Mice were purchased from the Jackson Laboratory (Bar Harbor, ME, United States) and conditional Cyp-D KO mice were generated as described below in the Methods section. Mice were bred and group-housed in the Biologic Resource Facility (accredited by the Association for Assessment and Accreditation of Laboratory Animal Care) at the Rosalind Franklin University of Medicine and Science. Lights were set at the recommended illumination levels with a 12/12-h cycle controlled via automatic timers and the temperature maintained between 68 and 74°F. Mice were fed high-quality commercial laboratory diets ad libitum.

NativePAGE Novex Bis-Tris gels (3–12%), NativePAGE running buffer, NuPAGE Bis-Tris gels (4–12%), NuPAGE MOPS SDS running buffer, NuPAGE transfer buffer, PVDF membrane, and SuperSignal West Femto maximum sensitivity substrate were obtained from Thermo Fisher Scientific Inc. (Brookfield, WI, United States). The antibodies phospho-AMPK (Thr 172), phospho-TBC1D1 (Ser 660), and β-tubulin were from Cell Signaling Technology (Danvers, MA, USA). Antibody for peroxisome proliferator activator receptor-γ coactivator 1α (PGC-1α) was from Santa Cruz Biotechnology (Santa Cruz, CA, United States). All other fine chemicals were obtained from Sigma-Aldrich (St. Louis, MO, United States).

Experiments were conducted in intact animals to assess the effects of conditional Cyp-D ablation in vivo on oxygen metabolism by measuring VO₂ and derived parameters at rest and during exercise. Two groups of mice, age and gender matched, representing control (n = 10) and conditional Cyp-D KO (n = 10), were subjected to a treadmill running protocol until exhaustion, during which VO₂, carbon dioxide production (VCO₂), and RER, total running time before exhaustion, total running distance, and work performed were measured. Fuel utilization was calculated using VO₂ and VCO₂ data. Five running sessions were performed. After the running sessions, mice skeletal muscle tissues were harvested, quickly frozen in liquid N₂, and stored at −80°C for Western blotting.

The Jackson Lab has cryopreserved sperm (stock 005737, Ppiftm1Mmos/J) that was heterozygous for the Cyp-D floxed allele (Ppif^Fl/+). This stock was originally generated by Dr. Stanley J. Korsmeyer’s lab (Schinzel et al., 2005). Jackson Lab performed a cryorecovery from the frozen sperm using wild-type female (stock 000664, C57BL/6J) as oocyte donors. A heterogeneous Ppif colony (Ppif^Fl/+) was generated by the Jackson Lab.

Ppif^Fl/+ mice from two different donating mothers were inter-crossed to generate homozygous Cyp-D floxed mice (Ppif^Fl/Fl). Then, Ppif^Fl/Fl mice were crossed with mice homozygous for the ROSA26-Cre^ERT2 (Cre/Cre) transgene (stock 008463; Gt(ROSA)26Sortm1(cre/ERT2)Tyj) to generate mice heterozygous for the Cyp-D floxed allele and hemizygous for ROSA26-Cre^ERT2 transgene (Ppif^Fl/+-Cre/⁺). Next, the Ppif^Fl/+-Cre/⁺ mice were crossed with Ppif^Fl/Fl mice to generate conditional Cyp-D knockout mice (Ppif^Fl/Fl-Cre/⁺) and controls (Ppif^Fl/Fl).

Genotyping was performed according to the recommendations by the Jackson Lab. Briefly, tail biopsy (~2 mm) from 21-day-old pups were digested overnight at 55°C using 0.5 ml/sample of DNA isolation buffer (100 mM Tris.Cl, pH 8, 5 mM EDTA, 0.2% SDS, and 200 mM NaCl) containing 5 μl of 20 mg/ml proteinase K. The samples were then heated at 95°C for 10 min to inactivate Proteinase K. Equal volume of phenol: CHCl₃: IAA mixture was then added, mixed gently for 5 min to allow precipitation of proteins, and centrifuged at 12, 000 g for 5 min at 4°C for phase separation. The supernatant was carefully removed, 2 volumes of 100% ethanol were added, incubated at room temperature for 5 min, and then centrifuged at 12,000 g for 5 min at 4°C for DNA precipitation. The supernatant was discarded, and the pellet (DNA) was washed with 1 ml of 70% ethanol. The pellet was then dried at 37°C for 5 min and dissolved in TE buffer. The DNA concentration was determined by absorbance at 260 nm and 150 ng of DNA was used for the PCR. Two PCR reactions per sample were performed: one to detect Cre gene product and another one to detect Flox gene product. For Cre PCR, the primers used were as: oIMR 3621 (F: 5' to 3' CGT GAT CTG CAA CTC CAG TC) and oIMR 9074 (R: 5' to 3' AGG CAA ATT TTG GTG TAC GG). For Flox PCR, the primers used were as: oIMR 5116 (F: 5' to 3' GCT TTG TTA TCC CAG CTG GCG C) and oIMR 5115 (R: 5' to 3' TTC TCA CCA GTG CAT AGG GCT CTG). PCR was performed in 12 μl of reaction volume using the kit (KAPA2G Robust Hotstart Readymix PCR kit) using an applied biosystems thermal cycler (GeneAmp PCR system 2700). The thermal cycler conditions were set according to the genotyping protocol for mice stock # 005740 by the Jackson Lab and the PCR products separated using 0.7% agarose gel electrophoresis.

Tamoxifen base (T5648, Sigma) was dissolved in peanut oil (10 mg/ml) and was given as 40 mg/kg as described by Andersson et al. (2010). Eight-week-old Ppif^Fl/Fl-Cre^/+ mice (Flox/flox/cre, Cyp-D CKO) and Ppif^Fl/Fl-Cre^/− mice (Flox/flox, control mice) received intraperitoneally 1 mg of tamoxifen per day for five consecutive days. Administration of tamoxifen induces activation of the ubiquitously expressed Cre^ERT2 recombinase, which stimulates recombination and excision of floxed Cyp-D in all tissues. Since Ppif^Fl/Fl mice do not contain Cre gene, they will not undergo Cyp-D recombination and serve as control.

After 2 weeks of completion of tamoxifen injections (i.e., on day 19 as shown in Figure 1), treadmill running was started to assess effects on oxygen metabolism (i.e., VO₂, VCO₂, and RER) in CKO and control mice at rest and during treadmill exercise using a calorimeter (Oxymax System, Columbus Instruments).

Treadmill running had two components: (1) Training protocol and (2) Treadmill exercise protocol.

Linear mixed effect model was used to assess the effect of genotype and running sessions and their interaction on metabolic variables treating running sessions as discrete variable (SPSS 24.0, IBM Corp.; Armonk, NY, United States). Student’s t-test was used to assess (i) the effect of genotype on metabolic variables collected at baseline and at exercise and (ii) the effect of genotype on molecular signaling implementing t-test (when normality test passed) and Mann-Whitney Rank Sum test (when normality test failed; SigmaPlot v.12.5). Strength of relationships between variables was assessed by linear regression and Pearson’s product moment correlation analysis (SigmaPlot v.12.5). Data are presented as means ± SD in the text and means ± SEM in the figures. A two-tail value of p ≤ 0.05 was considered significant.

Genomic organization of the mouse genome at chromosome 14 and the strategy for conditional KO generation is shown in Figure 1. As mentioned, Ppif^Fl/+-Cre^/+ mice have Ppif allele flanked between two lox-P sites which upon tamoxifen injection will undergo Cre-mediated recombination resulting in deletion of the Ppif allele. Ppif^Fl/+-Cre^/+ mice pups were genotyped to confirm the presence of Ppif^Fl/+ and Cre genes. Results showed that the products of expected size were generated by flox PCR (400 bp) and by Cre PCR (200 bp). Tamoxifen injection resulted in Cre-mediated deletion of lox-P flanked Ppif allele and was confirmed by Western blotting detection of Cyp-D levels in mice skeletal muscle tissue homogenates (Figure 1).

At their first treadmill running session, as shown in Figure 2A, CKO mice and controls had comparable VO₂, VCO₂, and RER. However, with subsequent sessions, the VO₂ decreased in CKO mice without a significant change in VCO_2. The lower VO₂ for a given VCO₂ manifested in a significantly higher RER that remained statistically significant for most of the sessions (Figure 2A). This behavior is consistent with a metabolic adaptation induced by exercise training in CKO mice.

All individual VO₂, VCO₂, and RER measurements obtained at baseline – before starting treadmill exercise – combining the five sessions are shown in Figure 2B. CKO mice had a lower VO₂ (p = 0.003) and a lower VCO₂ (p = 0.014) but of lesser magnitude yielding a slightly higher RER which did not attain statistical significance (p = 0.201).

All individual VO₂, VCO₂, and RER measurements obtained during treadmill running combining the five sessions are shown in Figure 2C. CKO mice also exhibited a lower VO₂ (p = 0.04) but without a significant difference in VCO₂ (p = 0.348), yet with a highly significant increase in RER (p < 0.001).

From the first and the subsequent sessions, CKO mice had a higher power and power/VO₂ but did not attain statistical significance (Figure 3). Yet, when all individual power and power/VO₂ ratio throughout the five running sessions were combined, CKO mice had a statistically significant higher power and power/VO₂ (Figure 3), consistent with increased oxygen utilization efficiency.

As shown in Figure 4A, at their first running session, the fuel utilization was comparable between CKO mice and controls. Yet, with subsequent running sessions, carbohydrate utilization was higher in CKO mice but the difference was not statistically significant. Likewise, fat utilization was also comparable between the genotypes at the first running session. However, with subsequent running sessions, fat utilization was lower in CKO mice attaining statistical significance in most of the running sessions (Figure 4A).

All individual carbohydrate and fat utilization at baseline – before starting treadmill exercise – combining the five sessions demonstrated comparable carbohydrate and fat utilization in CKO mice and controls (Figure 4B). All individual carbohydrate and fat utilization during exercise throughout the five running sessions showed increased carbohydrate and decreased fat utilization in CKO mice (Figure 4C).

AMPK phosphorylation at Thr 172 – indicative of its activation – was increased by 64% and TBC1D1 phosphorylation Ser 660 – indicative of its inactivation – was increased by 55% in CKO mice skeletal muscles compared to control (Figure 5). Analysis of relationship between pAMPK and pTBC1D1 (Figure 5) showed a positive correlation which did not attain statistical significance (R = 0.639; p = 0.08).

PGC-1α levels increased by 162% and GLUT4 levels increased by 103% in CKO mice skeletal muscles compared to control (Figure 6). Analysis of relationship between GLUT4 and AMPK demonstrated a positive correlation without statistical significance (R = 0.686; p = 0.061).

The power/VO₂ exhibited a positive correlation with the pAMPK/β-tubulin ratio (R = 0.753; p = 0.03) and the PGC-1α/β-tubulin ratio (R = 0.708; p = 0.04; Figure 7).

We have previously reported in HEK 293 T cells that Cyp-D silencing results in downregulation of mitochondrial transcripts and concomitant reduction in cellular oxygen consumption (Radhakrishnan et al., 2015). Further studies in constitutive Cyp-D KO mice demonstrated downregulation of VO₂ at rest and during endurance exercise (Radhakrishnan et al., 2019). In the present studies, we demonstrated that VO₂ in conditional Cyp-D KO adult mice at rest and during exercise was comparable to control mice in the first running session and decreased only with subsequent sessions. Decrease in VO₂ was accompanied by increased RER indicative of an adaptational response and training effect in CKO mice.

During exercise, in contracting muscles, activation of AMPK occurs consequent to AMP accumulation as ATP consumption increases (Winder and Hardie, 1996). AMPK activation occurs mainly through activation of the upstream liver kinase B1(kinase LKB1) through Thr172 phosphorylation of its catalytic domain in response to increased cellular AMP:ATP ratio (Hawley et al., 2003; Woods et al., 2003). Activated AMPK inactivates its downstream target TBC1D1 (Kramer et al., 2006; Treebak et al., 2006; Taylor et al., 2008; Ferrari et al., 2019; de Wendt et al., 2021) by phosphorylation (an interaction also known as “AMPK-TBC1D1 signaling nexus”; Chavez et al., 2008; Taylor et al., 2008; Vichaiwong et al., 2010; Treebak et al., 2014; Kjøbsted et al., 2016; Ferrari et al., 2019). TBC1D1 inactivation during exercise has been reported in mouse skeletal muscles (Taylor et al., 2008; Vichaiwong et al., 2010; de Wendt et al., 2021) and in human skeletal muscles (Jessen et al., 2011; Tobias et al., 2020). Activation of AMPK-TBC1D1 signaling nexus results in increased GLUT4 expression and translocation from the cytosol to the plasma membrane enabling increased glucose uptake. Glucose enters the glycolytic pathway and is metabolized in the cytosol to pyruvate and then to lactate by lactate dehydrogenase (LDH). The conversion of pyruvate to lactate is favored over alternative pyruvate fates given the near-equilibrium nature of the LDH reaction (Lambeth and Kushmerick, 2002) and a much higher activity than regulatory enzymes of the glycolytic and oxidative pathways (Rogatzki et al., 2015). Lactate then enters the mitochondria where it is converted to pyruvate by the LDH present on the mitochondrial inner membrane (Hashimoto et al., 2006; Gladden, 2008). Pyruvate is then converted to acetyl coA and metabolized in the Krebs cycle. Thus, the AMPK-TBC1D1 signaling nexus increases glucose uptake and enhances glucose utilization through glycolysis and the Krebs cycle. Consistent with this effect, we observed in the present study activation of AMPK-TBC1D1 signaling nexus (Figure 5) and increased GLUT4 expression (Figure 6) in CKO mice after exercise. Consistently, several other studies have also showed increased expression of skeletal muscle GLUT4 in rats and human after exercise (Ivy et al., 1999; Kuo et al., 1999; MacLean et al., 2000; Flores-Opazo et al., 2020). Increased GLUT4 expression has been correlated with increased glucose transport and utilization (Kern et al., 1990; Ivy et al., 1999; MacLean et al., 2000).

In addition, AMPK activation also increases the expression of PGC-1α (as also shown in Figure 6). The effect occurs by phosphorylation at its threonine-177 and serine-538 residues prompting self-dependent activation of its promoter and gene expression (Jäger et al., 2007). PGC-1α further increases GLUT4 expression and thereby glucose uptake (Jäger et al., 2007). In addition, PGC-1α activates mitochondrial biogenesis and other pathways, which could improve the exercise capacity after training. Several lines of evidence suggest that PGC-1α is a key factor in mediating exercise training induced adaptations in mitochondria (Olesen et al., 2010) in response to AMPK activation. For example, AICAR – the AMPK activator – improves exercise capacity in mice with concomitant increase in PGC-1α gene expression (Narkar et al., 2008). In addition, AICAR injections fail to increase levels of mitochondrial and metabolic proteins in skeletal muscles of PGC-1α KO mice (Leick et al., 2010). Thus, AMPK in CKO mice could induce additional adaptive responses, including mitochondrial biogenesis via PGC-1α, leading to enhanced exercise capacity.

In our previous studies in constitutive Cyp-D KO mice, the aforementioned metabolic effects were associated with downregulation of electron transport complexes I and IV activities consequent to reduced expression of subunits encoded by the mitochondrial heavy strand promoter 2 (Radhakrishnan et al., 2015, 2019). The expression of these subunits requires the interaction of Cyp-D with the mitochondrial transcription factors B1 and B2 (Radhakrishnan et al., 2015). The reduced activity of electron transport complexes I and IV would create an “energy-stress” (Gowans et al., 2013) during exercise prompting an increase in AMP:ATP ratio and the consequent activation of the AMPK-TBC1D1 signaling nexus. Other studies have also reported complex I inhibition as the mechanism of AMPK activation. In fact, the mechanism of pharmacological activation of AMPK by metformin involves electron complex I inhibition (Owen et al., 2000; Zhou et al., 2001; Mohsin et al., 2019). However, metformin abrogates exercise-mediated increase in skeletal muscle mitochondrial respiration (Konopka et al., 2019).

In our previous study in constitutive Cyp-D KO mice, we have documented a metabolic shift – i.e., increased glucose utilization over fat. Tavecchio et al. (2015) have documented in their study that Cyp-D constitutive KO mice had increased transcription of genes involved in glucose metabolism resulting in increased glucose utilization and impaired fatty acid utilization. In the present study in conditional Cyp-D KO mice, we have documented a similar adaptive response. In the present study, AMPK activation was accompanied by increased glucose utilization over fatty acids (as shown in Figure 4) that is energetically advantageous. Glucose oxidation per mole yields 6.3 moles of ATP, whereas fatty acid yields only 5.6 moles of ATP (Opie, 1991; Stanley and Chandler, 2002) resulting in a 12.5% increase in energy production with glucose oxidation. Thus, this metabolic switch enhances oxygen utilization efficiency during exercise and improves exercise capacity as evidenced by the significant increase in power during treadmill exercise for given oxygen consumed (power/VO₂ ratio; as shown in Figure 3).

Both constitutive and conditional Cyp-D KO had their Cyp-D knocked-out in all tissues. Cyp-D ablation was carried out at an embryonic stage in constitutive KO, whereas Cyp-D ablation was carried out at the adult stage in CKO mice. Yet, the CKO mice displayed similar effects – VO₂ downregulation and metabolic switch – as that of constitutive KO but of lesser magnitude. The metabolic switch in CKO mice was prominent at the second running session and thereafter, as documented by increased glucose utilization over fatty acids (Figure 4). The first and second running sessions were performed at the third and the fourth week, respectively, after tamoxifen treatment for 5 days. This hints to the possibility that the metabolic switch can be triggered probably only 4 weeks after completion of pharmacological interventions. The time delay for this metabolic switch – we rationalize – includes the time required for cre/lox-P recombination-dependent Cyp-D ablation and subsequent respiratory chain downregulation. In addition, exercise training induced mitochondrial adaptations involving PGC-1α could also have played role in improving the exercise capacity. These results point to the fact that selectively destabilizing interaction between Cyp-D and mitochondrial transcription factors to reduce oxygen consumption without affecting other functions of Cyp-D in adults could have potential positive impact on oxygen utilization efficiency and it could be highly beneficial under oxygen limiting conditions.

Conditional Cyp-D ablation reduced VO₂ and increased power/VO₂ ratio, demonstrating increased oxygen utilization efficiency consistent with our previous work in constitutive KO mice. Conditional Cyp-D ablation also triggered the adaptive response which involved a metabolic switch toward increased glucose utilization with concomitant activation of AMPK-TBC1D1 signaling nexus and increased GLUT4 expression in skeletal muscles after exercise. Our study raises the possibility that the metabolic switch can probably be triggered in 4 weeks after completion of pharmacological intervention. Developing tools for modulation of this adaptive response could be beneficial in a myriad of physiological and clinical conditions where oxygen availability is limited.