Loss of Ovarian Function Prevents Exercise-Induced Activation of Hepatic Mitophagic Flux

Abstract

Exercise effectively treats metabolic dysfunction-associated steatotic liver disease (MASLD) by enhancing hepatic mitochondria energy metabolism. However, the efficiency of exercise in treating MASLD in post-menopausal women may be reduced. Previously, we showed acute treadmill exercise activates hepatic mitophagy, the selective degradation of low-functioning mitochondria. Mitophagic flux is differentially regulated in female mice compared to males, possibly by estrogen. Here, we tested if loss of ovarian function via ovariectomy (OVX), which reduces estrogen, drives MASLD and compromised hepatic mitochondrial energetics, would blunt activation of hepatic mitophagy induced by exercise. Following OVX, 12–15-week-old female mice were placed on a low-fat diet (LFD) or high-fat diet (HFD) for 4 weeks to induce MASLD, after which half of the mice performed a single acute bout of treadmill exercise to exhaustion or remained sedentary. Two hours post-exercise, isolated hepatic mitochondria were examined via western blotting and proteomics for accumulation of known mitophagy proteins. After exercise, reduced basal mitophagic flux in LFD-fed OVX was restored to levels found in Sham mice. However, exercise possessed blunted capacity to promote mitochondrial recruitment of DRP1 (regulator of fission) and accumulation mitophagy-associated proteins (E3-ubiquitin ligase, ubiquitin, autophagy adaptor proteins, and autophagosome cargo receptors) in OVX versus sham mice on HFD. Mitochondrial H₂O₂ production, which putatively activates mitophagy, was elevated following exercise in all conditions except OVX+HFD. In summary, OVX reduces mitophagic flux, blunting the stimulatory effects of exercise on these factors. The impaired regulation of mitophagy following cessation of ovarian function likely contributes to the pathogenesis of MASLD post-menopause.

Graphical Abstract

Introduction

The concept of sex as a biological variable is a relatively novel area of focus in the study of metabolic dysfunction-associated steatotic liver disease (MASLD) [1]. Compared to other sexual dimorphic metabolic-associated diseases, few studies have focused on sex differences in the susceptibility and progression of MASLD [2]. However, mounting evidence of sex/gender-biased susceptibility to MASLD has shown that women possess innate protection against this disease prior to the onset of menopause [3–6]. Similar sex dimorphism for high-fat diet (HFD)-induced steatosis has been reported in rodents [7–9]. However, cessation of ovarian function (i.e. menopause, hysterectomy, ovariectomy) eliminates/diminishes this female-biased protection, and instead promotes MASLD risk to a greater degree than that found in males [10]. The simultaneous increase in susceptibility to MASLD and progressive depletion of sex hormone production suggests sex hormones (i.e. estrogen) may modulate the protection in women and female rodents. While studies interrogating the effects of sex hormones on the liver remain limited, data from the National Health and Nutrition Examination Survey (NHANES) III shows a significant reduction in the incidence of MASLD in post-menopausal women receiving hormone replacement therapy [11]. However, the mechanism(s) by which sex hormones impact MASLD development remains unknown.

Although individuals with MASLD possess expanded intrahepatic lipid storage, marked increases in the liver’s metabolic capacity, including fatty acid oxidation and tricarboxylic acid (TCA) cycle, have been found in both humans [12] and rodents with MASLD [13, 14]. Koliaki et al have reported paired increases in mitochondrial respiration in individuals with MASLD, likely a compensatory response driven by the increased substrate flux to hepatic mitochondria driven by excessive lipid storage [15]. However, hepatic mitochondrial adaptations during MASLD development are not sustained, with reduced mitochondrial respiration and increased oxidative stress associated with liver disease progression, i.e. MASLD transitioning to liver injury (inflammation and fibrosis) [15]. Importantly, participants studied by Koliaki et al. were predominantly women. This information, in conjunction with our previous findings of expanded hepatic mitochondrial respiratory capacity in female vs male mice, highlights potential links between hepatic mitochondrial respiration, sex hormones, and MASLD susceptibility.

We have previously shown that females possess expanded hepatic mitochondrial respiratory capacity in response to HFD feeding; however, these mitochondrial adaptations to HFD feeding are entirely absent in male mice [7, 8]. These data suggest an intrinsic female-biased elasticity in metabolic compensation to overnutrition. In skeletal muscle, the sex effects on mitochondrial function and metabolism are dependent on normal ovarian function, considering that OVX impairs mitochondrial pairing of oxygen consumption to adenosine triphosphate (ATP) production [16], and restoration of mitochondrial bioenergetics is only achieved after estradiol replacement [17, 18]. While these findings suggest that sex hormones are central to the maintenance of mitochondrial function in skeletal muscle, our understanding of the actions of sex hormones on hepatic mitochondrial function remains inadequate.

Selective autophagy of mitochondria, termed mitophagy, is a central means for maintaining organelle quality control and cellular metabolic health [19]. The activation of mitophagy serves as a protective mechanism to mitigate cellular oxidative stress induced by elevated reactive oxygen species (ROS) production and loss of mitochondrial membrane potential [20]. In the development of HFD-induced MASLD, decreased accumulation of mitophagy markers has been observed in mice [21]. Reductions in mitophagy likely contribute to the pathogenesis of MASLD, with genetic models of reduced mitophagic activity possessing exacerbated/accelerated progression of MASLD [22, 23]. The pharmacological stimulation of hepatic mitophagy has been shown to alleviate steatosis levels in HFD-fed mice, coinciding with improved parameters of mitochondrial respiration. [24]. In addition, previous reports show that induction of steatosis in humans and in rodents leads to compensatory increases in oxidative metabolism that are also accompanied by increased oxidative stress, a result that likely occurs due to a reduction in hepatic mitophagy [25, 26]. In contrast, we and others have shown that exercise increases hepatic oxidative capacity while reducing H₂O₂ emission. Activation of mitophagy with exercise likely mediates the positive adaptations induced by exercise (increased oxidative capacity paired with lower oxidative stress), which contrasts with the compensatory oxidative stress caused by the development of steatosis [26, 27]. Together, these data suggest the regulation of hepatic mitophagy is central to the maintenance of hepatic mitochondrial function and overall health of the liver.

We have reported sex differences in the protein expression of various mitophagy markers within the liver [28], suggesting female mice are more reliant on the activation of mitophagy to promote mitochondrial adaptations. Indeed, liver-specific knockout of the mitophagy receptor BNIP3 (BCL2/adenovirus E1b kDa protein-interacting protein 3) ablates the mitochondrial adaptations observed in female mice in response to the paired metabolic stress of HFD and voluntary wheel running [8]. Exercise is a powerful inducer of mitophagy acutely following an exercise bout [29, 30]. Reductions in the capacity for exercise-mediated activation of hepatic mitophagy may prevent mitochondrial adaptions occurring in response to exercise training that are obligatory for protection against MASLD and liver injury. Herein, we tested the hypothesis that loss of ovarian function impairs exercise-induced activation of hepatic mitophagic flux (rate of mitophagy over time) either on a low-fat diet (LFD) or HFD condition. We leveraged pharmacological blockage of autophagy with leupeptin, acute exercise, and isolated hepatic mitochondria with western blotting and proteomics methodologies to quantify mitophagy flux and mitochondrial-associated proteins that mediate effects. Our findings show that ovarian function is not requisite for acute exercise-induced hepatic mitophagic flux on an LFD. However, loss of ovarian function does blunt the hepatic mitophagic flux response when exercise is performed in the context of HFD feeding. These findings provide translational insight regarding mechanisms underpinning MASLD susceptibility during menopause and the reduced efficacy of exercise intervention to reverse MASLD.

Methods

Ethical approval

The animal protocol was approved by the Institutional Animal Care and Use Committee at the University of Kansas Medical Center and Kansas City Veterans Affairs Medical Center (animal protocol number 21–12-204). Experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals published by the National Institute of Health (NIH Guide, 8^th edition, 2011). Mice were anesthetized with pentobarbital sodium (100mg·kg⁻¹) prior to the terminal procedure.

Animals

12–15-week-old Female C57Bl/6J mice (128 in total) were obtained from Jackson Laboratory (Bar Harbor, ME) and group housed (4 mice per cage) at thermoneutrality (30°C) on a reverse light cycle (dark 10:00–22:00) with ad libitum access to water and low-fat diet (LFD; D1211 0704: 10% kcal fat, 3.5% kcal sucrose, and 3.85 kcal/g; Research Diets) for 2 weeks. At 14–17 weeks of age, mice were randomly selected to undergo a sham surgery (sham, n = 64) or bilateral ovariectomy (OVX, n = 64) with groups matched by initial body weight. Following surgery, half of the mice (n = 64) were maintained on LFD while the other half (n = 64) were transitioned to high-fat diet (HFD; D12451; 45% kcal fat, 17% kcal sucrose, 1% cholesterol wt/wt, 4.68 kcal/g; Research Diets) for 4 weeks. Body mass was assessed weekly, and body composition was determined via magnetic resonance (EchoMRI) at baseline and prior to euthanasia. Fat-free mass was calculated from the difference between total body mass and fat mass. Percent (%) body composition was determined by dividing fat mass and fat-free mass by total body mass, respectively. Vaginal cytology was performed to confirm the cessation of estrous cycle, as previously described [31].

Acute Exercise Paradigm

72 hours prior to acute exercise training, all mice were placed on a treadmill (76–0896, Panlab) for 10 minutes before completing a 15-minute acclimation treadmill running protocol, as previously described [30, 32]. At 9:30 AM on the day of study, mice received either a 250 μL injection of 0.9% saline or Leupeptin (4μg·μL⁻¹ or 40mg·kg⁻¹), a serine, threonine, and cysteine protease inhibitor, and food was pulled at 10:00 AM as performed previously [30]. Mice were then randomly separated into sedentary (Sed) and treadmill running (Ex) groups (n=8 for all 16 treatment groups) and all mice, regardless of group, were immediately placed on a sedentary treadmill. After 5 minutes on the treadmill, Ex mice performed a single, ~1 hour bout of progressive treadmill running with incremental increases in both speed and incline of the treadmill until all mice reached exhaustion, as determined by the animals touching the back portion of the treadmill 3 consecutive times (Supplemental Table S1). Sed mice remained on a separate, idle treadmill for the duration of the running paradigm to control the stress of being removed from the home cage. Following treatment, mice were removed from the treadmill and placed back into their cages without access to food for 2 hours prior to being euthanized at ~12:00 PM. The 2-hour post-exercise time point was selected based on our previous study using the same protocol which showed significant accumulation of mitophagy markers 2 hours following exercise [30].

Hepatic Mitochondrial Isolation

Livers were excised with half of the liver immediately snap frozen in liquid N₂ and the other half placed in 8mL of ice-cold mitochondrial isolation buffer (220 mM mannitol, 70 mM sucrose, 10 mM Tris, 1 mM EDTA, pH adjusted to 7.4 with KOH) and mitochondria were isolated as previously described [33]. Briefly, samples were homogenized using a Teflon pestle and centrifuged at 1,500 g for 10 minutes at 4°C. The supernatant was strained and centrifuged at 8,000 g (10 minutes at 4°C). The pellet was resuspended in 6mL of mitochondrial isolation buffer using glass-on-glass Dounce homogenizers and pelleted at 6,000 g (10 minutes at 4°C). The process was repeated with 4mL of mitochondrial isolation buffer containing 0.1% fatty acid-free bovine serum albumin before being spun at 4,000 g (10 minutes at 4°C). The final Isolated mitochondria pellet was resuspended in 400–500μL modified MiR05 mitochondrial respiration buffer (110 mM sucrose, 60 mM KMES, 20 mM glucose, 20 mM HEPES, 10 mM KH₂PO₄, 3 mM MgCl₂, 0.5 mM EGTA, and 0.1% bovine serum albumin, pH adjusted to 7.1 with KOH). Mitochondrial protein concentration was determined using bicinchoninic acid assay.

Mitochondrial Respiration

Real-time measures of the rates of mitochondrial oxygen consumption (JO₂) and simultaneous production of H₂O_2, were assessed using the Oroboros O2k-Fluorometer (Oroboros Instruments), as previously described [34]. All mitochondrial respirometry experiments were conducted in 2mL modified MiR05 mitochondrial respiration buffer with the addition of 2 mM Malate, 63.5 μM free CoA, and 2.5 mM _L-carnitine at a chamber temperature of 37°C. Lactobionic acid and taurine were removed from the MiR05 buffer to avoid possible interference with H₂O₂ emission measures. Coupled mitochondrial respiratory capacity and H₂O₂ production was determined under two separate substrate conditions 10μM _L-palmitoyl-CoA (PCoA) or 5 mM potassium pyruvate. Basal respiration was determined prior to the addition of 2.5 mM adenosine 5’-diphosphate (State 3). To assess the limitation of CPT1a on respiratory rates, palmitoyl-carnitine (PC; 10 μM) was added to the PCoA-supported respiration (State 3+PC). Glutamate (Glut; 2 mM) was added to pyruvate-supported respiration to assess maximal complex I-driven respiration. 10mM succinate (Succ) was added to both protocols to determine maximal complex I+II respiration (State 3S). Titrations of 0.1 μM carbonyl cyanide-p-trifluoromethyoxyphenylhydrazone (FCCP; uncoupled) were then added to all chambers to determine maximal uncoupled mitochondrial respiratory rates. Coupling control ratio was calculated from the equation: Coupling control ratio = 1 – (Leak JO₂/State 3 JO₂). % electron leak was determined by dividing H₂O₂ production by JO₂ under basal respiratory conditions. All mitochondrial respiration experiments were analyzed using DatLab 7 (Oroboros Instruments). All measures of Mitochondrial respiration were performed in saline-treated mice to avoid possible confounding effects of leupeptin on mitochondrial function.

Creatine Kinase Clamp

Mitochondrial capacity to modulate JO₂ in response to changes in free energy of ATP hydrolysis (ΔG_ATP) was assessed using a modified version of the creatine kinase (CK) clamp in the Oroboros O2k [35]. Isolated mitochondria were added to following the addition of creatine monohydrate (5 mM), PCoA (10 μM), PC (10 μM), CK (20U), and adenosine 5’-triphosphate (2.5 mM). Phosphocreatine was added serially to achieve concentrations within the chambers of 1mM, 3mM, 6mM, 9mM, 12mM, 15mM, 18mM, and 21mM. ΔG_ATP was calculated for each concentration of phosphocreatine and used to calculate conductance from the range of ΔG_ATP with a linear relationship to JO₂, as previously described [36].

Mitochondrial Antioxidant Enzyme Capacity

Isolated mitochondria were injected into Oroboros O2k chambers containing MiR05 respiration buffer with the addition of PCoA (10 μM) and PC (10 μM) to determine maximal H₂O₂ production rates under basal respiratory conditions, as previously described [31]. Assessment of thioredoxin/peroxiredoxin enzyme activity was determined following the addition of 0.25 μM auranofin and activity of the glutathione antioxidant enzymes was determined following the addition of 0.25 μM 1-chloror-2,4-dinitrobenzene (CDNB).

Quantitative Determination of Mitophagic Flux

Western-ready samples were produced from hepatic isolated mitochondria as previously described [33], separated via SDS-PAGE, and transferred to a polyvinylidene difluoride membrane. Membranes were probed using the following primary antibodies at a concentration of 1:2000: sequestosome 1 (SQSTM1/p62; Cell signaling, 5114), microtubule-associated protein one light chain 3 A/B (LC3A/B; Cell Signaling, 12741S), Ubiquitin (3933), and total OXPHOS (Abcam, ab110413) followed by a secondary antibody solution (1:10,000). All primary and secondary antibodies, besides total OXPHOS, were purchased from Cell Signaling Technology. Membranes were imaged, and densitometry was used to quantify individual protein bands (Bio-Rad Laboratories). All protein bands were normalized to total mitochondrial proteins using 0.1% amido-black staining.

Isolated Mitochondrial Proteomics

In a subset of samples (n=5 per group/HFD only), 100μg of protein taken from the liver mitochondrial isolates was prepped for proteomics, as previously described [37]. Briefly, samples were centrifuged at 10,000 g at 4°C and supernatant was removed. The mitochondrial pellet was resuspended in 50μL of RIPA lysis buffer (Thermo Scientific) with phosphatase (phosphatase inhibitor cocktail 2,3, Sigma, St. Louis, MO) and protease (protease inhibitor cocktail tablet, Rocher, Mannheim, DEU) inhibitors. Proteins were then reduced, alkylated, and purified using chloroform/methanol extraction. Sequencing grade modified porcine trypsin (Promega) was used to digest the proteins, and the resulting tryptic peptides were separated by reverse phase XSelect CSH C19 2.5 μM resin (Waters) on an in-line 150 × 0.075 mm column using an UltiMate 3000 RSLCnano system (Thermo). Using a 60-minute gradient, peptides were eluted from 98:2 to 65:35 buffer A:B ratio (Buffer A: 0.1% formic acid, 0.5% acetonitrile; Buffer B: 0.1% formic acid, 99.9% acetonitrile). Eluted proteins were ionized by 2.2kV electrospray followed by mass spectrometric analysis via an Orbitrap Exploris 480 mass spectrometer (Thermo). The chromatogram library was assembled using six gas-phase fractions acquired on the Orbitrap Exploris using 4 m/z DIA spectra (4 m/z precursor isolation windows at 30,000 resolution, normalized AGC target 100%, maximum inject time 66 ms) via a staggered window pattern from narrow mass ranges using optimized window placements. Precursor spectra were acquired after each DIA duty cycle, spanning the m/z range of the gas-phase fraction (i.e., 496–602 m/z, 60,000 resolution, normalized AGC target 100%, maximum injection time 50 ms). For wide-window acquisitions, the Orbitrap Exploris was configured to acquire a precursor scan (385–1015 m/z, 60,000 resolution, normalized AGC target 100%, maximum injection time 50 ms) followed by 50× 12 m/z DIA spectra (12 m/z precursor isolation windows at 15,000 resolution, normalized AGC target 100%, maximum injection time 33 ms) using a staggered window pattern with optimized window placements. Precursor spectra were acquired after each DIA duty cycle.

Mitochondrial Proteomics Data Analysis

Protein data was searched using an empirically correct library against the UniProt Mus musculus database and quantitative analysis was performed to obtain a comprehensive proteomics profile. Proteins were identified and quantified using EncyclopeDIA and visualized using Spectronaut (BIOGNOSYS). A 1% false discover threshold was used at both the protein and peptide levels. Protein-exclusive intensity values were assessed for quality using ProteiNorm [38]. The data was normalized using cyclic loess [39] and statistical analysis was performed using proteoDA [40] with linear models for microarray data (limma) with empirical Bayes (eBayes) smoothing to the standard errors [40]. Identified proteins were cross-referenced with MitoCarta3.0 gene inventory to identify mitochondrial-specific proteins[41]. Ingenuity Pathway Analysis (IPA, Qiagen) was used to determine the pathway enrichment of all identified mitochondrial proteins within the proteomics analysis using log₂ fold change (FC) and p-value to determine directionality. Expressional log₂FC and p-value were used to determine directional Z-score for each pathway, with a Z-score of +/− 2 deemed significant. To compare protein abundance differences across all groups, peptide intensity identified from the MS2 spectra was normalized to the summed intensity of all peptide sequences within each independent sample, as previously described [32]. MitoCarta3.0 was utilized to group proteins into enzymatic pathways for determination of average log₂FC differences across groups.

Statistics

Statistical analyses were performed using SPSS Statistics 29 (IBM). A 3-way ANOVA was used to determine main effects of OVX (OVX vs Sham), exercise (Ex vs Sed), and Leupeptin (Leu vs Sal) within diet. A 3-way ANOVA was also used to determine the main effects of OVX, exercise and diet (LFD vs HFD) within Leupeptin samples. A least significant difference (LSD) post hoc test was performed when a significant interaction was identified. Due to limited identification of mitophagy-related proteins in saline samples for proteomics analysis, the main effects of OVX and exercise within leupeptin samples were assessed via 2-way ANOVA with a post hoc test performed when a significant main effect was observed. Significance was set at p < 0.05. and all data are presented as mean ± standard error of the mean (SEM) (GraphPad Prism 10).

Results

OVX Alters Body Composition

Body composition was identical across all groups immediately prior to OVX/Sham surgical procedure (Supplemental Table S2). After 4-week dietary intervention, LFD-fed OVX mice had similar body weight to Sham mice. However, OVX significantly altered body composition in LFD-fed mice, promoting greater fat-free mass and reduced fat mass compared to Sham mice (Table 1, p<0.001). Regardless of OVX vs. Sham groups, HFD feeding significant increased both fat-free and fat mass (p<0.001). In contrast to the LFD-fed condition, HFD promoted expanded weight gain due to greater fat- and fat-free mass in OVX versus Sham mice (Table 2, p<0.001). This enhanced weight gain in HFD-fed plus OVX conditions appears to be disproportionally driven by increased adiposity, with OVX mice possessing greater % fat mass compared to Sham mice (Table 2, p<0.001). This aligns with previous reports of increased fat mass in OVX rodent models [16] and post-menopausal women [42].

Table 1.

Post-Dietary Intervention Anthropometries for LFD Groups

	LFD								2-Way ANOVA P-Value
	Sham				ovx
	Sed		Ex		Sed		Ex
	Sal	Leu	Sal	Leu	Sal	Leu	Sal	Leu	OVX	Ex	Leu
Pre-Acute Exercise Body Composition (g)
Body Mass	21.70 ± 0.52	21.65 ± 0.43	21.50 ± 0.71	20.98 ± 0.88	21.50 ± 0.39	22.35 ± 0.56	21.46 ± 0.60	21.60 ± 0.53	0.522	0.329	0.808
Fat Mass	3.88 ± 0.52	3.94 ± 0.16	3.69 ± 0.51	3.69 ± 0.68	2.77 ± 0.22	3.05 ± 0.25	2.66 ± 0.24	2.76 ± 0.24	<.001	0.449	0.698
Fat-Free Mass	17.82 ± 0.25	17.71 ± 0.30	17.81 ± 0.26	17.29 ± 0.39	18.23 ± 0.31	18.71 ± 0.42	18.63 ± 0.18	18.82 ± 0.33	<.001	0.411	0.981
Percent Body Composition (%)
% Fat Mass	17.63 ± 1.85	18.18 ± 0.49	16.75 ± 1.90	16.94 ± 2.39	12.83 ± 0.91	13.55 ± 0.86	12.30 ± 0.87	12.71 ± 0.87	<.001	0.387	0.644
% Fat-Free Mass	82.37 ± 1.85	81.82 ± .50	83.25 ± 1.90	83.06 ± 2.39	87.17 ± 0.91	86.45 ± 0.86	87.70 ± 0.87	87.29 ± 0.86	<.001	0.387	0.644

Data are presented as mean ± SEM , n=8 per group

Table 2.

Post-Dietary Intervention Anthropometric for HFD Groups

	HFD								2-Way ANO VA P-Value
	Sham				OVX
	Sed		Ex		Sed		Ex
	Sal	Leu	Sal	Leu	Sal	Leu	Sal	Leu	OVX	Ex	Leu
Pre-Acute Ex Body Composition (g)
Body Mass	24.62 ± 1.07	23.90 ± 0.79	23.94 ± 0.56	24.46 ± 1.07	27.85 ± 1.15	26.45 ± 1.05	27.09 ± 1.43	26.91 ± 0.62	<.001	0.885	0.538
Fat Mass	6.38 ± 0.97	5.20 ± 0.55	5.31 ± 0.53	5.64 ± 0.87	7.71 ± 0.71	7.01 ± 0.66	7.52 ± 0.93	7.51 ± 0.42	<.001	0.877	0.45
Fat-Free Mass	18.73 ± 0.28	19.30 ± 0.28	18.80 ± 0.44	18.84 ± 0.40	20.14 ± 0.47	19.44 ± 0.45	19.57 ± 0.57	19.40 ± 0.32	<.001	0.937	0.866
Percent Body Composition (%)
% Fat Mass	25.09 ± 2.96	21.44 ± 1.78	21.92 ± 1.76	22.35 ± 2.52	27.33 ± 1.35	26.13 ± 1.52	27.07 ± 1.95	2.78 ± 1.06	<.0.01	0.877	0.505
% Fat-Free Mass	74.91 ± 2.96	78.56 ± 1.78	78.08 ± 1.76	77.65 ± 2.52	72.67 ± 1.35	73.87 ± 1.52	72.93 ± 1.95	72.21 ± 1.06	<.0.01	0.877	0.505

Data are presented as mean ± SEM , n=8 per group

Effective Inhibition of Mitophagy

Because mitophagy is a transient process, interpretation of mitophagic flux requires the use of autophagy inhibitors that permit the capture of mitophagic flux. To investigate whether OVX alters the rate of basal and exercise-stimulated mitophagic flux, we utilized leupeptin, a protease inhibitor of autophagy/mitophagy [30, 43]. Classic protein markers of mitophagy associated with isolated hepatic mitochondria –SQSTM1/p62 (Sequestosome 1), LC3-II (lipidated microtubule-associated protein 1 light chain 2 alpha), and total ubiquitin – were quantified to assess the activation of mitophagic flux.

Leupeptin treatment effectively blocked mitochondrial degradation, as evidenced by the accumulation of mitophagy markers, p62, LC3-II, and total ubiquitin, on isolated mitochondria across all diet and surgical conditions (Fig. 1A–F, main effect leupeptin p<0.05). In contrast, saline-treated mice showed no differences in the expression of mitophagy markers. Because of this, subsequent analysis focused on leupeptin-treated groups. Within leupeptin-treated groups, we observed a unique interaction between diet, OVX, and exercise conditions. Under sedentary conditions, LFD-fed OVX mice exhibited significantly lower p62 accumulation versus Sham mice (Fig. 1G, p < 0.05). However, acute exercise promoted p62 accumulation on mitochondria in LFD-fed OVX to levels observed in comparable Sham mice. Regardless of surgical intervention, HFD-feeding significantly increased mitochondrial-bound p62 in all groups over LFD (Fig. 1G, main effect diet p < 0.0001). Under HFD conditions, basal rates of mitophagic flux were similar between sedentary Sham and OVX mice. Mitochondrial p62 abundance was further increased in Sham mice by acute exercise (Fig. 1G, p < 0.05), while this effect was blunted in OVX+HFD. Overall, OVX mice showed reduced mitochondrial-bound p62, which was a likely consequence of decreased basal mitophagic flux under LFD-fed conditions and reduced recruitment of p62 to mitochondria following acute exercise in the context of HFD-feeding. (Fig. 1G, main effect p < 0.05). Similarly, mitochondrial association with autophagosomal LC3-II was reduced in OVX versus Sham following acute exercise (Fig. 1H main effect p < 0.05). Neither exercise nor diet significantly altered LC3-II binding to the mitochondria.

Figure 1: OVX, Acute Exercise, and Diet modulate the activation of mitophagic flux.

Western blot densitometric quantification of mitophagy markers p62/SQSTM1 (A), LC3II (B), and total Ubiquitin (C) 2 hours post-acute exercise in LFD fed and HFD fed mice (D-F). Comparison of mitophagic flux activation based on densitometry-based accumulation of p62 (G) and LC3II (H) in leupeptin-treated mice (I). Representative blots for p62, LC3-II and total Ubiquitin. Data are represented as means ± SE (n=7–8 mice per group). Leu: main effect of leupeptin, OVX: main effect of ovariectomy, Ex: main effect of acute exercise, θxΓ: interaction of OVX and Leupeptin, θxΔxΣ: interaction of OVX, Diet, and Exercise from 3-way ANOVA. θ and Σ denote significant LSD post hoc for OVX and Exercise respectively. Significant values are p < 0.05.

Proteomics Analysis of Hepatic Mitophagic Flux in HFD-fed Mice

To investigate mechanisms that dampen the activation of mitophagic, we performed unbiased proteomics analysis on isolated liver mitochondria of HFD-fed mice. With the apparent reduction in p62 levels within OVX mice, we posited that expression of other mitophagy-regulatory proteins may also be altered. Hierarchical clustering and PCA analysis confirmed distinct protein expression profiles between saline and leupeptin-treated animals (Supplemental Fig. S1A-B).

In Sham mice, acute exercise significantly increased mitochondrial association of E3 ubiquitin ligases (Fig. 2A, main effect p<0.01), leading to a corresponding exercise-induced accumulation of mitochondrial UBB (ubiquitin B) (Fig. 2B, main effect p<0.05). This exercise-induced accumulation of mitochondrial-bound UBB was accompanied by increased recruitment of ubiquitin-binding adapter proteins in Sham mice after exercise (Fig. 2C, main effect p<0.05), including OPTN (optineurin), NBR1 (neighbor of BRCA1 gene 1), NDP52 (Calcium binding and coil-coil domain 2) (Fig. 2D–F). Interestingly, the accumulation of p62 was not different between OVX and Sham mice (Fig. 2G). Notably, the effect of exercise appeared specific to proteins involved in ubiquitin-mediated mitophagy, with no changes observed in the expression of receptor-mediated mitophagy proteins (i.e. BNIP3, NIX, FUNDC1, Fig. 2H). Despite this, acute exercise increased mitochondrial association with autophagosomal cargo receptors, LC3-II and GABARAP (gamma-aminobutyric acid associated protein), in Sham versus OVX mice (Fig. 2I–J). Importantly, OVX mice failed to exhibit this exercise-mediated recruitment of mitophagy machinery to the mitochondria, suggesting OVX may blunt the capacity for exercise to stimulate hepatic mitophagic flux.

Figure 2: Exercise-mediated accumulation of mitophagy proteins in HFD-fed mice.

Summed MS2 raw intensity of all identified E3-ubiquitin ligases normalized to total intensity of all proteins per sample (A). Normalized MS2 raw intensity of UBB (B). Normalized summed raw intensity of all mitophagy adapter proteins (C). Normalized raw intensity of adapter proteins: OPTN (D), NBR1 (E) and NDP52 (F), p62/SQSTM1 (G). Normalized summed raw intensity of all mitophagy receptors (H)Accumulation of Autophagosome cargo receptor proteins LC3II (I) and GABARAP (J) expressed as normalized raw intensity. Data are represented as means +/− SE (N=4–5 mice per group). Leu: main effect of leupeptin by 3-way ANOVA. OVX: main effect of OVX within leupeptin treatment, Ex: main effect of acute exercise within leupeptin treatment, θxΣ: interaction between OVX and Ex within leupeptin treatment by 2-way ANOVA. θ and Σ denote significant LSD post hoc for OVX and Exercise, respectively. Significance established a priori p < 0.05. BioRender was utilized to generate figure graphics.

Mitochondrial Dynamics

Because mitochondrial quality control via mitophagy is linked with mitochondrial dynamics, we next assessed fission-fusion signaling. The process of mitochondrial fission is initiated by the recruitment of a GTPase, DRP1 (dynamin-related protein 1) to the mitochondrial outer membrane, binding to the docking proteins FIS1 (mitochondrial fission protein 1) and MFF (mitochondrial fission factor). MFF expression was increased following exercise in sham mice treated with leupeptin. However, MFF expression was unaffected by exercise in OVX mice (Fig. 3A, p<0.05,). Similarly, FIS1 expression was reduced in OVX mice regardless of exercise status (Fig. 3B, main effect p < 0.05). While we did not observe significant alterations to DRP1 levels beyond the effects of leupeptin treatment, this is likely due to DRP1 not being constitutively localized to the mitochondria (Fig. 3C). Working in opposition to fission, mitochondrial fusion is regulated by MFN1 (mitofusin-1), MFN2 (mitofusin-2), and OPA1 (optic atrophy 1). In accordance with previous reports, no difference in the expression of proteins associated with mitochondrial fusion (MFN1, MFN2, and OPA1) were observed with exercise training (Fig. 3D–F) [30]. To understand the relationship between mitochondrial dynamics in OVX versus Sham, we directly compared the expression of fission-fusion proteins under leupeptin-treated sedentary (Fig. 3G) and acute exercise conditions (Fig. 3H). While negligible differences were observed between sedentary mice, exercise promoted a significant shift in DRP1 in Sham mice (Log₂ fold change (FC) = 0.31, p < 0.05), resulting in roughly 20% greater mitochondrial-associated levels compared to OVX mice.

Figure 3: Accumulation of mitochondrial fission proteins in HFD-fed mice.

MS2 raw intensity normalized to raw intensity of total proteome for the fission proteins MFF (A), FIS1 (B) and DRP1 (C), and the fusion proteins MFN1 (D), MFN2 (E), and OPA1 (F). Log₂ fold change of protein regulating mitochondrial dynamics between OVX and Sham sedentary (G) and exercise (H) mice treated with leupeptin. Data are represented as means +/− SE (N=4–5 mice per group). Leu: main effect of leupeptin by 3-way ANOVA. OVX: main effect of OVX within leupeptin treatment, θxΣ: interaction between OVX and Ex within leupeptin treatment by 2-way ANOVA. θ and Σ denote significant LSD post hoc for OVX and Exercise respectively. Significance log₂ fold change is denoted by *, p < 0.05. BioRender was utilized to generate figure graphics.

Mitochondrial Pathway Analysis

With the observed reductions in mitophagic flux in OVX mice, we next determined other mitochondrial pathways that may be influenced by OVX. Using the MitoCarta3.0 database, 836 unique mitochondrial proteins out of the 1,140 within MitoCarta3.0 (73.3%) were identified, matching previous reports [30]. Outside the context of mitophagic flux, the accumulation of proteins induced by leupeptin convolutes pathway enrichment analysis; thus, we focused our global analysis of alterations in the mitochondrial proteome between OVX and Sham mice within saline-treated animals.

Pathway analysis was performed by comparing differences in the expression of proteins in saline-treated OVX versus Sham protein expression under both sedentary and exercise conditions (Fig. 4A–B). Within the sedentary groups, OVX mice possessed enriched protein expression for pathways associated with mitochondrial protein translation/import, β-oxidation, and branched-chain amino acid (BCAA) catabolism (Fig. 4A). Following the acute bout of exercise, OVX mice possessed elevated expression of proteins involved in oxidative phosphorylation (OXPHOS), TCA cycle protein translation/import and BCAA catabolism (Fig. 4B). Although not the focus here, enrichment of these pathways was also mainly consistent in leupeptin treated groups (Supplemental Fig. S2G-I). Heatmaps were generated to visualize individual and average differences in log2FC of protein expression for OXPHOS (Fig. 4C), TCA cycle (Fig 4D), pyruvate transport (Fig. 4E), β-oxidation (Fig. 4F), and BCAA catabolism pathways (Fig. 4G) between OVX versus Sham mice. Overall changes in individual protein expressional profiled in pathways associated with fatty acid oxidation, pyruvate metabolism, BCAA catabolism, TCA cycle, and OXPHOS pathways within the mitochondrial between OVX versus Sham mice under both sedentary and exercise conditions are depicted in Figure 5.

Figure 4. Mitochondrial pathway enrichment in saline-treated mice.

Top 15 significantly upregulated pathways based on Z-score in OVX versus Sham mice under sedentary (A) and Exercise conditions (B) following cross-reference with MitoCata3.0. Heatmaps for individual protein log₂ fold change involved in OXPHOS (C), TCA Cycle (D), pyruvate transport (E), fatty acid oxidation (F), and branched-chain amino acid catabolism (G). Summed MS2 intensity of all proteins composing complex I-V of the OXPHOS system normalized to summed total mitochondrial protein intensity (H). Data are represented as means +/− SE (N=4–5 mice per group). Significant pathways cutoff of Z-score > 2 with a -log₁₀ p-value > 1.3. * represents a significant unpaired t-test (p < 0.05).

Figure 5. OVX versus Sham mitochondrial protein expression.

Log₂ fold change of proteins involved in fatty acid oxidation, pyruvate metabolism, branched-chain amino acid catabolism, TCA Cycle, and Complexes I-V of the OXPHOS system. BioRender was utilized to generate figure graphics.

Hepatic Mitochondrial Respiratory Capacity

Mitophagy serves a central role in regulating the functional pool of mitochondrial. Thus, we next assessed whether the impaired activation of mitophagy in OVX mice coincided with altered hepatic mitochondrial respiratory capacity in the same mitochondrial isolates (Fig. 6, Supplemental Fig. S3). Using palmitoyl-CoA as the primary carbon source for respiration, we observed no differences in basal mitochondrial oxygen consumption (JO₂) across all groups (Fig. 6A). While no differences in mitochondrial respiratory capacity were noted between Sham and OVX mice fed LFD, HFD-fed OVX liver mitochondria displayed reduced State 3 (Fig. 6B, p=0.06,), State 3+PC (Fig. 6C, p<0.05) and State 3S (Fig. 6D, p<0.05) JO₂ compared to sham livers. This reduction in respiratory capacity in OVX is attributed mainly to a reduced capacity to increase JO₂ in response to adenosine diphosphate (ADP) (Supplemental Fig. S3A, p<0.05) that seem to be independent of differences in Cpt1α-dependent and maximal-complex II-dependent respiration (Supplemental Fig. S3B-C). While no differences in maximal respiratory capacity (FCCP-induced) were observed (Fig. 6E), this was attributed to OVX liver mitochondria possessing greater spare capacity than Sham (Supplemental Fig. S3D, p<0.05). The HFD-induced reduction in mitochondrial respiratory capacity appears to be specific to respiration supported by fat-derived fuel sources, as no differences in mitochondrial JO₂ were observed between OVX and Sham mice using glycolytic fuel sources under basal or State 3 conditions (Fig. 6F–J). Interestingly, exercise repressed basal rates of mitochondrial JO₂ across both OVX and Sham groups (Fig. 6F, main effect p<0.01).

Figure 6. OVX alters mitochondrial respiratory capacity and conductance.

Saline-treated Isolated haptic mitochondrial respiratory capacity measures supported by palmitoyl-CoA under basal (A), ADP-stimulated (State 3) (B), ADP-stimulated plus palmitoyl-carnitine (State 3+PC) (C), ADP-stimulated plus succinate (State 3+S) (D) and uncoupled (E) respiratory states. Basal (F), State 3 (G), State 3 + glutamate (State 3+Glut) (H), State 3 S (I) and uncoupled respiration (J) were measured using pyruvate as the starting substrate. A modified version of the CK clamp was performed under palmitate-supported conditions (K-L) to determine mitochondrial capacity to response to changes in physiological levels of energetic demand (Conductance) (M). Data are represented as means +/− SE (N=7–8 mice per group). Ex: main effect of acute exercise, θxΔ: interaction between OVX and Diet by 3-way ANOVA. θ and Δ denote significant LSD post hoc for OVX and Exercise respectively, p < 0.05.

We also utilized a modified form of the creatine kinase (CK) clamp to determine if the reduced mitochondrial JO₂ in HFD+OVX mice was attributed to reduced capacity to modulate respiration in response changes in free energy of ATP hydrolysis (ΔG_ATP) (Fig. 6K–L). This data was quantified as conductance or the rate of change in JO₂ in response to predicted changes in ΔG_ATP. Paradoxically, while HFD predictably reduced conductance in OVX mice, LFD-fed OVX mice displayed greater capacity to modulate JO₂ in response to changes in ΔG_ATP throughout the course of the CK clamp (Fig. 6M, Int: OVX x Diet p<0.001).

Mitochondrial-Derived H₂O₂ Production

Following OVX, ROS production is commonly elevated and serves as a key indicator of poorly functioning mitochondria. In the present study, we measured mitochondrial hydrogen peroxide (H₂O₂) production under PCoA- and pyruvate-stimulated respiratory conditions as a direct readout of mitochondrial-derived ROS during measures of mitochondrial JO₂ (Fig. 7A–B). H₂O₂ production was corrected by JO₂ to determine the percent (%) electron leak. During PCoA-supported respiration, no differences in % electron leak was observed (Fig. 7C). However, exercise elicited a significant increase in pyruvate-mediated % electron leak (Fig. 7D, main effect p<0.01). Importantly, acute exercise failed to promote greater H₂O₂ production in HFD-fed OVX mice (Fig. 7B, D). To account for possible differences in antioxidant capacity between OVX versus Sham mice, we selectively blocked the peroxiredoxin and glutathione antioxidant pathways via the addition of auranofin (Fig. 7E) and 1-chloror-2,4-dinitrobenzene (CDNB) (Fig. 7F), respectively. Surprisingly, the activity of both pathways was similar between OVX and Sham mice (Fig. 7G–H). This was supported by minimal differences in the expression of antioxidant enzymes between OVX and Sham mice (Fig. 7I, Supplemental Fig. S4), although exercise did promote greater total expression of mitochondrial glutathione and peroxiredoxin antioxidant enzymes (Supplemental Fig. S4D, H, main effect p < 0.05). While H₂O₂ production was not assessed in leupeptin-treated mice, proteins within the glutathione antioxidant pathway were increased in OVX mice (Supplemental Fig. S4D, main effect OVX p < 0.05). Although OVX and Sham mice possessed similar levels of total peroxiredoxin antioxidant enzyme expression, Prdx3 (peroxiredoxin 3) did show altered expression with leupeptin in OVX versus Sham mice (Supplemental Fig. S4J, main effect OVX p < 0.05), suggesting that OVX mice may rely on greater turnover of mitochondrial antioxidant enzyme to maintain redox status.

Figure 7. Acute exercise promotes pyruvate-supported H₂O₂ production.

Palmitate- (A) and Pyruvate- (B) supported H₂O₂ production under basal respiratory conditions. Percent (%) electron leak was determined for both palmitate (C) and pyruvate (D) substrate conditions by normalizing H₂O₂ flux to oxygen consumption (JO₂). H₂O₂ production was determined following the addition of inhibitors for the peroxiredoxin (AF) (E) and glutathione (CDNB) (F) antioxidant pathways. The antioxidant capacity of these pathways was determined by calculating the change in H₂O₂ production following the addition of both AF (G) and CDNB (H), respectively. Log₂ fold change of mitochondrial antioxidant enzymes in OVX versus Sham mice (I). Data are represented as means ± SE (n=7–8 mice per group). Ex: main effect of acute exercise, θxΣxΔ: interaction of OVX, acute exercise, and diet from 3-way ANOVA. θ, Σ, and Δ denote significant LSD post hoc for OVX, exercise, and diet, respectively.

Discussion

The transition into menopause is associated with a roughly 2.4 times greater risk for the development of MASLD, independent of other metabolic factors [10]. Due to the myriads of systemic changes within the female body that occur during this period, particularly the gradual depletion of endogenous sex hormone production, it is believed that sex hormones play a central role in modulating risk or protection against MASLD. However, the underlying mechanisms remain elusive. In the present study, we sought to investigate the impact of loss of ovarian function on exercise-induced hepatic mitophagy. We also tested if OVX-induced changes in mitochondrial respiratory capacity and H₂O₂ coincided with alterations in hepatic mitophagy. This study revealed context-specific impairments in mitophagic flux within the liver of OVX mice. While exercise was capable of restoring mitophagic flux in LFD-fed OVX mice to Sham levels, this capacity to activate mitophagy following exercise was significantly reduced following HFD feeding. These deficiencies in mitophagic flux coincided with reductions in both DRP1 recruitment and mitochondrial H₂O₂ production. In combination, these events likely contribute to the reduced mitochondrial respiratory capacity reported in HFD-fed OVX mice. Overall, these data show a direct role of ovarian function in maintaining the capacity to activate hepatic mitochondrial quality control processes following exercise, which may be central to the innate protection against MASLD in women before the loss of ovarian function [7, 8, 28].

Mitophagy plays a central role in maintaining healthy pools/networks of mitochondria within the liver, with impairments in mitophagy linked to the development and progression of MASLD [19]. Activation of mitophagy occurs through two distinct mechanisms: receptor-mediated and ubiquitin-mediated processes. The key distinction between these processes is that receptor-mediated mitophagy utilizes specialized proteins capable of directly interacting with the autophagosome, while ubiquitin-mediated mitophagy is reliant upon a ubiquitin signal for activation of mitophagy [19]. This ubiquitin signal required for the activation of ubiquitin-mediated mitophagy occurs through the recruitment of E3-ubiquitin ligases to the mitochondria in response to mitochondrial stressors (i.e., membrane depolarization and excessive ROS production). Once localized to the mitochondria, E3 ubiquitin ligases catalyze the transfer of ubiquitin to proteins localized on the outer mitochondrial membrane, with the resulting formation of poly-ubiquitin chains permitting the docking of autophagy adaptor proteins (OPTN, p62, NBR1, NDP52). By associating with the autophagosomal cargo receptors LC3-II and GABARAP, adaptor proteins effectively tether the mitochondria to the autophagosome. Following engulfment by the autophagosome, the mitochondria are trafficked to the lysosome for degradation [44]. The capacity for acute exercise to stimulate mitophagic flux is well-described in skeletal muscle, resulting in notable accumulation of mitochondrial-associated p62, LC3-II, and ubiquitin [45, 46]. However, only recently has acute exercise been shown to elicit similar activation of mitophagic flux within the liver [30]. Despite the mounting evidence implicating defective mitophagy in the pathogenesis of MASLD, mechanisms that regulate hepatic mitophagy remain largely unknown [19]. Here, we provide evidence for sex hormones playing a critical role in regulating the exercise-induced activation of ubiquitin-mediated hepatic mitophagic flux. Although we leveraged acute exercise primarily as a tool to induce mitophagy, our approach also provides a mechanistic assessment of how exercise modifies hepatic mitochondria. Leveraging our mitochondrial-centric proteomics analysis, we identified a dynamic accumulation of mitophagy-associated proteins on the hepatic mitochondria of HFD-fed Sham mice in response to a single bout of acute exercise. This included critical components involved in all phases of ubiquitin-mediated mitophagy, including mitochondrial association of E3-ubiquitin ligases, extensive mitochondrial UBB, binding of various adapter proteins (OPTN, p62, NBR1, NDP52), and association with LC3-II and GABARAP on the autophagolysosome. However, OVX mice failed to display accumulation of any of these mitophagy-associated proteins, implying severe impairments in the capacity to induce mitophagic flux. Importantly, while basal rates of mitophagy were repressed in LFD-fed OVX mice, the capacity for exercise to stimulate mitophagic flux remained intact in these mice. This suggests a unique interaction between loss of ovarian function and HFD-induced MASLD in the regulation of hepatic mitophagy.

In line with our previous report [30], we found no alterations in the accumulation of all identified mitophagy receptor proteins (i.e. BNIP3, NIX, Fundc2, Phb2) 2 hours post-exercise. In skeletal muscle, knockout of the PARKIN (Parkin RBR E3 ubiquitin protein ligase) resulted in the complete absence of exercise-induced mitophagic flux, demonstrating the critical role of ubiquitination in the activation of mitophagy [46]. Without the ubiquitination of mitochondrial proteins, autophagy adaptor proteins no longer bind to mitochondria, resulting in drastic reductions in the completion of the mitophagy process [47]. Likewise, we observed the recruitment of E3-ubiquitin ligases to hepatic mitochondria following acute exercise, corresponding with increased mitochondrial-bound UBB and adapter proteins. Together, this data suggests exercise may preferentially engage ubiquitin-mediated mitophagy during the acute phases of recovery in metabolic tissues, like liver and muscle, to stimulate the removal of damaged mitochondrial components. However, it is important to emphasize that receptor-mediated mitophagy is still critical for maintaining hepatic mitochondrial function, possibly playing a greater role in regulating basal rates of mitophagy vs. those stimulated by metabolic stressors like exercise. We have previously shown that female liver-specific BNIP3 knockout mice cannot increase mitochondrial respiratory capacity in response to the paired metabolic stressors of HFD and voluntary wheel-running [8]. Thus, while our data suggest acute exercise primarily engages ubiquitin-mediated mitophagy, receptor-mediated mitophagy likely plays a more prominent role in basal mitochondrial turnover and training adaptations, warranting further investigation.

Beyond the context of exercise, the 4-week dietary intervention within the present study prompted an upregulation in hepatic mitophagic flux independent of ovarian function. While several groups have reported alterations in hepatic mitophagy markers following HFD-feeding, the findings are largely inconsistent, likely due to not employing global autophagy/mitophagy inhibitors that allow for quantification of mitophagy flux [48–50]. In humans, individuals with MASLD present with widespread reductions in the expression of both autophagy and mitophagy markers [25]. Thus, the accumulation of mitophagy markers following HFD feeding is likely a short-lived compensatory response in OVX mice, which are known to be susceptible to diet-induced MASLD [16]. However, follow-up studies employing a more prolonged dietary intervention are needed to determine how mitophagy flux is modulated by diet in the OVX chronically.

Mitochondrial fission is critical for successful organelle quality control by mitophagy because of the molecular size restriction for cargo incorporation into the autophagosome and lysosome [51, 52]. Mitochondrial fission is primarily regulated by the recruitment of DRP1 to the mitochondria, where DRP1 promotes the scission of mitochondria into individual fragments [53]. Acute exercise induces the translocation of DRP1 to the outer mitochondrial membrane in both skeletal muscle and liver [30, 54]. Following the acute bout of exercise, HFD-fed Sham mice possessed significant enrichment of mitochondrial-bound DRP1 compared to OVX mice, indicative of greater mitochondrial fission. Liver-specific deletion of DRP1 results in the accumulation of various mitophagy components, including UBB, p62, and LC3-II, in the absence of autophagy/mitophagy inhibitors, suggesting that these are compensatory responses to impaired fission [52, 55]. In summary, impaired exercise-induced activation of mitophagy in OVX mice may ultimately extend from a reduced capacity to recruit DRP1 to hepatic mitochondria and impaired activation of mitochondrial fission.

Elevated ROS production is known to be a potent stimulus for the induction of mitochondrial fission [52, 56, 57]. In the present study, exercise elicited increased mitochondrial H₂O₂ across all groups, except HFD-OVX mice. Proteomics revealed an accumulation of antioxidant enzymes in leupeptin-treated OVX mice, suggesting that antioxidant enzymes may undergo more rapid turnover in OVX mice to sustain antioxidant redox states. This compensatory response may prevent the buildup of excess H₂O₂ in HFD-OVX mice, effectively blocking the capacity for exercise to target dysfunctional mitochondria and thus reducing the subsequent induction of both mitochondrial fission and mitophagy. Furthermore, estrogen receptor α (ERα) has been shown to directly regulate mitochondrial fission by promoting the recruitment of DRP1 to the mitochondria in skeletal muscle, adipose tissue, and pancreas [58–60]. Loss of estrogen-mediated ERα signaling may further hinder the recruitment of DRP1 to the mitochondrial membrane. However, this requires further investigation into regulating DRP1 by hepatic ERα. Thus, the absence of exercise-mediated activation of mitophagic flux in HFD-fed OVX mice may be two-fold: compensatory increased reliance on antioxidant enzyme turnover and reduced recruitment of DRP1 to the mitochondria.

Previously, we have reported that females possess an expanded capacity to increase mitochondrial respiration in response to metabolic challenges compared to males [7, 8], while OVX results in reduced mitochondrial coupling control following prolonged dietary intervention [16]. Although hepatic mitochondrial respiratory function was preserved under LFD conditions in the present study, HFD-feeding resulted in significantly diminished mitochondrial respiratory capacity and respiratory conductance in OVX versus Sham. Intriguingly, we observed no differences in spare respiratory capacity across all groups, meaning that the maximal rate of electron flow through the ETS remains intact in OVX mice. This suggests that the impairments in mitochondrial respiratory capacity in OVX mice are not attributed to the alterations in the capacity for electrons to enter and flow through the electron transport system. This then implicates Complex V as the primary source of reduced mitochondrial respiration in HFD-fed OVX mice. In skeletal muscle, mitochondrial ATP synthesis rates are reportedly reduced in mice 4 weeks post-OVX, with estrogen replacement fully restoring ATP synthesis rates [17]. This evidence implies that loss of estrogen synthesis via OVX may result in reduced rates of mitochondrial respiration by impairing the functional capacity of Complex V. We are unable to directly correlate the inability to induce mitophagic flux and impairments in mitochondrial function in the present study. However, we have previously shown that reduction in hepatic mitophagy (due to loss of BNIP3) prevents compensatory increases in hepatic mitochondrial respiratory capacity in response to prolonged diet and exercise when compared to wildtype female mice [8]. Thus, the induction of mitophagic flux in response to metabolic stress may be crucial for maintaining mitochondrial respiratory capacity, ultimately contributing to innate protection against MASLD in females with normal ovarian function. Overall, this data suggests that innate protections against MASLD in female mice may be attributed to increased rates of mitophagic flux, which support compensatory increases in mitochondrial respiratory capacity.

Importantly, what remains unknown is how loss of ovarian function may be modulating hepatic mitophagy and mitochondrial respiration in general. It is likely that sex hormones, particularly estrogen, may influence these processes in a variety of ways. Acute estrogen exposure in cell culture have shown estrogen directly promotes mitophagy by stimulating the activation of Sirtuin 1 (SIRT) 1 and Unc-51 like autophagy activating kinase 1 (ULK1) [61]. Upstream of mitophagy, estrogen has been shown to directly modulate several components of mitochondrial bioenergetics, altering mitochondrial membrane fluidity, increased mitochondrial JO₂ while reducing the production of mitochondrial H₂O₂ [17, 18], all of which may indirectly regulate the rate of mitophagy. As previously stated, estrogen signaling via ERα has also been shown to modulate mitochondrial fission, a process necessary to complete mitophagy [58, 59, 62]. Thus, it is likely the estrogen may be modulating hepatic mitophagic flux through a complex, multifaceted regulation of mitochondrial bioenergetics, mitochondrial dynamics, and signaling pathways upstream of mitophagy both directly and through estrogen receptor signaling.

The present study is not without limitations. First, we lack an experimental group receiving estrogen replacement, limiting our ability to attribute the mitochondrial impairments found in OVX mice to deficits in estrogen specifically. Therefore, many of our suggested mechanistic insights are phenotypic, lacking direct evidence of mechanisms regulating altered regulation of mitophagy and mitochondrial respiration in OVX mice. While using isolated mitochondria with western blotting and proteomics allows us to measure the effects of OVX on mitochondria directly, we cannot assess cellular signaling pathways that may be involved in the observed mitochondrial phenotypes of our OVX mice. Our proteomics analysis does provide us with various upstream targets for future investigation. Secondly, although the use of OVX to induce menopause in rodent models is widely used, this model fails to replicate progressive declines in sex hormone production that occur during menopause in humans [63], possibly reducing the translatability of these results to the human condition. However, it is important to note that OVX produces systemic alterations that mimic those observed in post-menopausal women, including expansion of visceral adiposity, reduced systemic energy expenditure, and increased susceptibility to metabolic disease [64, 65], highlighting the value of the OVX model in understanding the physiological impacts of menopause. Moreover, we have recently leveraged a VCD model of menopause in mice, demonstrating similar alterations in mitochondrial respiratory capacity in hepatic mitochondria [31]. Third, while the use of acute exercise is a powerful tool for studying mitophagic flux, the mechanism of action through which exercise mediates the activation of mitophagy may be unique compared to other conditions known to modulate the activation of mitophagy within the liver, such as prolonged fasting, requiring further investigation to determine if these stimuli for mitophagy are also impaired following OVX. Future works should directly investigate the effects of sex hormone replacement on mechanisms regulating mitophagy to further assess the importance of mitophagy in regulating protection against MASLD in the female liver.

Conclusion

The present study demonstrates that loss of sex hormone production within female mice directly impairs exercise-induced activation of hepatic mitophagy. While exercise is capable of increasing mitophagic flux in OVX mice, this is mitigated by HFD-feeding and MASLD, possibly contributing to reductions in hepatic mitochondrial function. Alterations in both ROS production, DRP1 recruitment to the mitochondria, and mitochondrial dynamics may further contribute to these mitochondrial deficits. This novel investigation into the underlying effects of prolonged sex hormone depletion on hepatic mitochondrial function and quality control provides valuable insight into the mechanism regulating increased susceptibility to MASLD observed in post-menopausal women.

Supplementary Material

Supplemental Figs. S1-S4: DOI: https://doi.org/10.6084/m9.figshare.28800593.v1

Supplemental Tables S1-S2: DOI: https://doi.org/10.6084/m9.figshare.28800734.v1

New & Noteworthy:

Loss of ovarian function reduces hepatic mitochondrial respiratory capacity, but mechanisms are unknown. Here, we leverage exercise-induced hepatic mitophagy activation to determine if loss of ovarian function impairs mitochondrial quality control mechanisms. Our data reveal that loss of ovarian function reduces both ubiquitin-mediated hepatic mitophagy and mitochondrial recruitment of Drp1(mitochondrial fission protein) following acute exercise. These impairments to hepatic mitophagy coincided with alterations in hepatic mitochondrial respiratory capacity and mitochondrial-derived H₂O₂ production.

Data Availability Statement

All raw and analyzed proteomic datasets curated during the current study are openly available in figshare at https://doi.org/10.6084/m9.figshare.28473929.v2