging, metabolic, and hemodynamic stress. Approaches that enhance mitochondrial proteostasis or reduce proteotoxic stress may stabilize respiratory chain function and improve energetic reserve, whereas strategies focused solely on stimulating mitochondrial biogenesis may be insufficient if maintenance capacity remains impaired.

Limitations

This model reflects a metabolic-hypertensive HFpEF phenotype and cannot capture the full heterogeneity of human HFpEF. Proteomics provides a static measure of protein abundance and does not directly quantify protein turnover, import flux, or the dynamic activation of mitochondrial stress responses; future work assessing proteome flux, aggregation, and import efficiency will be important for establishing causality. Finally, defining when proteostasis defects emerge, identifying the major drivers of these defects, and clarifying how mitophagy responds to impaired mitochondria during HFpEF development will be important areas for further study.

Conclusions

In an aged, clinically relevant three-hit HFpEF model affecting both sexes, cardiomyocyte mitochondria exhibit reduced respiratory reserve and coordinated suppression of oxidative metabolism and mitochondrial maintenance programs. Preserved biogenesis regulators, together with reduced chaperone/protease machinery and increased mitochondrial ubiquitination, support impaired mitochondrial proteostasis, rather than reduced biogenesis, as the primary driver of cardiomyocyte mitochondrial dysfunction. These findings highlight mitochondrial protein quality control as a mechanistic axis and a potential therapeutic target for restoring energetic reserve in HFpEF.

Supplementary Material

Supplemental Figs. S1-S6: https://doi.org/10.6084/m9.figshare.33330294

Supplemental Dataset S1-S2: Data-Independent Acquisition proteomics results for old male and female HFD+L-NAME vs RD mice https://doi.org/10.6084/m9.figshare.33334206