Cyclophilin D deficiency exacerbates diet-induced obesity and selectively enhances skeletal muscle mitochondrial respiration in male mice

Mouse experiments

Generation of Ppif knockout mice has been described previously [4]. Ppif knockout mice were maintained on a mixed C57BL/6:129 genetic background, with unknown Nnt status (Jackson Lab strain #009071; B6;129-Ppiftm1Jmol/J). The study mice were bred true from homozygous WT and KO breeders, respectively, which had been generated from heterozygous breeders. All studies were performed exclusively in male mice, which were group-housed with a 12 h light, 12 h dark cycle (7am-7pm) at 22 °C and maintained on a standard chow diet (Harlan Teklad LM-485) or high-fat diet (Research Diets D12492 (60% fat, New Brunswick, NJ, USA), with ad libitum access to food and water. High-fat diet feeding was initiated at 20.9 ± 0.3 weeks of age and continued for 34 weeks.

Fat and lean mass were quantified using NMR technology (EchoMRI, Houston, TX, USA). Energy expenditure, respiratory quotient, food intake, and locomotor activity were assessed using a combined indirect calorimetry system (TSE Systems GmbH, Bad Homburg, Germany) as described previously [8]. Mice had ad libitum access to food and water and were acclimated to metabolic cages for at least 24 h followed by the simultaneous measurement over a 72‑hour period. Intraperitoneal glucose and insulin tolerance testing (GTT/ITT) was conducted in mice subjected to 6 h of fasting and injected intraperitoneally with 1.5 g glucose / kg body weight (20% D-glucose (Sigma) in 0.9% saline) for the GTT, and 1.0 U insulin / kg body weight (Humolog, Lily, Indianapolis, USA) for the ITT. All studies were approved by and performed according to the guidelines of the Institutional Animal Care and Use Committee of the University of Cincinnati, USA.

Mitochondrial isolation and Seahorse^™ respirometry

Mitochondria were isolated from quadriceps muscle, brown adipose tissue (BAT), and liver. Tissue samples were excised and transferred into 2‑ml microcentrifuge tubes containing 1 ml of Isolation Buffer 1 (IB1: 210 mM mannitol, 70 mM sucrose, 5 mM HEPES, 1 mM EGTA, 0.5% fatty‑acid–free BSA). Samples were homogenized and centrifuged at 700 × g for 10 min. The supernatant was collected and centrifuged at 14,000 × g for 10 min, and the resulting pellet was resuspended in fresh IB1 and centrifuged again at 10,000 × g for 10 min. This wash step was repeated using Isolation Buffer 2 (IB2: 210 mM mannitol, 70 mM sucrose, 10 mM MgCl₂, 5 mM potassium diphosphate, 10 mM MOPS, 1 mM EGTA). The final mitochondrial pellet was resuspended in 50–200 µl IB2.

Mitochondrial oxygen consumption rate (OCR) was measured using an XF‑24 extracellular flux analyzer (Seahorse Bioscience) following a 10‑min calibration at 37 °C. Isolated mitochondria (15 µg per well) were loaded into XF‑24 plates in 500 µl MAS‑1 buffer (70 mM sucrose, 220 mM mannitol, 5 mM KH₂PO₄, 5 mM MgCl₂, 50 mM KCl, 2 mM HEPES, 1 mM EDTA, 0.1% fatty‑acid–free BSA, pH 7.4). Sequential injections of substrates and modulators were used to define respiratory states: State II (5 mM glutamate, 5 mM malate), State III (0.25 mM ADP), State IV₀ (1 µM oligomycin), and maximal uncoupled respiration induced by FCCP (300 nM). OCR traces were used to calculate basal respiration, ADP‑stimulated State III respiration, oligomycin‑inhibited State IV₀ respiration, and FCCP‑stimulated maximal respiration. Respiratory control ratios (RCRs) were determined as State III/State IV₀ OCR to assess mitochondrial coupling efficiency across genotypes and dietary conditions.

Citrate synthase activity in mitochondrial isolates from quadriceps, BAT, and liver was quantified spectrophotometrically according to the manufacturer’s instructions (CS0720 kit, Sigma-Aldrich, St. Louis, MO, USA).

Statistical analysis

Statistical analyses were performed using GraphPad Prism (V.9 or higher) or SPSS (V31). Longitudinal comparisons of body weight and percentage weight gain, and glucose excursions in the GTT and ITT, were performed using Two-Way RM ANOVA followed by Sidak’s multiple comparisons test. AUC values and differences in weight, fat and lean mass, fat and lean mass gain, food intake, respiratory quotient and locomotor activity between genotypes within each dietary condition or light/dark phase were evaluated using two-tailed unpaired t-tests. Energy expenditure values were assessed by Analysis of Co-Variance (ANCOVA), using body weight or lean mass as covariates. Results are shown as means ± SEM.

Higher weight and fat mass in adult Ppif KO mice on SD

To investigate the impact of Ppif-deficiency on the development of adult-onset obesity, we conducted body composition analysis, indirect calorimetry, and glucose homeostasis testing in 20.9-weeks-old Ppif WT and KO mice fed standard diet (SD; week 0; Fig. 1a). KO mice exhibited significantly higher body weight (Fig. 1b) and fat mass (Fig. 1c), while lean mass was comparable between genotypes (Fig. 1d). Although food intake remained unchanged, locomotor activity during the dark phase was significantly increased in KO mice relative to WT mice (Fig. 1e, f). Energy expenditure, adjusted for lean mass and assessed by analysis of covariance (ANCOVA), was comparable between genotypes during the light (F(1,13) = 0.948, p = 0.348) and dark (F(1,13) = 1.884, p = 0.193) phases (Fig. 1g, h). Glucose tolerance, evaluated using a glucose tolerance test (GTT), showed no difference between genotypes, with comparable glucose excursions and area under the curve (AUC) values (Fig. 1i, j). However, during the insulin tolerance test (ITT), blood glucose levels at the 120-minute timepoint were significantly higher in KO mice compared to WT mice (Fig. 1k), although the overall AUC values did not reach statistical significance (Fig. 1l). Together, our data show that Ppif KO mice on SD display modestly increased body weight and fat mass, elevated dark-phase locomotor activity, and a slight impairment in late-phase insulin tolerance, without noticeable differences in food intake, energy expenditure, or overall glucose tolerance.

HFD exposure escalates adult-onset obesity in Ppif-deficient mice

We next divided Ppif WT and KO mice into two weight-matched cohorts, each receiving either SD or HFD, and monitored body weight and body composition over a period of 34 weeks. Under SD, body weight trajectories and percentage weight gain were comparable between WT and KO mice throughout the 34-week-monitoring (Fig. 1m). Under HFD, KO mice showed significantly greater body weight from weeks 28 to 34 (Fig. 1n). After 34 weeks, fat mass gain was greater in HFD-fed KO compared to WT mice (Fig. 1o), while no significant difference in fat mass gain was observed between genotypes under SD. Lean mass gain was also significantly greater in HFD-fed KO relative to WT mice (Fig. 1p), with no significant difference between genotypes under SD. Collectively, these results indicate that Ppif deficiency worsens diet-induced obesity, leading to a greater accumulation of body weight, fat mass, and lean mass specifically in response to HFD exposure.

Indirect calorimetry does not reveal differences in metabolic homeostasis in SD-fed Ppif-deficient mice

To determine whether the initially observed increase in adiposity in Ppif-deficient mice on SD at week 0 was accompanied by alterations in energy homeostasis, we repeated indirect calorimetry monitoring after 16 weeks of SD feeding (Fig. 2a). Body weights were comparable between genotypes (Fig. 2b), whereas fat mass was moderately increased in KO compared to WT mice (Fig. 2c; p < 0.05). Lean mass remained comparable between groups (Fig. 2d). Food intake patterns were similar between WT and KO mice over 72 continuous hours (Fig. 2e). Likewise, phase-averaged food intake during the light and dark phases did not differ significantly between genotypes (Fig. 2f). Respiratory quotient (RQ) monitoring displayed the expected diurnal oscillation (Fig. 2g) and comparable phase-averaged RQ values during the light and dark phase between WT and KO mice (Fig. 2h). Similarly, locomotor activity followed a diurnal pattern in both genotypes (Fig. 2i), and both light and dark phase locomotion were comparable between groups (Fig. 2j). Energy expenditure (EE) traces over 72 h demonstrated the expected higher activity during the dark phase in both groups (Fig. 2k). ANCOVA revealed that lean mass was a significant predictor of the average EE (F(1,13) = 8.880, p = 0.011), and genotype also exerted a significant main effect (F(1,13) = 4.927, p = 0.045), visualized by the regression plot of EE vs. lean mass (Fig. 2l). When adjusted for body weight (F(1,13) = 7.165, p = 0.022), the significant genotype effect was no longer present (Fig. 2m; F(1,13) = 1.177, p = 0.298). Phase-averaged EE, after adjustment for lean mass, showed a trend toward a genotype effect during the light phase (F(1,13) = 3.661, p = 0.078), and was significantly higher in the dark phase (F(1,13) = 5.733, p = 0.032) in Ppif KO mice (Fig. 2n, o). In summary, these findings indicate that Ppif KO mice on standard diet at 16 weeks show a modest yet significant increase in fat mass and energy expenditure, but no detectable differences in body weight and lean mass, or food intake, substrate utilization and locomotor activity, compared to WT controls.

Ppif-deficient mice are more prone to develop high fat diet-induced obesity

After 16 weeks of HFD feeding (Fig. 3a), Ppif WT and KO mice showed significantly increased body weight and fat mass compared with WT mice (Fig. 3b, c). Lean mass was comparable across genotypes (Fig. 3d). Continuous 72-hour food intake monitoring (Fig. 3e) revealed comparable average food intake during the light and dark phases (Fig. 3f). Respiratory quotient traces over 72 h showed comparable oscillations between WT and KO mice (Fig. 3g); however, the phase-averaged RQ during the dark phase was significantly lower in KO mice (Fig. 3h), indicating a higher utilization of lipids as fuel substrate. Total locomotor activity over 72 h showed comparable patterns between groups (Fig. 3i), and phase-averaged locomotor activity during both light and dark phases was not different between genotypes (Fig. 3j). However, total energy expenditure (EE) over 72 h (Fig. 3k) appeared higher in knockout (KO) mice compared to wild-type (WT) mice. ANCOVA, after adjusting for lean mass, confirmed a significant effect of genotype (F(1, df) = 7.709, p = 0.017), with lean mass being a strong predictor of EE (F(1, df) = 19.597, p < 0.001). Furthermore, a significant Genotype × Lean Mass interaction (F(1, df) = 6.547, p = 0.025) indicated differing EE scaling between genotypes, which was visualized by differing slopes for the linear regression of total EE versus lean mass (Fig. 3l). Of note, adjustment for body weight in the ANCOVA did not yield significance between genotypes, indicated by largely overlapping regression lines for EE plotted vs. body weight (genotype: F(1,12) = 0.280, p = 0.606; body weight: F(1,12) = 12.210; p = 0.004, genotype × body weight: F(1,12) = 0.237, p = 0.635, demonstrated also by largely overlapping regression lines for EE plotted vs. body weight (Fig. 3m). Lean mass-adjusted ANCOVA analyses and regression plots (Fig. 3n, o) further revealed higher light phase EE (genotype: F(1,12) = 8.020, p = 0.015; lean mass: F(1,12) = 23.906, p < 0.001) and higher dark-phase EE (genotype: F(1,12) = 6.018, p = 0.030; lean mass: F(1,12) = 13.041, p = 0.004) with a significant interaction between genotype and lean mass (light phase EE: F(1,12) = 6.944, p = 0.022; dark-phase EE: F(1,12) = 5.020, p = 0.045). Taken together, these data reveal that HFD-fed Ppif KO mice develop greater obesity than WT mice and exhibit selectively increased dark-phase respiratory quotient and light- and dark-phase energy expenditure without differences in food intake or locomotor activity.

Comparable glucose homeostasis in Ppif WT and KO mice on SD and HFD

At week 15 of SD exposure (Fig. 4a), intraperitoneal GTTs revealed comparable glucose excursions and AUC values in WT and KO mice (Fig. 4b, c). In HFD-fed mice, both groups had more pronounced and higher glucose excursions compared to those on the SD. Although the AUC for the KO group seemed higher than that of the WT group, there were no statistically significant differences detected between the two genotypes (Fig. 4d, e; Two-Way Repeated Measures ANOVA left panel: p = 0.088; t-test right panel: p = 0.12). ITTs at week 17 of diet exposure in SD-fed mice revealed a comparable decrease in blood glucose following insulin injection, with no significant differences in AUC values between groups (Fig. 4f, g). In the HFD-fed ITT, both groups exhibited a blunted glucose-lowering response relative to SD-fed animals, yet AUC values again were not different between WT and KO mice (Fig. 4h, i). Collectively, these results indicate that Ppif deficiency did not significantly alter glucose tolerance or insulin sensitivity under either SD or HFD conditions at these timepoints.

Elevated state 3 respiration in quadriceps mitochondria from Ppif KO mice

To determine if the metabolic phenotype of Ppif KO mice was associated with differences in mitochondrial respiratory function or mitochondrial content, we conducted mitochondrial respiration assays and measured citrate synthase activity in isolated mitochondria from the quadriceps, brown adipose tissue (BAT), and liver of WT and KO mice fed SD or HFD for 34 weeks (using four randomly selected mice per group; Fig. 5a). At the time of sacrifice, the surviving Ppif WT (n = 6) and KO mice (n = 5) on SD had comparable weight, fat and lean mass, respectively. Ppif KO mice on HFD (n = 8) had higher weights (Fig. 1n), fat mass (KO: 20.74 ± 2.31, WT: 10.45 ± 1.77; p = 0.0041) and lean mass (WT: 22.83 ± 1.08 g, KO: 27.95 ± 1.36 g; p = 0.012) compared to their WT counterparts (n = 6). In quadriceps mitochondria, ADP-stimulated State III respiration was significantly increased in KO mice compared to WT mice under both SD (Fig. 5b) and HFD (Fig. 5c), while no additional significant differences were detected at other respiratory states. Respiratory control ratio (RCR) values in quadriceps were not different between genotypes under either dietary condition (Fig. 5d). In BAT, oxygen consumption rate (OCR) profiles were comparable between WT and KO mice under SD (Fig. 5e) or HFD (Fig. 5f), and RCR values were similarly comparable between genotypes under both diets (Fig. 5g). In liver, OCR profiles under both SD and HFD and RCR values under both dietary conditions were comparable between WT and KO mice (Fig. 5h–j). Citrate synthase activity was not significantly different between WT and KO mice across all three tissues under SD (Fig. 5k) or HFD (Fig. 5l) conditions, indicating comparable mitochondrial content between genotypes. Across all tissues and dietary conditions, liver mitochondria showed the lowest citrate synthase activity, while quadriceps and BAT exhibited higher and relatively similar levels. In summary, these findings indicate that Ppif deficiency selectively increased ADP-stimulated mitochondrial respiration in quadriceps under both dietary conditions, without broadly altering mitochondrial coupling efficiency or mitochondrial content across metabolically active tissues.

CypD, mPTP sensitization, and tissue‑specific mitochondrial respiration

CypD regulates the Ca²⁺ and redox sensitivity of the mitochondrial permeability transition pore (mPTP), lowering the threshold for pore opening and thereby influencing cell survival, ROS signaling, and metabolic flexibility [3–5]. Consequences of CypD deletion are strongly context‑dependent. For instance, CypD is essential for necrotic cell death in Ca²⁺‑overloaded dystrophic muscle [9], but dispensable for denervation‑induced atrophy [10]. CypD deletion was linked with the protection of mitochondria from metabolic or Ca²⁺ stress [6, 7]. Our findings align with this model of conditional mPTP involvement. ADP‑stimulated respiration in the quadriceps of HFD-fed Ppif‑deficient mice was elevated, which suggests that reduced mPTP susceptibility may enhance mitochondrial stability during sustained nutrient excess. However, our measurements were performed in isolated mitochondria and should be interpreted with caution. Likewise, due to the need to process multiple tissues immediately after euthanasia, our respiration assays were conducted using a common protocol optimized for liver mitochondria. Consequently, FCCP concentrations were likely suboptimal for skeletal muscle and BAT mitochondria, limiting interpretation of maximal uncoupled respiration. Similarly, BAT mitochondria were assayed without GDP-mediated inhibition of UCP1, precluding rigorous assessment of coupled respiration and respiratory control ratios in this tissue. Accordingly, future studies employing tissue-optimized protocols and complementary approaches such as permeabilized muscle fibers will be required to validate and extend these findings.

CypD and mitochondrial fuel selection

CypD loss‑ and gain‑of‑function have been linked to altered metabolic substrate preference. Proteomic profiling of CypD‑null hearts revealed broad remodeling of the TCA cycle, branched‑chain amino acid catabolism, β‑oxidation, and pyruvate metabolism [11]. Other studies report that CypD deficiency biases toward greater reliance on carbohydrate metabolism and reduced fatty‑acid oxidation [12, 13]. In pancreatic β‑cells, reduced mPTP opening enhances coupling efficiency and insulin secretion [14], further implicating CypD in the regulation of metabolic substrates.

Our long‑term phenotyping aligns with this metabolic shift. Ppif‑deficient mice accumulated more fat mass on both SD and HFD without consuming more calories, indicating altered metabolic partitioning rather than hyperphagia. Together with the observed increase in skeletal muscle oxidative phosphorylation, without accompanying changes in mitochondrial content, this pattern may suggest a preferential use of glucose‑supported respiration. Whether such a metabolic role of CypD is consistent with the AMPK‑driven remodeling described in other CypD‑null models [12], remains to be determined.

Tissue‑specificity of CypD effects

CypD action varies markedly across tissues. The heart, which expresses high CypD levels, shows strong CypD‑dependent regulation of mPTP activity, Ca²⁺ handling, and susceptibility to heart failure [15]. In contrast, skeletal muscle exhibits relatively low CypD dependence at rest but stronger involvement during activity‑induced Ca²⁺ flux [16]. Liver‑specific Ppif deletion increases visceral adiposity and glucose intolerance under HFD, whereas muscle‑specific deletion has minimal effects [17].

Our metabolic phenotyping and respirometry data mirror this tissue‑selective pattern. Only the quadriceps showed increased ADP-stimulated State III respiration in Ppif‑deficient mice, the BAT and liver did not. This skeletal‑muscle‑restricted enhancement of oxidative phosphorylation may explain why EE, adjusted for lean mass, increases in Ppif‑deficient mice, yet adiposity still rises, as non‑muscle tissues likely exhibit altered thermogenic, survival or storage properties upon Ppif deletion than muscle.

CypD also has direct roles in adipocyte biology. High‑fat diet induces adipocyte cell death through a CypD‑dependent intrinsic pathway [18], and CypD loss protects adipocytes from lipotoxic necrosis [19]. This protection may favor adipose expansion. Together with our findings, this suggests that increased fat mass in Ppif‑deficient mice, also under SD feeding, may reflect enhanced adipocyte survival, reduced inflammatory turnover, and greater lipid storage capacity. However, this hypothesis requires further experimental validation.

Energy expenditure changes upon Ppif deletion

A major discrepancy with earlier reports of reduced obesity susceptibility in Ppif-deficient mice arises from differences in energy expenditure analysis. Devalaraja‑Narashimha et al. – like us - [19] report elevated energy expenditure, but they normalized VO₂ to total body weight, a ratio‑based approach now recognized as statistically invalid and prone to misleading interpretations. By contrast, applying ANCOVA with lean mass as the covariate, as recommended by Tschöp et al. [20], we reveal that lean mass is the primary determinant of EE and that genotype modifies the EE–lean‑mass relationship in our Ppif-deficient mice. However, this significant lean mass x genotype interaction does not necessarily translate into higher energy expenditure in KO versus WT mice when lean mass is comparable, as was the case at both indirect calorimetry time points. This may partly explain the apparent discrepancy between the calorimetry and body weight data, which at first glance appear difficult to reconcile. However, our calorimetry data - together with the increased lean mass observed in HFD-fed KO mice – identifies skeletal muscle as a key tissue for future investigation and raises the intriguing possibility that CypD may play a previously unrecognized role in age-related skeletal muscle remodeling and potentially sarcopenia.

A limitation in our data is nonetheless the lack of causality between enhanced mitochondrial respiratory capacity, altered energy expenditure and higher adiposity. Future studies are required to clarifying the contribution of these mechanisms, potentially including direct assessments of substrate utilization and metabolic flux, and indirect calorimetry measurements at earlier stages of HFD exposure. Such indirect calorimetry studies, which are subject to technologically inherent limitations in detecting subtle energy expenditure differences between genotypes, primarily quantify oxygen-dependent oxidative metabolism and may thus not capture additional processes contributing to energy balance, such as substrate utilization, metabolic fluxes, futile cycling, or other potential energy-dissipating pathways [21–24]. Accordingly, a potential contribution of such processes to the observed phenotype remains an interesting but untested possibility.

Discrepant reports on obesity susceptibility in Ppif KO mice

Previous studies on the metabolism of whole-body CypD knockout mice have produced inconsistent results. Ppif‑deficient mice either displayed protection from diet‑induced obesity [19] or a higher propensity for adult‑onset obesity developing under normal SD conditions [25]. The latter phenotype was further corroborated in liver‑specific Ppif KO mice, which developed increased adiposity and glucose intolerance on HFD, whereas skeletal‑muscle–specific deletion produced no differences in body weight or fat mass [17]. Our results align more closely with the studies reporting heightened susceptibility to obesity upon loss of CypD. Differences in genetic background, gut microbiome composition, and housing environment remain unavoidable sources of variation across CypD‑knockout studies, including our own, and may interact with genotype in ways that are difficult to fully control. Such cohort‑specific influences likely contribute to the discrepant metabolic fates reported for global or tissue-specific Ppif‑deficient mice.

A clearer understanding of this susceptibility may require deeper analysis of Ppif functions in metabolically active tissues. Such studies should entail adipose‑tissue biology in Ppif‑deficient mice, including adipocyte size distribution, lipid‑turnover kinetics, macrophage infiltration, and crown‑like structure formation, to determine whether the phenotype reflects cell‑autonomous adipocyte effects, altered adipose immune dynamics [18], or both. Given the divergent phenotypes of liver‑ and muscle‑specific knockouts, adipocyte‑specific and myeloid‑specific conditional Ppif deletions could help resolve the tissue contributions to whole‑body adiposity.

Competing interests

The authors declare no competing interests.

Clinical trial number

Not applicable. This study did not involve a clinical trial and was not registered.